<%BANNER%>

Impact of the 30-Baht Health Insurance Policy on Hospital Drug Utilization in Thailand

xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101123_AAAAEZ INGEST_TIME 2010-11-24T05:03:47Z PACKAGE UFE0011544_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 2891 DFID F20101124_AAAANG ORIGIN DEPOSITOR PATH kanjanarat_p_Page_175.txt GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
c909336fb2bc6ac277847a0eba3c7f90
SHA-1
1a57c4ee7245063038cde871f7f7fe8156c1c26c
13936 F20101123_AADHYU kanjanarat_p_Page_162.QC.jpg
50299ec4b61eec0638e9382c0593153c
0e9bf8bb9c5023659aded9c08115f024b08b95e7
10415 F20101123_AADIFD kanjanarat_p_Page_002.jpg
4bd8ac246135a11c787805b61c591215
ee80e1e2139b750d8608d44f89ae867ea09eadaa
5380 F20101123_AADIEO kanjanarat_p_Page_076thm.jpg
d2b163dc4e097b52aeb773be43a18f6e
e7805b802c582471ac5cec4c460308261b72bac7
44576 F20101123_AADHZI kanjanarat_p_Page_107.pro
ccac9e0f3b792a7cbf21fe0597ca2f53
471f92863b34ed5dc672e45f6b0287758100d554
2538 F20101124_AAAANH kanjanarat_p_Page_176.txt
3db91178f256a458883f2fa2fc6c4c3b
6e1882817c559c41469c544751f6822fffa0dfd6
1772 F20101124_AAAAMS kanjanarat_p_Page_160.txt
a27ce9a0109cf84ae3f8183bdc3d0881
e9d3d495c322b6d49914eb21fdbcc1eb26ba755f
6202 F20101123_AADHYV kanjanarat_p_Page_143thm.jpg
ae379b950ef74982da2fe7793f4d312e
459dcc06f169d9e03a75b1dc9d7ae552a13dad11
57468 F20101123_AADIFE kanjanarat_p_Page_004.jpg
6239b61143b989038ca4b6a6e71055cc
3cec1bb814d31b55c76803d44946ac3601d1ece7
2032 F20101123_AADIEP kanjanarat_p_Page_066.txt
52e69f058c8f413a287e57978bf05add
c61d8e27e91bbc0c542fa25a754157c7289207d8
58116 F20101123_AADHZJ kanjanarat_p_Page_127.jpg
eb8fa0b1e671f53af0642ef8952f923c
efd3be3e9d6e82ac2002cce30bf57ccbf15ec3a0
1675 F20101124_AAAANI kanjanarat_p_Page_177.txt
007fadc116526e6e68159a96314fecad
57c6c733c6bf843d8c09f3710bb7b37c2cf5de58
1110 F20101124_AAAAMT kanjanarat_p_Page_161.txt
d61937fbbd1dad3482cb9357e0d9eb9e
a76d8449b00fe083824e1204b08dfe0898373bfa
15545 F20101123_AADHYW kanjanarat_p_Page_152.QC.jpg
463f68006b18cbc2eb1a75d61daf3b5a
c6475a22254b5189150e76fa700f8bc90b749cd5
1568 F20101123_AADIEQ kanjanarat_p_Page_057.txt
7fc52dc7c739969ee7c755f06e528179
355fda2cc182de601b240c5a858152fdf314a189
30036 F20101123_AADHZK kanjanarat_p_Page_011.pro
2ff2ecbac725415b7f7ff3cf63ee0723
1a4e11875257b7ca7391b0051d581c60cbebe809
2218 F20101124_AAAANJ kanjanarat_p_Page_178.txt
a18e99cdc2bbca10cd86eab750e55e4e
146a9b1cb5ed59448a51cc608649a7ae357d78d7
1144 F20101124_AAAAMU kanjanarat_p_Page_162.txt
6cdb064c236efb691753ac382092f869
9993eaa88fb0e8625404a0f37c6037a2e1d10632
44521 F20101123_AADHYX kanjanarat_p_Page_064.pro
1f962a018e4a64fa6bf7c672df3579b3
a97f58e4dbcca92a8b95951101a855fd50becfb9
19457 F20101123_AADIFF kanjanarat_p_Page_005.jpg
35f5bd69cacff77d6dc3b69be2898af4
4088608e56b333b06d18ac69c2cd7f4c5c1cde0e
6211 F20101123_AADIER kanjanarat_p_Page_138thm.jpg
101afead21dcffd6735526598a379d89
cfc7fb278980c4ec560d5da6237d24e9cdd5e33d
1053954 F20101123_AADHZL kanjanarat_p_Page_022.tif
de63fb1d031dea98d8df0cdf93e0f12a
692b97d9856927a25df25e7c46641bfe51f7d853
2717 F20101124_AAAANK kanjanarat_p_Page_180.txt
9975b48c918c0561d426d200b83359df
ee0870d629f3f162e59fdffb7cc0009f22925722
2063 F20101124_AAAAMV kanjanarat_p_Page_163.txt
49db40bd80a86ff5291447bd6a4d6dcf
7c143c603d8f8980ee88f56e7012074e6c93fa61
70076 F20101123_AADHYY kanjanarat_p_Page_017.jpg
996ab02ecd09721520ff8234391e82ae
dd0863bb42944851bcf21a3978f62ea7f2b3df77
71210 F20101123_AADIFG kanjanarat_p_Page_006.jpg
ff8e8aea6bc804c15899fc425d382749
978c2c75c24a2006e8ecf266237a609f022db354
13792 F20101123_AADIES kanjanarat_p_Page_086.QC.jpg
dd19f909dc691979af23837ed4d95238
61439effbbbf8a82628fe0f545bfb095016fb6a7
1891 F20101123_AADHZM kanjanarat_p_Page_123.txt
2c493b6ff44ae9dfe8dc20eea9e6921d
c0c3f4407facc7714b6469d8afc4694266cb1ff3
2379 F20101124_AAAANL kanjanarat_p_Page_181.txt
b596278c2c5ac6b73e1fc40f770875b7
2bbec251e5b35764e30776759a50b51de70d9fdf
1234 F20101124_AAAAMW kanjanarat_p_Page_164.txt
e24a4c505c6d386f16a0275c82c4886c
4eec3806fa7359844c084bdc819bf5cf67501621
96364 F20101123_AADIFH kanjanarat_p_Page_007.jpg
8e1998cd3b6381343ea8f95a9f2318f5
c3dddb7f5d11e13b31bae580e815198f91ee6f6d
15120 F20101123_AADIET kanjanarat_p_Page_166.QC.jpg
16c100926cf540b436d2257b2db8f9c0
bf4218febbc48bb85a12f9d7d91820c0636efd2a
F20101123_AADHZN kanjanarat_p_Page_038.tif
b8d7b2fb894b2f65f288a79fc0b5a5a5
e4b8dc6639776a16cd2b6cf566769351518b7722
7028 F20101124_AAAAOA kanjanarat_p_Page_045thm.jpg
09308cb3857c45dfa7e96ff179e6f647
732180bdea98f4a065da0f26af8e5e8621b68902
2454 F20101124_AAAANM kanjanarat_p_Page_185.txt
7aa86ad8ba1fe726d2baf2c13465d2ad
eb77f30bf27bd73e986f4b2ee086540c83440650
1262 F20101124_AAAAMX kanjanarat_p_Page_165.txt
32e64753b2d47b4bd9176ca266a00e22
1970c3f4ae7f1006653f23abcc283e908992e6f0
67883 F20101123_AADHYZ kanjanarat_p_Page_068.jpg
e528ed80530708cb058e5caa5460a69e
d44fa1f258f6ac32cd1cd32b55f77be4588771dc
105548 F20101123_AADIFI kanjanarat_p_Page_008.jpg
c89c4f48aeda96cf497d9ce0075fa076
c4b006261f76a630f03f0bdc0b7c2d9018ae19ec
104898 F20101123_AADIEU kanjanarat_p_Page_104.jp2
5f85707244bb970c808aca02cdc7b21e
c44daea4891194dc64ed0cfad313028e5e2b1cbd
29247 F20101123_AADHZO kanjanarat_p_Page_009.jpg
be4103dd419d026261bff24d2845c44a
776835e1b1d4133b37b81e225c88d159160d7419
2274 F20101124_AAAAOB kanjanarat_p_Page_005thm.jpg
b90ed62699dc0f32f2bad641befe9473
f8cc92c3e8903d132c922981a5296d2d8b5bf737
2257 F20101124_AAAANN kanjanarat_p_Page_186.txt
30c7342a2004f4c06c4770253792f5ab
10114cd83bb1edc145f3ab6b3d1a97e7f4566d3c
1241 F20101124_AAAAMY kanjanarat_p_Page_166.txt
88a9aa2de6c66b81f905a3e6126d62a4
b9bf9dbef2e48553ecea939df9b436ea0ef5385f
77716 F20101123_AADIFJ kanjanarat_p_Page_010.jpg
eb4074441af4284d29dddd8f0c8a618f
2a908c9f66aa32c2c2fb443e8f04a5798e5b792e
6256 F20101123_AADIEV kanjanarat_p_Page_046thm.jpg
3f0fc0a67a3a6ef9a5c70ba6dd0baed8
500651b3b930e5732ecd42c3f1112eb17fbad5ec
642907 F20101123_AADHZP kanjanarat_p_Page_086.jp2
25d11e3a2b04f1920392e9dd1b255755
e4916a0e25c7baa438df92e2fbdf5760d04505ef
8532 F20101124_AAAAOC kanjanarat_p_Page_156.QC.jpg
1fcccc40ad4cfef8d649799645e912e6
3b540f9a33b6594de923c1b4b223fadbb76d0887
2444 F20101124_AAAANO kanjanarat_p_Page_187.txt
7a4c51d52a91615c141adec2c8db5efe
c540bab61b4544a5fc85e0cfe35428db80c25932
1280 F20101124_AAAAMZ kanjanarat_p_Page_167.txt
b6f80e059295cccbfc0d1dcb362afe3e
83975fced2c0ec2a69193e994688485119c769d2
82935 F20101123_AADIFK kanjanarat_p_Page_012.jpg
5f407ed9e6ce1b9bbb0d91cfd15485ba
6531ff235b2d668c8512ae66606abf567ab82464
55073 F20101123_AADIEW kanjanarat_p_Page_186.pro
88e9a9418a5c87e2cd8fd5b425cb2574
e29cf6fc2ec71d03e045eb7bb314b53dcdd09501
108890 F20101123_AADHZQ kanjanarat_p_Page_096.jp2
b888bdabb72fe87205b7cd6381b7aae9
04f35860d4045dc79a68e42aa0ebae8a11bfbba7
6744 F20101124_AAAAOD kanjanarat_p_Page_145thm.jpg
aad4b8497e212bb4d578d2fa8c5a301a
4a79c2336b777b6d9f6ae3869f7853d2428c4e96
2425 F20101124_AAAANP kanjanarat_p_Page_190.txt
2ecf9bc90ee017b5e5b21fa401fd384c
ce94968af078161399c2e85c2a239fe5cf52fc6a
102647 F20101123_AADIFL kanjanarat_p_Page_013.jpg
f7437950f0c7626a09997adff3e81e06
21bd833676c0cde43338343953bebc2942efdd83
13742 F20101123_AADIEX kanjanarat_p_Page_122.QC.jpg
48a31f08f6f5f3bf3c39153c96028b3f
226da6c9e785b0a2fb391dd933ced9c1b8582f2d
1051972 F20101123_AADHZR kanjanarat_p_Page_045.jp2
ca0e3f2d410f6df1efe54e41203f22fc
45fdf1a68f50fc81d5ba5fc1b13b60045260ce9f
70217 F20101123_AADIGA kanjanarat_p_Page_034.jpg
9324dcfec13df1b2fd9514f129de9db1
61d9a6256ad967240661b4498837cf0e2215f8b9
24116 F20101124_AAAAOE kanjanarat_p_Page_116.QC.jpg
277a07c2181c9b6e66bc65fd965dbf6e
830cc426ee3d1d8a661eeb9a68aad3c9eba02ec5
1698 F20101124_AAAANQ kanjanarat_p_Page_191.txt
0844086e80d4e37996fea61d50a87e1b
05284bdd7522f519fe553ae0945d3fe26d100ea2
56217 F20101123_AADIFM kanjanarat_p_Page_014.jpg
104e6672c274597bf685885f1e95599c
8650ca05a43f16f291d25fec434ecd7a7d8cadf0
751940 F20101123_AADIEY kanjanarat_p_Page_167.jp2
83548f2710ba665326680690bb0c00de
22c84dfb78283a38d46a9a87dbe6fc38aaba4597
60629 F20101123_AADHZS kanjanarat_p_Page_016.jpg
5adc0f06cd4dbaf4ff952d96508fdc9a
0d8e9d368451cd764b2013d9f10074ac85174462
72587 F20101123_AADIGB kanjanarat_p_Page_035.jpg
4564a48e6bab668c90e9fe6f62cb59b4
6786543e93d38398e9292fbb12b2e22f6e138a77
2251 F20101124_AAAAOF kanjanarat_p_Page_169thm.jpg
22e0dfa93caf557108909618a02b4334
c59398ca48530f0371e1edff306be6bdea14841d
1037 F20101124_AAAANR kanjanarat_p_Page_192.txt
fa0f6c75e430bc0f4cc3c57d7d1766d3
793c883269918b7c2464ecb7746a8cc34c55f7e5
69426 F20101123_AADIFN kanjanarat_p_Page_018.jpg
995bd60a55b52e8501bb79be472fe38a
b0671bb8bbc0c71db35813b1dbc5a7c8f385d9ff
224601 F20101123_AADIEZ UFE0011544_00001.mets FULL
fbb0942e27ec74d1dcfec9e66dc81b1b
e415d418b0468d8d518718b75fa26fca11ad0547
6227 F20101123_AADHZT kanjanarat_p_Page_128thm.jpg
7613c4b9e429daca5cb5f2ccd8c7aa30
33c1e8a1dfcb937bc27cb38e2631153e5b2c5886
61967 F20101123_AADIGC kanjanarat_p_Page_036.jpg
55d6fc24c93a7adb2ac611ca7b190140
1f280676b6c34fe6fc560105af7af3c27b34c57f
22377 F20101124_AAAAOG kanjanarat_p_Page_060.QC.jpg
302d1a0d734df31f92589881f0c313bb
ad8eccf809648c76f78e0b34fe0bd46a25a245be
2448 F20101124_AAAANS kanjanarat_p_Page_001thm.jpg
bdbeea593fd2a125bf2ccce09afc5d84
fc88c49a7142d7352c0fc69a017160e0341c3c8c
90561 F20101123_AADIFO kanjanarat_p_Page_020.jpg
b462bbb5a7278c5756bdd879cf2ebd38
a03c1817ecf8aa3c08fbec421fb1363fb377c2b2
26177 F20101123_AADHZU kanjanarat_p_Page_184.QC.jpg
394ef840198a78f9daebf03069975688
13e8e93c778f4bdef696c2ae65b64947711306a6
72953 F20101123_AADIGD kanjanarat_p_Page_037.jpg
e0df3a10a5f0f2a0f86f32a951d07e39
80e6ce4171476c25e1f4aaf8450dfb94827d6948
28485 F20101124_AAAAOH kanjanarat_p_Page_190.QC.jpg
ae0420bee9e3782582e649c3e285a744
633ecba8be44e60c200ee784e99803454716554d
46755 F20101123_AADIFP kanjanarat_p_Page_021.jpg
22a689b3d65fc145b7495a67f0bdc906
f009fb83e0f072b923709e33bcaeaab4fbb6f5b1
1051923 F20101123_AADHZV kanjanarat_p_Page_125.jp2
3688c662dce1339dcfd55444ce31fca9
f7039ec1d9c7a8f147886a4c3f6747c8663ac18f
74288 F20101123_AADIGE kanjanarat_p_Page_039.jpg
dde49945758ecd379fb32e4601b304f9
20a43c1a66459c9e744f0379a3d1926106f42cc5
6503 F20101124_AAAAOI kanjanarat_p_Page_066thm.jpg
ae89b0333f6cda1e44c13dbbc42471e1
45bad7f8b7ef3fd8a338f30597ef736c80284ace
20449 F20101124_AAAANT kanjanarat_p_Page_170.QC.jpg
abd5e425daf85efcabc7b3dede3809c2
938ad9faced24c3f2455242ae0ce45ad6ac825a5
61866 F20101123_AADIFQ kanjanarat_p_Page_022.jpg
3d47d2057212cf7cc37f68116ba4703c
a8002cbb42af351067e762cb765488d02f20c100
104642 F20101123_AADHZW kanjanarat_p_Page_087.jp2
e74045bef816b2d853868eee2c66eb74
6de3d1d962a23d3af88b558f0bf26ac7f7c82880
71258 F20101123_AADIGF kanjanarat_p_Page_040.jpg
09cfdb21662e2fce7c501fc94a86a44b
53553f29f0a92673b02c338c3b2fd9279ba68648
6546 F20101124_AAAAOJ kanjanarat_p_Page_182thm.jpg
3b5dd70d89ff080901a51ce1e1f32bc9
dc42fc5fc94f7ed9c77091fd2a38a07cf86ba7c7
6360 F20101124_AAAANU kanjanarat_p_Page_056thm.jpg
d220665b359dc9eec252cece15eb13cd
652b96d5b12faa28303bd943575dd14dbe36663f
67277 F20101123_AADIFR kanjanarat_p_Page_024.jpg
bb008805d54c96a73d78e121223e8c50
79f45a9defef9fed09602698a316a967225855a6
109721 F20101123_AADHZX kanjanarat_p_Page_146.jp2
c7915894705e818b53c9f017e1aee88f
261b55745928608893b8164df5b474e8d5f9f9e4
20084 F20101124_AAAAOK kanjanarat_p_Page_151.QC.jpg
432ea8cd3a59dfd32f9bdece7e090d02
b4e0d8fffe743f7c2e7eae68aee447a07352c2f2
5353 F20101124_AAAANV kanjanarat_p_Page_170thm.jpg
73460a50d6838fc243df3949d17bb715
3beb19a4c72e9ed4878f450e70349e37ea1cc283
66241 F20101123_AADIFS kanjanarat_p_Page_025.jpg
810239eec2c67e3ec2146f21a2bfedf4
b1f1e1557b5276b7bd173d48a9a08e0383cf3004
614 F20101123_AADHZY kanjanarat_p_Page_094.txt
e87d756567c9d061bc7d01c2f29b671e
646f913741a3be545e2491c264a8cce50997001e
39766 F20101123_AADIGG kanjanarat_p_Page_041.jpg
665ebce2a3d859d5afda42289de6c3de
50212c5d40f40f6fed5ea2b13b29f5b68ae8e875
6942 F20101124_AAAAOL kanjanarat_p_Page_185thm.jpg
d7699349e3bf55b3145f09025c761483
95d06eee2819b99958499162a6d84e683045b5b2
6781 F20101124_AAAANW kanjanarat_p_Page_187thm.jpg
9f3d5cfa2843769f0312d47c26e0d532
f703a5ec1d6670a39eb731303733e79ff6319612
70022 F20101123_AADIFT kanjanarat_p_Page_027.jpg
736ccfb5f0ab41ee118a76c4779eadcb
b9b37fb88c3f4801f210adbd05356e0284cb1c74
25000 F20101123_AADHZZ kanjanarat_p_Page_041.pro
e91036f3e102a02987afb696b5a68cfd
410ba8f7d277f131ece379589742cc57ce189bdc
39918 F20101123_AADIGH kanjanarat_p_Page_042.jpg
d7919e1525a9d460192616a706f69438
27c5b3d47e78f3448b6cdbe3802a9181a3c16585
15060 F20101124_AAAAPA kanjanarat_p_Page_021.QC.jpg
97bd27870b1289ec3417e122a7d078f3
9714f79eb96f8fdd419833bed524fd418d640789
7141 F20101124_AAAAOM kanjanarat_p_Page_119thm.jpg
df53bd751ba256d28673ae304d38ae47
30dc14248b291665e6c0a3ec2e16268fca6f28bb
23546 F20101124_AAAANX kanjanarat_p_Page_148.QC.jpg
8951f457add6c486e9363e8a855df9be
ab5dfb2a2c1ecf3029b8c8b5e2dfb9ddee811d5d
66323 F20101123_AADIFU kanjanarat_p_Page_028.jpg
870f465ccde0670df42c3aff6a7e1352
67240c68408edd74959b5ab49230bc5c89077594
25563 F20101123_AADIGI kanjanarat_p_Page_043.jpg
ade9fc5ab1c0bffc7dda1a9062532d20
3a88ccca4e2ebcfc942280b9b9558bc86137974b
20794 F20101124_AAAAPB kanjanarat_p_Page_075.QC.jpg
aa37107106d9f7f99ef00fade65d406c
4e18b98a168c5229aedb172b890717a0c410b8d5
15806 F20101124_AAAAON kanjanarat_p_Page_117.QC.jpg
e17547c33db7703795dabdd3c32bbb0f
1fdca7a3f051fe067993fa67d35e17c00e05961e
23697 F20101124_AAAANY kanjanarat_p_Page_145.QC.jpg
1d870c9c43d7a90c0e223d19d082e6f5
5d1a8e489904b500151e5a0953579c4c5b6346be
72916 F20101123_AADIFV kanjanarat_p_Page_029.jpg
0d78fb1992074b01536885a2468138f2
0f6e5965834d2ff53b500e143d35d69efdd64ff4
71427 F20101123_AADIGJ kanjanarat_p_Page_044.jpg
4921ad4a2cfe061aa8a2a326bae065fa
1d469b6afca064a5a35f7008e2b5213d73e8b979
5929 F20101124_AAAAPC kanjanarat_p_Page_134thm.jpg
476235ad4369bd02dd9ac7d5c6bf4ca8
0313fd9300fec4057c0323a8787b4dc3b6a7f7dc
5105 F20101124_AAAAOO kanjanarat_p_Page_014thm.jpg
3206ab1433f0559a10b5ac23a960134b
63173c0dc4aeedf6c47b0ef86f5c3be932d2bb74
3518 F20101124_AAAANZ kanjanarat_p_Page_106thm.jpg
3314f10cd3711f66274e13309e64db34
9a11dfb169f8ea995355f4d1a42981faaef00deb
71057 F20101123_AADIFW kanjanarat_p_Page_030.jpg
5eb93a3b75dd8288d792e17360e05afa
5c9898cd43a52a66c82f19437a1a3401ab21fcbd
89077 F20101123_AADIGK kanjanarat_p_Page_045.jpg
de4c54bb8499b745a6a8eb1d64723a9e
e8e1798b52ff05e5474c5c002199fbb45515c852
22075 F20101124_AAAAPD kanjanarat_p_Page_136.QC.jpg
eb0e707803c425885d3390fb9a3eae17
ea64ebcb1442686dd56c11a8932573a5ff99e47d
5212 F20101124_AAAAOP kanjanarat_p_Page_032thm.jpg
994b06166201a1674162d65fb1e58484
9885886733644c59ca0af1c525b764e11596921d
50370 F20101123_AADIFX kanjanarat_p_Page_031.jpg
d41e35fc8401bca3240a1f9f49d153b3
ca9fed922625816318380a59f9165f41808b76b7
71127 F20101123_AADIHA kanjanarat_p_Page_069.jpg
3cdfc425f9fb703b80e0a8eaa637077e
c79139f211e6d30603f6758a2107e58d88da0d1d
69356 F20101123_AADIGL kanjanarat_p_Page_046.jpg
0724c15e6e707f3452f37713871f4ce7
690baa752df22a9a8a7b820556098ee4925f2f33
23304 F20101124_AAAAPE kanjanarat_p_Page_019.QC.jpg
a7965cb7017d884111efa808335124a5
73ceb9e140b744e85497da1cd137266d0a01328d
13487 F20101124_AAAAOQ kanjanarat_p_Page_192.QC.jpg
c1eb6538d09584584e4eeb19acca9f3d
ff9ca471a738dfc527a5c6c1cccf0335254e0738
50734 F20101123_AADIFY kanjanarat_p_Page_032.jpg
918b565888b4882b92c6941b8b761c8f
f3947ccd096a2fd1933103e615c8e380878ddd3d
65241 F20101123_AADIHB kanjanarat_p_Page_070.jpg
f8334a2317154cb0645265611d2d2398
5508a4457f7a9ca5fc614fd23e04ecffbe5048c6
76430 F20101123_AADIGM kanjanarat_p_Page_048.jpg
3ce7674da3092d486c81125d7b183f13
dde8e36d82bcb65f618ac6d3196dd88022384cc9
6618 F20101124_AAAAPF kanjanarat_p_Page_035thm.jpg
03a1a550409cc33c304efd4f838e32fb
26183f96b6f243ae70a20112299abf8554e0dda8
3801 F20101124_AAAAOR kanjanarat_p_Page_162thm.jpg
ff9bbdbf14dc6588069196a0c7798e07
c23ba2bfc28259913e29291998c4564a6e61134b
72760 F20101123_AADIFZ kanjanarat_p_Page_033.jpg
41753167f2341f570a62d4d7c47ff3fe
c2523ebbc50ec4cd88fe1516cc5136fe3d8dd25b
23212 F20101123_AADIHC kanjanarat_p_Page_072.jpg
3e605eb4f1937db770e964209150f33c
74ab1c31cc1d0a7d445ce8fa50ff1f6be222b3c7
73861 F20101123_AADIGN kanjanarat_p_Page_049.jpg
572038024d3d705f60932d0a7162b4de
596e6390a88d3cafb49d4a6251f159c9aeed2201
5845 F20101124_AAAAPG kanjanarat_p_Page_010thm.jpg
3fcf32608a38f58868e659999b30119b
fca47cbba2b14c521d684b446910b865db7a46b3
7223 F20101124_AAAAOS kanjanarat_p_Page_013thm.jpg
a04cd31517d93819031fb771c218bb56
ad5093c2e80ddc2a6f80294a2ab0d8c728250cfb
62781 F20101123_AADIHD kanjanarat_p_Page_073.jpg
433b8a4393b98595ecf2929175c22a53
7fd4440f87bd5da2d0eda2d71bf0a0758ae6e4d9
68428 F20101123_AADIGO kanjanarat_p_Page_052.jpg
142ba83663492b0c89b524a034496bb6
be98ee795d9761ed0faf8d33984e09c0ee479029
25752 F20101124_AAAAPH kanjanarat_p_Page_045.QC.jpg
1efbe3a94422c249f64d21e785c8b7c4
3d1dcdc4649bcee2b72a3d6b1dfebd59cf5708a9
5818 F20101124_AAAAOT kanjanarat_p_Page_103thm.jpg
5a8549daa1d4814b212c3d4dc48d6026
210e6628c9ea8e4ed2102153aca6a1684a71414d
66416 F20101123_AADIHE kanjanarat_p_Page_074.jpg
dca78ddcef60eb56a12fa37e7f102cb0
76f018f6c23fb28e8733f37e3b44264572167a00
66773 F20101123_AADIGP kanjanarat_p_Page_053.jpg
053ff4b2209077b1c04986017b255c3c
20190f5653a70ddedb4bda023b285e3d42c03d21
6767 F20101124_AAAAPI kanjanarat_p_Page_135thm.jpg
8735f47dee8e2376936fc4c249eedd48
0c5d87c31659a988c0e734a30f5de6d00ce04da4
64264 F20101123_AADIHF kanjanarat_p_Page_075.jpg
5002632b7d58ed4ed3680e8a9e7897ef
54d7f00e782e1e8ed416f9a978be4bc94a9c8162
72386 F20101123_AADIGQ kanjanarat_p_Page_055.jpg
9028384c3fd711fe81c5556b83ea6ceb
b96d9d9277a2cf25d097dadcedd4344d8c993af4
6205 F20101124_AAAAPJ kanjanarat_p_Page_008thm.jpg
1af999ff527aa54dfb0c4abb08288f0f
d359481c6d60a7bcfb3874db2cea3f1ec8583829
13616 F20101124_AAAAOU kanjanarat_p_Page_088.QC.jpg
6c62f88461b3aba75575074f7b4e7f57
23ea30e1cb88c76264876664a8a231e15fe8ef6f
56119 F20101123_AADIHG kanjanarat_p_Page_077.jpg
b5428ccb70b06aa61ec46386231f18fd
499af201438387da7260ca8a1c368f2c00f10e7d
54540 F20101123_AADIGR kanjanarat_p_Page_057.jpg
42b66daf40433e375e73beac0e7ca804
a3d63a5833f1ef06f347b64a999b202c647be7ea
6071 F20101124_AAAAPK kanjanarat_p_Page_087thm.jpg
cfd31869c288d441a7168055d66d24ad
eaccae0b0e65a23b416a9f529eef3038d1d73ee9
25033 F20101124_AAAAOV kanjanarat_p_Page_173.QC.jpg
de81e5b54e7590eed6b3154daef9f863
7fae611680f47a41750f3de6c72598382723d959
65982 F20101123_AADIGS kanjanarat_p_Page_058.jpg
d9da85b260bc447df5ab4fc366bbfc78
71462da3caa540ef9afc58c0db2f3124f0d8d32a
6448 F20101124_AAAAPL kanjanarat_p_Page_005.QC.jpg
12189eec9d389a2333a9523c4889e30e
ca87bfdd1ddcc8f74706e534f34fc59fcb124698
23569 F20101124_AAAAOW kanjanarat_p_Page_033.QC.jpg
8ec55e950a762e1d484a60eb0e3368c8
4468fa4b0cddd98bd835ce3c721123725ab1e110
51648 F20101123_AADIHH kanjanarat_p_Page_078.jpg
fe0a068a17bdf417dec8cf9d3b1af677
51276f5d532210a3b2980289dbc19ce56ee0087a
71664 F20101123_AADIGT kanjanarat_p_Page_059.jpg
651bee198bcd9e86952a190507d34ec2
8f13affaf9b31c5d5648b7167a07dc8644fe4fba
23230 F20101124_AAAAQA kanjanarat_p_Page_039.QC.jpg
ad44c1712dddf6a381c525499af01861
a876ea2276c0902c234e85f312cf9a10213abde5
23203 F20101124_AAAAPM kanjanarat_p_Page_051.QC.jpg
60e9d720e67041a077d725c4ed54728f
9a6aa42a61b2a7cfc72d2fbb1e610dcdb7f58753
13392 F20101124_AAAAOX kanjanarat_p_Page_041.QC.jpg
3d100bde9db215fc14c47f7d1c935171
3f1d72b1f140ae568564862753ab296afdc37c05
46659 F20101123_AADIHI kanjanarat_p_Page_080.jpg
f30bdd8c1239ee6b31a4de849d4f2b4c
2a45ecd1845b8900d0ff038fd7bdb4d170d3b4a6
70703 F20101123_AADIGU kanjanarat_p_Page_061.jpg
79928335b73d26be377fdab85d132c0d
45ef795d85aff8ba5827a51c64598def80de46e5
6466 F20101124_AAAAQB kanjanarat_p_Page_173thm.jpg
3cba1c88ae05f9a75740b97172fb7ff0
1a6a93a400e7d9a6cdae4a7c1c39ad864ddba65e
23675 F20101124_AAAAPN kanjanarat_p_Page_093.QC.jpg
3acf9c837837973673968471866189b0
ff1ba3f4cff3d1d082a6c46f8b6a80818e9ec0fa
6261 F20101124_AAAAOY kanjanarat_p_Page_147thm.jpg
6fdba93865061dd1082dee2529d41da9
e2311ddf6856ef70f2d3c6fd87c7b9225776544d
73524 F20101123_AADIHJ kanjanarat_p_Page_081.jpg
132d8850ba27e726bfbc1486f13cd12e
c350a9fb013d9b37246e26c125aeaaef14dfa548
69158 F20101123_AADIGV kanjanarat_p_Page_062.jpg
f1b6fbf4ac73f2a3b8e55fe699aa5d89
13274b8deb3067fdbfa662c8e1181a590a051c62
20927 F20101124_AAAAQC kanjanarat_p_Page_095.QC.jpg
9792ceff7025559c3b88e866de345ce9
23aa734a9587784a66f061d31d7e005dc97bf454
4885 F20101124_AAAAPO kanjanarat_p_Page_157thm.jpg
3a46d15b9445582477d57e0d3f3ef701
4e866ccd853a965eeb3abffb6cf0c01802696e43
5871 F20101124_AAAAOZ kanjanarat_p_Page_007thm.jpg
3a2b3020cd5ad321ab298611f4aed8fc
8f073cf3589d4af3aaec99689d91715084623797
59504 F20101123_AADIHK kanjanarat_p_Page_082.jpg
9c462b0294e3a3468be6ed0dff342f35
aec41e6df008f1d96cb48331ae1a84c1db4640df
30427 F20101123_AADIGW kanjanarat_p_Page_063.jpg
69533437dbf764e7f8903fa3c8d40213
9165201e6490ab03c8248ca7c0b58fea67ca578c
23843 F20101124_AAAAQD kanjanarat_p_Page_050.QC.jpg
bc8b4eaaafac353df49c4f2181f13f86
7d7b4fa538cc80c923c658281adabbd67afc5dcf
6051 F20101124_AAAAPP kanjanarat_p_Page_107thm.jpg
13b2e471a8619d0d875d3ca5c207f079
4beb3332c10e9a836b6bd60812cd726ad6f5a391
72852 F20101123_AADIIA kanjanarat_p_Page_101.jpg
12468839c7320edc5e1353ddef9574a9
84d4022e4a9d28d57929e65ab50d3365cd5d1172
41212 F20101123_AADIHL kanjanarat_p_Page_083.jpg
131be3456c96054ff4a1806bcbda68ee
5c317b616e11f8ace9749435451a7791242f3430
64566 F20101123_AADIGX kanjanarat_p_Page_064.jpg
8a1f0d819b62c2fcf55d70d30d46d75c
2807914bbab388a12c42946b3b64af57b38698d9
22666 F20101124_AAAAQE kanjanarat_p_Page_065.QC.jpg
cd425e28c93a24dd9d170060360eba70
98ac7dfa41e966deab3e2082255bb528b589a54e
19158 F20101124_AAAAPQ kanjanarat_p_Page_100.QC.jpg
ce56071f27b0d960b3b76b8913b78675
e125bdf721b4bcf27f91e289a155dd6dfce4d622
67709 F20101123_AADIIB kanjanarat_p_Page_102.jpg
354e37c7c98424da3b8d5397a83fb6bc
6e6c17a3cb3fe6e116e8e65ab32b6a88ef2cf974
49361 F20101123_AADIHM kanjanarat_p_Page_085.jpg
26ac4e597259359baaf09a371a335bd9
7cfe119c8fb071f4fe53fbbfacce05aaf09464b2
73181 F20101123_AADIGY kanjanarat_p_Page_066.jpg
bdd40eec26b961b60dfccd2d5bbc67d1
f282a2f1cf8ff606fea8511684659eb452357fc0
7523 F20101124_AAAAQF kanjanarat_p_Page_001.QC.jpg
c19ee5e572a6bd4564113f671c21ff0a
bac373f6a331dc3a30b659f7830b8573b85b84e0
21806 F20101124_AAAAPR kanjanarat_p_Page_104.QC.jpg
da12af252f2968791d1623f3d6d45f8c
1743e9d3cf76e109c5e88ecbce858fbcb5084e70
65124 F20101123_AADIIC kanjanarat_p_Page_103.jpg
9bd2cbf4614988f2b8353d24abc61312
5661c273b5f01cd751894df2a88b98d77f694d85
43765 F20101123_AADIHN kanjanarat_p_Page_086.jpg
76fa0bad4ce23f0ec39e121c80868ea0
d395d6b3efcee609942c91e4f8862716f19aa5a1
73145 F20101123_AADIGZ kanjanarat_p_Page_067.jpg
ace2c38fc62781b9d485c05fab4c67cc
bb3b81425e5d77d108b3dacf6296ccbd151c43d0
7152 F20101124_AAAAQG kanjanarat_p_Page_180thm.jpg
5316958736763d80dd4e4e7810bf9349
dfece79a0934c682850c9ccd321c4adba79374bc
22243 F20101124_AAAAPS kanjanarat_p_Page_138.QC.jpg
95e309da91f0589769a3cedad084c191
703b107b3e54ad13fdeed124474e2a44164761a4
69635 F20101123_AADIID kanjanarat_p_Page_104.jpg
3cf780f89a6ac8917671bf6bb376980a
38988f5a6e3a65791ce19e6502a874d32be90921
68298 F20101123_AADIHO kanjanarat_p_Page_087.jpg
25cb8b3d9f4f210515e681f147016633
78b5cf61959cb02ef3ed61d0bd98aa7bd42fc53b
23452 F20101124_AAAAQH kanjanarat_p_Page_027.QC.jpg
ec7e6aa3bc27902eb44ee5a0af8bf8e7
a5da58bd9a54eba959d4c47b56f2c27e977e42a8
22851 F20101124_AAAAPT kanjanarat_p_Page_026.QC.jpg
cedda59b51711e1159b3a7acdccfc069
775ee961219623651e2f7537644bd595d87f7d4f
38930 F20101123_AADIIE kanjanarat_p_Page_106.jpg
0a367359ddde326a7efd82714a45eb32
0f0b6d6e0601d60bba3067f47d2e802c8c6a549b
45234 F20101123_AADIHP kanjanarat_p_Page_088.jpg
5aa91d14350d33736f3419b6c398f84c
c9581b9fbadee149e5897c8d04a971cda8d5c85d
5582 F20101124_AAAAQI kanjanarat_p_Page_022thm.jpg
d8952605aa557cf64fb5bea497657515
2033108e34c959f110a67273cf493601fda01a99
6201 F20101124_AAAAPU kanjanarat_p_Page_025thm.jpg
fe49532b63aa705d38159fd6671075a9
8567f5665f1421b9e1a147877d3e784b554e08ea
70776 F20101123_AADIIF kanjanarat_p_Page_111.jpg
d0f2fb6fedaabc7ebe3b3a6c7867e527
bf1129f4dbae8b6d622d09977ea4d70ff5899c90
69555 F20101123_AADIHQ kanjanarat_p_Page_089.jpg
115453f29305cd2446d2419a0298de6b
9352e836c09bcb4e4d93be8037a7dea7935c60ed
4783 F20101124_AAAAQJ kanjanarat_p_Page_077thm.jpg
8ec90c5643a389811e9a6f2b719587ca
1a290adf820dc3cb51536d1b6a7c09f7177d6217
61198 F20101123_AADIIG kanjanarat_p_Page_113.jpg
978a137f564e8d1906f11826e30aab42
dafb16182d8330626cc2bff54855b161ac6e00d4
42531 F20101123_AADIHR kanjanarat_p_Page_090.jpg
f6ce1a9d6ef4aa45383ade0f3846166d
57e2f354763ee3618fa1adaf18abe67c4d0232df
23467 F20101124_AAAAQK kanjanarat_p_Page_061.QC.jpg
9f7109a7d1d6acb068d4cecabb99c3ff
0a923486e9011b3e6c25715f674187f6f3bf140d
27555 F20101124_AAAAPV kanjanarat_p_Page_115.QC.jpg
039bd5d4cd9f043e9386aefdf8ed5908
36ddb68aa994a08f9aafef3bf2a48399d438de67
78515 F20101123_AADIIH kanjanarat_p_Page_114.jpg
7cc9617ad556b0c4abe0fe484e36514c
19e170261ce4a5f1213f1ad23e346f9bd9a44fbe
72755 F20101123_AADIHS kanjanarat_p_Page_092.jpg
483dfbcd69ec6b2c8aad00cb14e0b091
41f27f44846cbf2515f3cbd3677b23d487097084
12857 F20101124_AAAAQL kanjanarat_p_Page_124.QC.jpg
6c06d2a27666f818029b01defb72f27b
be6ee2f6fb20dcfa4db498a879431fe8780a9276
4000 F20101124_AAAAPW kanjanarat_p_Page_192thm.jpg
a409210bbfb2efc163bbe7e7c55106ad
8c632432ec77f5c30ad50ce290e43bb770133d6a
73342 F20101123_AADIHT kanjanarat_p_Page_093.jpg
1cf42ee7763a57fc3466ed4c23e8a524
30b3f316389c3d6b2d7a07b4fb426b3c90449cec
3338 F20101124_AAAARA kanjanarat_p_Page_002.QC.jpg
dce7338faad92ded7ab6d61b533649ed
89d11f0ed470741342c4eb64097da4cef81e668c
18384 F20101124_AAAAQM kanjanarat_p_Page_127.QC.jpg
585a27ea0b80a4bea98b18ddec13b215
7fb87dd44c0fcd8f9383a77a39a3688d50d72b9a
5763 F20101124_AAAAPX kanjanarat_p_Page_097thm.jpg
0891ea37d10e4a79915fc4160a4cd1e9
46f65015dc72702cf937bcab2a624d6e9b4d2f08
101153 F20101123_AADIII kanjanarat_p_Page_115.jpg
c9e79ad21e3ab16fd2e0eb1de2ddb297
cd03dc3c054dafed9012740db2d259f4a440f1a7
26377 F20101123_AADIHU kanjanarat_p_Page_094.jpg
08a01381c5b9cb6489950356eebe7915
84d76062d38610a21a33778ad4aa260990b8d2a0
6260 F20101124_AAAARB kanjanarat_p_Page_183thm.jpg
59c27cb5dee1bf115a16bf010551558b
3205b86aa9ef11a3894d2d77f0bb1a31bdb55e8a
19573 F20101124_AAAAQN kanjanarat_p_Page_074.QC.jpg
efd6148282903561a1b965f8cb00984e
a480db398289e54429ec8c98613d520864785f44
22549 F20101124_AAAAPY kanjanarat_p_Page_034.QC.jpg
10167f7263b6122befde188fe3dba30c
9cc3613101f6fbf0dbb0c411cbb285831d7e3047
78889 F20101123_AADIIJ kanjanarat_p_Page_116.jpg
3cddffc9993ebd5ef71f1d50bea44c39
4d0ed08869893bb703f7803f83fe36d9399af36d
64414 F20101123_AADIHV kanjanarat_p_Page_095.jpg
057a9c0e44c9ce088784e6c921281855
bd9c28d4a285dcf21768449528b4716ff56dd80b
8837 F20101124_AAAARC kanjanarat_p_Page_094.QC.jpg
55365e18efe8c07ec80b8d7fc445bf4f
2cd3aebb549bbf1040c5295be1a374113d40be3d
6516 F20101124_AAAAQO kanjanarat_p_Page_029thm.jpg
bc757bbcc402b7e084521f4c18056542
24d8d6f2a7779ec6e13f8b3ad24f20fca470c81d
3226 F20101124_AAAAPZ kanjanarat_p_Page_083thm.jpg
50c214bbdadfdb8dc136d88887787b61
fdb20f464ae1fb53f71c89b67f96605f754fc457
61125 F20101123_AADIIK kanjanarat_p_Page_118.jpg
dfe0fbce9806011fe758c970a8b6372b
97ef954aae263d8500ff3efb29163605c993776b
71982 F20101123_AADIHW kanjanarat_p_Page_096.jpg
737fd9d4922648e22952e83a6e5da7b4
c692ab12da62c7e27acf3b97f6a9e21107403293
6265 F20101124_AAAARD kanjanarat_p_Page_084thm.jpg
c76f245bc20ee1af876c10061122e345
b9dd6231616599cf173e196f22189b5093110bf4
6871 F20101124_AAAAQP kanjanarat_p_Page_114thm.jpg
5ab554b612db94a8a76d8dbe13ebf59f
d6fb38e56c62438590963c075ccfa82166e82fbe
103665 F20101123_AADIIL kanjanarat_p_Page_119.jpg
1812624265a76e8b03ff90061fff0478
13d70c2f3f5b86a6eb87b3ae39c3e1bebdf7bbd7
57777 F20101123_AADIHX kanjanarat_p_Page_098.jpg
0fefe6fba208f6a3b2caeaa671c65925
845d0e3857a8bcb17838b9cef36e60189e7891c6
73757 F20101123_AADIJA kanjanarat_p_Page_142.jpg
611288dd96462fd62d45834da431dc6f
102df664713925eb1fde2d532a15e44523873b21
9764 F20101124_AAAARE kanjanarat_p_Page_132.QC.jpg
805e62eadf7881b2ed0bc89a5237c4c4
f0097db5f363afd00f9190eadf1b0f494da09bf4
3970 F20101124_AAAAQQ kanjanarat_p_Page_124thm.jpg
4384b304897f634fc078fa7a1d58602c
caa1fe336c73999fd35153db1d003ec5c60a1c2f
61586 F20101123_AADIIM kanjanarat_p_Page_120.jpg
e4f52ae6faba9baade1e10f373aac8cf
aa9aa03cb2ff63c4fffe7f3b6a144500b5b79be5
48537 F20101123_AADIHY kanjanarat_p_Page_099.jpg
06ee06805735cb1159e1d9bf45890936
7da81cb97e2a7fa18693d0e55cef6672d63bbedf
65908 F20101123_AADIJB kanjanarat_p_Page_143.jpg
09479ddf22bb63de90fdcebc66426226
5a536684f0cdc475118bf3e6bca68c6c682493b5
10657 F20101124_AAAARF kanjanarat_p_Page_042.QC.jpg
b26fe06c86f597cf4069d20806a5b61c
bd2647fccd8e9cc8e4a64086cb4b30e19b88d8a8
6734 F20101124_AAAAQR kanjanarat_p_Page_069thm.jpg
46ff0e74297acc392f2b821844126591
0c551f62658725305c497c2d2076384d351a0757
64582 F20101123_AADIIN kanjanarat_p_Page_121.jpg
1e3b5e45066349a5f7745f381944bcfe
1a336435064d50561deb506086d1b3d674b59db5
59755 F20101123_AADIHZ kanjanarat_p_Page_100.jpg
d6ba0033635a75ccc2fbc2b9c3f05523
b302af63c93e4608174fee17e881a8fc94356666
71633 F20101123_AADIJC kanjanarat_p_Page_144.jpg
0f1c8909f04c884cbd412a3d0761b27d
17a993349c3d3e4ff114cb44cbd4b0f26df660e9
F20101124_AAAARG kanjanarat_p_Page_137thm.jpg
3f02e27810c1e93f30ab1bc157a584d2
42968debb611e261c175f13122a344fb9a7180e0
5637 F20101124_AAAAQS kanjanarat_p_Page_073thm.jpg
f46f8f114b4421dd84a235f32d8354d1
f7b9c40eec1e21d0c113251075faaffee85729c6
43790 F20101123_AADIIO kanjanarat_p_Page_122.jpg
5559e49689aee3f2081f814b61ca67a0
402cf3149afea9822d9020ead3c0f8ec63696b73
72650 F20101123_AADIJD kanjanarat_p_Page_146.jpg
42466e9b0649a951891ec55ff94efc39
86f98a5626ecda219b3b377c4f67fc9a2ebaf5f0
16669 F20101124_AAAARH kanjanarat_p_Page_032.QC.jpg
3b04cc1b78b929050952c9750736e233
f573ffb57d63352fd5c8ed2f33c59b3f937d5c6a
6086 F20101124_AAAAQT kanjanarat_p_Page_064thm.jpg
2543e99e48ab1bd877a133557819f47e
f7e44baaf92147c69f4e4ceee3f5c89576e8f457
44804 F20101123_AADIIP kanjanarat_p_Page_124.jpg
7582db9e3abd46d1a61841e271ff9abb
6fcbd4c0f3220240fa6e2b211e38f9b23ee58676
67502 F20101123_AADIJE kanjanarat_p_Page_147.jpg
5a8e4a8cec89a91941f180479e412b46
f7d451546912a45d28cd31da2925ad1a5e8abd56
6535 F20101124_AAAARI kanjanarat_p_Page_055thm.jpg
68f2e1dce56e167bcc0647a02a4c4192
25fd9624ec5460b9e7906fee2e4bc16e97045dc6
23977 F20101124_AAAAQU kanjanarat_p_Page_101.QC.jpg
691ffced2fc80ff5512b7d4378cda65d
a3633cdbd599cb31ce8aadb017cdad470cc80a78
57990 F20101123_AADIIQ kanjanarat_p_Page_125.jpg
f70f6d7b9570ef40be0ab6d3b664de28
1491dc15016b70338b532a085b3b6b619377a1fb
70552 F20101123_AADIJF kanjanarat_p_Page_148.jpg
f22e74879626921e3244afd14809c7d3
3ce74f72e38e74435329c9e36e57d2f43b56c772
5778 F20101124_AAAARJ kanjanarat_p_Page_111thm.jpg
785152b9cc6e3c72943ffd8fed5eda6c
bb029319930bd92b4e0b821c56335c96be173aa8
6136 F20101124_AAAAQV kanjanarat_p_Page_174thm.jpg
bfdcd77e04a8c3f441ccce1cfe665f33
846e9199cfdc64106b9d9cd0a28f3d6de6d46806
40290 F20101123_AADIIR kanjanarat_p_Page_126.jpg
b64655b764ab15f80af5dc6c9b5de02b
2f0c14bf343511357f55e5d001a9ec53cd2dcbb0
68009 F20101123_AADIJG kanjanarat_p_Page_149.jpg
7b138918463427fc603e9e31191ab2ee
0bfa1abe2d0d1766530a265052a4b18107c3426a
4074 F20101124_AAAARK kanjanarat_p_Page_131thm.jpg
f9ac3987cbea208c752fb5f85c307f87
fc73cfb4b506f9a630e48b53691fa951d63c445b
65442 F20101123_AADIIS kanjanarat_p_Page_128.jpg
747c8ec5ea89f0c34cc5e6b34d500bcb
6d311353368ea39ed8f45abb8296e6a629282434
40932 F20101123_AADIJH kanjanarat_p_Page_150.jpg
79cc0efdc9273754f5cf2269b93e0145
0719baf5826e14b6de3365b8bc50e0061fafbc03
22927 F20101124_AAAARL kanjanarat_p_Page_030.QC.jpg
e43481f9b3e5cdc09f3cbb9e22db922c
b420ad2fd9a699b6cd41771e6efb0ea4229257e0
6348 F20101124_AAAAQW kanjanarat_p_Page_019thm.jpg
e95b1cf065e0ebc4181e1c6ac8ec120e
77df5eff692effcf209c1230210773e8ffd449ac
63202 F20101123_AADIIT kanjanarat_p_Page_130.jpg
bbabc8bba1c5aa7187da8c9b2ab7f562
cdb05ff3a9232051bfe4820771090ac64ddffc0f
35550 F20101123_AADIJI kanjanarat_p_Page_153.jpg
32e6483c551a82b828a0f36d245f2b3d
d9a74d543381efb67095b46d574fd56932da64f4
21212 F20101124_AAAASA kanjanarat_p_Page_058.QC.jpg
6b21353a6a8861f6039e422c8ad248ff
de9cb3c47b7a58b201ad7be2328cc4f69412c06e
6101 F20101124_AAAARM kanjanarat_p_Page_149thm.jpg
c54123728d829f9f6437870eb2d12ceb
6cc2671cc6059041165723bb77b4408808df24dc
8329 F20101124_AAAAQX kanjanarat_p_Page_009.QC.jpg
74a467b99e798613df07ef583e8291ff
494daed855d61b2e8e69ed99bd937c47fd6ed50b
45392 F20101123_AADIIU kanjanarat_p_Page_131.jpg
e3bc68b2c3a9fb57ee1ae31ce10a173d
392a8e545e30ff108e7bb860441f3370614741f2
6164 F20101124_AAAASB kanjanarat_p_Page_053thm.jpg
02a33f3d92f351dd42c3d878388a5a90
da91afa7ba2d5cd7e1a6120b72cf377417373a33
6703 F20101124_AAAARN kanjanarat_p_Page_050thm.jpg
2248b7b4b69721a2fbae99532b69cca3
4e3cc587dce289501bd38ce620a6d380ecf4512e
18644 F20101124_AAAAQY kanjanarat_p_Page_091.QC.jpg
a98337600a8bf5159f540b7b9e4d67ae
b02413b325aabe78f3be9474c020518a0b97cbc3
31774 F20101123_AADIIV kanjanarat_p_Page_132.jpg
0f0b95a924e9ed3f16facb0f7d10d544
fd8ffbb43f26d97db8f50ff9137c268c1f419377
53284 F20101123_AADIJJ kanjanarat_p_Page_154.jpg
a3d262f4af54c54bee489ba60d0bee0e
7484857e0807776e2bb6212be7ff40fe18ba88b4
20284 F20101124_AAAASC kanjanarat_p_Page_121.QC.jpg
f6c6c62f78625859cda0c00f16a2e641
3469d2facef0f1077400740d08f69ee3ab50cbfe
19207 F20101124_AAAARO kanjanarat_p_Page_113.QC.jpg
fc07c3fb7238c46ae625bdd663e5ed26
ef4909ced0e18510c6f77c22e8ea2817aa033610
6253 F20101124_AAAAQZ kanjanarat_p_Page_047thm.jpg
792dce1360a7b181694bd62dd19cda9c
e55c99587093a5da6bb12228ad8f85b299a50d47
65032 F20101123_AADIIW kanjanarat_p_Page_134.jpg
acc12b6e133e21ec7426fdcd83978dc1
b1d772d6e33bfa1d66dd861c0599c0a98db71568
53452 F20101123_AADIJK kanjanarat_p_Page_155.jpg
9454ea8a75c14d0c9a9fd2e96b96776a
0f59b3f6a49132e95c93f4a19fe4cb3025102927
6701 F20101124_AAAASD kanjanarat_p_Page_049thm.jpg
1e5da304f08b071ab5a7b261503301ff
bec86fd7b242168612661b010830c6a6698685f2
7750 F20101124_AAAARP kanjanarat_p_Page_190thm.jpg
eac4e1b6b21d4fa795e3f8f6bc5b029e
23e0593fe0e0a03871abf1adb3dc41b4edac3ff2
94553 F20101123_AADIKA kanjanarat_p_Page_179.jpg
08727fba999d8a0a693d2891a7bc71fb
8a8f44cf75a90b320a678a65b4606f164658e59d
53414 F20101123_AADIJL kanjanarat_p_Page_158.jpg
4b83e117e163fdb46a31e62a171e5992
b165e8d8641797886b423e84b8770104757f094d
74535 F20101123_AADIIX kanjanarat_p_Page_135.jpg
5bd9fcc798d3bdc271bde29d556f0b95
f32534dc0355425b124c1e8301341b9f5ee5e0ec
4526 F20101124_AAAASE kanjanarat_p_Page_122thm.jpg
fca53799562f639906da8e4cf19d35ba
d768fb5c4ccc6fbc00372a4a58849e0571b0677f
6161 F20101124_AAAARQ kanjanarat_p_Page_054thm.jpg
97e961045c1a143dcf019189d09e6253
0cd132f8cf816bae21056bb9e36bdade6ecdda8a
95253 F20101123_AADIKB kanjanarat_p_Page_180.jpg
c784701237156d11743ff2a0c5a5e1eb
11d0fa4efc5149181dc579d8b54e96efdcf65c06
41471 F20101123_AADIJM kanjanarat_p_Page_159.jpg
100264d6f9c45acd7635db6fadb9f208
47ae8efb863a522329c01257e44ba9a7f6c7312d
68667 F20101123_AADIIY kanjanarat_p_Page_136.jpg
f161bb5fcc769920a0c60d93176cb731
ad3bf43fae815259419a6b59f36fde6642322902
5576 F20101124_AAAASF kanjanarat_p_Page_118thm.jpg
d975e0ef0460c1d8e228b43141a0b365
dc5751686b6c78ad0ded93141670564fb1854163
23732 F20101124_AAAARR kanjanarat_p_Page_055.QC.jpg
5834cc7ca559ce352c09fcbef960f012
27ac4fdf465bc3429d365d244c3df3f43a7fa937
82704 F20101123_AADIKC kanjanarat_p_Page_181.jpg
d2e65c0c03e77dd5bfe4191fc4633d0a
e832408055a15eb6a43ebaaed70956e3e87d8ec5
55698 F20101123_AADIJN kanjanarat_p_Page_160.jpg
04b15596994692499faee6f40248b31a
9afc1fb7dc8d44bc758425ee05941c9219709b8c
68073 F20101123_AADIIZ kanjanarat_p_Page_141.jpg
aef50a5a6245585e880a0231f65520ab
698549976ade6d3ca42da8fb26c1ee4c2a848ad5
6178 F20101124_AAAASG kanjanarat_p_Page_034thm.jpg
7511ccca1aee173f6cd311340e744e67
5240feda3c9b45895b338b73ec93dadf1e0538bd
6199 F20101124_AAAARS kanjanarat_p_Page_140thm.jpg
75a0c4b965b3f30500338b95e57a44a1
109ed93141676f37ed001df925da04cb931704c1
87410 F20101123_AADIKD kanjanarat_p_Page_182.jpg
78f3586bb88b1b8f5fa8dada0dcac3ae
dd166326a3dc66c3896b98f69f1f9eac44fe4618
61516 F20101123_AADIJO kanjanarat_p_Page_163.jpg
fb9ce37150b2360a117e1efdb3f3d8db
b5e148f72428f28309635e584580fa0521364693
5415 F20101124_AAAASH kanjanarat_p_Page_117thm.jpg
f86db9cf8e62b3e20440be979ccad95c
867774dc14bfa6fff552e6cdb4b0a10379ffce5b
17190 F20101124_AAAART kanjanarat_p_Page_077.QC.jpg
880e68dcc0e0075474754c5d8289d79a
8fde6b4ee070824ecd28dd68f732018c5602c859
77481 F20101123_AADIKE kanjanarat_p_Page_183.jpg
8c3340588296b03152983e071224259d
f941774689ef5b27645cde66fb61275279f5953a
57853 F20101123_AADIJP kanjanarat_p_Page_166.jpg
9c2e9e78207988be1349e93fd60d05d0
b82db5f414de69a0df234453c558db0bb6d94364
5417 F20101124_AAAASI kanjanarat_p_Page_130thm.jpg
bbea8d0d738aa471beeaea356fce083c
671c742dac23b4e0cbc32bd62aeaf5260b50f0c6
13510 F20101124_AAAARU kanjanarat_p_Page_079.QC.jpg
bec5bca5e273f4ca54b1dcc0abecc4fa
856dff374e5455e60e000c930eeb691552d57032
80434 F20101123_AADIKF kanjanarat_p_Page_186.jpg
f572ede1b47b18e5ecdcbb2a7c0af9b3
8050e311ee8dfd5e7c4b5ed65a463f7fc4121430
58853 F20101123_AADIJQ kanjanarat_p_Page_167.jpg
6faba6fa856841ac61d29e9be83a84e2
8915f9fd5f45086afc655746e1bcdc0f39106212
2895 F20101124_AAAASJ kanjanarat_p_Page_132thm.jpg
1a9e505658e74485c6d2d59cb8078622
cf0a7be22747ca7e9503a323f54c10a6fb0202a8
6657 F20101124_AAAARV kanjanarat_p_Page_142thm.jpg
c341ca809a475a4c71234669d773d69c
807fb87ea50a69cb2f7ba5a1963df8816adb403a
90077 F20101123_AADIKG kanjanarat_p_Page_187.jpg
55d2462adf6c2a1acb85c0cf39377c85
7114b41adfb0d0cd16439245130159f5fe938d49
26246 F20101123_AADIJR kanjanarat_p_Page_169.jpg
192feabe4382b42f5e9322e5ad822ba0
44dfb26706239ee6c4e0086774fe8b2466c88211
6317 F20101124_AAAASK kanjanarat_p_Page_065thm.jpg
5f18c15b08d54a131f9272b97db89ec3
68aa9848b9e0acfb3b9b9a31bc77c833b8a3e454
21448 F20101124_AAAARW kanjanarat_p_Page_084.QC.jpg
08dfd28261b5c6dc991a0205dd698a0e
156f08e971f189e113caee404fc82a54050a1fff
92998 F20101123_AADIKH kanjanarat_p_Page_188.jpg
8d8bc41a9800d21998042e126363eacd
32b46de8ee3a79391abffb936dae66eb10b77cec
72376 F20101123_AADIJS kanjanarat_p_Page_170.jpg
94998dde21acfa88e74a51ef98e7a64d
4a82de60ff6309ae292d431f0bb68f0ac5a3515e
15567 F20101124_AAAASL kanjanarat_p_Page_165.QC.jpg
d5a85532b28c9dab9de2dcbab4d28860
872a5c9be56ad3c6cdcf9b8cbe029d217867dbc7
85158 F20101123_AADIKI kanjanarat_p_Page_189.jpg
8019ca1195937d2bd4d66df104640e9b
6f51f62917b81ce69a29eed79249a91cf1a4a819
91165 F20101123_AADIJT kanjanarat_p_Page_172.jpg
19a7a2c9f9e4f920d7d11d5d39f37153
aaf4fd0119ec5ac8ba2b92fcd56bcf2b4623e244
20929 F20101124_AAAATA kanjanarat_p_Page_064.QC.jpg
1651b367e38b51561fdb9d83ba355dde
431e253a0b7fc60c7ad370da5907c3d22671f6c6
17507 F20101124_AAAASM kanjanarat_p_Page_006.QC.jpg
5fb8de37dbbe2a4ce4c674d5faf10d15
a3496c1af404642a16f236eb08967a8443faa8a8
3649 F20101124_AAAARX kanjanarat_p_Page_011thm.jpg
3a1332bed55d3fd7aedf67456f6b4742
174030f13cc0e9312389f7e707b24372f61ceb02
71273 F20101123_AADIKJ kanjanarat_p_Page_191.jpg
72edfbef71686da34088109c446ab20a
49156a52ff295d986fc9488e1b4a1011c732dce7
90843 F20101123_AADIJU kanjanarat_p_Page_173.jpg
48731ee677236970d21f8d94f7823f7e
d7abe51f417e0e27b5d97bcd89ae79ea183c40d9
18018 F20101124_AAAATB kanjanarat_p_Page_130.QC.jpg
febaef9ddbb527904ab62ef1bd3749e3
f8ad1fbe3b352ee229d9c4a28495444faf5d887d
3864 F20101124_AAAASN kanjanarat_p_Page_041thm.jpg
294087b3bbe418ad6ea8fb480f26e40f
5bc69a0ef56f33a577eeb24c19c8fd0431a8525f
16248 F20101124_AAAARY kanjanarat_p_Page_085.QC.jpg
e9c49f3caf9b75d854006356894a9f12
6ce99844d9b2b4748f536d7591193736b76de7b7
79896 F20101123_AADIJV kanjanarat_p_Page_174.jpg
39d22375284ae941d98390b32096f7fe
9dcd34748501f494fc3f6b3b2ec23f92087e44bf
6126 F20101124_AAAATC kanjanarat_p_Page_104thm.jpg
155e62575496b7d2ccd8eaf94351b41c
ae7c50163f1249c243403256b7b849fea21df3cb
18236 F20101124_AAAASO kanjanarat_p_Page_057.QC.jpg
bab2e233f6c293c9fda1b73f2e339762
769e7cb6ca1e75f108dee2f1fcdc497a2ceb5462
4402 F20101124_AAAARZ kanjanarat_p_Page_133thm.jpg
762a6d4b4522d4e64baa8164962a1266
fe090c149accef22f6aa6f016f2b77eb26cb8822
41840 F20101123_AADIKK kanjanarat_p_Page_192.jpg
3f4f63b123588ba1e330dc0f5ca2620c
5af734f092f691063d3999f8fd48badbde9f4060
93214 F20101123_AADIJW kanjanarat_p_Page_175.jpg
de73789cd55655673c746936fd45c267
28abb96ef85734e2b731367447d60d5b6e896cbf
4224 F20101124_AAAATD kanjanarat_p_Page_088thm.jpg
28dd234a0e290f10d90ae2b0e8995f4d
c83fc25558807f983c3153fc64acfb1a8c544704
24578 F20101124_AAAASP kanjanarat_p_Page_048.QC.jpg
c6f09e81ace396d85d176600c75ee865
084e7699465e16d5bf7971675c5f8dab47b231d4
24151 F20101123_AADIKL kanjanarat_p_Page_001.jp2
cedc71d63de879f4849ab11c823ed60c
2d246d2b0f436c9dfd4a32bd9f518e6b80f2ba6e
81803 F20101123_AADIJX kanjanarat_p_Page_176.jpg
bd5532d061a14e301e480c491288232f
f2ed1988d3e0a54693222e3aaea7e41017d80b2f
109299 F20101123_AADILA kanjanarat_p_Page_019.jp2
fa0234c9e5da1332d40fdc222d7ac302
4f663f0d39e992bcb50763453b8194e540ec7080
6305 F20101124_AAAATE kanjanarat_p_Page_176thm.jpg
c3709cc641a07a04206f749895e9c08c
4a98bc1073e9cfae5b8b543b22cce15b74fdaba0
28430 F20101124_AAAASQ kanjanarat_p_Page_013.QC.jpg
6bdcb2ba196d849019cec16a03e39d87
fb391c8bd53c34626566ab22edd920f77c3064c4
5865 F20101123_AADIKM kanjanarat_p_Page_002.jp2
cbffc84666a1b0d5e122d7e5d2208430
9eefa9f3ae021edf125c289790babd39d3c2dedf
59225 F20101123_AADIJY kanjanarat_p_Page_177.jpg
bb9bfa3bedc4406a5d21ee658a8da916
975b0e82a3afd6b2f3bc653dec87002452503796
131896 F20101123_AADILB kanjanarat_p_Page_020.jp2
ae5350291070e0eab042de22f29a986b
cda72848d32205ff8615be592bc3abd8805358c0
26962 F20101124_AAAATF kanjanarat_p_Page_180.QC.jpg
1f89c918efc66b6f622fe6991406c24a
d7c66cd851e3adeb8331919daa305ccff29f93c0
22675 F20101124_AAAASR kanjanarat_p_Page_102.QC.jpg
227cc5aa81243f1bc4fe9f0271efd2d4
1ae7cd9afd15278290013e6c4094ae0c8c905291
83237 F20101123_AADIKN kanjanarat_p_Page_004.jp2
5270d9796fa3d48791711e84abc81c07
3a20da1c595224d38903d41c75b93fa8ec7f219b
74625 F20101123_AADIJZ kanjanarat_p_Page_178.jpg
ab83bc39b80916eac4d70ef667e521b1
961fbd8f1eeaaef9e3e3aaf9b451c79371c54ebb
66258 F20101123_AADILC kanjanarat_p_Page_021.jp2
daa3155cf9bdb10b88731451d23fb883
076cfa3bc1db4497d772d7202326d375f340bba2
22312 F20101124_AAAATG kanjanarat_p_Page_174.QC.jpg
952662a70e90fa53cfff9a333a2cd3fd
2b407670ef5f008f967c8aa073861d1b428a762d
2726 F20101124_AAAASS kanjanarat_p_Page_156thm.jpg
cc8def3db478f4dade2c22806823ee1d
fdda00054584cf9d818ad458fae5cb97b6e1f191
21770 F20101123_AADIKO kanjanarat_p_Page_005.jp2
feab657bba91cfa8b517257447b4581d
897caa0be53f5c8c1eef23edb535aee428f0167b
89626 F20101123_AADILD kanjanarat_p_Page_022.jp2
9e8316fc2a7a2cdb796869baf8f64b41
3363fea323be9ff4e3fc248ea25c0c340eca5987
22528 F20101124_AAAATH kanjanarat_p_Page_067.QC.jpg
646bfe2cb57f4187ffdd306a9cf5dc64
074453ab0ac75bfbc89496d9748f11c37ccaeea9
4367 F20101124_AAAAST kanjanarat_p_Page_155thm.jpg
6e735b3ee7fd91ac90a6165a18cc6d04
31b5df560b46128c803c4a69f8d0817a38e6cda7
1051984 F20101123_AADIKP kanjanarat_p_Page_006.jp2
908650f2cd2b7452c25ebea022814422
7c681d53e04751927f35715b582d3b75af5c3442
878958 F20101123_AADILE kanjanarat_p_Page_023.jp2
2dde139126ed2d52a8807e7ae6889aa7
d6936280daeb97c48f9d0e2e1d1c9b71b489c77a
22491 F20101124_AAAATI kanjanarat_p_Page_178.QC.jpg
84c9e9cfa356eba9cab0170aff0e76d0
457b17b9bfe28d3e961c84d52e585130ebd67544
3375 F20101124_AAAASU kanjanarat_p_Page_042thm.jpg
d4f4edf36f18b756d9872467a5560f70
fb625f7b69b8583e242ddd5bca86146032d6b6d0
1051974 F20101123_AADIKQ kanjanarat_p_Page_007.jp2
25a2d25b28efe3f779a1e6113ac2c7b7
0af540f7ff7d6a0b3fcbfd711bf69bbdf73949bd
106668 F20101123_AADILF kanjanarat_p_Page_027.jp2
48d06897c50a857c16d758d940766617
af90907b6bd38f779abc939e95253551872f700c
5920 F20101124_AAAATJ kanjanarat_p_Page_075thm.jpg
c8b20b0487d697147236c2a733c288be
85b13379324393d63c94a5c3344d29f6209f396d
21869 F20101124_AAAASV kanjanarat_p_Page_025.QC.jpg
6e71275a157f5029e770ea3e8ddc3516
8735454bb8ed0c23385106e38da1eee060aadf1c
1051956 F20101123_AADIKR kanjanarat_p_Page_008.jp2
e62cf8387bc82ea81aef49bdf94133bb
cb7959db19ddd87e84dd86dcba0f64d697834288
109908 F20101123_AADILG kanjanarat_p_Page_029.jp2
20aab722c719cb855dd02502839e66d1
5f42da3822daccad7b852396abd58cf270a3b060
23651 F20101124_AAAATK kanjanarat_p_Page_092.QC.jpg
97bf840abc0a38cf2680baa8099d7a1c
7fe0ecdc038a2e23e4489ab31d842fb73c6a8200
6745 F20101124_AAAASW kanjanarat_p_Page_189thm.jpg
920724bfa6a386670df40aedce979af8
582c5af1a2d3c0b41ca01eb1e63ae960e9c9360d
1051976 F20101123_AADIKS kanjanarat_p_Page_010.jp2
65db7249134ab4a41abb4f991372bcaa
b8b655e689676780c1f7d953a9278d32db45e50a
71735 F20101123_AADILH kanjanarat_p_Page_031.jp2
53f954f761cf7f3ad435e742e7c4d79e
f5e5bfc4803149c7a376edc6cd6fd8c0e6fbeadf
6220 F20101124_AAAATL kanjanarat_p_Page_081thm.jpg
46ca328f13dbdb279555359786a7a8f4
5d07ef573ab5fc75f294f94939dd6822f4158d23
12563 F20101124_AAAASX kanjanarat_p_Page_011.QC.jpg
1a1fe0e88ca4f30971bdd312a76925e6
6d98d888cbbbe9fe00a8d0dfb1de1843092cbdda
1051909 F20101123_AADIKT kanjanarat_p_Page_011.jp2
6571e8a1be65b4936b891851ed2aba37
efaea1737709ab1adafc80faf25c60845ff6d313
73327 F20101123_AADILI kanjanarat_p_Page_032.jp2
5fe1d13c5f713361209bf747581ee3b2
61b11506cd4387c69bd1b074bc1a43af0310734e
6366 F20101124_AAAAUA kanjanarat_p_Page_096thm.jpg
bd2ae6230681e5465bb2ee3217122658
e9040bcff4d987eabf25d22f2412399461c6fd53
15588 F20101124_AAAATM kanjanarat_p_Page_177.QC.jpg
821e1aa0ef0089085bb11732f53c991d
d6c9d3a65e78067c2864455d8422705349138748
1051971 F20101123_AADIKU kanjanarat_p_Page_012.jp2
e2acb49c27535981573e204996c501c2
e51cb0d01a4a46cb68ca746a7a486ceb5acfdda0
110200 F20101123_AADILJ kanjanarat_p_Page_033.jp2
12c5a5677f6941bf854077f1295fbac3
29ee2a4acfb50df980093cff505197163b52a0fa
4699 F20101124_AAAAUB kanjanarat_p_Page_085thm.jpg
8f27dd445975ad583854d6e0b2468ab7
5d25472585d62671e8336f879c274cb20d8a27b8
5629 F20101124_AAAATN kanjanarat_p_Page_091thm.jpg
01fcf865f40934dd39b1f003a47ece12
731edbe9f63183f410fb999956cebcb8868bedbd
6419 F20101124_AAAASY kanjanarat_p_Page_037thm.jpg
8443efa60e5fb83bcc72fac47f52bb13
08b4c07a6f2943ef87126c68d2f8f460912c796d
1051985 F20101123_AADIKV kanjanarat_p_Page_013.jp2
6a0df6fb6afcece7ddb8d41c80e9ecac
96d90d484fa4bfa15b82e3eecce5d66def502c8a
107479 F20101123_AADILK kanjanarat_p_Page_034.jp2
956c941825f1d11094578767c1ca8e22
139f0074edaa102a075916c484f82686f218f713
13504 F20101124_AAAAUC kanjanarat_p_Page_131.QC.jpg
f9dca737601cad95eeab18d7758426d1
44209e6d01eb0ff5c5dacb02b3b1a9eac625eb4e
21714 F20101124_AAAATO kanjanarat_p_Page_149.QC.jpg
3316102171809714f7429595e4ba940b
53840383607a4c34bf13304120792f68e41849ae
5897 F20101124_AAAASZ kanjanarat_p_Page_036thm.jpg
be51a8a74a45ccd5cf12840da47467e8
c2656b73b6854c8d620fab50678df620bfb8d3ae
83061 F20101123_AADIKW kanjanarat_p_Page_014.jp2
b2e089e9fe16b6ced5ee2bda9abd9d4b
fbdc2fdcf27a239e184a4292f0323f2ef80b955c
20003 F20101124_AAAAUD kanjanarat_p_Page_191.QC.jpg
d8ac53f61cf705e5a734a672b8938cce
2fbb1d9ab4ce0d076a21ab5c1df39d5f2f4ea828
19626 F20101124_AAAATP kanjanarat_p_Page_082.QC.jpg
8c40c540155ec36bfdc155a2992227f1
edc4e42d9003e7315ae215faece53daa96c8d41f
88811 F20101123_AADIKX kanjanarat_p_Page_016.jp2
b9989637e741108a26b1ab98fb1759e0
90f3c9d465de416605b1461f2bd0b0f39e7fe50a
107414 F20101123_AADIMA kanjanarat_p_Page_056.jp2
5fa344af94cb37971888a2bd6f2e5813
50df178ab7f49191dc25ff666f7c4b6455e78e19
110096 F20101123_AADILL kanjanarat_p_Page_035.jp2
1e6640e2a6ac2b4b15ea5d55bd780f3e
4b4696fab42537f497a61083a7fd2b24931b54fe
6392 F20101124_AAAAUE kanjanarat_p_Page_068thm.jpg
9d142172874ce2c4ae463faaf9020fe8
d9b69e11b6b736327ebcfd046a698efa5bd955c3
23939 F20101124_AAAATQ kanjanarat_p_Page_066.QC.jpg
b066dc505854d24c53e2ee006a865c55
9bcf2a2930136a61e18fff3c528cc04228faa6f0
107547 F20101123_AADIKY kanjanarat_p_Page_017.jp2
9302afb63182f7615677e87aee2d5869
52e05f9f944c3721b3a380670890ba13022330a8
73964 F20101123_AADIMB kanjanarat_p_Page_057.jp2
ff6dd281db27f74140a3dcc721b7c15c
529b5e9286c393390d241ccd4515d17ab8f6ec87
95194 F20101123_AADILM kanjanarat_p_Page_036.jp2
177df4c88cf98b9ea4c5c06e98374cb5
68a5b30f936dd88b1039b994485046d680d32522
6523 F20101124_AAAAUF kanjanarat_p_Page_112thm.jpg
fc63cc5b6dafb65b0a7087033519a76e
27c4177fbdca37cd26660011624a6c46f2689757
3491 F20101124_AAAATR kanjanarat_p_Page_159thm.jpg
37a58019d78f4eca8e122a3903606df7
bd207069aac2588d743ed7c4074a12c000d0e5b2
104785 F20101123_AADIKZ kanjanarat_p_Page_018.jp2
e7eb4f49261aa9c449cfc5eb9e68b6f0
fe65721e620ee7a905950811d481639d563a4e18
107029 F20101123_AADIMC kanjanarat_p_Page_060.jp2
c9e93c86f4400b8f7e9dc274b8b523d3
1f5f9170f85681e9a9727f63d8e07d6d21135298
112248 F20101123_AADILN kanjanarat_p_Page_039.jp2
0115f66b1cee192b1a52ac213551c034
9cebeb7e6cb2e94c757b735f54d3f94ba2cbfddb
6313 F20101124_AAAAUG kanjanarat_p_Page_123thm.jpg
ac6dc94bf78b1962cd16ac78332e594f
13f1baf51dcaaccc58419f53f61c16a1d7a21238
F20101124_AAAATS kanjanarat_p_Page_067thm.jpg
1be08cd4677b076fe15bdae389939096
98db9922974cd4dd5096c3ca987d3701949c83ab
109211 F20101123_AADIMD kanjanarat_p_Page_061.jp2
19db26ad42106e5a9faaf06f51b4a085
5f7c6c1f6682d4f491fcc87c6e084ff25511b99f
106007 F20101123_AADILO kanjanarat_p_Page_040.jp2
2f9615a06eb32e7d4cbf38bfa9fd3c3e
2758e8f8b1fb84d9c923abe38cac2f46924ac5f4
23357 F20101124_AAAAUH kanjanarat_p_Page_176.QC.jpg
0f03776ce28ff080765285b104fd1159
fb7b5d9320cc57bc00d084675771aa2463f141ca
12596 F20101124_AAAATT kanjanarat_p_Page_080.QC.jpg
3a47dbc18336c9cad8b5033e54ebde80
2f1bfdf9d692b0352b1a5c96f64b0d602483903d
105109 F20101123_AADIME kanjanarat_p_Page_062.jp2
21938a4441ad1e0db7893f6fa00e2ab4
cee416f8107c33f1bd547de2d9275abe39ac55b8
58110 F20101123_AADILP kanjanarat_p_Page_041.jp2
357ced701122bfc713488da9cf3ae63e
1270f3e5ac56200fc415084a03287ca670170757
21034 F20101124_AAAAUI kanjanarat_p_Page_143.QC.jpg
aeae37012bc9b0dc2bb7cd261e95afb4
9e59f4304c823901a5e0ee02c903040955cbdc64
23673 F20101124_AAAATU kanjanarat_p_Page_029.QC.jpg
b34f334ad2c3733ace3238872eb1bbe9
b21d282b94d221bbe8237c589864afaf236c303e
97188 F20101123_AADIMF kanjanarat_p_Page_064.jp2
d5aa1f2d5d3102e42301c3d1ba134144
60c266cafbfb912dc6fcfbeb287b9af7d08bfa30
875217 F20101123_AADILQ kanjanarat_p_Page_042.jp2
a935abb89834d2aa14c66998064355c5
9e4801aff499722a26b60988a1aba129f15c9741
6556 F20101124_AAAAUJ kanjanarat_p_Page_139thm.jpg
d8cb23a8e3faecd3c1e0567236d17da3
8ff16f35106757ccdf4c9943fd7181ad1a06d828
5694 F20101124_AAAATV kanjanarat_p_Page_074thm.jpg
14a0dfa0b1ed6d06da32d3ede5882a9f
eb430e42c3a1e7006941d6eea2ce656aaa53d81b
106902 F20101123_AADIMG kanjanarat_p_Page_065.jp2
f2ce37f72d4406bfa261aae8e661f5c7
8b4fd9283bd483bc9d13e8686223a975f34b65b4
103686 F20101123_AADILR kanjanarat_p_Page_046.jp2
9f3275972656b0e336900dafe3c880a2
667bf430219fc3375423c3137a5a5494ad9eb48b
5946 F20101124_AAAAUK kanjanarat_p_Page_028thm.jpg
40599f6e2724841ee0f9db37870164ab
0d5548a131f8388d6c2c8af0295a1ed1784410f4
7361 F20101124_AAAATW kanjanarat_p_Page_072.QC.jpg
f887a75e0a79a26a0c75baee78e76d3b
99ea9fb2473bad46cf17eaa23bc328b41846e80c
110799 F20101123_AADIMH kanjanarat_p_Page_066.jp2
6daf5a923fcd3fd44d13d14e8c2463f0
79fe08a308e49ee9cc26c60a6bcb5644f1d3748a
116621 F20101123_AADILS kanjanarat_p_Page_048.jp2
6db176f1eebd8e56aaa2d5f56bc5147c
22be5ce4e39855b9eff0876d2e43bfe25b14c45d
24505 F20101124_AAAAUL kanjanarat_p_Page_137.QC.jpg
5a5c8d80b103c4d0fb935532971af905
ab8ed984151d3f65c0dc42c573d00e375af53371
6450 F20101124_AAAATX kanjanarat_p_Page_061thm.jpg
6093809d5d72ddb4b709927f90b2b8c2
f4c336863bf57adc031c26bfdbb1101676c5c47b
109061 F20101123_AADIMI kanjanarat_p_Page_067.jp2
9f173ce655a309cdd7968151d545bce8
8bc74cfd2dff50b41626f7d27b2c82f1ce22e917
113192 F20101123_AADILT kanjanarat_p_Page_049.jp2
8adcf513e230a259ede4a1cd525ed84d
3024163c2e8b93e8e6cd36e4a313529a7379a172
6277 F20101124_AAAAVA kanjanarat_p_Page_175thm.jpg
b801bbbd2f108c56d6b6496eb1122589
ed937da70f98e3d0819d87a80822507e98fecbdc
22596 F20101124_AAAAUM kanjanarat_p_Page_089.QC.jpg
afa1514b0f93bf520a99bd5b9fb40756
3ea052ad5c014763e538365fe4c2d493addfe6c3
6395 F20101124_AAAATY kanjanarat_p_Page_060thm.jpg
ad2e8e6411d1a26eb08b6016f3fe288c
872626fcb897465ae2819094c074a4967681254d
109184 F20101123_AADIMJ kanjanarat_p_Page_069.jp2
78c5b346e886d9d4816f243be9ef7403
2cd0c62195629824b321e732170eb90e7b07e5ab
111388 F20101123_AADILU kanjanarat_p_Page_050.jp2
2f582337951074894fd37ec07f7971f5
46670192fa15475565df15d64ba01a1986dee532
6039 F20101124_AAAAVB kanjanarat_p_Page_095thm.jpg
8f7228c7142b9879168068e02babf7c0
5e37675576b5f70451d053320500228d17caf253
9951 F20101124_AAAAUN kanjanarat_p_Page_063.QC.jpg
e489958d9741a20f049edf14da36b129
d3b08d347d7cbe7283297ccb498e31b7dfd16203
99876 F20101123_AADIMK kanjanarat_p_Page_070.jp2
d3ec1fed6976ce3c2bcd0618651a478b
486d3ba44f9ee197e91b23ac09a9fa4e5539125f
112305 F20101123_AADILV kanjanarat_p_Page_051.jp2
d75b14c7c71d0ccd1bdf7d2cc65f7a8b
e20e902d1821399fd1de63fe6c21bc7196ec8ecf
18829 F20101124_AAAAVC kanjanarat_p_Page_015.QC.jpg
90be26bc8afb709d4779d137e58fc790
0af4fe7785eb80cf75890c4a3aa49d2e30a8846e
4215 F20101124_AAAAUO kanjanarat_p_Page_090thm.jpg
601d7ef328a6773953aaf1f2dec9b0ba
b611cfe07e27053f9c06e46374f902721481eac8
6368 F20101124_AAAATZ kanjanarat_p_Page_059thm.jpg
47986a6a5f63f998c7ad069cfa9444f8
e2c977f9552d71275c9fc8a745912003f673bd0c
91555 F20101123_AADIML kanjanarat_p_Page_071.jp2
5657dca10f6b8b78f36e0ce2b1b6d07d
584526724b4384d1be90dcb6c0a851fae8a9956e
106679 F20101123_AADILW kanjanarat_p_Page_052.jp2
434235d8f58610c0f7afd645fe5d2beb
e495f87d1d78cb13ff34d150f2b6d4482cd84bd5
20968 F20101124_AAAAVD kanjanarat_p_Page_070.QC.jpg
4e3d2826aba86b21cfc41b2b6303246d
4f2009d8599612042256b1e9f197b3321b65c580
5384 F20101124_AAAAUP kanjanarat_p_Page_191thm.jpg
9e5e5a4847f7381cee9aa983bc96f9dd
443c187b139b81de5db54bc53fe34e0af8872dca
111354 F20101123_AADINA kanjanarat_p_Page_093.jp2
ea62b1c38b21a8f52ef753fda0aac6c6
749b2c563d30d47055b7d28e31c2ebc21b0f0bb5
102315 F20101123_AADILX kanjanarat_p_Page_053.jp2
b3fb0cd0561533bf26c2503ee83f6a3e
e285a00ce1fdeac219fb357e4c7e1dec8449a637
7067 F20101124_AAAAVE kanjanarat_p_Page_179thm.jpg
53a2b68db9c1a0404eff3639e483ba1c
d1cd1524e547f3640573e057d3e4de8671e5de1b
20282 F20101124_AAAAUQ kanjanarat_p_Page_097.QC.jpg
263bc80235e0175198192826c21548bf
a8131d4f4601c64412738065cc5045f48f57c771
34025 F20101123_AADINB kanjanarat_p_Page_094.jp2
4af3e7eb15a7fc8d8589b6f36913a4b5
49d11dbaca40c551e643c693629489ec51e1d6c9
27752 F20101123_AADIMM kanjanarat_p_Page_072.jp2
cfe9071327c9932cac2cba5e4127899b
2de8f957dea5fbfbada9c4d12f3e2ba20093cd9f
101266 F20101123_AADILY kanjanarat_p_Page_054.jp2
a4cd41f68c6287a8b6e9c54de844d3f5
80834b3aafb6f77e864ce98ddf86b83f0efaa181
6538 F20101124_AAAAVF kanjanarat_p_Page_051thm.jpg
99c74c211163017e6a67c05bab7c92fc
fbb5906a21c5efeeb1cabb286cbccb2ab5ac5b37
24149 F20101124_AAAAUR kanjanarat_p_Page_114.QC.jpg
f4693287372f8e32b50bebc5df4f2daa
6cffa0cf48b68a701a8822e40893c1cb90e4c2fe
97154 F20101123_AADINC kanjanarat_p_Page_095.jp2
38df3c087890d9323a25ebe8212a97dd
832030a81ae0e2f08321909f30d5e15e496697f7
1051948 F20101123_AADIMN kanjanarat_p_Page_074.jp2
a67a1f00faa2afc34b51672c452a5a6e
e13df8ae54f326b56c2a4eb27a3e490ed5f0cccd
112545 F20101123_AADILZ kanjanarat_p_Page_055.jp2
65629a55a9c64c56ea29e04af78a6c7b
b207fbe00957d8b9fcc5fb549a9ec8b246321c1f
13843 F20101124_AAAAVG kanjanarat_p_Page_161.QC.jpg
c3277e95aa2a4378afb01128a464e71d
9063c1e941aa06e175f09dae1741a00e6bd124c3
26036 F20101124_AAAAUS kanjanarat_p_Page_188.QC.jpg
a0887c66be5e20dc752f67ebc5822a46
8a93352aeaead688a0e4a799c6a073f3e9b3fa8f
91207 F20101123_AADIND kanjanarat_p_Page_097.jp2
a023a72d4f6f8943381d51f1cf3e254b
324ca5535ca75dab3d1950c9fe213c56e0824827
895035 F20101123_AADIMO kanjanarat_p_Page_075.jp2
49e2eac340509a1314bf4482f01c08b4
2e3321cd544736a50162e5d1a2826ee940f5a020
F20101124_AAAAVH kanjanarat_p_Page_167thm.jpg
5907f4fcea3e0ddcde998806557e64d7
be04d9a1d4d9fdbd29124474caf451666125d545
18730 F20101124_AAAAUT kanjanarat_p_Page_004.QC.jpg
e47616feaf70149346dd13155f90333a
f586f0375a63789e5a6de63ef301e84464088511
87950 F20101123_AADINE kanjanarat_p_Page_098.jp2
848c844ea39ed94dd7b9920dd16fe0a7
71e805beb01d2d3325801683fb31c4ff1855a63e
88303 F20101123_AADIMP kanjanarat_p_Page_076.jp2
9c3b5a1923df11d7651b4608b530c841
de591d8ba7c5b90c1b8009cc2ec5615b714f510c
20251 F20101124_AAAAVI kanjanarat_p_Page_071.QC.jpg
c759d4f3a2827b2dfed4102051b88b5f
602b0dfdedfa51347fb90f2c261aa469387d33d5
23073 F20101124_AAAAUU kanjanarat_p_Page_144.QC.jpg
2884f8a407399c5c876a9c35a0bef5cb
76535fb3daa704f061b3d571d26c02e564310207
68881 F20101123_AADINF kanjanarat_p_Page_099.jp2
b34be08e57ada28bf22c3dac0f29b269
4ea0c7279b06188136d7a0d94f419e8beaacb392
73084 F20101123_AADIMQ kanjanarat_p_Page_077.jp2
d3fdb0b3e5e9c6c618b9848afe551f6b
403c65e7df29360f327b6a328a9400dfb8984dc8
11998 F20101124_AAAAVJ kanjanarat_p_Page_106.QC.jpg
fe15b7eae6dd8a89124ae13be945336d
f4299574af8cb5ddd8a306621b378042c5281436
21757 F20101124_AAAAUV kanjanarat_p_Page_147.QC.jpg
36a1ce481ade4e315917752d434409c5
c1d51bf0c69dd6719a5a88ec958e93eb8c360520
96509 F20101123_AADING kanjanarat_p_Page_103.jp2
da530f645d6b062ce5a8331f0c200a87
54a887f965e048bf48f46f197701544f4b235058
75472 F20101123_AADIMR kanjanarat_p_Page_078.jp2
1fba46357988399e79696d27e28ece91
215ccd4a2f7c8944e18cbe251ea5d8bdae0da782
19946 F20101124_AAAAVK kanjanarat_p_Page_120.QC.jpg
f0cf50d88a441c1d132801ea92b97cf0
731b2c6bb5ba986389ff9fd378f639376650d1fa
22726 F20101124_AAAAUW kanjanarat_p_Page_056.QC.jpg
ad62295ae319100fa6a230c9f5dc2f75
7b743df248e4f3106e6d264f00e90a3a02949a4f
89261 F20101123_AADINH kanjanarat_p_Page_105.jp2
fc9c1ac60fab07efed0cbb1b62375f8f
83ca8635c46a081648196e4779a2fe3887280a8f
1051977 F20101123_AADIMS kanjanarat_p_Page_079.jp2
a12884d1dd0d930cc0f443f2c18ea4b2
519d188559791855952445c798949608c5d719be
7049 F20101124_AAAAVL kanjanarat_p_Page_043.QC.jpg
758f8957b04769809417c19103f89e7a
c5a63a4d849509886152d353a4dbc74725e6030b
23250 F20101124_AAAAUX kanjanarat_p_Page_038.QC.jpg
bf4271d56ad9f8f501116dc011082278
461293b7e7611719210a7316d919ad66c2dfaeb0
449832 F20101123_AADINI kanjanarat_p_Page_106.jp2
469d84b71743631922024455d31f7cda
b933945223a353fda36e881305c49b6695620395
991310 F20101123_AADIMT kanjanarat_p_Page_080.jp2
769797948eabe5b25463793308b66a55
fa4b93483c9da90e552d3fd67d17808592355cbe
6457 F20101124_AAAAWA kanjanarat_p_Page_101thm.jpg
a663fd012910d141557cea3a12fec6b7
d0f21096f5992f1c0c2a6cbe658526366bf1b6ae
6354 F20101124_AAAAVM kanjanarat_p_Page_144thm.jpg
282c94e7537b0b758b2c7e79b0852e73
7219020ad1903c753bb5e599a6a79c66f2b81024
20900 F20101124_AAAAUY kanjanarat_p_Page_103.QC.jpg
cc92a47aadaf99fde8b95f272443ae07
311218585079d454596f09b13b3fc50a243af589
98603 F20101123_AADINJ kanjanarat_p_Page_107.jp2
4265717b328c1f4b1071d76f5faee976
d0eb40d9b312d63107d6c9984969002f3ba160cb
1014417 F20101123_AADIMU kanjanarat_p_Page_081.jp2
91b0f235ec10fce0c06521045159381d
5d6ebd2099e481d8118b5a5c1691268b11aeea81
22392 F20101124_AAAAWB kanjanarat_p_Page_010.QC.jpg
a10e8803eed79804fb5b4fed62b05b6f
860cf2c6937d2356ec758dd7de625e1dff994978
3728 F20101124_AAAAVN kanjanarat_p_Page_080thm.jpg
416eb013c1c2791eca40c4195852bbc7
a5251f8e800322eccb375165886a54806a4d8368
25854 F20101124_AAAAUZ kanjanarat_p_Page_008.QC.jpg
1443f8e8876f61ecbc45ce0b32a6989b
5a1c3981fff46fcc7c79ae7f1bb6e00a26eddedc
624297 F20101123_AADINK kanjanarat_p_Page_109.jp2
3ca73547c10ac315fd60f4164d62ba68
d0712370edb4b69ea72bebe5dc3cf6928ebf4cea
661243 F20101123_AADIMV kanjanarat_p_Page_088.jp2
cdb90dd8ef9e9cb260987af6d3a78389
2e521fce014f7238d90c357ddd6112cdc9488951
17049 F20101124_AAAAWC kanjanarat_p_Page_108.QC.jpg
7bbd9f6667e54cdd2a274c5873f5e9cf
d0c774c49657b8cd1d675ee5cfdb828cc675523d
4684 F20101124_AAAAVO kanjanarat_p_Page_099thm.jpg
79b4cc21c88007980ec71b0c7f9ea45a
f5771ba0a1a8f8022a5cd1f3b9ef82f04500ddfa
824049 F20101123_AADINL kanjanarat_p_Page_110.jp2
4644362b8b4c7ce7eddf400a20ae009d
475a428e3625b2edc7ae633e28506451f346bbfc
105936 F20101123_AADIMW kanjanarat_p_Page_089.jp2
3834630e27f0e5bac3445a10c1c2a80f
a6cfa2ce4ff7aedc7ca14ee15001e737ab4c4745
23346 F20101124_AAAAWD kanjanarat_p_Page_069.QC.jpg
316a72f8ac2d3f8ea56d6962ebcf4b32
4c9cb7d0748ab91d5f0a1601574c13714a712628
22591 F20101124_AAAAVP kanjanarat_p_Page_012.QC.jpg
ab4b93ccb393df51ea4889a7d1baf028
3ebfa92a0e05a1b423552376b51d64b7576faf08
953434 F20101123_AADINM kanjanarat_p_Page_111.jp2
4cbfb6e1d513d2409e28fb290f3aa751
9fbbf9e61df6c606e0cff84dfcc5f1bf3722c85f
661084 F20101123_AADIMX kanjanarat_p_Page_090.jp2
06934468a5cdcae959afe32dcee2d8d2
e78675149abc51d06f95aebaaa517b623600de56
925390 F20101123_AADIOA kanjanarat_p_Page_131.jp2
1d568d79a30d955a65237838c3010e73
af03a558bd5d5595bd852bedab5c5c530e6a4010
4667 F20101124_AAAAWE kanjanarat_p_Page_006thm.jpg
e6240b861aa02230cac019a35cce78bd
d2dde22c112260b431a1c77bc8f9be8b4e448dea
23211 F20101124_AAAAVQ kanjanarat_p_Page_017.QC.jpg
2ce02b2f2319ff2c049344f14ff2de8a
81b523103572471e10d36a832f2a23bc6f434737
862492 F20101123_AADIMY kanjanarat_p_Page_091.jp2
4994d0f89934ba36f5a997b6554a1525
0b7cbbfd897cb860d955147075d85afa1fc41d72
623626 F20101123_AADIOB kanjanarat_p_Page_132.jp2
0188d26a9f557eda933e67adcc80c2bb
464ab819562943bfb244a12f5379eb2362071ba9
22081 F20101124_AAAAWF kanjanarat_p_Page_046.QC.jpg
1d0495664f992c369170752eb52d0360
02756417812384fc914f900707cd13a878f91d9d
25087 F20101124_AAAAVR kanjanarat_p_Page_187.QC.jpg
0301d31a830b7c74cb3f0d24b2fd3e75
453d3a49f08d406fc9a4427be86d885009982c87
1051912 F20101123_AADINN kanjanarat_p_Page_112.jp2
580b321c3581c81e809072e8081a3931
22d617ade769e20cd5994596a2a410d4d14eff7e
111415 F20101123_AADIMZ kanjanarat_p_Page_092.jp2
0b0a82a0672ddd6a0c0808bd6086f062
6456c41208727eb923773693fcfb81ad4ee928e3
66780 F20101123_AADIOC kanjanarat_p_Page_133.jp2
3f832b2b2dc6ad7d8ef69b0725d982e2
83036da80d34cc91f4de3d7c677614e5eb5e5f4a
24870 F20101124_AAAAWG kanjanarat_p_Page_181.QC.jpg
234fed684b383b6042cd29f2075e0272
dfc5c6b49e713a482d44eff087d9e43a11d4c933
21173 F20101124_AAAAVS kanjanarat_p_Page_111.QC.jpg
1c240e038ceaec3aabac459c4b006042
a4a101c45928d22ec4331f916f5b818a9ae7e6a7
946930 F20101123_AADINO kanjanarat_p_Page_113.jp2
b64cb2415838b0200ada3efc7a7d50f5
1821b6f829a9deb5c144f86787abffa9a80fb14a
97313 F20101123_AADIOD kanjanarat_p_Page_134.jp2
098fe570340cdf650be4644eca5d875e
a4f6f7d106a458b0a9c4f4e02f73b324b7419941
4772 F20101124_AAAAWH kanjanarat_p_Page_125thm.jpg
5016b0b625ee75c1cc5d1d2d8077fb53
b013ab09a324bc368ba4702bf1b6ce614a7a8978
5684 F20101124_AAAAVT kanjanarat_p_Page_100thm.jpg
e11d106b445307c7176da9588ad721e9
16154ce668b10e8c39ae8e9a36a0a2618290dcf9
1051962 F20101123_AADINP kanjanarat_p_Page_114.jp2
96072d57e3cb8af5aacc106871cc02ba
440f48f8df3208323fd2095310252222cee15da8
111806 F20101123_AADIOE kanjanarat_p_Page_135.jp2
93f265e636d618547bbb1bc72d89d8b8
5d45c4afb25cb48138643fd31138ae7e3e6c0c03
7314 F20101124_AAAAWI kanjanarat_p_Page_169.QC.jpg
e06215f53fb21fb2ccc129e0872e66ea
7531a8530fe2917dd0444fa16ff370a1ad42d59b
5979 F20101124_AAAAVU kanjanarat_p_Page_121thm.jpg
2c5e9fa3b101f7b8138bc9b473e6e195
17ae4cab48b4ca8568792ed4a5fd4f7567537851
1051951 F20101123_AADINQ kanjanarat_p_Page_116.jp2
36e9ebbb332a4a8f2de10be19c93f241
3fb590fde49c53540718bb5add2eeae48b1438a6
112480 F20101123_AADIOF kanjanarat_p_Page_137.jp2
3be5e795936c2f99c87aca0a8899d74c
30ff5f7fd4410e591ad94d6cac2d6365e3d6518c
13185 F20101124_AAAAWJ kanjanarat_p_Page_150.QC.jpg
c123ed2bba7537b25502e4f1dbf7012a
c27eaf777f94742bcf09997ea2518c93d8dd70ef
7045 F20101124_AAAAVV kanjanarat_p_Page_020thm.jpg
64410ca3cbb9efbe9f911d892cda82a4
04d8262cf7d77aca20ed7fb304b3e1702af09cb4
748261 F20101123_AADINR kanjanarat_p_Page_117.jp2
b79f40eadf85c48a4432acdfd6003f7e
3d0521c5dc8f86adc85942a776419e3746cb04cc
104217 F20101123_AADIOG kanjanarat_p_Page_138.jp2
4821e2ce7c3c551a9fa46192f93b1ac2
a4dab0994948bd805e574d6fb84fdfb6e9c12d00
15088 F20101124_AAAAWK kanjanarat_p_Page_158.QC.jpg
89f52d2ab9b9388559ebfc0962a8b62c
62cafe5f243cc953851ea5ac973a1789649ae48b
23401 F20101124_AAAAVW kanjanarat_p_Page_040.QC.jpg
c206172cdd590e3e831ae046fb8b6816
2c59a7831e5e24a448634e3bf4b9fc2723612a64
F20101123_AADINS kanjanarat_p_Page_119.jp2
52757d40a0d27df0b5084250c776faef
d5133f69ce4456ca1c75b3b7f68105aaa3dc80e9
108067 F20101123_AADIOH kanjanarat_p_Page_139.jp2
4b62c37b022b1d593ea3777fba3226ec
532bdee6fdbf6316376f9d7144eccb2f3daa38e4
6471 F20101124_AAAAWL kanjanarat_p_Page_172thm.jpg
8837a9eca9351ed5951176acf6be645b
c14b734da7afff3d778a95507dd04c128a7ab109
18780 F20101124_AAAAVX kanjanarat_p_Page_076.QC.jpg
efbbfc1c2ae9970190c6cb971e37adc7
d0e24c8e08ad23d517efeba6860465ba11b223b6
900264 F20101123_AADINT kanjanarat_p_Page_120.jp2
4c25e1e9e3d7dbe9ae2d620a0ad8c1ee
fbc5a175fc437422e0d13248047d7049ef13ef8d
103621 F20101123_AADIOI kanjanarat_p_Page_141.jp2
939c7d4a8e084b92b81a581eb0227178
d849d057074d68c85a83978196b491a5b5c9b372
5656 F20101124_AAAAXA kanjanarat_p_Page_151thm.jpg
15fef5241594d578e7e54c321545853a
3fad15b8c4e073ad767754a00d7d6854f3aeed6e
4072 F20101124_AAAAWM kanjanarat_p_Page_160thm.jpg
ff3f054db5ae753636aa8dc25b72809b
fb8cc71d63d5c381dfb037bc38ce1d866a92814e
6482 F20101124_AAAAVY kanjanarat_p_Page_026thm.jpg
1e08a9d5b9d76c6b7feae1c6cc9c92b1
acf5eb8571db5761dc90ce087d89a3e4fca15e02
112474 F20101123_AADIOJ kanjanarat_p_Page_142.jp2
cd694d7671b099b0d7d73ded072d0939
6f0da12df529cf7c62d2074ba83fdb6b0354d548
1035269 F20101123_AADINU kanjanarat_p_Page_121.jp2
1824330edd4797cad9d3e2175011b12a
4e7d7335cf944532b642e257e091125f54c31b16
F20101124_AAAAXB kanjanarat_p_Page_049.QC.jpg
4a958dfd7610395cadac3c82eeee807d
0bdf971a6f06659f364f1716e58df1b7aacd2988
21338 F20101124_AAAAWN kanjanarat_p_Page_047.QC.jpg
caefba692da7888efca4753bf6278b84
4e591054b6e8c45d57d7a8405d9b4f192a03edd1
6423 F20101124_AAAAVZ kanjanarat_p_Page_186thm.jpg
da0f4a1a8f38ad5e5a83272f412fc5e0
77be41a4c6116e25c7c09c5a566969851c82f1f4
99530 F20101123_AADIOK kanjanarat_p_Page_143.jp2
c785ff654284076d54fa91195851d667
898eb557d2d6f5c4d766e720769f68193ab28b5d
102057 F20101123_AADINV kanjanarat_p_Page_123.jp2
47cfd946d550c47b9122f74f654c0a95
1dd48de0d3f88a616284e878fb89a359fc60669d
21591 F20101124_AAAAXC kanjanarat_p_Page_134.QC.jpg
f7890d54013be13aab3d77ccf6fa5e6d
fab383c5edbabddd01c770c054cf0ed5332bc136
6537 F20101124_AAAAWO kanjanarat_p_Page_148thm.jpg
c477aec740e3e7050d5bd4f7b2c48f85
3b98be4cf001394cdeb3d6b2169aea6d4ba009b4
110278 F20101123_AADIOL kanjanarat_p_Page_144.jp2
3d4155cbe7d4bd523b752afe87fffb8c
eb49b6bf020c9aaacac786e9c13e2b38741695a7
575767 F20101123_AADINW kanjanarat_p_Page_124.jp2
cf35762c163d1d098e157dfac1317c4f
bdd5b0856308667dae763ee19df9eac9985ceaa6
24353 F20101124_AAAAXD kanjanarat_p_Page_135.QC.jpg
72a4b9d1f90e2677f06e0ece5e0df31b
b03a812204de804a563ecd665fab696d15c9bdef
23425 F20101124_AAAAWP kanjanarat_p_Page_096.QC.jpg
7cbcde11516e6b88258af72333a2b21a
03b1253455de048966e3b6760af75f1c90786aed
732199 F20101123_AADIPA kanjanarat_p_Page_168.jp2
8794225ea024e91bde44adc947822b88
674721cb35a5edd4620dd3c32f2302f45b1fd3c8
110867 F20101123_AADIOM kanjanarat_p_Page_145.jp2
2f4da9eca52ec1a6550c9207a6a4ce45
443ad303f93a51fa5a872c6f2f3b6944f59f8d95
816538 F20101123_AADINX kanjanarat_p_Page_126.jp2
b80f312c8935a8158ec981f8906d7090
f9f840de50c7bc25ab6df5b841aded482fdf4bd6
6415 F20101124_AAAAXE kanjanarat_p_Page_044thm.jpg
9899abe2625e7d8ba98a712f9a3862c5
1cb96e417254c5b96538bbbfe0d69fe2fa0d5a4a
5226 F20101124_AAAAWQ kanjanarat_p_Page_110thm.jpg
a874d2ff85e89a6027d5297242228ddf
abed2784efc5bcdde9b01b9a7c45aa5764c128f8
258842 F20101123_AADIPB kanjanarat_p_Page_169.jp2
540e77bde856f9ced8ca0bbd6d639017
fd63e6ef98f6f1ded947d60a890ef949b9f6c54b
107086 F20101123_AADION kanjanarat_p_Page_148.jp2
117d49c30aad18bdfe42a317c1211e4e
e9e593d6e356facaa0a3fa4209b3d34612085038
972384 F20101123_AADINY kanjanarat_p_Page_128.jp2
c2c734b90538105eb4b7687928daeba3
51b2714086d70b011f1714a296cf3c79f095b665
19491 F20101124_AAAAXF kanjanarat_p_Page_110.QC.jpg
491a5d3ae96568aa5ce9d55a513a983c
1daa5ee7ac7205b094dd9834a503e42bb3ea3ac1
11868 F20101124_AAAAWR kanjanarat_p_Page_126.QC.jpg
c6bd20fe3bffbe65a622e5e8bd0bc4c8
e898651a353807104a21b9d254e492f74ec73a35
107444 F20101123_AADIPC kanjanarat_p_Page_170.jp2
f19978550c30573a9c8c3033a8d5feae
6555d1521db6bcbd3494e40ae40066a92766eb2e
844662 F20101123_AADINZ kanjanarat_p_Page_130.jp2
f7d1d6ea4a251a68c77a6e1aa295d1ee
c2e3009992ba49b334f50feb4a0ce02dc1301ece
4236 F20101124_AAAAXG kanjanarat_p_Page_163thm.jpg
f42c7f2437d840f331185385640bb368
f698f18d8cd1185b5ff96b464bf54e5066201940
5899 F20101124_AAAAWS kanjanarat_p_Page_120thm.jpg
dd3ef82ffc131cd25a9f75e98951d7a2
01d26303b121b41bc3e28cf8018fa87f40485b7e
159130 F20101123_AADIPD kanjanarat_p_Page_171.jp2
cbf6a7ddabef1b761a37ee4bce408b53
7af1274603af8c94e6e024c587d9bbb73183307f
103156 F20101123_AADIOO kanjanarat_p_Page_149.jp2
e8a3480560f92aa785125c5a9ee4494b
67320b3baeed1823584381b68fbb930a341a1657
24345 F20101124_AAAAXH kanjanarat_p_Page_175.QC.jpg
24377f217b57f7908fd9d757dc547e43
6a1b0ea9840c35579efad110193af879f82c388d
22244 F20101124_AAAAWT kanjanarat_p_Page_068.QC.jpg
5c000ceb608b5fe670ab492cde6ad9f3
2ec9b30ff4e7d6516d90a7688ac84274300c85d8
135588 F20101123_AADIPE kanjanarat_p_Page_172.jp2
6bd8602cad7aab9cbabcb979fc7d6790
b20be72be90c8443d1c4710f062e4041ba2cf40d
57931 F20101123_AADIOP kanjanarat_p_Page_150.jp2
7f55e10e0421312489572e0c0915d0ae
945613f0404eb081d3aa6c0f5bf2b161e1a50b23
25271604 F20101124_AAAAAA kanjanarat_p_Page_116.tif
f762657d02d51af19f40a9c2b4a4545a
b3e8f12a9f7655e214a9547e6731ba857c1b967a
21923 F20101124_AAAAXI kanjanarat_p_Page_087.QC.jpg
4cf822f89f7b147b524c4220df96831e
37df9ea59361665702b4d89baba8103b4f8e672a
3165 F20101124_AAAAWU kanjanarat_p_Page_153thm.jpg
47b1ae76ecf0b1b8578ce227fce86552
9ae28cf2d1f5ebd2218b6c586ec3383beee6a094
136070 F20101123_AADIPF kanjanarat_p_Page_173.jp2
e155c426a420463a4ccd789b6d1435ba
364e878a536722091c5fdb94da40ae58c1c87c7e
67757 F20101123_AADIOQ kanjanarat_p_Page_152.jp2
f6e4f35ef8267371d37899dbba26a77a
c903fb546dfdb9addc370b9906110a7d8aece1ba
F20101124_AAAAAB kanjanarat_p_Page_117.tif
f91a2960e314ac0a8c9c89b67f7b69b0
9aa131976ccb9388b11b989ff24ba59f2b0de2e9
3727 F20101124_AAAAXJ kanjanarat_p_Page_003.QC.jpg
568455821afcc9dc01240b21f9d9fcd8
9dddcfe85ee3163f0f12f03366b2639432365d6e
6273 F20101124_AAAAWV kanjanarat_p_Page_141thm.jpg
a9a219cc34d57629feb3f49e96af5b4b
f7472a88f080a69376dd177d4d6a2ffe80491c31
118073 F20101123_AADIPG kanjanarat_p_Page_174.jp2
807e5f0d98581df35648296853d8d3c9
95498f2ca3c5210fa5bee35c2e427962f8319037
1051969 F20101123_AADIOR kanjanarat_p_Page_155.jp2
90120a534ba419305cbef01d03472859
f7fd6370a9c6f0fea9cc6cb0cdf11b31f6a3b3d5
F20101124_AAAAAC kanjanarat_p_Page_118.tif
9dfed9697e0c197a8a5b422e6a4d72ef
7970476f4c98ff5241ffcd5f236378adec32fa7c
10853 F20101124_AAAAXK kanjanarat_p_Page_153.QC.jpg
dc2c7cf4c8f049914e3635c216385a40
754928348cfee4e5f097f3572acc2fe801fcf686
27425 F20101124_AAAAWW kanjanarat_p_Page_119.QC.jpg
830b99c360acfe4988552741754026a6
05d439d33346e5846a40e7a95318dc924b74fa51
134294 F20101123_AADIPH kanjanarat_p_Page_175.jp2
dd411962238d5f78375bfc7f5156b8e2
ebc3dd74e6020277b28dc152e7d11ba17306e1be
1051936 F20101123_AADIOS kanjanarat_p_Page_157.jp2
6179eea8a502141b4b4ce3a898439d39
6226ff947240b1f1385303578f45d44efc0bab32
F20101124_AAAAAD kanjanarat_p_Page_122.tif
b4f677852efc201a957968644bd8e197
799d467496b5a031284a7453759ae2b0cb912d1c
3463 F20101124_AAAAXL kanjanarat_p_Page_126thm.jpg
141f9a49aa250f41dae7fbc9f303ce6a
f8a5295df2af1f49f47e2661d22e95c5ccbe9df2
16939 F20101124_AAAAWX kanjanarat_p_Page_125.QC.jpg
fda537a5bd76f8af24853a5c5a8e20d8
aa83025729784b46d7bf5fefbc79fd90d84a8774
122452 F20101123_AADIPI kanjanarat_p_Page_176.jp2
0d360f11066563bd18ac8acfe8e82bba
f458532f0d6616f0083d4280130065031a2e0306
669766 F20101123_AADIOT kanjanarat_p_Page_158.jp2
961f48d6bb0b04d45943f573980a6eb2
c06c6e33758122984706b97da6ec6b0ef3fb62c0
F20101124_AAAAAE kanjanarat_p_Page_123.tif
9484674a03967ff7559bc4f1470b8265
d0fba77c6f39d1f4d732f6a6e86a50d08b9b0933
3808 F20101124_AAAAYA kanjanarat_p_Page_161thm.jpg
6ee9695fdd3e18cdb2048f0b9be1c226
62f4d955789b31d1d621320e75a6f31e75a9996e
6452 F20101124_AAAAXM kanjanarat_p_Page_024thm.jpg
6af273bcdcc1b0b995d673eb32b222c5
ca8bdcbb5e7239b43736bbc78af0bfa6f18b637a
12320 F20101124_AAAAWY kanjanarat_p_Page_083.QC.jpg
6b67d6017b48d908573349dd65512eb3
7e4161ea9069dc5fa75bb3526c994372387cf6fe
84842 F20101123_AADIPJ kanjanarat_p_Page_177.jp2
91d366ccb4f76f74b5dbe7d76d8da90b
7a2967cb7a52bae4d00a28b6f8aad5dbb02300bc
495110 F20101123_AADIOU kanjanarat_p_Page_159.jp2
67fbb816f4da3c1f8f7c73ce968217ac
e30b41db23b42961eeafb23989259eb130bb4098
F20101124_AAAAAF kanjanarat_p_Page_124.tif
8101e37aad0a36769bb6eb805af6cfd4
91060903bc02e2bcf99fe73d17e236db4219a515
4116 F20101124_AAAAYB kanjanarat_p_Page_158thm.jpg
ac1dc6cc8971c0fae032b4f2bd85ad0e
fec77f9d8336139dc467d796c1fafcd017bf7b05
21843 F20101124_AAAAXN kanjanarat_p_Page_054.QC.jpg
5cbe33b4475885f803b84df762febf7d
24fe490ce876d8a0f63d47a1ff0824f339e96d9c
5480 F20101124_AAAAWZ kanjanarat_p_Page_016thm.jpg
bd21965a3f6c0c08cacc32b4cb533ff9
e0475d5a828d386acce8f9c17ae6c3702ecdb882
114138 F20101123_AADIPK kanjanarat_p_Page_178.jp2
14e4f960604468ce9c610fb96cc30e3e
10b1668aadd13c499a18aab7958e1c14ae54951f
707212 F20101123_AADIOV kanjanarat_p_Page_160.jp2
347d0ef93b6f0dae0bfc7c412636c2a5
2acc9a0c65f38638dbdd27082d3376c8f6a8cdf0
19558 F20101124_AAAAYC kanjanarat_p_Page_105.QC.jpg
994d60180b31c91f0c93beb0377f96eb
2f879df12248d6ad7755e2248bd674520ab8a7a4
4198 F20101124_AAAAXO kanjanarat_p_Page_177thm.jpg
ab6b8fb6f79cc891a4f8ddeec27ccf39
f03ef1a4f82bf534d6c75b1418c460b0b989bef0
141447 F20101123_AADIPL kanjanarat_p_Page_179.jp2
33a6eb8b66cde48f0a0851c805816aa1
36316eb98d0e4def0de88a372ab5b5613be1e03e
636196 F20101123_AADIOW kanjanarat_p_Page_161.jp2
a283eef9b666d877d66a3b404c8c85fa
74139aaab7b0408aad6e3e9b2ffa5c84b5355051
25265604 F20101124_AAAAAG kanjanarat_p_Page_125.tif
80716332937661d3b3953cf8723942e8
12ad58412e1bb85a18c0b9d34dd2cc4d22971a05
15569 F20101124_AAAAYD kanjanarat_p_Page_164.QC.jpg
27c10d7e382d22dd78c480e8c934b8a8
5ca035c9c72820b10e1f5847a587d1ad35ca43b3
4276 F20101124_AAAAXP kanjanarat_p_Page_086thm.jpg
00016a6bad17e7feb4158e3e07de31b7
243fedf5572d63fe02f8e35c1e688ef42019e5ef
140930 F20101123_AADIPM kanjanarat_p_Page_180.jp2
80b4685320603bae75a0108105c68ea9
374fd3a692063b368e5bf9448731f2c78a7d9302
782351 F20101123_AADIOX kanjanarat_p_Page_163.jp2
f0038c65149739de60ef986b1bad2a65
c28ef29f5d3b55b787be3d76cfd0551144f48a67
F20101124_AAAAAH kanjanarat_p_Page_126.tif
9a68572253b6f8570a45c444f71d3f21
b34d0f11275cf1467144f71a6e28bb7d6fa8199d
F20101123_AADIQA kanjanarat_p_Page_007.tif
64ebbc7a8a038dc8bb627557a0513a32
bbad2e18be33a51a2e3866e7b8f0db32f0f9ee86
291358 F20101124_AAAAYE UFE0011544_00001.xml
90189d87c556c3b6fcaa74c241cbe627
c7aaa5f530ca9d948c40e76b3980923ef92c2c6a
F20101124_AAAAXQ kanjanarat_p_Page_128.QC.jpg
774c83b0e670eecc03206dff4117d51f
f8bca00ebe0106592ff4ef822a7b746b9077c357
125809 F20101123_AADIPN kanjanarat_p_Page_181.jp2
3ebe5db1e0c41be239c29a3d4c58b1a9
42a758d26363d45122cf4c51754a28060e9462d8
726621 F20101123_AADIOY kanjanarat_p_Page_164.jp2
3a2d4936a76b60efea6d1e3b48f1169d
1ef40435bb486bac7943b1c936a3780463271cc8
F20101124_AAAAAI kanjanarat_p_Page_128.tif
c324e766d3389fd0ae189583e030e029
e31d731529565edaafffa9a3fe7e7e69a31bc1d8
F20101123_AADIQB kanjanarat_p_Page_008.tif
34711cf0279695e3331d81d81dafccff
5c789506d019d72c455c5894e8f3a83f7eb50d55
1557 F20101124_AAAAYF kanjanarat_p_Page_003thm.jpg
92cd0eea2ce1c11da21e3502d83644e5
5040690d50a7fd7b2d2f8b94fd508a3416a99f8b
24840 F20101124_AAAAXR kanjanarat_p_Page_182.QC.jpg
ccd98bc9ff1e4649ec794501cda58558
afc29fee109756606ba9766d99ced6d1c08f8af9
126365 F20101123_AADIPO kanjanarat_p_Page_182.jp2
b12f0d9e54cfa6616c3b868375031bf0
c63369c76ead6ed4cabbc5713484f5ce8ba03517
731580 F20101123_AADIOZ kanjanarat_p_Page_165.jp2
d54a894c86e84ab4462a3e725d4ad396
6cf5143d90540f039283d9ef1ba15addfc1cfb01
F20101124_AAAAAJ kanjanarat_p_Page_129.tif
b5e2c6e742fc5923c9b4998040f9872d
b57df68c4d1b8426c2103e9bec63b5ce4ccfbd48
F20101123_AADIQC kanjanarat_p_Page_009.tif
90c7591952e725a9f8b0a1b72c1cb456
6171161ce7ef8dd993f6016f7b475918a6c23f3b
19757 F20101124_AAAAYG kanjanarat_p_Page_022.QC.jpg
7c51d017342cb292f930ce73fdbbcf72
e93ebc70705852059e41cda1f31000f4843af8d2
6306 F20101124_AAAAXS kanjanarat_p_Page_062thm.jpg
66b9cb06730cfc9e9bc026aec73b06e7
b1871b129da93139ce682d803856a94754cc70ed
F20101124_AAAAAK kanjanarat_p_Page_130.tif
a12b99db41b46f73a4cbf61badd7e0e3
8aaff7fea7119e8d2ea51bef7fef1a8236007e3c
F20101123_AADIQD kanjanarat_p_Page_010.tif
6db1777fb67859b46dd4ddd71443558e
52b59c53da18db83ed3cb8383228d127b5b60075
5745 F20101124_AAAAYH kanjanarat_p_Page_023thm.jpg
e806afc05ab38ac16606e55792e0dd0f
899d96ee452642181722343c09cd2f22984ac9c5
27905 F20101124_AAAAXT kanjanarat_p_Page_171.QC.jpg
20f4b33a937bc5e9a714e6f87ee3a962
3085882d129366ef13c7a5c9bbabef646a51b407
112933 F20101123_AADIPP kanjanarat_p_Page_183.jp2
bfc4efdbd41982bf3967a430baf3d149
8443982c21c5e0f238683853db53934946236dd7
F20101124_AAAABA kanjanarat_p_Page_150.tif
53bbbf8fa990b11cd021824a7f63d966
7eff89c0cd1eb10f11da2bdb092a99232020f8fd
F20101124_AAAAAL kanjanarat_p_Page_131.tif
b11c206487022a8bff6d49a8310da70d
cdf364c5ebc0cc2d51c6cf8f42262185c59e631e
F20101123_AADIQE kanjanarat_p_Page_011.tif
86efc54dcec1e844062f1a9e1eb8f812
0fc5eb8db33ed3644b42ab57fbcd13ac7f2cd2c3
23909 F20101124_AAAAYI kanjanarat_p_Page_035.QC.jpg
87d90b43eb0a97cc953b0a8719e411ed
809a37eb51eaeb0533fde12b68c138d3bf13663c
11974 F20101124_AAAAXU kanjanarat_p_Page_159.QC.jpg
9da87b3c33e815c9ea32375549247e26
65f4f53d946a771d06a0d181f94871f4dbbe874c
F20101123_AADIPQ kanjanarat_p_Page_184.jp2
9f525eba7742fc827582691a3ec03597
87d12cc3dd75d9b0e8e9bf9227cc29d74529a57e
F20101124_AAAABB kanjanarat_p_Page_151.tif
3e35c25421d554a1843637ce9b21f4c0
7674bec303f8e350d64a4b545d4a0e616c919737
F20101124_AAAAAM kanjanarat_p_Page_132.tif
b90f9715f24ab65cb1d2e6601f735e9d
86ec47885c2fa90a2ef781766b89f049c3f40e04
F20101123_AADIQF kanjanarat_p_Page_012.tif
88ee88be7011cf7f2caf2f6fb3acc7f1
0880cd6c5d09ba27e29dc1ef1f9c2bd18366f3a9
F20101124_AAAAYJ kanjanarat_p_Page_037.QC.jpg
bc91e5c13726b3e9ab7aeace29fdeb88
060348b5c075633702949a1dc22ca282344a8835
6498 F20101124_AAAAXV kanjanarat_p_Page_030thm.jpg
82b3d2fe7ebf5e800a568cceee7dcdcb
ba30b50075efeb4262305f72048894016c98f06f
118427 F20101123_AADIPR kanjanarat_p_Page_186.jp2
a22583a43c5b0ccb7e970bb0d88649ee
95e9289987307a58d79cc3bda3f190c4dcd99f7d
F20101124_AAAABC kanjanarat_p_Page_152.tif
8a3f4af3ededfa977296ad829ebba1f2
40f95e9151eca33bed828659df366f3ef1f56cfe
F20101124_AAAAAN kanjanarat_p_Page_134.tif
bf7163659ac2f9e54859881e96fb6e65
59e1e7f39437098253e1b85fa4453d1b07ebc515
F20101123_AADIQG kanjanarat_p_Page_014.tif
868a959e41c2fbf54e19f5763980c47d
243b83cfc59efc6b082faacc282e9959ec211330
21308 F20101124_AAAAYK kanjanarat_p_Page_053.QC.jpg
6a75a23c075cb9e41ef34f80f31c0509
2aa81d61e203838fa44564db1b46053929df5cd5
15178 F20101124_AAAAXW kanjanarat_p_Page_160.QC.jpg
39baf3ec3704f519d6316f889cb54b6c
99a6c31af4b9239c0899f86d00111dab3623fa18
128352 F20101123_AADIPS kanjanarat_p_Page_187.jp2
cf93f3e0e89b0fb88a8bdee9f41068b9
9d499dbef9a927d6263ce939a0b033670eea41ac
F20101124_AAAABD kanjanarat_p_Page_153.tif
a977b44b4604e62295f699b0077fc471
34495e2147a314973fb86d4e03622be768a9daf9
F20101124_AAAAAO kanjanarat_p_Page_135.tif
cc27479cefe0f30c9c1aac0fb2386d21
fde9c1a281ed187cb0306c2dd1ba0af0bbe781ac
F20101123_AADIQH kanjanarat_p_Page_015.tif
50c5c6af4e3fa1afc6d1278d4f21ca33
7811592240841048103389d93c78215a2c486c4e
3089 F20101124_AAAAYL kanjanarat_p_Page_063thm.jpg
eb8be8e3030504a977e3dc90b493a306
8370c5b9085614cf132c6509eab695e2e7687f03
6329 F20101124_AAAAXX kanjanarat_p_Page_136thm.jpg
3f74f2169c51b58b296bff018abe36fb
82c6817216a2137929284522c44e670cc2fca6fe
129833 F20101123_AADIPT kanjanarat_p_Page_189.jp2
4e97d84dd72451e19503ed459b9baa95
06fce102ad8923be15dce046a4598ffb0761bc25
F20101124_AAAABE kanjanarat_p_Page_154.tif
eb4888f77af891c20995c59d57213270
2c9ec08d2ffd198c2f8e1041cb92a021e6906bac
F20101124_AAAAAP kanjanarat_p_Page_136.tif
0e75876ff0b0832f1b8e8748c01ae536
c9aafb4e347cc27a76463640b545c28a6d882478
F20101123_AADIQI kanjanarat_p_Page_016.tif
a2a5384d37457893a7d3a80bca557897
6efe798b1e881d303e0192435d647fb4898d5f08
5605 F20101124_AAAAYM kanjanarat_p_Page_071thm.jpg
b06539d829b9dc48c65e1211cfbde59e
fd0970f80cfacb7067b46a237bdc336fdb3e2724
5261 F20101124_AAAAXY kanjanarat_p_Page_057thm.jpg
19d4c41946cfa36ef4ca24cb4e85e3a6
3af372badfc2345c82b6da4c8765fd681908ca56
1051982 F20101123_AADIPU kanjanarat_p_Page_190.jp2
d51e9206589f5e0b300c7ea63ecd5ce8
db94e5769835a74a43ae58b46d478055449c19f6
F20101124_AAAABF kanjanarat_p_Page_155.tif
1421cf7683b2355381da6be0d3122605
9e2b90285d012ceb3391d8d50087d9c7c1ed4d57
F20101124_AAAAAQ kanjanarat_p_Page_139.tif
5d10125c490047482c88785cb7ba99d8
183d3767d07c94bac91a3590016c842105abf0eb
F20101123_AADIQJ kanjanarat_p_Page_017.tif
3632a3758845eff38667f1a7088cfbc9
3e7caa7a5b33370ebf57b21f668a2ee18a463760
F20101124_AAAAYN kanjanarat_p_Page_113thm.jpg
cab24f8fdfa1990ff0ce3a0a6f5ea820
d6d8e54fa3113be8af85be9352d737427c9149fd
21564 F20101124_AAAAXZ kanjanarat_p_Page_112.QC.jpg
6ff45511b91ab8b74c1f86f4cbf28723
2702d6114bbba406552cfa5f0d5acc43a9054668
952686 F20101123_AADIPV kanjanarat_p_Page_191.jp2
ca4f66d73c41dbc16d59b836010272ff
fdd0023e0eeb98d98bf5914f326a654e9a249db3
F20101124_AAAABG kanjanarat_p_Page_156.tif
b55405bc9c0b96c4bb40eec7882c9cc7
b3b90346826b430b400a3eebd5bb7b158edaab1a
F20101124_AAAAAR kanjanarat_p_Page_140.tif
23e3a5c654ef14a561652fa7e470a71f
5128b1874cad944a15b56ab0e314166496d6da7d
F20101123_AADIQK kanjanarat_p_Page_018.tif
d933c02084974f160a8fe4256bb516d1
277ac67d4ffeaa82ca4bbcdc30389afa1d4eeace
7020 F20101124_AAAAYO kanjanarat_p_Page_116thm.jpg
c61c9b31f9d20fa32f45b0042c83736a
43814fc33a31c79ffd717ebd73084fb228fe4152
F20101123_AADIPW kanjanarat_p_Page_001.tif
cfdb44a70defe2462c31451f60d19aca
a6ece3157257b3c42b33ca880ecee64e91698e05
F20101124_AAAAAS kanjanarat_p_Page_141.tif
7a73f94163de689ffbbd49357f4f86da
88f947c404ce18300e577c5b0e83a295e00f72c3
F20101123_AADIQL kanjanarat_p_Page_019.tif
e06c54312b9c7a0f6f3dd41e7b4911df
da842d0aaa2557b68391cd9af87a96541776e6a2
5405 F20101124_AAAAYP kanjanarat_p_Page_127thm.jpg
ba953dfd38516cbca79a94eef352e1bc
5929420d4cb580bc763039e76767d5a70e913f9e
F20101123_AADIPX kanjanarat_p_Page_002.tif
66d98abf5e6286fe1a41a0a76d9d7c49
0ae5733c08a005d4f091862cb65910aaa03e50bc
F20101124_AAAABH kanjanarat_p_Page_157.tif
71062853730050851f266d9beb063331
a1ac2f1d59570a7e3ce69c9cbcef40b807c1e2d2
F20101123_AADIRA kanjanarat_p_Page_039.tif
ab99f412cc232d0fe024141f3130e639
665b58a3ffed2ed76285251f21d3b342d48b3d56
F20101124_AAAAAT kanjanarat_p_Page_143.tif
ae0e188deb91870c275cff37e60ff5fa
cda8e7e995a0f9bbd4291d2a214f7f53db1933bc
F20101123_AADIQM kanjanarat_p_Page_020.tif
c38d24da175f247ed0bb8ee468fdef50
04339a6bb7adb47627ae9709afe977835acbb60a
6463 F20101124_AAAAYQ kanjanarat_p_Page_129thm.jpg
01323e4aebd042729fc9d2a07f07adba
b8261803a41717cfc288dad77f3eb97bbc800561
F20101123_AADIPY kanjanarat_p_Page_004.tif
5541661d30a145216e1c4668e6a35aa4
0d6706ba5e5b4d5d7ea18f8c8268fd99307fc553
F20101124_AAAABI kanjanarat_p_Page_158.tif
cc0d88fb0cc5d2531489d52a9b24681f
e40e6106201ccf9c3fcae07bee0a5cf019a538a3
F20101123_AADIRB kanjanarat_p_Page_040.tif
f33478fb0ea143df36369a05060349b9
05bcbec4975ccd9b14c20c055438034a32f9e58c
F20101124_AAAAAU kanjanarat_p_Page_144.tif
eef07c191fe0b05dac52fb5f42d7e463
6fbc07cd864fe643c7f7971222c5dddf802a3d9b
F20101123_AADIQN kanjanarat_p_Page_023.tif
7aa198f8eb34c8fd81e9e7ae6160b40b
c5abcf7f082e5dcecbba767a2aa33ad21d9007ca
23326 F20101124_AAAAYR kanjanarat_p_Page_139.QC.jpg
8bfc313fe28c96583d4ed510886fdbae
e866cbccb65b54aea52dce4ec91de00614f6658f
F20101123_AADIPZ kanjanarat_p_Page_005.tif
2c3e2476a15071151a40812bbe7e3f5e
7fad9e079edb5930fbd70e9d1bb213a04e58c973
F20101124_AAAABJ kanjanarat_p_Page_159.tif
8ab8cb978802ca8e664e583be2d8adac
c67ecaa34a297f01c4a371a1a8ec103c90aef784
F20101123_AADIRC kanjanarat_p_Page_041.tif
726824beebf2c2107f0cd607c38a128f
c75e37b5c7407157a7b799996017223f556622ed
F20101124_AAAAAV kanjanarat_p_Page_145.tif
0d165e3af20dcb644fb4377f8d3e670b
df91156acfe48061b9301e6ee015003ff4fc69e9
F20101123_AADIQO kanjanarat_p_Page_024.tif
4ae0d7dd88225fb030e23fa5b26fd5b3
b0a8bbc29b31bdba27c88575e20401f5e0bd04af
4693 F20101124_AAAAYS kanjanarat_p_Page_152thm.jpg
aadbf31a43f7311b6df25554713939ae
5afa0a1c336f9e75a076a7c6692d15071a355610
F20101124_AAAABK kanjanarat_p_Page_160.tif
b4e1c1c2c6f987e6ff9a0c1fb39c88a7
0b6d6daf24563f7a7aff3c2dc4e5ea848cd79a79
F20101123_AADIRD kanjanarat_p_Page_042.tif
53f993c8833ac94b555d99360ef0b7da
5a1e7c18c4698635014aac0b19e775ed9bc05a7d
F20101124_AAAAAW kanjanarat_p_Page_146.tif
9e8b50526f29fcbe1feab3214ed2bd63
385aa872e5c88e9489313e7dbd7dd35caa008f9c
F20101123_AADIQP kanjanarat_p_Page_025.tif
54c4a8992e29b5e4776660b86f0f38e6
57d2d4307cce1199c78dfddeca7191bd89c124a2
4265 F20101124_AAAAYT kanjanarat_p_Page_154thm.jpg
0eb9bddd126cedc3be448505fa76fd6c
3ee3ffc049904ef9f2552ab6e07365900484e2a0
F20101124_AAAACA kanjanarat_p_Page_179.tif
993e1a094e0953f40f9870461f15cefd
59593462d40cf336fe01728407f5534b8efcebaf
F20101124_AAAABL kanjanarat_p_Page_161.tif
c869e596026063f6f532162ef8507a19
f57a50b00c611271bdfae77a9d7923c7a218ce86
F20101123_AADIRE kanjanarat_p_Page_045.tif
38ade0046a80423618602eeda2ca350b
868996ae64d967bf2d090f7700bc9e0c9c168389
F20101124_AAAAAX kanjanarat_p_Page_147.tif
e922e18cd46fea2fe533c4dfad701e94
e0f4cd358914216b16a4ce84c09d6cc2a2c7ff39
16067 F20101124_AAAAYU kanjanarat_p_Page_163.QC.jpg
b07f4056e4cc8e72e174d28aac83eb3c
fe4dd54ca0d38fcb97013e8f9e90b909680a94b4
F20101124_AAAACB kanjanarat_p_Page_180.tif
9055744410c8f0b324e13c9c061aef22
78d95ff34772caf45329edb4a3fe4c51d719c8ab
F20101124_AAAABM kanjanarat_p_Page_163.tif
9edc757ede9e61dbd1fcfc460d30c5fe
2a83aa4f3bf2b21d49cc2040c9e5a12ef28abcf4
F20101123_AADIRF kanjanarat_p_Page_046.tif
6e8f1f451cfdefab267bb3c43eef97ec
b79e4da1db9f4a530454cd4fc660782e302b0620
F20101124_AAAAAY kanjanarat_p_Page_148.tif
b342f558a98f05da8298d2cb8ae6b1c9
c08215204aeae9dc3ae59ad061db4b1766a8a4b6
F20101123_AADIQQ kanjanarat_p_Page_027.tif
5e13d3a1bbb141393f988efc02edfac8
76ab416052b1bca126db6df5dbfbb3d23a9f3958
4157 F20101124_AAAAYV kanjanarat_p_Page_165thm.jpg
de47458ac707468f1129cdfd9b555fa5
ccf1e5dbe9e88a30060b251453759d776bea4d29
F20101124_AAAACC kanjanarat_p_Page_181.tif
64d83e2cb4d16a61066aa61e9d63fbc8
b98ada319599848c3abcd3c30c9e801b3fc7b38d
F20101124_AAAABN kanjanarat_p_Page_164.tif
6aebb440d74a716232c57cb25f009bc8
c4a30e0b737be8ae148576cd8bca7a73a7d904fb
F20101123_AADIRG kanjanarat_p_Page_047.tif
1c6ff5687d7f2bd103f5307bf9ead444
5bc0091a9648493280087a6245f744e6c951047f
F20101124_AAAAAZ kanjanarat_p_Page_149.tif
7bf09bfbb0fffdd0553b6f84450f9b4c
6a4f630fb5caa481f7b54d74932ca7510862c98c
F20101123_AADIQR kanjanarat_p_Page_028.tif
ed5589b551273aaecc1e7697237d8615
72a54384f71ad80f07464fc78af6c70b1feffce2
F20101124_AAAACD kanjanarat_p_Page_182.tif
1e1c81f6ba55ef8af9d6414a898f2e00
a449180f979e6b07b11a944d74c0bd97b1bf6282
F20101124_AAAABO kanjanarat_p_Page_165.tif
2472f3b5ed73a42b66d2feb3c8df7763
407a8180d2c8923dacc73a3e86809bcd64d67434
F20101123_AADIRH kanjanarat_p_Page_048.tif
0f0dcdaf5c81aa4cd2f09a526ec8b94a
12016e0826e4142f0ab00e0d0f50a83ce3030cf9
F20101123_AADIQS kanjanarat_p_Page_029.tif
0da0ad8ce4cbac9e70e3935567de3815
c0173be88b46ac33fa0be469ba2425ad1ba21466
F20101124_AAAACE kanjanarat_p_Page_183.tif
5e76a5b8958fb2de6fabdd156d3b62a0
27060b3b993d2211a22251696318228ff77dfb06
F20101124_AAAABP kanjanarat_p_Page_166.tif
6094ec51d72a4d32f572266e6b80da8d
cef59ce267970d72b7e4554be24f0735b9127d08
F20101123_AADIRI kanjanarat_p_Page_050.tif
0e014de381ea5ed6cc13f361c31fe1d5
1239baa14e397db19edad8ae83268806fedcae2c
F20101123_AADIQT kanjanarat_p_Page_030.tif
178696e992146f26b7baddca0bafd18d
9ddf77919d6b91facabe17cd89dcd5a1d6b8040b
F20101124_AAAACF kanjanarat_p_Page_185.tif
c3e46178cd0002922653fc7d2b1d239a
f3586fe93eb108e5a0d24a7a518a51f17f990fa0
F20101124_AAAABQ kanjanarat_p_Page_167.tif
77cba9c4ba328fa0b7f699d3b647d43d
ad9d903ecb7255d53abe748661c7e9681b712805
F20101123_AADIRJ kanjanarat_p_Page_052.tif
daeee9461d2cc402c46e2c99bec14c70
e75870fc899527c90db37f8f8e305f458c4fd958
F20101123_AADIQU kanjanarat_p_Page_031.tif
08279b8b962a740adba9e0f571487623
d986cf0bbb5132cf3ec5d15cc0164842bb26ecad
F20101124_AAAACG kanjanarat_p_Page_186.tif
862aa53570551534281609a4ff80357f
b6616ab393b7a47e9ab7773b99ce0d2e139566c7
F20101124_AAAABR kanjanarat_p_Page_168.tif
2f8b494f515978be1fbb07d72d1116c8
633b039c5302c966c107ff8ac417cf70da3d9d9e
F20101123_AADIRK kanjanarat_p_Page_053.tif
a18a541b5c8e8b37de83e3516f378db6
639f53325e99a8a643221404c59dc990cb3ab758
F20101123_AADIQV kanjanarat_p_Page_032.tif
213b56972bb8d017e90f41f19c593634
42a1f186fbbcebdc507bac9caefd6f5d3eb176b9
F20101124_AAAACH kanjanarat_p_Page_187.tif
95b0fcdde482b82b929f213170a7c388
ba90186040a6870dc86c8181ee67936dff1f2138
F20101124_AAAABS kanjanarat_p_Page_169.tif
388bd1182a6996b9570bb9708580cef4
5eb63f311f0ec5dbef3886067372e523335ced24
F20101123_AADIRL kanjanarat_p_Page_055.tif
959dd73a578876b991764dc01700be73
680d28a3e05637079c99020a7ad3beccf9cc3ace
F20101123_AADIQW kanjanarat_p_Page_034.tif
33202750e28063896d2b5da94ac0b614
8a8930626047e9737b7cd62bd99d84a32572c425
F20101123_AADISA kanjanarat_p_Page_074.tif
1536900364c9c19267d4658436bcd44c
9079cf52a76ba3f423010208b565de09df8f532b
F20101124_AAAABT kanjanarat_p_Page_170.tif
8efef732bb740c9699845ca5c628bca4
f889040d04bed2dcb4bb6a9a257543bddce97606
F20101123_AADIRM kanjanarat_p_Page_056.tif
a593e5142bf48f80b9723fe5f1db6a32
98e4eb742383d8a9c1d510dc0ab8c4e326dc3798
F20101123_AADIQX kanjanarat_p_Page_035.tif
3bc9d7ccae37ad92309ae985f3384e1d
0df40cbc86c96e74eaf8e72fcb98538fcae2e242
F20101124_AAAACI kanjanarat_p_Page_188.tif
0bc485b5c812269d044b2ccf9d6b7ce7
8f56bb3375afb7a2c3fefb71b2c91c1a087842f0
F20101123_AADISB kanjanarat_p_Page_076.tif
e9b76065277ac6fadbc184e323c30620
c8b9465c81989577b5da9722015c822dd8e38e54
F20101124_AAAABU kanjanarat_p_Page_171.tif
80fd444783aaa523d7cf1542ad9bbbb4
df6980cc29402c900c2897742cfb201a15cac22c
F20101123_AADIRN kanjanarat_p_Page_057.tif
664826bcb60529085452ae239b3f4c73
b328a968d9b1ceae1cac254f13bebb874d2958d8
F20101123_AADIQY kanjanarat_p_Page_036.tif
f342ff4f443c633c6020bd1ea8771416
5eff73a058d5e2823d01f0323dbd0f9328dbfbdb
F20101124_AAAACJ kanjanarat_p_Page_189.tif
3d426958c663f8f2f2752bd3c2cc109a
890ed80729edb8070ea47ef2a3548397ee906862
F20101123_AADISC kanjanarat_p_Page_077.tif
5562d29b74a37cad8c86a32d7ff718e4
bf6d8afb9da264113c92e9f491f852f803bb812b
F20101124_AAAABV kanjanarat_p_Page_172.tif
aceba2dbd21ecd718874349788b83f26
5009e37843f2fd0389023d4ee051971866657293
F20101123_AADIRO kanjanarat_p_Page_059.tif
fb561d8915f9e2450b22b2e1f487ae7c
3ce16f58827fc8500f28416491b7ed235e589764
F20101123_AADIQZ kanjanarat_p_Page_037.tif
7e2da8ecb65e156d066e66602373d087
23f18e1da640d99e77621b914c4708bfcf8cff80
F20101124_AAAACK kanjanarat_p_Page_190.tif
e75abba4851aa6017de380e35ecfcf77
af92bbfd8f2ef25b1e85553d2baa5cf86c2929af
F20101123_AADISD kanjanarat_p_Page_078.tif
71f6f2b70afb41c5f0143f7e0d6a6619
c2378ef497d509ae8619317ab2e7463bb41e68e1
F20101124_AAAABW kanjanarat_p_Page_173.tif
7e78a65b60f41d0d563c0f155712dcfa
ddd10ddff6873ad31708c4c7818a4547345254ea
F20101123_AADIRP kanjanarat_p_Page_061.tif
4aad9835809b468907065ab407850b67
ecaea0898167d49f37e45d1d7e1201134584f861
F20101124_AAAACL kanjanarat_p_Page_192.tif
439040fe4bc3139b346bcda51b5517e4
9179b732b7f48c832b7d8a5d2cd453e12009b6e6
F20101123_AADISE kanjanarat_p_Page_080.tif
ec52481b787feb8bd5081d5d7bdc106f
9351e1cde36b902472be0fc5e3a76fd4f6b5afb6
F20101124_AAAABX kanjanarat_p_Page_175.tif
7f6a717229d0dabd11dd09f536119616
0cc0f63f068497094c628e471fda61ae01a413bb
F20101123_AADIRQ kanjanarat_p_Page_062.tif
b654164381987a8413acba9b1a20331f
1170c0b86c78f53cd36816bd4ade02b5df15659c
29569 F20101124_AAAADA kanjanarat_p_Page_021.pro
b28f2df525e7b8845cd443b4d9ae7007
a2f97682677256ab2041bf7594c537eb91559e34
8333 F20101124_AAAACM kanjanarat_p_Page_001.pro
7f71e372d9e4eb35353bc6562b4f0fa7
0a0e9138e4981b7a9767dd16fe116069966e3291
F20101123_AADISF kanjanarat_p_Page_081.tif
e5e64f8df22d3d0d5398fc7bb8b16993
83befdccd880c7111ad4f1ead943e2614e2aa6d8
F20101124_AAAABY kanjanarat_p_Page_177.tif
4409ca0ddf3a1973cb1dae582da4f5b7
2dd8e95945eca6eca0a7883763a4c7a1d8299482
40617 F20101124_AAAADB kanjanarat_p_Page_022.pro
4566158fbd85b58509cebd2a1fafcdf1
ec8149ec42f34ada127341a473297c536cf4304a
F20101123_AADISG kanjanarat_p_Page_082.tif
312fb4807f7e000b034300ad3648351f
1c2f22c4a2d9c155ebe1cd6b33fa5eafceb23874
F20101124_AAAABZ kanjanarat_p_Page_178.tif
3bd50529295f57a265e3d77d47c5b543
167c42340b1898085f7645e73772ffd2b84d3c18
F20101123_AADIRR kanjanarat_p_Page_063.tif
206deba03dc358a7a99d44281a5665c2
beca43210fcbda80e81c4827d1f9fb6e778d5f35
31675 F20101124_AAAADC kanjanarat_p_Page_023.pro
71df847a2c6a2f0885660f8daa568033
9e7e46082b12741f931ef6b0df1da20d82152964
1256 F20101124_AAAACN kanjanarat_p_Page_002.pro
248420c14cf955d3aae5c7a474cf66a3
0c4d700c1566fbfdc84dce93e8c30b0ff015c9f3
F20101123_AADISH kanjanarat_p_Page_083.tif
90bae047bf3b797eb5fce6ac67e444fb
78d10bb36d3b304b00141bf093b7143dd1e1c7d9
F20101123_AADIRS kanjanarat_p_Page_064.tif
ab1afd53190d06851d7a4ec7461bc212
e498f4e8894566a13bf86f1dcfeda517ea5bb9a7
46124 F20101124_AAAADD kanjanarat_p_Page_024.pro
5cb4684bf40ea30c4f535cef2fd7f1b6
97e7cf72a4fc5390a61542d06207cb6a079a9774
37515 F20101124_AAAACO kanjanarat_p_Page_004.pro
faf764cfb313b097249a6771e348f23e
b33123fcee16a2414bb5684cfea9925f8131e134
F20101123_AADISI kanjanarat_p_Page_086.tif
94ebf71827a6fde42f024a779d3ecde6
f4808d724cde3e31ffc511c0639a57739141f90e
F20101123_AADIRT kanjanarat_p_Page_065.tif
c020b88a010218c021a8f8c2bc426b4e
38932c0e6f65f27d84f6f830ca37693264944c6a
46071 F20101124_AAAADE kanjanarat_p_Page_025.pro
ef0f8da731855d31184613336128d670
8d774f145175ddf45eeafed346d6a9176f44f74b
76471 F20101124_AAAACP kanjanarat_p_Page_006.pro
a79eb225b6b05812872c5d926129ef0a
842dea64cc248f21288ce7534854f24c90fe2639
48874 F20101124_AAAADF kanjanarat_p_Page_026.pro
61431813e5138c32f4291698e629bada
a63a75279b73670d7cfdab61277813918fb51a81
F20101123_AADISJ kanjanarat_p_Page_087.tif
dec3691714fc9ecd757fcafaeb1bb57f
692d381faa9b99ebec59a85f9dd45e42f4ac3a46
F20101123_AADIRU kanjanarat_p_Page_066.tif
e059022b5f119ad983635b369d07b560
c30c7b4dba8c0f25691ef6e7ad9e831dd796a30f
105758 F20101124_AAAACQ kanjanarat_p_Page_007.pro
68012689fe44a8ef41932b2b4793c4c0
09b6dc6351516f6d2c743e7c9eb9c7dc06d8212a
49362 F20101124_AAAADG kanjanarat_p_Page_027.pro
99a5e7833a71eaf59ef1d816b30a0271
ef349982551087d6b63df349ebdd8f699f9e8f53
F20101123_AADISK kanjanarat_p_Page_088.tif
9d492c9c927a63a78d2d1473a5d58edb
50eabe21c8b6d7081cb91784540b4c01ec37b526
F20101123_AADIRV kanjanarat_p_Page_067.tif
d94c1e35502588eec9dd69bec3b4e132
f5cbde84f20b26e7eab10b76d39f47f8d012e9db
116083 F20101124_AAAACR kanjanarat_p_Page_008.pro
23f299c947beab57180e7d2e45b73541
4f9906cc6cf6a6719680b89092ee972d88a27be7
46669 F20101124_AAAADH kanjanarat_p_Page_028.pro
561d3c26c2afd52a4409c8472e4e5d2a
10ff0cf583b5d507ffdbbcd0f112e34fce93d113
F20101123_AADISL kanjanarat_p_Page_090.tif
eb4eca58bcc64eebdac11008e25aacab
baadbccaf0c35593ed11a1d39e03d9ab66065729
F20101123_AADIRW kanjanarat_p_Page_068.tif
0f7fbd895805e990a7021f8aa954492a
61fb264c2ee26a4ab80a656095465e350dfbdbce
59386 F20101124_AAAACS kanjanarat_p_Page_012.pro
35e82eaffcc98ed949740460633cd177
2965d3ff4edae4dac954bc02c6761805ea5387dc
51749 F20101124_AAAADI kanjanarat_p_Page_029.pro
e08d61743199b56e3386a68ec017695a
3e31aac3b314d0833384b38c30d836c806de6cdf
F20101123_AADISM kanjanarat_p_Page_091.tif
1b2bb090dd5834fa39dce033af3d0204
15be89a831aa2cf39d0e17eab73a27ba0d4b6cfb
F20101123_AADIRX kanjanarat_p_Page_071.tif
a223b572f635b2e750d6f2a893601056
e9ac7e5ad2a581a930a00c16a26f786b00fb011b
8423998 F20101123_AADITA kanjanarat_p_Page_106.tif
8e7686ba9c7a465ede237efe80f8b216
a3bf6bd9fede3bfcc7f96d21b860b9165ddb51c7
78324 F20101124_AAAACT kanjanarat_p_Page_013.pro
e5a862cabda561f920783fe65bff578f
e48d188e0ec2a847c0d6e4b4bc375bf016fc3ce6
F20101123_AADISN kanjanarat_p_Page_092.tif
33009cee2dfd21fff9d3c281647944fd
789444937d1e40a9fc93e2da1fa43708bbb3aa91
F20101123_AADIRY kanjanarat_p_Page_072.tif
0913a59ea89342763bc2de78a6ff25bb
a37e953be62df7e20544e135b1528138405473da
F20101123_AADITB kanjanarat_p_Page_107.tif
61f02e91763ce867b161caab385d7a1c
4a2e56dd84b610c80b3ec373f036566f4d7ad867
36157 F20101124_AAAACU kanjanarat_p_Page_014.pro
beadfa679728692ef3b9b7006d0567ce
a24b42b5ad3b5a22b88755fda5d64c4321bf3cc0
50854 F20101124_AAAADJ kanjanarat_p_Page_030.pro
67205e6a7d325905f596e27e0c2d0a74
d4a2835bc31163df18b4e38fbdcf7d23e0023880
F20101123_AADISO kanjanarat_p_Page_093.tif
dd22af513d5aef958405abf6ea6a8557
b26dee70789cf268b44ec1d69c206f133d293655
F20101123_AADIRZ kanjanarat_p_Page_073.tif
388cfd2fbf52ff95cb13fe78879b7389
c17b4f8a6cfa0f3ef321d121d5339daa285a7c54
F20101123_AADITC kanjanarat_p_Page_108.tif
ce92da62e8dbb5f1362073db76398f45
77fa02c3d68e7bff782f80f718de5a5a2ad6d3de
38469 F20101124_AAAACV kanjanarat_p_Page_015.pro
732bc51072730e2f97a5c5cadf13b368
41a81fae749f1dad38f0f18fed3b0e78e7c0258b
32502 F20101124_AAAADK kanjanarat_p_Page_032.pro
dab0424bad289500914d79c139c2cd91
2964498f5d58e32419e4fe1e631044aeeaada25c
F20101123_AADISP kanjanarat_p_Page_094.tif
a453fcf98d0d816ef746cf53d2ac9035
866ad30d444602aafc58f4bfcf63bcded8fcdbed
F20101123_AADITD kanjanarat_p_Page_109.tif
711f89e3c26859ee10d419656e92012a
1603a10b900490868792d384f8a93bced7f57750
40523 F20101124_AAAACW kanjanarat_p_Page_016.pro
3ff19489f49f34d5d800f065eb2312a0
8d1b4cbd90caa55e7c5990a09bc99adf75073543
49044 F20101124_AAAAEA kanjanarat_p_Page_059.pro
9943e61ff16c30ee8dc0347322767d41
2a7c239b34ecc0416cff20be053e3bc03019beb4
50275 F20101124_AAAADL kanjanarat_p_Page_033.pro
0fbecaf7af60c5fa36850ab5f190b18a
8966e312daf215ef27ed6b9fdc872ea48d99bda7
F20101123_AADISQ kanjanarat_p_Page_095.tif
78b412857e85ff0b0c8c13a94f036c4c
37f06d813c648336f55c89786296da3baaefe328
F20101123_AADITE kanjanarat_p_Page_111.tif
285efdb02c82fd7d74fac0eedc4b94d8
19edd2b551b9882c4a53ef9406ef120eaa7a7f48
49616 F20101124_AAAACX kanjanarat_p_Page_017.pro
0ab2f30b78faa0ee0f16a5bb83662a49
acd9338e53de6dcdaf6aaaca947808c73a7f0043
47456 F20101124_AAAAEB kanjanarat_p_Page_060.pro
798ad04735baeba40da513e3ae1feb3f
3613e73f18318fcfd2b04b9686135e0f9bf7232c
48687 F20101124_AAAADM kanjanarat_p_Page_034.pro
d5c552cf5b7edf8c6f304b210b56b421
0a3a8b6fc5b0d8a72822f32f7fb0500bbfb66329
F20101123_AADISR kanjanarat_p_Page_096.tif
6b924a93df1b66b76ca324be38757523
044c7bf50f87aa7da30e94f923235ec33a3a6f0b
F20101123_AADITF kanjanarat_p_Page_112.tif
4f3c83a4c58ed1be74486b320768ed39
bb1c0478b445caa554fa9dbe390c95bc3ece08ce
50496 F20101124_AAAACY kanjanarat_p_Page_019.pro
bdf8822346c8f08a7f175fc58f23a966
66efff0153028c57e88731e61409ce1c112359b8
49079 F20101124_AAAAEC kanjanarat_p_Page_061.pro
75c272410ccf5c2e54c9eca0bfbc726b
005b0eb4e36d191875b328d59302484483aed58c
51271 F20101124_AAAADN kanjanarat_p_Page_035.pro
9c80ab8e06796cc6ec7f827d1103159e
29c7ec420fcea924e4a9c854adab492f8a3f76a9
F20101123_AADITG kanjanarat_p_Page_113.tif
0e9e6e2444befb23f710b7bfa4d12469
e20a0d2672a2664f5231176cbfa05164b242780e
64731 F20101124_AAAACZ kanjanarat_p_Page_020.pro
b66e51a7f30175e3b89ccda8342228da
ab685a980fb483e12828399c20fe5def0bccf0df
47017 F20101124_AAAAED kanjanarat_p_Page_062.pro
c274a5b75b004bb03dc5b608f9b04aec
120a3c0b5c224c9f3e000c1d0a2da987903c772a
43144 F20101124_AAAADO kanjanarat_p_Page_036.pro
93f2101aa810d32ad60bb6680df53da4
6e6019573bb092f7e630f2b6eebf96869f4277d7
F20101123_AADISS kanjanarat_p_Page_097.tif
e407bdd31aa3e651bb6d4aa17bf43ef2
95828783a21ecb737ff1d22a8ad2966b1f66fb8f
17685 F20101124_AAAAEE kanjanarat_p_Page_063.pro
34a4bb6d73d0aa827063f3cd1646f8f4
a8b70d60772da7e3a63cf46dc2870494abf1e889
48482 F20101124_AAAADP kanjanarat_p_Page_038.pro
6c026c36d7216ed661f065a97e52d99f
efd29371216de7d82f9691a82ff7f5cb00c9c7a2
F20101123_AADIST kanjanarat_p_Page_098.tif
c7d25248958bc35329a4657508bf99a3
fb369685ca63e0a6fe06ecf5267bd847506f699c
51628 F20101124_AAAAEF kanjanarat_p_Page_066.pro
e636f9634604517ba98fd8cc4c250b98
80762d1e9fd0d96f0e9d60adb3e6f95b3f0be94e
48603 F20101124_AAAADQ kanjanarat_p_Page_040.pro
b8b62ecffbe4ca144245e6784607c7bc
251c74e0c472ba3684eba34d8a12de1351960126
F20101123_AADISU kanjanarat_p_Page_099.tif
1e13188bb7f75116d2f971b9bf8be5d1
9ad492131fefa9946aeb8703ecf9ff30cbd5dc2a
46917 F20101124_AAAAEG kanjanarat_p_Page_068.pro
96d044c958fef2139319b0b2c071ca10
da3a9d6ca6a626de508dbd20e4663bbf48ffdbd4
50534 F20101124_AAAADR kanjanarat_p_Page_044.pro
bd84fe08675b3a3d5ee1d6048549d42a
dc5699f626f1939b5e29a91a426838cc0fee82d0
F20101123_AADISV kanjanarat_p_Page_100.tif
91f383bb4e0f65852c96e65839ab9dbf
4ee36c8bbc06c94f9cc00aa3e0bc930459685376
49672 F20101124_AAAAEH kanjanarat_p_Page_069.pro
ae27079ccfbaa9248f14389ac47d3156
10b5218fef728d55bc046c1b1622e757ff10f448
65165 F20101124_AAAADS kanjanarat_p_Page_045.pro
ec7aa3b94681c078392b6df14426168a
72b23660d209521bb6c0be6191151227f6f51745
F20101123_AADISW kanjanarat_p_Page_101.tif
4082b5ecfbcfa97aa9ef5eb4a22f16e9
8d0caf69fe08cb60ff9e6ab64ea1c69d331d9a1a
44472 F20101124_AAAAEI kanjanarat_p_Page_070.pro
9d9fd8b4d7133287b25e73880de1b1fb
380363c5935a7797a8310c362cb40e2a514d250d
54018 F20101124_AAAADT kanjanarat_p_Page_048.pro
6c6ca8e4d3dcef341fd93e1d30dfad11
247c16ef0e14305bfb2b01eb495bd1142dbd0c50
F20101123_AADISX kanjanarat_p_Page_102.tif
06c698fae927037ff3f22a9a1244a850
8dd07992304c9631040e7f5d8c65700fdd776c4b
41239 F20101124_AAAAEJ kanjanarat_p_Page_071.pro
8aa7c753d2f89a10e5de6d08557c0bd7
c8c0256f9f8fb075c9dc9a55dd272cc22be16215
51339 F20101124_AAAADU kanjanarat_p_Page_050.pro
233c829502096a1e3d671ce6fb1c2fe6
b2e12d3e6c75ff13529aec9e1e34a4e4ef45572a
F20101123_AADISY kanjanarat_p_Page_104.tif
1cd389cd67843a5ca161363b9f2fe823
5c3569a527da4a7c31c64610bab7126bdd51db83
52328 F20101124_AAAADV kanjanarat_p_Page_051.pro
6fb39f9003406c5fca4d77beee1567d5
94d463e491eed26894bbfd58a1bd22c1595b47e5
F20101123_AADISZ kanjanarat_p_Page_105.tif
eb3bd02af1aa54566d1cdc9b7150740d
b92695e0817741657129c16cfb85e0f1f333ffe4
10955 F20101124_AAAAEK kanjanarat_p_Page_072.pro
80bcc5efa746348dc92d2190bff21381
74d1a723d231d2f4e760b855e29721f296026527
50909 F20101124_AAAADW kanjanarat_p_Page_052.pro
277f47938af74f2846d7d73fb250294d
cbe81617564833104437751bf906b373565c58f0
18153 F20101124_AAAAFA kanjanarat_p_Page_090.pro
1bd81b007707f8cbad98fcdf840c732c
ede33626adc132944c9b08237b37534c4c433154
42125 F20101124_AAAAEL kanjanarat_p_Page_073.pro
0267b4a274980e22f2abca60761fca6d
ff54cec87da81e89e7e1d2bb46d6c9c3a949f29d
48008 F20101124_AAAADX kanjanarat_p_Page_053.pro
6c7e55a7a49a9e17f21cb6c048d64760
555b75deaf79a91faa796aec91e38101c82550ac
34154 F20101124_AAAAFB kanjanarat_p_Page_091.pro
4d34615e260f0f1d96bb846038480875
7d0a8a42aabc31a2d7d51c7727d3f8cae90f2174
25654 F20101124_AAAAEM kanjanarat_p_Page_074.pro
0a78b88e70bc760f0c79ef4f49b21438
7e56d4f370b2a391774868c68dc5efdc5d2cd618
51433 F20101124_AAAADY kanjanarat_p_Page_055.pro
00283422de55a232ebe9660f7c23563b
84af2b5d0b5f6c954913fd108eda8679b358523f
51599 F20101124_AAAAFC kanjanarat_p_Page_092.pro
66de83862e2ff01cbcefca4834a2e2aa
94f1dcc4576bc84856ae0bc8db888dfe44b02c2c
39684 F20101124_AAAAEN kanjanarat_p_Page_076.pro
e24479143437dadcfda4a4220f915311
f9c563acab11953de163c3533120494b6e032760
33817 F20101124_AAAADZ kanjanarat_p_Page_057.pro
c197d9d183d6d3cded5a3b2acbbe5e7d
7234a3d05870e6424380bf81d1ec08d06ec361e5
14272 F20101124_AAAAFD kanjanarat_p_Page_094.pro
9fa286fbfcd71f5c7b1ee5bb13ba2e2e
650df7b2896735012f3ee5c2e8a71548aec9a23d
24252 F20101124_AAAAEO kanjanarat_p_Page_077.pro
4f5b136fd7ed9cb2bbea953404904d4c
4ccd77dbf6c9ac05e32c6836bf18ba071c47d658
44853 F20101124_AAAAFE kanjanarat_p_Page_095.pro
f38a173ba7c9752144d65149c23a5042
4fa5e9ea3eff2882cc3822d40930282680fcc2b3
32938 F20101124_AAAAEP kanjanarat_p_Page_078.pro
a003fb53de99a5df451869f8dea834b2
995e436858504b9d079f56da013a8b2ffddee60c
50349 F20101124_AAAAFF kanjanarat_p_Page_096.pro
09bc189dc35a6b8bb9c04513e61c4660
9ed6ceb2da3eb0d41972a3c6ceb978857b3f301e
52527 F20101124_AAAAEQ kanjanarat_p_Page_079.pro
4f41e5ae2ad2c6a6aad83cde395f2347
d96493be35305b6c97cd9583c5a8f354c2b964a0
43925 F20101124_AAAAFG kanjanarat_p_Page_097.pro
4001828e89efb54fba676eb577297d31
6890165a21f4431b7d6bd087f4c85e6b34007a80
49735 F20101124_AAAAER kanjanarat_p_Page_080.pro
3f6eb72acafd8c1e844545eb77a32dfe
3ea076e04843603cddb1dfbc239c9188a25ef1d8
30162 F20101124_AAAAFH kanjanarat_p_Page_099.pro
29060bfbf94d945fe8b2c452475fda78
802d30beb45ec5bd707587b32bf0234e9337893c
48781 F20101124_AAAAES kanjanarat_p_Page_081.pro
bfc5aa2f0523e7c3ae6aaa07471c53a5
82fac529bf3bb39e287b19d9e0d3806decc2f5ef
50923 F20101124_AAAAFI kanjanarat_p_Page_101.pro
47edad4ce3aa181d5141113aefac13bb
52cc7c0424e88dd8c67efdb07a73bdd54aa5fc91
40170 F20101124_AAAAET kanjanarat_p_Page_082.pro
a37a7bad5cde0c752f46ea78840e1fa8
0293dd273e767521bfc0055b172fd2a65e830769
46962 F20101124_AAAAFJ kanjanarat_p_Page_102.pro
c41c76d88161f19bc98d2e076eb21f25
ba8700e91cb0df7a7ae266ab59a6dd090de93f9d
41009 F20101124_AAAAEU kanjanarat_p_Page_083.pro
c6936d0768483281df5f3b198015f0ba
d2cbb739d46014438a2a1313c3247234679d6c99
43971 F20101124_AAAAFK kanjanarat_p_Page_103.pro
e5a460dd59bdd908f9407c442a221968
149c45e57f9620f889aa1a80636dc5bd69f80045
45733 F20101124_AAAAEV kanjanarat_p_Page_084.pro
db4ddcbc11a35614c4ffb3bfa27183d5
87cf760fa21bab3d859b4e45fade4f9458306a31
32091 F20101124_AAAAEW kanjanarat_p_Page_085.pro
28f27ddb2ce1b4b05b7b3b8cc801a0f9
ee80fccf6e96b78dc1c8b819ef99438cb9cbf6aa
34327 F20101124_AAAAGA kanjanarat_p_Page_128.pro
49438ee31e86872da6e2601a09775712
ad39eb104982258ade64ab9416a5acae7f0a95cb
47098 F20101124_AAAAFL kanjanarat_p_Page_104.pro
8b016401031fb031d3a0bdb02852472c
73ac915cdd72f64f4aaa2d3392d8625fdb3a336f
46872 F20101124_AAAAEX kanjanarat_p_Page_087.pro
44c3a54b595bb4fdbbbbeb7e9a46655c
c6965ab49fd8318ce09ac2eeee566904cb58060c
35308 F20101124_AAAAGB kanjanarat_p_Page_129.pro
debb148d7a288dfc08020efb236ace4b
0b220a945909722a9db00bac39674564cbb07a32
40579 F20101124_AAAAFM kanjanarat_p_Page_105.pro
b5f6e713852d259db45bf28eea5d727f
241025f39d8fe169d91d1ba612a8265c58d33992
18045 F20101124_AAAAEY kanjanarat_p_Page_088.pro
167efbe2a4b8b996a3d1f02a74fdd6fc
d134c376a33e8b5788a7a135127495eb6723c581
46718 F20101124_AAAAGC kanjanarat_p_Page_130.pro
bd00f3c036db1fded1d688a80a44909d
12eb82a5ebce837d66e50182ee7400b3a634d5e2
18563 F20101124_AAAAFN kanjanarat_p_Page_106.pro
32ad19e47ea20440501b3a5ce084c43f
50abe6237c12ad1fbed0db2bebc7c0d484f7dd52
48272 F20101124_AAAAEZ kanjanarat_p_Page_089.pro
9d0c08debeb3f20c505e74a0fa91ce4e
f5f51d92a66b0848a2e7b46af01833580bd156d6
68469 F20101124_AAAAGD kanjanarat_p_Page_131.pro
f7344387e84b10fc890cab3f1a95dd73
4dd750133571756a3b0121ce83bbb22b362a9f91
32629 F20101124_AAAAFO kanjanarat_p_Page_109.pro
a2ae2357da2c3a2ab400cb8ba4f22920
e515beaa0b13cfdab70d5e6893db0f6d6d2605a0
43162 F20101124_AAAAGE kanjanarat_p_Page_132.pro
7b539da58f31b82d3a2d76b5bbd657b7
bd7e480b53f737cb9980f5c49a2d0c1bfbd2883d
44749 F20101124_AAAAFP kanjanarat_p_Page_111.pro
1ed841d8f95bae9dd083f8e3cb26c696
af55ca0f596b29b422eb0008d62a24a706834d2e
1852 F20101123_AADHSG kanjanarat_p_Page_095.txt
a40920869090b9a5da516852884bbe49
736796c48444069b228e344effb2ac24c082a280
29268 F20101124_AAAAGF kanjanarat_p_Page_133.pro
19c9a9b343b40a7d854ba8697f87e175
d9e5c9fd33641badc22155d94b3ffdf9d8659a8f
35385 F20101124_AAAAFQ kanjanarat_p_Page_112.pro
bfc8551c258ae8bbe6ce43a63a7533b6
69e724e75ee7f4df6f098f8e083c7a5293f4e85b
17762 F20101123_AADHSH kanjanarat_p_Page_014.QC.jpg
a68ee3a877b49a5b2274285309bc4f7c
48b00a733b0e0db38229f04430ffad11d8b16f95
43991 F20101124_AAAAGG kanjanarat_p_Page_134.pro
de9c2ad723aca1e41318e3aedbc9447e
06400f7d9fba42cfe1d94f795fd82e1c6e835a54
57313 F20101124_AAAAFR kanjanarat_p_Page_114.pro
a2b2e772d17993e22c76d4505147ca05
a6731394fb3c485b02c759e2475cb4c19bc6c08a
3741 F20101123_AADHSI kanjanarat_p_Page_109thm.jpg
b490cd40260d91817ae2941287e1b758
fbab8ac7e6a8703bb36d2a7df69bc7518738cd27
47310 F20101124_AAAAGH kanjanarat_p_Page_136.pro
f1a4c5b48cdddd57a1ae67b6c32af8a3
126310b2caa6a415ef995ca8cc0e660dd7dec41e
46088 F20101124_AAAAFS kanjanarat_p_Page_118.pro
312ed054c54d273d5d722e7142b4e774
b75894e60013aa8c5e61ad23c97a51d3c724d186
18068 F20101123_AADHSJ kanjanarat_p_Page_118.QC.jpg
fde322e8969be2d1d06bbe45b580b6ff
6692cf9f60fd955996fc6207cd268d1639bea28b
52838 F20101124_AAAAGI kanjanarat_p_Page_137.pro
bac7fe95c502aae7c5b2b0377d5e3f18
82d9c9a5ce8e8522fcb0ea6240f10096dee7b060
87534 F20101124_AAAAFT kanjanarat_p_Page_119.pro
1fd6ac41c70e686c67299c60b01ccb29
eab4404d0219ac2a8dff4ac9c79d522664ff4c00
1993 F20101123_AADHSK kanjanarat_p_Page_017.txt
4530820eeb454852856d148e2274e6cf
2c35683916de9f6798628a23ba7f58fdc0700d81
48560 F20101124_AAAAGJ kanjanarat_p_Page_138.pro
5d33530317c6d146a186da346cee1445
acc0387649e2d4d0708e3675604c1d01f7e6dc5e
35644 F20101124_AAAAFU kanjanarat_p_Page_121.pro
4ca471995abeefdc30e0d13f99cc421e
71c1684e873fec574a2781af6def346b48e5fa4d
4377 F20101123_AADHSL kanjanarat_p_Page_007.txt
acb1e036300f7b277c4dc4670bcd79b2
315c40d4f81741f4151adfbab5f7e2ce9906c091
50268 F20101124_AAAAGK kanjanarat_p_Page_139.pro
1fd11dc6977ed8497089e9890cb96219
c78827aab18017253dcb4d989365cd35c179fb84
14604 F20101124_AAAAFV kanjanarat_p_Page_122.pro
2cd2a8015b5f166d4ad639323836c17d
7acc8708871d9fd048a243c83bc63c248d1b3ec1
F20101123_AADHTA kanjanarat_p_Page_191.tif
f83ecf866c087dd83aea6294b2a68fb1
ce6d8983bcd292b9b1435dde52409db90e4614fc
37853 F20101123_AADHSM kanjanarat_p_Page_075.pro
8792bd7cb946113397c09f2eeb38273b
b4fc0f461c2abe2f7752f1981b48a18ff42ec182
46716 F20101124_AAAAGL kanjanarat_p_Page_140.pro
95f2d65e687e3401dc1cfc024df149ce
cf9d4c6a785cff8bb184b71dfbfda5e415565604
47672 F20101124_AAAAFW kanjanarat_p_Page_123.pro
45231464678b2c57a8513783093d7610
29dd39382c9f53aa3a1b688916e020bfbcd761a4
19723 F20101123_AADHTB kanjanarat_p_Page_016.QC.jpg
13d91621ce1a2264731d0b341ea531ea
6c07b1fcee3d6eeaf6359c174b43cd28af91625e
8514 F20101123_AADHSN kanjanarat_p_Page_005.pro
c0ba9dd1dcbd642889f0a34e7dbfa032
e2596e91e193792cea9223275dcc8420ba995dad
11261 F20101124_AAAAHA kanjanarat_p_Page_156.pro
4899d7d7f62327e68be34e285d5df2c0
8350ebacb70f7bee2121c201c5c7d69feea1fcdb
60077 F20101124_AAAAFX kanjanarat_p_Page_125.pro
b4d4d8f42227e8f3353d6d026434a739
08b721f3268ff9ce092d3ae485b412b5e586c2eb
39432 F20101123_AADHTC kanjanarat_p_Page_110.pro
2e5d725ce75fb7d99140f70dacef3e24
55aa7482f6e718674e7faa6ab1a2cd32ce6fe324
1074 F20101123_AADHSO kanjanarat_p_Page_150.txt
1c13847fd24a3145f4ca5506c90f1132
5acb384966714a011c213211188dfccc2b4ab704
41092 F20101124_AAAAHB kanjanarat_p_Page_157.pro
a34de9b8172944cd7bd44c5cbede728f
dcb0040ff4c0165fabefe76e037b6e5ddfe87e6b
47762 F20101124_AAAAGM kanjanarat_p_Page_141.pro
5915c49c566f1558f3573fa14d33d14b
4a8d147e3463e61b578755af1fb94cf5ea7b6989
40800 F20101124_AAAAFY kanjanarat_p_Page_126.pro
88e9348277144b07fed346855d454f07
2eaaa3ea01e6d9c829981ddfd15815544a6fa73d
F20101123_AADHTD kanjanarat_p_Page_070.tif
7242836309f1bc0499a05ec9eb4c61d4
262268bb919bfc030075927abd840532c9402236
6347 F20101123_AADHSP kanjanarat_p_Page_038thm.jpg
74b163c862c33596507255354461b11b
b6ef52d99d1a6216d91dd7e8e81996611e400dae
31119 F20101124_AAAAHC kanjanarat_p_Page_158.pro
1f1635d2e70ca255aa72ed969242fb79
f5ccefd9121f95caf6ea02729a4f506b60548a0d
52223 F20101124_AAAAGN kanjanarat_p_Page_142.pro
95a5b299569657de7c8a1af6b895dab7
b2015139f00843381d347a6e48862c045ff277ff
27754 F20101124_AAAAFZ kanjanarat_p_Page_127.pro
d3621551b2918739a27cda0ab48c835c
23a0aa1a65e403529ef029ea639767fe1342626a
47291 F20101123_AADHTE kanjanarat_p_Page_133.jpg
3faddfc614dba2f24261553b28ccc0bf
d268cbadfd1fdc289acb08e0b330e1fa2a5fceb3
69749 F20101123_AADHSQ kanjanarat_p_Page_060.jpg
4fe327dd09ef3d5392aed51fd12db81f
a7897bc562f3d2097fae94d9f5028c4f449519ec
20885 F20101124_AAAAHD kanjanarat_p_Page_159.pro
b4240188f90be1cc962c797d7193e626
306738e4e7d19b5e495dd350a5b0861a7432022b
45270 F20101124_AAAAGO kanjanarat_p_Page_143.pro
583172555f4de09639b75f4d6609024d
c507609b2d81d46ad44866a941bd75b8240810a7
105141 F20101123_AADHTF kanjanarat_p_Page_026.jp2
bf68503ec75800464bad362c5d4e7e86
92653805c201c6c39a862963698340fe6afe1b5b
73495 F20101123_AADHSR kanjanarat_p_Page_108.jp2
0d0f1ce85c49fb646180a795a151e480
597c958761cf8076deb6efd16866601b42fe0e7f
27410 F20101124_AAAAHE kanjanarat_p_Page_161.pro
674e4d053f9d8a806555b701acdabca1
7aab35c4ab8f026be35552a568a37a8f369ef712
50078 F20101124_AAAAGP kanjanarat_p_Page_144.pro
a7743f6a4cd514210900bc84024079b0
5147df8073244dd010caa707594b1698d56dda8f
6127 F20101123_AADHTG kanjanarat_p_Page_089thm.jpg
26842696ead993fe3ac4d7885980369f
5f439fa7b26f325d926640064b2bb4c1057b1606
26316 F20101123_AADHSS kanjanarat_p_Page_020.QC.jpg
81132edd382ff968808b47b247f7d856
b666ef6c4a383115df5b7f9c38b0cde14de15861
27612 F20101124_AAAAHF kanjanarat_p_Page_162.pro
fabddf7fea0692245683e08ecbab9a13
6848ce95fc559d482873b7d8cc4e0d0507981eb6
51632 F20101124_AAAAGQ kanjanarat_p_Page_145.pro
cfcc1c59c096b29fa96812823be21f76
23116d4c39682a86ba515ab2453f6ba124e78cb4
22140 F20101123_AADHTH kanjanarat_p_Page_024.QC.jpg
fa0ccbbab89c88254db3a6786bb6e516
1a68466ef22823525a0e5ff5a91368532f482a30
37360 F20101124_AAAAHG kanjanarat_p_Page_163.pro
4b410d83d26388e951259fe199425c7f
ad158b439fee00b178a51b2626cb611b7e3dbe66
50694 F20101124_AAAAGR kanjanarat_p_Page_146.pro
682dde507fa2504bde76d22cba14f925
2ff48167911166b98ce0320a16b4a6da9dbdfe9c
26740 F20101123_AADHTI kanjanarat_p_Page_179.QC.jpg
a1cabd5145233d7d43d2429ee2658c84
17907eadf0196640c7a61e873139ff9da3a4ca44
101009 F20101123_AADHST kanjanarat_p_Page_025.jp2
c66f657a46c2ba6bfa212012882ac28e
537df06120b69d91bc0c1642ee7f4745baa661e8
31587 F20101124_AAAAHH kanjanarat_p_Page_164.pro
6dd22fc444b4bd8979a66f6e3694715f
d423702d6a91fb80fa2e66904214cfa5a8d7f40e
47600 F20101124_AAAAGS kanjanarat_p_Page_147.pro
e1f433b850545c7b74f6a1aedbbc0baa
02850b29ed5725d8484d0f26fffe654e4bffccc2
3863 F20101123_AADHTJ kanjanarat_p_Page_079thm.jpg
b99534bf82fb0574eca6dbc6f11752a5
5784de2b65c7fd6f0574a4b0300e2b4477d710d2
F20101123_AADHSU kanjanarat_p_Page_097.jpg
688a25d01ae48c16949ced21094e35ab
66fc4bce85fa0a9128c600d524247d3cc8377b45
32333 F20101124_AAAAHI kanjanarat_p_Page_165.pro
05dd5d029fca4d1a16dce7f5d541a777
51c3baa4956f174a2af20d3fddd9dd4bf8a54e4f
49685 F20101124_AAAAGT kanjanarat_p_Page_148.pro
4e169af4816ccdada9311ec5ab98a6ae
4a63aab6aebca21e3900a6a93fd54e4bc43c888e
F20101123_AADHTK kanjanarat_p_Page_060.tif
113e97bee06602e519e86904e27d96c8
e74cf228881c781993ba99bc8fd81463218421bd
F20101123_AADHSV kanjanarat_p_Page_120.tif
45778f23b0bbd84307353df472fe8dbd
530416a260bff39e9ef976eb1d90051928f55b77
31647 F20101124_AAAAHJ kanjanarat_p_Page_166.pro
f153ec707f9be86c7d3e78f42d839b0f
31d5ff053492e88b3a94c27f77dfda25e51bb396
25743 F20101124_AAAAGU kanjanarat_p_Page_150.pro
885e08373348ec935de05730743d5da0
2cfaef6ad694de5ddfed296d3df1cff1831a9b36
20950 F20101123_AADHTL kanjanarat_p_Page_107.QC.jpg
00fc8a996777b591a8d6760ad8149679
91cc915c5600c38d11b254e83c32b1e937f4b0d5
2622 F20101123_AADHSW kanjanarat_p_Page_020.txt
1a44196b2baddea2c79f1316be4cf3fe
1844c99073f1d84587d0f983c401c44d0d415e0e
32355 F20101124_AAAAHK kanjanarat_p_Page_168.pro
7fc56cce8d850a24ff59cf285eadc7e8
13da25d9399e6f3e90223f0c9132400c49a4c04e
42366 F20101124_AAAAGV kanjanarat_p_Page_151.pro
2cdb7cdcfd8a42f18c4c14df3efb6fce
103980c626edc76eba8dcd1b104d09c8b2392ac1
1942 F20101123_AADHUA kanjanarat_p_Page_061.txt
999a6d5455da9b89253b596dc65b9ef8
556ceda6e0d9692505eb88c41bb347564d770f89
F20101123_AADHTM kanjanarat_p_Page_003.tif
2a443257bb99409b32ba1854ad5dc8a6
7ab82d768153346b0bd198e99d83f0a1f73d015d
1108 F20101123_AADHSX kanjanarat_p_Page_117.txt
0c1e49dceb74635c18a28b62324c3796
3688b2640bb3beefd2bdebbc85502a94be461a5a
10861 F20101124_AAAAHL kanjanarat_p_Page_169.pro
1c6d201c12675c22451f7ba26180b6d6
d496645eae3c14ec86f72fe547873d7b72eef3cf
29309 F20101124_AAAAGW kanjanarat_p_Page_152.pro
7d9ecf6d14b9544be515987ee26608da
5e9f2ac19ca7848716aa5f0b53282bf70e3761f0
68118 F20101123_AADHUB kanjanarat_p_Page_123.jpg
ffb35fba7a5c93255c9dba6d1b274b30
cb337532177b00a0a192fb1ed0be215af256813a
5280 F20101123_AADHTN kanjanarat_p_Page_004thm.jpg
7104e6931d81eb8d64bfe415f7ab74b5
fa7ff07aa7a58fd8f6223941e0436f1ad8937742
90523 F20101123_AADHSY kanjanarat_p_Page_082.jp2
b9f7f2e3c246b33027d2fb27c1cfda63
c67e29d5a48ea1cacbe247004eda2c678e41fd9f
41315 F20101124_AAAAIA kanjanarat_p_Page_191.pro
6eb17b6dbef925b4c0e1f9a64470e8bb
3dbd0efbfc0b6004e9e55b3a66f1790b8d6726ec
52274 F20101124_AAAAHM kanjanarat_p_Page_170.pro
628046c1a42f152e935fb032e2796b6b
4bbcf4d0dea3d9d9181669b75fc21e79bcffc42a
17468 F20101124_AAAAGX kanjanarat_p_Page_153.pro
b29133519b5e7af9f76787336b05488e
0ef023ae52cda7d1a8463807e3de84e3d0d3d059
6367 F20101123_AADHUC kanjanarat_p_Page_052thm.jpg
eb96b7da5f92f7c988e62d8a226db8ff
56694906d8d831f71ae83780ca1e239a819b08bb
F20101123_AADHTO kanjanarat_p_Page_138.tif
4f2ac98ab26070856efa3b9ed397b54a
e85e33a659106579ed8b769f6c11dd7cb623dca3
22362 F20101123_AADHSZ kanjanarat_p_Page_052.QC.jpg
52c6772c401996d24accc4bed6a461e5
f223f190c7911a9504796d587820035661d927a3
24833 F20101124_AAAAIB kanjanarat_p_Page_192.pro
7dec1053817608d2be3745f2c124b9a0
665be8bfc710da5c18de1d52198192a61decdb6e
19129 F20101124_AAAAGY kanjanarat_p_Page_154.pro
b3c335d2761d52c92013010dd0fb6425
c4d89b9ff4c27611434ea389189102d09840b4b9
1826 F20101123_AADHUD kanjanarat_p_Page_058.txt
009c5ffe76c1ac449211c96eb8fe27ff
d75ca475985f4b546106c3e60f1878a4996b87dc
6269 F20101123_AADHTP kanjanarat_p_Page_033thm.jpg
5733f51f40d026c373fae9a150f1ed4e
fb4c140e736dbfc889b906ff3a8e906b03f047bf
463 F20101124_AAAAIC kanjanarat_p_Page_001.txt
458fc29acd5d723063cca8615ac2e490
8c1874337eb64259905c8c3375e1a198a19ba5ec
80493 F20101124_AAAAHN kanjanarat_p_Page_171.pro
a92f72e73b175575abd0238bf8d10f01
85eb678fd1cbae7d3edbaf090557f9e55c08ae33
41006 F20101124_AAAAGZ kanjanarat_p_Page_155.pro
b7ce808031872e05b4c62d74a885e8f7
fe734e443e0b4d1c36c0af4e0737a4d5137a9160
F20101123_AADHTQ kanjanarat_p_Page_075.tif
21ee971cbdb67c87c9bd307aea3d4cd5
2264f1cbc71c07df6b9ecb7d4c40cab84a0d0a45
50765 F20101123_AADHUE kanjanarat_p_Page_067.pro
cf437a06fd9ecae4249f1c9ebd9138d6
933346b9fbbdfb918841262b177829a798fd5991
117 F20101124_AAAAID kanjanarat_p_Page_002.txt
e14301a66063a5d5b2482b957cfd7d54
9799bc29af1920ba81a682b03f704e07c9fd99d8
68672 F20101124_AAAAHO kanjanarat_p_Page_172.pro
49ed5728ee0861332f82c5236a1aa795
f061bb4f36619db922299ce2fae70805e8f4f02d
137332 F20101123_AADHTR kanjanarat_p_Page_188.jp2
12e81e0322e8f7424368721ac3b9caad
7003f890166101030713319c8190d86c5ad2374f
F20101123_AADHUF kanjanarat_p_Page_121.tif
b01f55800c7db198b472da80b3dd9788
c6097f53612be06a90686bd242e514999b50e75b
200 F20101124_AAAAIE kanjanarat_p_Page_003.txt
7ee2c8a4c80c1eda780389a41527ab5e
8b4692b9ffbfa6600e2005e777341a6b847bb644
69065 F20101124_AAAAHP kanjanarat_p_Page_173.pro
03b2c9e3b8713ffe1f644beec3803539
972b543446004824e39dc0f7fba4232b7840c8e8
F20101123_AADHTS kanjanarat_p_Page_114.tif
1a186ae7b0ecf7ad0ca64e38c9455b7a
124cb663d3a4431a6f11a75b50f625d59ea8cf2c
68843 F20101123_AADIAA kanjanarat_p_Page_056.jpg
a149b99b396edc71f8393fa146f0228a
d0921bfd86c7c7bd630f499f5699c361913ce62f
15037 F20101123_AADHUG kanjanarat_p_Page_154.QC.jpg
24f74b61e124c2649f0b93e47eee6524
08fa58f2401798eaf560c475bdbc0cc1e3c01529
1535 F20101124_AAAAIF kanjanarat_p_Page_004.txt
3a86ea5bc09587aa08c1c5461218b426
67e7b6ee0da72d0c90dc9b9fe6318e2b1c011ef7
57888 F20101124_AAAAHQ kanjanarat_p_Page_174.pro
613db41d2052c7d436c8f500b36797a2
9ffd0cd81da7edc02c3e9dad104fa20e48d8e370
92579 F20101123_AADHTT kanjanarat_p_Page_151.jp2
c4b10e1767f0a0d9bb066023044b2a84
b89dbbc5322495d81391f0212943ba78499f6071
2860 F20101123_AADIAB kanjanarat_p_Page_173.txt
336ca9a4d7e8090b338d0f10e5c13770
70e8a5c769abb94e2383d09560cf0c14edbd50e2
12624 F20101123_AADHUH kanjanarat_p_Page_003.jpg
2e608fba60261fea897c75275c0a6aae
aa47086575f2619694012dcce5476e2d3605e23c
353 F20101124_AAAAIG kanjanarat_p_Page_005.txt
ae41b43c5a19ae3b9c138f9b964034a9
2a01b889ffc351b30c6b5b2606069eca1e05b60c
68316 F20101124_AAAAHR kanjanarat_p_Page_175.pro
dcfdbc1d3294bfa9be795af8505fdfbb
4f27a504d29eb7628371c04acf9eabd96ba5f5cb
724111 F20101123_AADIAC kanjanarat_p_Page_166.jp2
68a38fea67b4d644fa2d2be5d62b9c84
04382bfb6157089ef64eaf35901d756f30c4458b
58657 F20101123_AADHUI kanjanarat_p_Page_181.pro
517f5549abf71ce83254fa92872c8264
7a0b774ea75a2872a793f89e0926b346884c2c71
3220 F20101124_AAAAIH kanjanarat_p_Page_006.txt
d3683b6efc0d779887111290c75a759e
8df8d9bc671579ea845ae400a6e64391b30413a5
54425 F20101124_AAAAHS kanjanarat_p_Page_178.pro
ce75741e85ad0d711501e9bf5f4ed9a2
78f33fc298edbc7202e17e006687e27f160361cb
60507 F20101123_AADHTU kanjanarat_p_Page_187.pro
7fca44e5a39569b1b1ea8b46a6c9335a
ee745dfe1f640f5213349a782215c9b45abd98af
2736 F20101123_AADIAD kanjanarat_p_Page_094thm.jpg
baac67c09b5a548ae633138b52badb2f
f3367ebfb9b6b25e25cd9845f7005c57296d47b5
24927 F20101123_AADHUJ kanjanarat_p_Page_189.QC.jpg
f7ea34431189ce26af4707328c8ec957
b27a66a47d4eab3858b2a09c215eaefae590e70a
4692 F20101124_AAAAII kanjanarat_p_Page_008.txt
741e3d1f815f3db46e8a92cf88596cc6
88e8c71707cf2fe7f7351b15fa50efd8097dcd08
67569 F20101124_AAAAHT kanjanarat_p_Page_179.pro
5794cb03794aae5b17712645b3ca3057
3d6b6feb7983f4733a2679f6d51a3629e91f0807
F20101123_AADHTV kanjanarat_p_Page_044.tif
1b670e0b6e5629c8b89e2ba92ce3f021
c57f7616b2f9bd0b7407ef8b2256cf96967d5f46
F20101123_AADIAE kanjanarat_p_Page_162.tif
9941aea89d09c49bf397e620506fa4a7
3d1ab40808fdccd0218361eca4cdb5b394b09d1b
57916 F20101123_AADHUK kanjanarat_p_Page_076.jpg
f1a93ca929e1110fb47b90dfa184c565
576f6830cb4731573a1228651660086c8c8c6b0c
1209 F20101124_AAAAIJ kanjanarat_p_Page_011.txt
fe06d5e922cbb95b4722f0c62b3ba50e
f0ab86e0b2e34e2d97c4f9868e12aff1f035cd27
67170 F20101124_AAAAHU kanjanarat_p_Page_180.pro
23a507ab4449bdedbbd25362aa2bdeee
6d55201a76df283676d9b2ed4bc7f262387ef23d
1776 F20101123_AADHTW kanjanarat_p_Page_103.txt
08b5a8bef07f7d4205ea781a9715ecd0
9c1228b4162288adea6040aec933678593a3a737
1998 F20101123_AADIAF kanjanarat_p_Page_030.txt
63b8ae837afe38d28ded8f23f3ad2dac
290ccd200162972b8912b2a9f69413e548fc8c87
1661 F20101123_AADHUL kanjanarat_p_Page_129.txt
78a4a69e085b39ef4c2b5d06a78c9c64
362818e7031de851a8c5b324a20a62b62c8f510f
2417 F20101124_AAAAIK kanjanarat_p_Page_012.txt
87cf9782e1315d0577e6b5333e70401f
3f8a3bf12194256063fa50d3bf9d5effe5a942ec
59599 F20101124_AAAAHV kanjanarat_p_Page_182.pro
b84a81d7c6c7b0bd61a6ae4a4f24ba12
0fc8a39d6e5f5c6799a4167fe7ff214f68975f46
101413 F20101123_AADHTX kanjanarat_p_Page_147.jp2
ecac272cd5e31c41c95cc160d17ffd4e
3c275c5d353c6661086ea082ec8127605757ec22
6732 F20101123_AADIAG kanjanarat_p_Page_048thm.jpg
b33951a39b0de1858d49b10ec3c97094
699e8336f08823bdc11c22fd0ec9e8f9cfca8585
15465 F20101123_AADHVA kanjanarat_p_Page_167.QC.jpg
fc538305d7c0700fcc3a1d02706c0179
bd879fcade014c5131d4c2c83dfb09c733e21321
50086 F20101123_AADHUM kanjanarat_p_Page_116.pro
79bd10f2fa4a26225d77cda24d6bc318
73ab6c3715417439871b22606c260a4c15302faf
3123 F20101124_AAAAIL kanjanarat_p_Page_013.txt
14a9cf639b9578730e323ec408d9e150
b06fa9513c5b54776effc026bf0bdd3b535340d2
58643 F20101124_AAAAHW kanjanarat_p_Page_184.pro
c06bb1bcda050ca62fa1fd93d11440c7
52ee3ed078745072148cb505cd3a4ae78b3ef723
66706 F20101123_AADHTY kanjanarat_p_Page_084.jpg
b27616ca28998e1cec37a576e395e558
c9d160d9e5fe59c9320193a544eca785b2f2f12c
52380 F20101123_AADIAH kanjanarat_p_Page_037.pro
7c676019f024241f51483446e40dbf62
8f7d932f64189999863cd5eba7859c4784ebaaa7
5779 F20101123_AADHVB kanjanarat_p_Page_098thm.jpg
685397ce2432880c35d5f52c0bfba150
d76b9b5039494484c2f37dcfe4406644c97abd2e
62804 F20101123_AADHUN kanjanarat_p_Page_151.jpg
7488e4969dd3466958e91fae720f88d1
a1890f6358c6ff48e7375a334dceb7c65eb29fa3
1992 F20101124_AAAAJA kanjanarat_p_Page_033.txt
310fe1d10966cfa69b9aaedfc18f9480
fdbf3235f7a91f033afd56c642a3f45daff32701
1607 F20101124_AAAAIM kanjanarat_p_Page_014.txt
017527069b0ac4661d2aa1a207ac3c15
c4a10510f0c05d2039df86915be1f84420f99565
64872 F20101124_AAAAHX kanjanarat_p_Page_188.pro
6ddfae54a4718e1939c4b6c49c2e34bc
4ea8d45b00cc277366df8d50e771e56c41d9ea9a
F20101123_AADHTZ kanjanarat_p_Page_127.tif
898af364becc9d2a8e195fac2ec64c14
6a3f01af5efa37379bd056474c9be78855b014e5
20006 F20101123_AADIAI kanjanarat_p_Page_073.QC.jpg
5794e065d8736c469bc5a6bb74817a23
851aa30103276fc3a8009769d9787d4d48de19e6
111484 F20101123_AADHVC kanjanarat_p_Page_037.jp2
eaa813b89450526f6d31f5cb2e931660
b81d651fec45f7e54c50557b3b9ea922bff313ec
24137 F20101123_AADHUO kanjanarat_p_Page_142.QC.jpg
b6e180e12e584e86ba35349ca612fdde
bec7b433f97c975cafe4f62e068bd751f076d4a1
1927 F20101124_AAAAJB kanjanarat_p_Page_034.txt
874d2c31607d4ef11de8a3bbd458ea6b
da855d75dc9c2cce44f9061bf3f481ae8ef43235
1547 F20101124_AAAAIN kanjanarat_p_Page_015.txt
c76f441ed27d6a6618b6b0f50217839c
d0db099186a0e7c24dd3857141d32785d0d9169f
60600 F20101124_AAAAHY kanjanarat_p_Page_189.pro
7983a9e6fea68817fc8883ccd387be0c
dc0e22a51ab5c365e36524953986a07c38689a4d
41215 F20101123_AADIAJ kanjanarat_p_Page_177.pro
99344ff29d1905df147bb4a0ed692081
ccb24a77594caad108e895ed449f4ded4d6c4436
F20101123_AADHVD kanjanarat_p_Page_051.tif
1fbbeaa978615debf2d18e5e28ad9c03
f67fb60a67cdf0c919ad447e5368a0068d1d7b9d
72153 F20101123_AADHUP kanjanarat_p_Page_050.jpg
ba45b70ed921a88a9e3a2b93a2bedf9b
71f2b5e85dd10adae60e0a2d79710c285e14b444
F20101124_AAAAJC kanjanarat_p_Page_035.txt
a6889d40df17ff5ff226973fa03822b4
60904e8dc5d245e15e5ccc84c7fc02dc8f7762e1
60087 F20101124_AAAAHZ kanjanarat_p_Page_190.pro
1f61e09673795b1c7ccd57b3e71a7461
25353a3fc753ba224789e37534e1a991b4fcb9ba
62310 F20101123_AADIAK kanjanarat_p_Page_110.jpg
7bab01d2569f0c9cf1826bcdf40faec0
da75388c61dbfb3edf9c53325668a8a949d5b1dc
33076 F20101123_AADHVE kanjanarat_p_Page_160.pro
0efecde4dc559e6664ace6806f4f4fa1
17bc4fe9dee77f90c475a32c6cfc25ac9c25a968
58614 F20101123_AADHUQ kanjanarat_p_Page_164.jpg
b2f4440a4b8e089ee92609bdc0966424
3a022fde1bf9380d589d817eb40f7be39abb3d3a
1717 F20101124_AAAAJD kanjanarat_p_Page_036.txt
91d2490a72524fbc06797d4dead2ce27
b1f0c3f9c827e4a6b4091ce03999fd04b3aa111c
1715 F20101124_AAAAIO kanjanarat_p_Page_016.txt
68e562f1ff5031514bcb79903923cd1b
3fcf4f3848a2cade357b3dd1e63956bbca6e5761
703904 F20101123_AADIBA kanjanarat_p_Page_122.jp2
9e8e108599cd4708d707ef5fce704e10
60eae4144264f3855c6a4ffff79f0da64854fa75
48443 F20101123_AADIAL kanjanarat_p_Page_046.pro
535daa2a379fab9a35a344ba385a5717
d901bda25b686d6d60f9ab9000d3a1d1c3c10410
F20101123_AADHVF kanjanarat_p_Page_006.tif
b5ac95c8bfe2976fbd0c058afe3b1d7a
2e95beb51ea614697770a3365ba9d33acd6c85e5
6680 F20101123_AADHUR kanjanarat_p_Page_181thm.jpg
a339258217d05b7ec318c7f4fb255e84
ef1df28e573ba9844c80384b6dacf2921bc27536
1961 F20101124_AAAAJE kanjanarat_p_Page_038.txt
dc61c6135ad3068f5627e7416a5bb1d9
2934d70f7e85f533ea55a6e6511c907ef34bced1
1913 F20101124_AAAAIP kanjanarat_p_Page_018.txt
53c1c2e93df54fe4970a743f807b9b00
a95143e933cae32c4bc24db8226b77ce60f1e268
72069 F20101123_AADIAM kanjanarat_p_Page_051.jpg
fc72cef333b61f851d7ec5bcc18e670b
71bfc963630cdcf63cdc5fe1a12e767c12190186
62461 F20101123_AADHVG kanjanarat_p_Page_071.jpg
22617057c803be744263ad8468018e8a
3cff05c0509c23fdc46a814df5dafd20764322ee
60182 F20101123_AADHUS kanjanarat_p_Page_010.pro
d473f84ee9c2bbda6336ee97dc79c706
f26bc8908bbfc615fc6a618ef534188cbe0e37c5
2125 F20101124_AAAAJF kanjanarat_p_Page_039.txt
b9c98e23dff61ed941f98a9440305411
352485f3109fac82451d9309e96778637a48c837
2023 F20101124_AAAAIQ kanjanarat_p_Page_019.txt
cb38547b211aef48cc57f4ed593ffe29
b77634841e9bebbe01236db36b90e19c350d54f5
F20101123_AADIAN kanjanarat_p_Page_084.tif
338da6034b891d860bd55af48c036627
4fd906d5bf10cfd5436b62d45aa170cec4f82f3d
18636 F20101123_AADHVH kanjanarat_p_Page_098.QC.jpg
6030a9b6acc6050c3162c50a0887109f
5e7c193f54ac6e775a641e4d123665cbc1ee6e82
50771 F20101123_AADHUT kanjanarat_p_Page_162.jpg
8518f0d6551deea3e8201cb72059b347
d5a779186e3e0cb3abaaaa9a6c640b356df33829
59869 F20101123_AADIBB kanjanarat_p_Page_165.jpg
ea47a46c6c9397a8595b91f9ac83b330
3722cb03470f5ea450faea5014c21e3f8ad4d59b
1924 F20101124_AAAAJG kanjanarat_p_Page_040.txt
ceed5b702efc41ff09922e7fa31d8317
5a061fcf55c16e3e6c24df20b4f05a6e91f22fea
1711 F20101124_AAAAIR kanjanarat_p_Page_022.txt
20731456db5f65d1dcbd7fb220f0485e
f90ba0b7f00b092905d0867e395d998fc71649ee
48271 F20101123_AADIAO kanjanarat_p_Page_018.pro
474ac7abe20862366f4c41f54c46f157
ac858c95a3f2f0017e3bd4362f827993ff62dbe8
F20101123_AADHVI kanjanarat_p_Page_021.tif
969f3b94e5d29153a886ca2e2f4b9d31
97800cc0c5e57d1c610b774a3985328f88182cd9
2471 F20101123_AADHUU kanjanarat_p_Page_132.txt
a422ccf0950bd76fdaa92112c5de9844
8bcc92cb3e7889e6edf778743db8b52090ba1b10
2342 F20101123_AADIBC kanjanarat_p_Page_072thm.jpg
ed8fe572e3e10e8ba1ef1adbd94f4931
dbf2835d8f24d3b792c5cbc1bc88c7f779e5332c
1000 F20101124_AAAAJH kanjanarat_p_Page_041.txt
daa5c9af728a30bafb9bdda971e0179a
54c4635426e685af30d0f67f7ca651efbf324bc3
1827 F20101124_AAAAIS kanjanarat_p_Page_024.txt
241baec71569d2fe606bbc9e51031142
a8bf8bd0f3f7a61f3c0b49145644ad01958e5f24
F20101123_AADIAP kanjanarat_p_Page_069.tif
56954703d2121264815f57947cafff3b
e2b8f4c3ccfb642582e8fe759e984c42503b3fbc
110274 F20101123_AADHVJ kanjanarat_p_Page_101.jp2
e9fd2de794c57bbb59de4465019c4932
7be9390e7829344f93a69b6d0624097125e799a7
106586 F20101123_AADIBD kanjanarat_p_Page_171.jpg
6c7154ccab25c8a49fcee911d072fdcf
4d83d205746ef69c0e1b19c5d26e7bd9904a63df
1895 F20101124_AAAAJI kanjanarat_p_Page_042.txt
5d62916a2fff0832a8985ecfcdbceb53
24553311a0499bdd52ec0aae9cbc9f5d0cdbbc01
1834 F20101124_AAAAIT kanjanarat_p_Page_025.txt
33d15147def156c401ab7de1c2318e82
d464eaa69fcbb58348486c0fa61194dab359d054
100911 F20101123_AADIAQ kanjanarat_p_Page_028.jp2
b5a289ea9ed73be4d4bff33935edf483
2b0559f2c323174fb94620ebfcad63a2051f0a50
45693 F20101123_AADHVK kanjanarat_p_Page_047.pro
a8130b86377b8d90892883ef0718088d
e8718026c749fe0524ecd3ed24e2e958d07eb8d6
69531 F20101123_AADHUV kanjanarat_p_Page_112.jpg
930bd8779d9993d819edac7d574e9e02
9dfde2ab3712a050b8faeec4eed51d92cbf5eb42
F20101123_AADIBE kanjanarat_p_Page_176.tif
bd7ad909b26d440ea4d08b613a3078c4
d3e67ca1daabfe1e5810ec077da55a0d94ab7127
1999 F20101124_AAAAJJ kanjanarat_p_Page_044.txt
f874359b4f2eb575f63b4921b5a7a419
5cba7ae3bfb8a4501bacfc14ecd30c396cd4cb91
1931 F20101124_AAAAIU kanjanarat_p_Page_026.txt
56dd26a6920ab55e6c36674978e314e3
c6568939eaf7cea34275947ed2f8fcefdac8f7b3
2378 F20101123_AADIAR kanjanarat_p_Page_184.txt
29a93ed561f468265c33babbb50af323
04205378957afad9156242594de97042e0800f25
F20101123_AADHVL kanjanarat_p_Page_110.tif
1ad9920ffa3007de92bc141a4f3c1cb1
c661e7230db490e4354204b148fa44ab18b193af
6233 F20101123_AADHUW kanjanarat_p_Page_039thm.jpg
aebad1570f108e8c1372d10bb17f2dd6
17e8d012b2df794f55e7d87f01e0492c04567145
849935 F20101123_AADIBF kanjanarat_p_Page_118.jp2
9c57b16d4c8e6a779884bd70be930360
1caa1e2d7dc65c9c3c064d73c65700e9e312656a
F20101124_AAAAJK kanjanarat_p_Page_046.txt
e4a2c44a19583d83e1cb69beedb54a1e
1d00abc5adfc08acd10ed230ea73360b3e229655
1975 F20101124_AAAAIV kanjanarat_p_Page_027.txt
2ee6ca031d8fe428a827934b1f4ef15b
827c7ddfcc594e651a36af2c965599211c614cae
F20101123_AADHWA kanjanarat_p_Page_058.tif
54068103f44b48621a533b8fc2e5add5
11df746ccc66b5cd4553dd44695f33c2f1f9f027
2258 F20101123_AADIAS kanjanarat_p_Page_043thm.jpg
6ff7aa5d1d1928a5190224a9f7226156
a784b35af16d6881df6b2a0842e9d77adb7090bb
7135 F20101123_AADHVM kanjanarat_p_Page_184thm.jpg
df75a0de14e43a695e95e7268e513983
5187fe77d6676333c8626b1e596a1e1db348644e
60824 F20101123_AADHUX kanjanarat_p_Page_176.pro
f155122c673e8e411770085f32f84c45
ea15f43e80858ec32b1bbc21f4fd976ba8fd22b8
22478 F20101123_AADIBG kanjanarat_p_Page_141.QC.jpg
0c68984323ec1769d603e34bc230f0a4
85a83ef208dcc58f4e0e60612819989648b693a6
1814 F20101124_AAAAJL kanjanarat_p_Page_047.txt
0a011ee0cfb69e5093c3b681cc5b543d
bf96017aaedaf2731257fa0821f9ce13cdd43c51
1847 F20101124_AAAAIW kanjanarat_p_Page_028.txt
ada2799aa20790d94310200139738fe6
4670f920325c12dd0ac6163707fad41e4f3147b9
9389 F20101123_AADHWB kanjanarat_p_Page_003.jp2
945debdf33e9751c10bb228de5be2b57
489a61329c399a7ef77433d2f2a0ae9cb7a8a251
16031 F20101123_AADIAT kanjanarat_p_Page_099.QC.jpg
f8c12e9f7508daec4f196f3deb502d71
b4b22a0748a2d62cdb8008c117c7e8c6783d5ff3
65411 F20101123_AADHVN kanjanarat_p_Page_054.jpg
5840d625247ae1e256b93c71617f90c6
de529d2150cf26eb3142e04a4e7aca5b27e8dfe9
71154 F20101123_AADHUY kanjanarat_p_Page_129.jpg
492a3c663c96a5f4f0605099cdadd4f1
af3437d934db140b7f517f118e9e2c15e21210e8
22450 F20101123_AADIBH kanjanarat_p_Page_081.QC.jpg
fc4ad2d12ba7fef20c6664be6a057e37
ed5ede6d2c38eef98f369dffdad447299baf91ab
1867 F20101124_AAAAKA kanjanarat_p_Page_068.txt
9ce9c2822008f616a291fd71ea1318ef
b6ed46dbc9dd759560b9eaeb73049516d72b0d84
F20101124_AAAAJM kanjanarat_p_Page_048.txt
c0bb9c684bf7f12bab80adb762e71637
c6b2c309bcd7a4214499ab4c1ce184a82d929f2c
2049 F20101124_AAAAIX kanjanarat_p_Page_029.txt
f6d368416642cc22cf032bcf319b9909
b3daf06702e1de93102dfb513678d81126626916
104097 F20101123_AADHWC kanjanarat_p_Page_190.jpg
f30ad851b61e5f527663b0b2c30530ce
87311e28fd7a2df6d3513a319443636b18995118
F20101123_AADIAU kanjanarat_p_Page_013.tif
8f98c9034f4bf007cfb879858afe31a6
a6047baa08a792b35598882698403731672a8515
1918 F20101123_AADHVO kanjanarat_p_Page_149.txt
2fce88284c084fdb5001b0600acc51ac
bb4f31d8ea9dd05b329e4d7ec3b7e7b150ecc68d
6592 F20101123_AADHUZ kanjanarat_p_Page_146thm.jpg
3fcd66f4a0ff6d212e6867753e980073
26856cabb07051beea21050205bd24ef1dda8daa
21689 F20101123_AADIBI kanjanarat_p_Page_123.QC.jpg
b295a3f6db96c95f0600dbeb4b6b3617
bf45f75933149e8601ce92978df0e9e49460e027
1953 F20101124_AAAAKB kanjanarat_p_Page_069.txt
5ca3c51f6c1b51f6d15fa2bb4dd64971
4eccee9674775034c64c07dc23591b5a5e5e5692
2072 F20101124_AAAAJN kanjanarat_p_Page_049.txt
b9a5bdf0c723fc1c92e11cabf821788b
d7e1a543ddfd5e199c1f4f11182968ae873e9c9c
1666 F20101124_AAAAIY kanjanarat_p_Page_031.txt
5282b2d3307a1155a147281ce0674d1f
d5d7e0a0b47c681a4e515c995be7ddce7ee31b98
522284 F20101123_AADHWD kanjanarat_p_Page_156.jp2
f1c2f15e68677665e448acbc3089a336
937e8e9c92e21bff8bcd5a0f1004fd49c37cec9b
F20101123_AADIAV kanjanarat_p_Page_089.tif
57fba9637fcde1d1c95fc01de69d76a3
fb38f0c5070c41ad4aed2ced89e5d54b0d29d7db
41982 F20101123_AADHVP kanjanarat_p_Page_063.jp2
f799e17690715e08ba8d81a9a451a517
3aa41044d1e43a0e51e14844e5bb05082fe7022d
64147 F20101123_AADIBJ kanjanarat_p_Page_157.jpg
ab527ff4f887fcbab64e671674999289
865ac58ba60f3b80045486a839ae9be760b52bb4
1778 F20101124_AAAAKC kanjanarat_p_Page_070.txt
f4f96f3b8d40d980368ef8b0687f518f
30300a74d8381702385745361339060d5c0c1fe5
F20101124_AAAAJO kanjanarat_p_Page_050.txt
fa89886c68f532d5d1a3587002b970a0
9a7e11561dccd973d10aa82c785ab73628f44828
1562 F20101124_AAAAIZ kanjanarat_p_Page_032.txt
4bf83236a3cd199f9288ecf7a75a07da
8a488a9c1aec0d8e898deeb38bc074f06b01c567
1051960 F20101123_AADHWE kanjanarat_p_Page_154.jp2
56135d794e5b9c1a34751540d653d4f2
6064cc28480087793a00d1e1725a4219cb8e3452
1392 F20101123_AADIAW kanjanarat_p_Page_108.txt
2a7c84a5e7f2e7a791034ec164042552
a4ee6a7b979692679efcd7695ae68f5d65802812
43339 F20101123_AADHVQ kanjanarat_p_Page_011.jpg
6336096b4a0c7133e8988a1cd47a2653
65128f7b0dc9630bd2f234b5485869d54d7420b9
32680 F20101123_AADIBK kanjanarat_p_Page_167.pro
4a100f9f5088470629c65869a48c9f69
757c9652a3a63c6d43f99e52e971546d30ac39f6
1663 F20101124_AAAAKD kanjanarat_p_Page_071.txt
f6d08a063812895049419c826eda318f
a71e0d657349ecaaa195583ba6711ba45bc02d5e
2067 F20101123_AADHWF kanjanarat_p_Page_037.txt
e24a490859da4882113934c2a6c44b4d
54eae0dd3db343dff407bd5094e4f54881f3662f
39397 F20101123_AADIAX kanjanarat_p_Page_100.pro
d2c09733c9351562ea952bdacb3c1a00
30dea8d758f3972300613a9cda3afe69cb4e9acd
1463 F20101123_AADHVR kanjanarat_p_Page_023.txt
f08ce9ad39ebcb368a8697853f0377f2
119b954ed52132e1d33d01d3fb75e8a9bc978480
F20101123_AADICA kanjanarat_p_Page_115.tif
2394d8b8414a891fa627fbfbf13dc5ac
107ab3d9e006bc28ea0c5b2e072fbcb6066fd078
F20101123_AADIBL kanjanarat_p_Page_168thm.jpg
ede5d7a9464c7b8265dce1d552d58828
b7d5ff8ccc50bef0972933d418d751150fce1161
1769 F20101124_AAAAKE kanjanarat_p_Page_073.txt
00e77615875bdcbc26c7b3e750ccaa88
d3aa9a661ca38c83976e1c49a679df1823b7722e
2079 F20101124_AAAAJP kanjanarat_p_Page_051.txt
1ba2ced6ba58e09cfb9389e1d5dc7002
82fa04851924e92784b0d9d0824ef857408efb20
105288 F20101123_AADHWG kanjanarat_p_Page_038.jp2
f4cb4f3b9ea3500db0a24b7f6740334d
4e1fea8db6a81cd8755b2fd4a92b4033c1661dbb
83599 F20101123_AADIAY kanjanarat_p_Page_115.pro
2ec448ead9409c93b00f80e0f2d4c7c9
bb792ee49dfbda678b3fe66f76c44bb2760d1bf3
639999 F20101123_AADHVS kanjanarat_p_Page_162.jp2
904ed7d757bd0094f9423a06c09d9066
2819e5b7d6d96a06a6c9428ab2f0dff147a394a0
98847 F20101123_AADICB kanjanarat_p_Page_047.jp2
0728d16a33087806ccf746c3e08344e7
2b7b75eac94050d245363b59cda269998c114cd0
2160 F20101123_AADIBM kanjanarat_p_Page_080.txt
84ecfec487a4b96eaf9b1580cb3ba45f
22310c9ccb3a2809830f4aaff872c3c4c77a1d56
1285 F20101124_AAAAKF kanjanarat_p_Page_074.txt
9176103a7dea7142cb9e6949e698111c
2444da3f2587772bbe10033737f2993c0fc6243d
2019 F20101124_AAAAJQ kanjanarat_p_Page_052.txt
52e54dc866dbd6fc6eba0e8d406dd504
9cbdaade506551be19840ef56cb5db72d5895054
60357 F20101123_AADHWH kanjanarat_p_Page_185.pro
c5beda7edae7a81d218cc1caad62acd0
28ebc5bc3a68e9cb4e5b5cf07df3f302a8d7f8e3
2452 F20101123_AADIAZ kanjanarat_p_Page_010.txt
aee7cb4e64e57daf9df0967076e71057
807e4c58d615e37a047e024d935ba97d159d6e3f
58252 F20101123_AADHVT kanjanarat_p_Page_168.jpg
6f5987f0f3e79251e637557595b83792
bb83b5f8c47b6e90507901b17a97427107dde2a0
53229 F20101123_AADIBN kanjanarat_p_Page_039.pro
02799bdc03b662a7f1e686311752861f
4bf4d449da138fa79c388726ea57d51994fcaaa1
1696 F20101124_AAAAKG kanjanarat_p_Page_076.txt
98604a2d43d79590f7a556314bc35178
e48a38d3cc59f0374480b1677578db4691a66d63
1912 F20101124_AAAAJR kanjanarat_p_Page_053.txt
84e984773d2a0bf5ef60178b639896e9
23b394ce73f7148ca293a62749c26da785f6d9b9
4168 F20101123_AADHWI kanjanarat_p_Page_164thm.jpg
0d121ab7a64a6e8fd5825a657d056975
5651631da38ee91f7897dc1e78565ebc82fd438c
16125 F20101123_AADHVU kanjanarat_p_Page_031.QC.jpg
d2aafe4ad8042d57e1e62a4fbd34fa1d
e885a883eb03ea7ed5a14629dabdf271be3b118b
46714 F20101123_AADICC kanjanarat_p_Page_152.jpg
302a26ea11fc5634aa351dbc79278a0a
b71918a20fe2a7bfb2494547f072387b72fa1bc5
2957 F20101123_AADIBO kanjanarat_p_Page_003.pro
62880f022d55105bbda13bf52ae46204
0cf7a2a49b30e0882cba0deaea96522f29fd46b1
1351 F20101124_AAAAKH kanjanarat_p_Page_078.txt
42b88f2c4ed1aa4c1c4d7566b57bab93
ebd11f963eba428e373b215409d2110781c9b230
2021 F20101124_AAAAJS kanjanarat_p_Page_055.txt
4f6be57aaca3c071d720d189b3a64181
1a9f69476e44ee9ecb892828d4c1d86411955125
F20101123_AADHWJ kanjanarat_p_Page_026.tif
2b4a99ff6f2d23383eb09b9bc98558f5
6cf1a8feff89d95948d4b306bd9b2dd708409133
F20101123_AADHVV kanjanarat_p_Page_043.tif
ea2121b3f0c7239a8e6e420de4d369a0
ac38b7f1e2e788570010a8c495987028f16f6e86
6365 F20101123_AADICD kanjanarat_p_Page_092thm.jpg
077d803d05b1d37995114d8b79d1b39d
145227b0665d53ecc51115ab1d593a2296c8c631
6944 F20101123_AADIBP kanjanarat_p_Page_171thm.jpg
976ca66fdfbf1a086b7493e4be2ac740
f7941947e514736e66b4d75c95245a9d7cffc36a
2255 F20101124_AAAAKI kanjanarat_p_Page_079.txt
771063c664c2da31c3a4dfdfade55cda
f10b2aa00a5740739b999ffcb0838c8e6dbb000d
1958 F20101124_AAAAJT kanjanarat_p_Page_056.txt
e511c0749652d8da0ccea69ec7b8c37b
716ff96eb832cbe60a36eada50fcdf5ae39b882f
F20101123_AADHWK kanjanarat_p_Page_078thm.jpg
3991d0261290af39d1df98c59fc284d7
dcf5da0cb830a96685e68f917f3c33e6dbf7d2f2
F20101123_AADICE kanjanarat_p_Page_085.tif
dbc14bdc845d4671e21049bd4bd6c8d0
ddc79caf278eaaf1ca1f40da2fe257d6debd4312
F20101123_AADIBQ kanjanarat_p_Page_184.tif
5ae677b7852a4991bdde06fb49d34497
2b70147a86ee9a38e6f67f99304a20ab3507a948
2052 F20101124_AAAAKJ kanjanarat_p_Page_081.txt
bb49caea3587033791ececba07eaa532
395c2ee5efbfc73ff6553533925c6a72b2670376
1877 F20101124_AAAAJU kanjanarat_p_Page_060.txt
aef2a57c1256dfb88484711c55e7e74a
1b85f01a5e840d8336063d5af4931085af3732bd
636682 F20101123_AADHWL kanjanarat_p_Page_009.jp2
71ed664084080d9658584bfc37e26a86
eba29d5a3fb8e4f66cc4e57925c86612c34092e6
65998 F20101123_AADHVW kanjanarat_p_Page_047.jpg
18e1e782811411b356e457e98b52c678
93d1768165cddc0d88525e9826fb698825717fda
797988 F20101123_AADICF kanjanarat_p_Page_083.jp2
6b555df0c102dda2a3c485a35608bdf8
240c69cca0f15f82b83daa5107f9819f42ff07b6
4463 F20101123_AADIBR kanjanarat_p_Page_021thm.jpg
b73789c132f92a7cd3d5a971d64ad824
6021fd104c24f38494df51c1ba7cd135b40b8318
1614 F20101124_AAAAKK kanjanarat_p_Page_082.txt
57f6b2ddc8286028fdeadd00327a33d1
56e984b9b05733334e2dfb88421196186bd7199b
1860 F20101124_AAAAJV kanjanarat_p_Page_062.txt
a216df62f281f6e1926655e5e45f995d
1d534969d00a4f2f221c6953c78f9dec71f7a4ee
58416 F20101123_AADHWM kanjanarat_p_Page_192.jp2
a1ee4f6be6975280dab9cd66c0624786
0887b97904febdb7d9bc1364c58d39709a5a878a
86822 F20101123_AADHVX kanjanarat_p_Page_100.jp2
7b1449e7493d40ece2862fe2f8bdad76
d49f0bfb3390bfdbad734354466417da8b451c92
22175 F20101123_AADICG kanjanarat_p_Page_129.QC.jpg
e08a7b00a8a143fa77dd7c8ba9b63660
a907db5c8f2103e09ff6d5ecf71b21fd8ff92a1a
F20101123_AADHXA kanjanarat_p_Page_141.txt
f8a6aeb0bac1d9f229a239d62c4d3125
950837e20d22e033963c5ea77ece722efc07cddd
5825 F20101123_AADIBS kanjanarat_p_Page_082thm.jpg
33796218d074ac6022771d3a0f773049
ccd96d0d8d850757e9212845a84d361f36b8babd
F20101124_AAAAKL kanjanarat_p_Page_083.txt
64ebf8e39a3d808f4262ca71cc18d651
9078e7b29627a7aebb80b33ebe89542bd03274a0
715 F20101124_AAAAJW kanjanarat_p_Page_063.txt
ae086355aa1ec2bd31e13cfbda78773c
e6012785ba26ce4eb076feb3a18030861fc23b52
57964 F20101123_AADHWN kanjanarat_p_Page_015.jpg
c58b5ba6b8e1c11564a91d5f92dbe80f
13d17698a4d0e3338d3873c617c3792a49f80ad5
2720 F20101123_AADHVY kanjanarat_p_Page_009thm.jpg
034046856cdebcb3dbd7efce54851509
13bb9f5afdb7b4ce0c50d69a6e28b81825c63757
17451 F20101123_AADICH kanjanarat_p_Page_157.QC.jpg
d3aa1a62d8894b1309094e55b6b73734
1fa46391663de86c57588acc0aadf048540d5a6f
F20101123_AADHXB kanjanarat_p_Page_040thm.jpg
02d1593acfa43425e18eacae1455f410
f75be99323315d148a87194edf6d5b081afb660d
6332 F20101123_AADIBT kanjanarat_p_Page_017thm.jpg
f4f894dc8eb876c5d47ba8d78b017fe0
0ba1dd458efa55ea45ca8c0ab9fbf02406e99002
F20101124_AAAALA kanjanarat_p_Page_104.txt
2e92e8824f13fe500742c86c378c2b16
f5f98c877045625bf3a1913097d5882dc55bc954
1851 F20101124_AAAAKM kanjanarat_p_Page_084.txt
2931cb767a47998c229de1834a9b87f5
e46c699427a0962bf835bdd35447f16dcf3ceb37
F20101124_AAAAJX kanjanarat_p_Page_064.txt
30a3815a80c6d77581250aa8e71a0d9b
aac04e86da2b3d7027f1936dbb198793f09a729c
6157 F20101123_AADHWO kanjanarat_p_Page_178thm.jpg
343382159ea4c5bf52ade9bbaeb9e110
973e78a11d93476bd33bf28c2ee0c040dfc44b46
68888 F20101123_AADHVZ kanjanarat_p_Page_138.jpg
a6d261380c0ba3a8a4150f0c06a94b2b
d61035e2e20cf9767a9d4361aba7b48557ed0c94
6346 F20101123_AADICI kanjanarat_p_Page_018thm.jpg
0dd7e450f543588e146435ac289f403d
c6036587ea30c59dc990dbbd92edcf9184a49b9c
1863 F20101123_AADHXC kanjanarat_p_Page_087.txt
a8157f72bacee63121bef5c982675c31
509fefc0d23e130b772d2744cb188c7ac954fa39
863953 F20101123_AADIBU kanjanarat_p_Page_127.jp2
697c5c2b58a4397592acb40e23800e6e
29645c578ae004fb7e208a5f91f87b18384ceb6f
1680 F20101124_AAAALB kanjanarat_p_Page_105.txt
897539125945acebd3377dbb1f4888b8
f9bbc30df487498cb341677da729ffcfdd353eae
815 F20101124_AAAAKN kanjanarat_p_Page_086.txt
173956e096713ace2dc1ccd275b8ecdd
5dfa149b2d3abaded6efaade6d09729934f8093d
1939 F20101124_AAAAJY kanjanarat_p_Page_065.txt
4ee3d3adee2a55f9356fb73b1480f89d
2762858d172db5de8d44861c63f4742e532a7eee
49023 F20101123_AADHWP kanjanarat_p_Page_109.jpg
7fccdccc4e7e16c94afa0dff320cdeec
c4edf9f7857629c56781d18b45987d7b913b37eb
1015 F20101123_AADICJ kanjanarat_p_Page_043.txt
7079fdb5b31d1c2a0e47fdc04b4a4d20
9391ce1cc8517afb19a10364ab5e3085ce31a480
23329 F20101123_AADHXD kanjanarat_p_Page_183.QC.jpg
d7d005f48cfb5e652c888bcccc33d7f2
88fe9ec333f213e6d77fa25541f8b90bb30f3614
74616 F20101123_AADIBV kanjanarat_p_Page_137.jpg
09f9f25ce92ba46b4371ff2cef40138b
3a098bda74b61151d84573cb4fc3658b5b8cc7dd
867 F20101124_AAAALC kanjanarat_p_Page_106.txt
466c5c4d5ea48441a73d275a3f3805f3
b18b0f5278cb02eba741f22435bd1bb373675c5b
1910 F20101124_AAAAKO kanjanarat_p_Page_089.txt
cceb47b9b98155dd4f4cc26e66e28397
97beed747023957905856c1085df331354e3abbb
2109 F20101124_AAAAJZ kanjanarat_p_Page_067.txt
ea1d71942b9603d6055e5db92b684631
324cf54e1974a2c5fd3c33358d5a5ecfc78fa7ae
6917 F20101123_AADHWQ kanjanarat_p_Page_115thm.jpg
c9ae43f9095918b0504077bef37fe108
a217c30045e433fe79bc4a860ad26937d53a06b1
F20101123_AADICK kanjanarat_p_Page_142.tif
1a9bcd64c72146cb68c45050879fba44
017b4f716b45187c6bab52749c0914cded29c907
25609 F20101123_AADHXE kanjanarat_p_Page_185.QC.jpg
039896633eb1ef1d3916644f231bfff5
4d04065529818a1bd47b80d196ea35211215fb6b
4065 F20101123_AADIBW kanjanarat_p_Page_150thm.jpg
f85180d27e33a3fa1aecc267c23043bc
98a654d49f579df089ba04b562490e004f40c9ed
1807 F20101124_AAAALD kanjanarat_p_Page_107.txt
a3e95900561c7db3f3af724cdee9d486
9d01372a484675a059995cd138954f517576e0ba
872 F20101124_AAAAKP kanjanarat_p_Page_090.txt
7a8474870f499150821ea858362409cf
9ce6469c11a3844d3ea0301b11c6cf059f72a277
1227 F20101123_AADHWR kanjanarat_p_Page_021.txt
9bb61d327396f3f5ee219abe447e17d0
59e544f9d3935e2a520b472590526d625d8459f0
21672 F20101123_AADIDA kanjanarat_p_Page_028.QC.jpg
ae2751c1dd2aa2997e362c022ff002b7
a865b40cd8934faa0b443d01b04799a207abe5a7
F20101123_AADICL kanjanarat_p_Page_103.tif
077f0a4f236f1fd9c2497826761f92e6
6cf25f5f190f645087d2ecbdc10aa10c56a16614
17241 F20101123_AADHXF kanjanarat_p_Page_086.pro
8df5900ca8a54918b4d57e4c1fa63e10
baba556bd10f67e3ff173d3e70f78f544330b6f6
23341 F20101123_AADIBX kanjanarat_p_Page_146.QC.jpg
acfd098f1f14bf609d0ce181f1d5103a
7a7424ad49607cfaf7d5c50f5ca04649e95c13fa
1734 F20101124_AAAALE kanjanarat_p_Page_109.txt
779166d3a52fc577437c66264403bf0f
5c68aa9924553189a7a372c3ecdfb053bc0c3d6b
6072 F20101123_AADHWS kanjanarat_p_Page_012thm.jpg
4bce10cff72519bf6deddffca17e948e
48b17b58afea341094c0ac9a293d44fa01b1393d
51450 F20101123_AADIDB kanjanarat_p_Page_093.pro
2459197c905d2019afff63b5c38aa9c8
1397e448e08fe1d3c2b596331b24b8b142f7db6c
398896 F20101123_AADICM kanjanarat_p_Page_153.jp2
1a2e2d1a8007bed77ffdbfc8bd150ec7
09bbf2962b77024b3874374ad96caec00155373a
2614 F20101123_AADHXG kanjanarat_p_Page_188.txt
88d6470d8c9557cfa1880b9357ba36ad
10279cb16eacbefdd650659644982c1854b4b227
33686 F20101123_AADIBY kanjanarat_p_Page_031.pro
6fe20b093fa410bc872b51f2b923db09
2f595ec4dc405808a0ebe30556559a3b8a9cc7a0
1881 F20101124_AAAALF kanjanarat_p_Page_110.txt
2ff04affbedbb9301da8b2442ecef7a9
557200233323c0e2d7fcd2ef6145f0b1af41e642
1818 F20101124_AAAAKQ kanjanarat_p_Page_091.txt
5944a333116051e1fef3d745c1cc7ec6
059d67fcaf83497100f042781c01b62b37111eb1
30377 F20101123_AADHWT kanjanarat_p_Page_124.pro
9db56bcfbc19704a38641215d0fe565d
534c5601c0af95a9c11f475b510a0f12aad5f6d3
52780 F20101123_AADIDC kanjanarat_p_Page_049.pro
885b3f1596b627ab7bf8ab0eb0c1d983
894588135908d8ee71cfcd717088bdc4ab8d5fe3
1387 F20101123_AADICN kanjanarat_p_Page_002thm.jpg
fd7ba4550ef58660703e221e80b1b06c
a99fd8ac893473f56cf30b023074173948046b1e
43759 F20101123_AADHXH kanjanarat_p_Page_098.pro
ba1a3e5eb95fafd47ef3a701dc23f6df
c3f5e39123ee8095e5b9e6506e274c986b65b6e4
2147 F20101123_AADIBZ kanjanarat_p_Page_116.txt
6277576160c90808d7e0bf430f1282b9
bf92f9cdc0663c6554d6a6d9d7ecab53caca6ba4
1556 F20101124_AAAALG kanjanarat_p_Page_112.txt
4080ef1411a53ea7da58b3b401f13009
80bd9c1477c94e5a4ff0ce940bc037532b1aad08
2064 F20101124_AAAAKR kanjanarat_p_Page_092.txt
f180a673c2888d8eb3137d3089ad73e3
3fb2914058da34665ef6076010385685f18ee344
13161 F20101123_AADHWU kanjanarat_p_Page_090.QC.jpg
9d6884170ac274327749419a29c2d5f6
903d623b45b78358a16036b9fd828d32e12f453c
20255 F20101123_AADICO kanjanarat_p_Page_036.QC.jpg
2ca95f764a23e7dd081447a3fe592164
b7934739081eb590f2d681ee0c8cf899c30649f2
1031829 F20101123_AADHXI kanjanarat_p_Page_129.jp2
1d599ac1b914260fcd1ae53ab2948ad9
12fb6abcd2558ca80a4208d637281513e8672b2e
1460 F20101124_AAAALH kanjanarat_p_Page_113.txt
1a5154d2b677d2dd60b66c344760c2ca
2e710c8a66f62aafd987a8b5cd9abcab06a17fd8
2026 F20101124_AAAAKS kanjanarat_p_Page_093.txt
d50877e358f075f87bc29692712b7170
a386feb8d2129782576729e8b5b21cd518dc34bf
107867 F20101123_AADHWV kanjanarat_p_Page_044.jp2
c9ffd696e1ef49bbcc8878dde11e63ae
3ada68911344f1258088bdae8e8c68bd1f4a6f51
1651 F20101123_AADIDD kanjanarat_p_Page_075.txt
4fa5a2b0d4c1205a7ff8e66f95ec20e6
be90666d4e82080e370b4d0537143b2468256030
15690 F20101123_AADICP kanjanarat_p_Page_133.QC.jpg
3c4014cde5edcc57fdb3ea6e7a7bf8d0
39a083b13ec178593229cd2a471ac7a8623ffc9a
65141 F20101123_AADHXJ kanjanarat_p_Page_107.jpg
56b1cab421d9a1d15d6b6d61ba55a851
906456cbbc31c6589906a3c2a07894205c9f48e6
2500 F20101124_AAAALI kanjanarat_p_Page_114.txt
1eaefe953d98d47306387a4e110d3e3f
2d764b4d4fb39f5f6c1603243b5185a4fb1955c2
2020 F20101124_AAAAKT kanjanarat_p_Page_096.txt
e02fcb765b8ef7f891822b30b0223ea6
04296fb176a259bc8b8495e19bae836989f3ae7b
85718 F20101123_AADHWW kanjanarat_p_Page_015.jp2
4c500535dd3953c3b6bc8d6a82b2a4a7
b0ad015c642d2e3dba92304927f1ca221e7969ef
59928 F20101123_AADIDE kanjanarat_p_Page_105.jpg
d6d179c78645ab037c6c8def0c10ca73
b041e151d5e28f3057d1a9164f605ede2b52fab0
14775 F20101123_AADICQ kanjanarat_p_Page_155.QC.jpg
6872c2ee7e47ba30dc059cac505a1e0c
4782aed0a39f0d65d56b177924f7aa203ea7c6c4
103619 F20101123_AADHXK kanjanarat_p_Page_102.jp2
4f480f469a2a53c13ec85dc0cdf5f5d3
f1c47bc2dd590ac6741f9a39c1695137c80c32cc
3763 F20101124_AAAALJ kanjanarat_p_Page_115.txt
7c8771b1f3ff5d9e8b015321b16916f9
cc8e68b7c766d5794825ee406b69b6654e25e651
1766 F20101124_AAAAKU kanjanarat_p_Page_097.txt
7b0c040bfa297ec49839b0dda5ad91c8
63ab132eed7845d0e598b6e773001e5942902db8
73307 F20101123_AADIDF kanjanarat_p_Page_145.jpg
7e6dafb8e38a87fbbf69daeac3c03aa4
a532667ac25cf5694a4b53df374a6c7b65e6ddb2
24456 F20101123_AADICR kanjanarat_p_Page_172.QC.jpg
4ffe952e2abf4c58d26fb3587e38b7c8
c78938838e70de6f79d4dd6225556778e391f7f4
22088 F20101123_AADHXL kanjanarat_p_Page_140.QC.jpg
02f35543dab1d500f978c354b467ac07
8cbf4258407e1bc8e6cc65afd9c33f02dec3ed4d
2078 F20101124_AAAALK kanjanarat_p_Page_118.txt
32ed940ba57c2ab6b86f9d01ef3c48a3
7023b2ee615425f54bee03e05e0eed7253173e75
1765 F20101124_AAAAKV kanjanarat_p_Page_098.txt
b301880a412394810fa521b64b975f0c
a25f1a8799b482f20f468cb44dead1e2ab10fc22
1841 F20101123_AADHWX kanjanarat_p_Page_054.txt
803c5feda1f197c4741a835ee6e5b09e
ff0d40a3a262cd0c9d6a0eabbd057302d51439eb
49087 F20101123_AADIDG kanjanarat_p_Page_065.pro
ac9ffafc1ef140e031576acc74a6f247
9a2c9d186e2ce4e0144f371c4ba9a3f39aade61c
104537 F20101123_AADHYA kanjanarat_p_Page_068.jp2
8c116febcf66fa40770cb2476a794805
9e191abb6bc6908e5df2f86cc3114e9f10fc6bd3
102727 F20101123_AADICS kanjanarat_p_Page_024.jp2
b0b952fef7a8b9fd1f277f7d29056c5f
51280224a4ee1516530f19db47944d79df1fa9e7
484 F20101123_AADHXM kanjanarat_p_Page_072.txt
302033439a1c68f34c4e978627631d90
a45498ec43544b4e5444b4994fc3b4ca0be68f52
3971 F20101124_AAAALL kanjanarat_p_Page_119.txt
65093fd302dcddf64ca38251ac29912e
425c6a5d4494d57bda0ad064f386cbafc37cb831
1251 F20101124_AAAAKW kanjanarat_p_Page_099.txt
4849610255eb807471f850538a7050be
11dd9ab18511f5782aa917fc800d7f8a7af0f4ab
60901 F20101123_AADHWY kanjanarat_p_Page_091.jpg
adb8d1730e32b209c196c4221250a6a0
eee69c8e5e02240c93e8a69fa64a8f015bd443bf
F20101123_AADIDH kanjanarat_p_Page_049.tif
30e8920ee6e32b2d8db5d3fcc31e83ed
211671130644e6ca53236383fab05afb76a9a4e5
102950 F20101123_AADHYB kanjanarat_p_Page_084.jp2
ed1b8c7834eeba2ffe0af54f126ab680
6b811c06f60acf944a860087555d0d22d67058a0
4964 F20101123_AADICT kanjanarat_p_Page_031thm.jpg
fcfc5e13859885b5038cc1c648c9fcd3
aaa688b8e013188356e96048d6c930fc62eb5de9
1938 F20101123_AADHXN kanjanarat_p_Page_059.txt
ac8d8807c8351b17c5537c6d5b7dafcb
93137f3957845d270df4543f2ad730ab7d08f464
2011 F20101124_AAAAMA kanjanarat_p_Page_139.txt
c7448e6ce12a3f35b39e542b0b95f036
22152ab1aa39992ac01a19480f630c6be7a60cc4
1326 F20101124_AAAALM kanjanarat_p_Page_120.txt
c0c1ce35ac480a6f9ca8e238def17eef
5510e70ecce76799ea1e3c0240a8a15b28bf7c23
1636 F20101124_AAAAKX kanjanarat_p_Page_100.txt
ee5e42054b4b32b4130931432c133a35
85123cbd6999419ceb28565361a8b67c1bf2e468
5107 F20101123_AADHWZ kanjanarat_p_Page_108thm.jpg
b426285feeeeeda18cbb89e0402d5e06
f1e815b009ee673fb622f8b3cefc9c5d2ac1b920
101375 F20101123_AADIDI kanjanarat_p_Page_140.jp2
876881487674322d426dc9934cd5edc6
0a68f6784686b6039a6409f7e62a1565219e257d
F20101123_AADHYC kanjanarat_p_Page_079.tif
bd99ea47c561ebce7d699b8f30948250
bc6d8e02de9425e51dcaba6d70ed26d77e2d66bb
51498 F20101123_AADICU kanjanarat_p_Page_108.jpg
7829fac9e70847c4db3f48b48ecfddbc
c59c9df12ad201c39ea36055f09a541c5bf45bd2
2171484 F20101123_AADHXO kanjanarat_p.pdf
aa0dc9149615a0c46a1e10023edb6c65
3d387524e393fbd52a87fc821828292b7f67d9da
F20101124_AAAAMB kanjanarat_p_Page_140.txt
14266fa188203c88a6b7a9a61f0a8ea2
f9df1fbeb76ed0d8761af4e9ba716cf8adf3fc8b
1590 F20101124_AAAALN kanjanarat_p_Page_121.txt
88da511fc35dbcc18c20c72f05e68bb8
4c579dca324d999bd747a163a7d9b694b8dfc101
2005 F20101124_AAAAKY kanjanarat_p_Page_101.txt
f50ab30ea78962b28d2cf5138064022c
de1070e37148e323d434dd52944a26d9bbfd0c0a
96533 F20101123_AADIDJ kanjanarat_p_Page_184.jpg
370f5094ca07ec6b9186cbc3c0887d72
bfad2b3cf43c19c7c6d17616f5d15785178e8db5
F20101123_AADHYD kanjanarat_p_Page_054.tif
0613cd9ff95c5a164acddc4addbeebcb
f15756d65081e3b728e3d14af781c56c4bd8842c
61080 F20101123_AADICV kanjanarat_p_Page_023.jpg
fd8edabac77b04123d9046ff57ce7876
93929abfa0baa7f69fbbe492468a7bbe30f75ee1
72620 F20101123_AADHXP kanjanarat_p_Page_139.jpg
ab3fa32d552737b892b180c1af6097a9
ac11b9ae3478acf21f9bbc862291b4f0e49d72a9
2073 F20101124_AAAAMC kanjanarat_p_Page_142.txt
8a3faf1f467584f59822bcc38f009577
b78ffa7c1e6d0b19813013a1a8b1c1818fc7f787
907 F20101124_AAAALO kanjanarat_p_Page_122.txt
250f1d86a61b80eaff762b345371389f
60d50b144f3d2de4585de8043b5456ae4b554375
1935 F20101124_AAAAKZ kanjanarat_p_Page_102.txt
d67f8acb2dde921a1dadd23c899e08f4
39591ccb0a7643a1630bc5c4e93d83474a314865
1513 F20101123_AADIDK kanjanarat_p_Page_128.txt
4ec131494e614cf03742dce13f7c333f
c3ebee409967b23066208933086f44ac1fa11121
72172 F20101123_AADHYE kanjanarat_p_Page_085.jp2
97c4ba9a934401405882072d1b3cae94
4bb1035add4a94ccf1db0479bb18f31c4f124f1c
69690 F20101123_AADICW kanjanarat_p_Page_065.jpg
de57eb87046aaa5e2ea5972e24889cde
127951f92a922f4dfe5a6e852d9d469fecc48b46
101464 F20101123_AADHXQ kanjanarat_p_Page_058.jp2
705dd65c54d1ed94665699e9800956fa
a82bc989aeb3d82d7c7693a6a4e45e46f5d9a64a
1797 F20101124_AAAAMD kanjanarat_p_Page_143.txt
584010650a359dfa6b3cac292b4c6c5e
ce3d9c69482acc3ccf4d52d58cdbb9ba1e1501fb
1529 F20101124_AAAALP kanjanarat_p_Page_124.txt
e4f885dac0ecc439609a431e13c9c271
755a05ce45fd4381cb0f58ad54855a9e71d56b04
2734 F20101123_AADIDL kanjanarat_p_Page_179.txt
485e2b9f0f86ad60e14e54b1d68d6618
4465f8b68a12b3d5d0cd4425c8914d637576e5e7
1801 F20101123_AADHYF kanjanarat_p_Page_111.txt
47c56b67e4c475750e1526708793f6c1
73c28dca322aecff5aa21bdb7e3f5bcae9a3802e
F20101123_AADICX kanjanarat_p_Page_119.tif
25de9b253c5f527f483bee6fe8304b53
e00de4e1cde1a7fa9f69435d89913d1743cd63ed
2457 F20101123_AADHXR kanjanarat_p_Page_189.txt
0d6fc9df40414b8e7b8e24eb992faddb
d4776e6966268d6a458a120fbc2f5d5783999546
5743 F20101123_AADIEA kanjanarat_p_Page_105thm.jpg
129032bcfcc90bdc5bd89c23f72ea6e6
ebe4f1c433496fa43dd2073da7f2416a2f6d7c3d
F20101124_AAAAME kanjanarat_p_Page_144.txt
86c9aef6070f8ca8ff336942eeb7a9eb
88f736c3b7575765080ca17c3911036741073188
2648 F20101124_AAAALQ kanjanarat_p_Page_125.txt
766d44888200f7b1ac7a9aade1b67271
cadb8ddba59427922662d28a4b08b829a423390b
13736 F20101123_AADIDM kanjanarat_p_Page_109.QC.jpg
a577b5ef871819e75468a3d2260e8c1d
62580345682241aa869792e0a727eed17a4d0e26
23714 F20101123_AADHYG kanjanarat_p_Page_007.QC.jpg
fcae0f051c15527b32a6c3b3d7a1024b
7ed5b5c3c48b3d2003e32c43cec069926014a5bd
910 F20101123_AADICY kanjanarat_p_Page_009.txt
1889bf9a4cf702837d692bd1e563def3
628533c6e5385e1e258ab38c35f57b38e469f2ff
F20101123_AADHXS kanjanarat_p_Page_137.tif
423993e81d88feca1a5c3b38cd7aa7ee
bb7ea23f716901fdf04608a8c9a530f8304e0a2e
52142 F20101123_AADIEB kanjanarat_p_Page_183.pro
9a2485fdcfe03d160f653d31990b2ec7
57897ea264ee90d0aedbb3039395406f598c7dc5
F20101124_AAAAMF kanjanarat_p_Page_145.txt
8da6c4f2ba28af01294c75979307b8a9
c6cc8c468d5d6d8bedd6339f6604c22255e3e46d
6489 F20101123_AADIDN kanjanarat_p_Page_027thm.jpg
0b3546523523c7cf910e00838cce13f7
6f65d2e936b5a9bc1ee2ce60e1e3cee592ca4c8a
30732 F20101123_AADHYH kanjanarat_p_Page_120.pro
783209eedecd64b1ca99d919048c3431
6c272e4113a075b9d31133d5c219be9a34a385e3
87163 F20101123_AADICZ kanjanarat_p_Page_185.jpg
4d33653d86a0b9e7ff7701be74e2bdd2
31c6442b7260630c5bad70c530a6898ce26f5bc3
6627 F20101123_AADHXT kanjanarat_p_Page_093thm.jpg
ddea410b04e5a2488af8542ccfbad30e
5dbb216a195967876dee3c70e7a3bab33d572c32
3994 F20101123_AADIEC kanjanarat_p_Page_166thm.jpg
ae6f336fe7109b7f9b821c978d4e3e7a
93c86b09a596b4252ef543a1b092145101a802f1
F20101124_AAAAMG kanjanarat_p_Page_146.txt
9eadaf172324e44c14b494a3b4ec3330
687e9f4add75e0217374bbd3d818100d050be2b2
1832 F20101124_AAAALR kanjanarat_p_Page_126.txt
0eff5bda44c84f46beb7f410c976e80c
3d8d8303dd8ac268cfc2944476b3c0ab5f474671
F20101123_AADIDO kanjanarat_p_Page_133.tif
105e72755b8c75bbe7dc8fd6c2f911c5
d57e2647072b050261c66e50af54839e3fd9d181
1051943 F20101123_AADHYI kanjanarat_p_Page_115.jp2
18b92ab9fa70c64806eeafae9dcc8fd3
3ead5f2d0e4b68c374077281e5b1befda2056d40
30707 F20101123_AADHXU kanjanarat_p_Page_156.jpg
44f34132649e43503efc0a82d35dc290
b5a368f5671dfe85ab5c9892dc607246c88d454f
49601 F20101123_AADIED kanjanarat_p_Page_056.pro
7f733e423488a5150e8ce4efa8063916
afec0a268560cb2453b726749b947486837a14c6
F20101124_AAAAMH kanjanarat_p_Page_147.txt
5949a5974bdf5caa91f0d1a874226af4
30ebed03dd54b83fbbbe6e1efad4129afdf0ef7e
1229 F20101124_AAAALS kanjanarat_p_Page_127.txt
cef048d5e422c764fa2a01b30cab71e9
5fa1bb93b6c989f563b73312a66a9fe21e64619e
22041 F20101123_AADIDP kanjanarat_p_Page_018.QC.jpg
2e4a0a130632a4a16a4ddaa8397f7f0b
e2606725cd8d6e0cab8e20eca3f75ce33c26d9da
6349 F20101123_AADHYJ kanjanarat_p_Page_102thm.jpg
1f94ec9126a6c6846a030d0431395e9f
5ceb9561e9f00f36bf39dfa1ec39a194d3da95a3
52851 F20101123_AADHXV kanjanarat_p_Page_135.pro
3f01617ce483f9e70b5c1c2aaa7c5817
0067fa4ef0b47fc1c9d663432d7657e81a1e78a6
1991 F20101124_AAAAMI kanjanarat_p_Page_148.txt
311318b29a16b80ae69ddbe56409b7df
69a32f56ef06cbe2bc0e9f370d30c21e6df6cae9
2145 F20101124_AAAALT kanjanarat_p_Page_130.txt
657d375376cb59a6d15b917d127e0ca7
d6beadc14e0b68520e436f9e67ae68ef5e2962df
7276 F20101123_AADIDQ kanjanarat_p_Page_188thm.jpg
03f0a982bf8891e7e77e6d9210151dee
0f97a0fd212135b1ce9331127f217dba06b4b4a3
129816 F20101123_AADHYK kanjanarat_p_Page_185.jp2
c65d4c871198ceeac4cb1573a0d93d44
0f2e9c25e4ebf7997bbd49b206fadc00eb1a32fe
2077 F20101123_AADHXW kanjanarat_p_Page_137.txt
73812c30e50428c230b8f5e2be133f51
5d1755134523f59fe109c4ae4a51fdd5501ecb1d
46886 F20101123_AADIEE kanjanarat_p_Page_117.jpg
f86b8efbbf0395d8600512c0c375f7e4
6cdda834743eec774cc2ee6ef7fed183879ca57c
1798 F20101124_AAAAMJ kanjanarat_p_Page_151.txt
681890fbb8835c949ed630ee67118653
9033bc908ecaccf89327a20a65633337feca041a
3962 F20101124_AAAALU kanjanarat_p_Page_131.txt
ef4a9e1eafe41bcdf41b83ada10b4e0d
95dc50fe139f4d365657a6615fb676bccc7d24d4
68737 F20101123_AADIDR kanjanarat_p_Page_140.jpg
dd70f27ae70aa71832574bcfd02f7809
ad035cfe917bf13f2cb42a03e29278bec1978a1c
15298 F20101123_AADHYL kanjanarat_p_Page_168.QC.jpg
c14c0d9431653f4f5d8631bc6ef17571
cfc11a93ebb47e7ce891b875320c58778826db73
22852 F20101123_AADHXX kanjanarat_p_Page_186.QC.jpg
d48a4dc0e8bfb260a79f662e429232b6
0805e22a7524c827a704a15f25bbc5568b3ae0d1
46089 F20101123_AADIEF kanjanarat_p_Page_054.pro
eed80eac5e07e1925ab74d3e0ab789ad
a44596b0fa06d134bd963092b5071f2e4715e59c
1210 F20101124_AAAAMK kanjanarat_p_Page_152.txt
f2c91584912b76413497f0eec0565a7c
164a9765df8dc8198db0fd0e3d31c49e26291a50
1205 F20101124_AAAALV kanjanarat_p_Page_133.txt
fff098f3828c2703a26b4b84c0f21038
ca2d0d919ebd0f2f5e51949c43a9885d034e21c6
5980 F20101123_AADHZA kanjanarat_p_Page_058thm.jpg
2450e8d0653277eed9ee4991bacb9df7
537ac9cf704a284daee48c8b0e16d96cdffcabdd
F20101123_AADIDS kanjanarat_p_Page_033.tif
bdb19b4895d3836f07ba086a5ba79a9d
f21ff46c08e8a1a051867d9da1fa8d5d5e0dd50a
47534 F20101123_AADHYM kanjanarat_p_Page_042.pro
c4456d43f712fb19d992baec25a20c4d
f1ccac85d7ce2b83e0e480bef06a5dd78cab6d6c
92536 F20101123_AADIEG kanjanarat_p_Page_073.jp2
08ec4722a6cb539cae4fd8667ccaf471
f8dfa812b2412b327b93ce46a057026afe24f55b
881 F20101124_AAAAML kanjanarat_p_Page_153.txt
810e79f14fbeb019890830884ac3687f
896813a0e4f7ecbd476311d467dc9656efd08067
1820 F20101124_AAAALW kanjanarat_p_Page_134.txt
f2cba4f6012901c4d71ea5618db3182b
6ba7b3435a18aa99dbc36c35f70b63e79baf9dc9
29806 F20101123_AADIDT kanjanarat_p_Page_108.pro
bd558c11a0d304da69d45ccaae2a6ccf
724635f44401b341210e9339b9a9adad0c89b24c
70119 F20101123_AADHYN kanjanarat_p_Page_026.jpg
4fe1d524ec5d9d277acb46c5b0ab51aa
3bf7dbd4b48333e7e453b4e06ebeb6fb60ad0587
F20101123_AADHXY kanjanarat_p_Page_059.jp2
6f2ae35c78103a8dc2a2545b71f36ef7
d6f70352c706b5ff025b9cc04e81904ab9512fbd
1056 F20101123_AADIEH kanjanarat_p_Page_077.txt
821187e8a7389b0eefd7469d8496fad3
fdca152e508e83da639de5fe535545450c1ceb38
F20101123_AADHZB kanjanarat_p_Page_182.txt
5c8c4c64aca2549c1841abedd3e2220b
dfadd3660e1cb4877b21ad96fbd816e43a5be208
1269 F20101124_AAAANA kanjanarat_p_Page_168.txt
8be81063d54b124a6db792011796aa73
7670127fc27983de1d51dc1d892f2f40060a259e
F20101124_AAAAMM kanjanarat_p_Page_154.txt
1d918fba774154f3d9f8f7d25c331b2d
6f278b1ffc051da157bf93414edc997dc48dc756
2080 F20101124_AAAALX kanjanarat_p_Page_135.txt
0d880e96510b96c89665ef37b439139a
3cc6bbcbcce5d66274262b3fc234b8621adc1b38
1288 F20101123_AADIDU kanjanarat_p_Page_085.txt
c079aae9e1ee87507ffb6e0847ac2b0b
1747ddc59f3724f0daffb73dfe9625be0af403d5
24380 F20101123_AADHYO kanjanarat_p_Page_117.pro
74bf720af5bdf70db2bf8000e957f585
9c2b7687531f34ffadc5016b30b1b4c01c7d735d
46135 F20101123_AADHXZ kanjanarat_p_Page_058.pro
da5fbb75d853c4ae60f3be48ca2e59ec
e7f34be5e05275506fdea04aa69066c75e0e110a
108665 F20101123_AADIEI kanjanarat_p_Page_030.jp2
50efb1a7c16a539fe94592742d6aafda
1c34600d6bd93872f05d0d57c3c355479e3ed6d8
925 F20101123_AADHZC kanjanarat_p_Page_088.txt
6b60bb1f53e659d87c4ab4b39f8373f4
c7fd0e573f6184769734ced9ff48ba94c2aae332
426 F20101124_AAAANB kanjanarat_p_Page_169.txt
4ab2546e71776a80c616ff1835dcf297
d05d49bb346c11cc46c811d56c5f1e60e6b9e9bf
F20101124_AAAAMN kanjanarat_p_Page_155.txt
d62b68c06001a3aeeedb9e8659649243
b7f4581e0b13c1ff757fc9d5c9b28c492b3581cc
1869 F20101124_AAAALY kanjanarat_p_Page_136.txt
4b8445ffe81bf5c9db9d3ad8efbee039
c847efd23cc89bb1d081f8e2173755455d080d67
F20101123_AADIDV kanjanarat_p_Page_174.tif
7c985cdc4a9c6885e2f3215dd66bd6d0
52062d494c5780392e1693c73b33f891521eb006
69678 F20101123_AADHYP kanjanarat_p_Page_038.jpg
e654b164cd263a8deab621b572d66c46
d64eb47803332bd04f1524839513627ddeb13c12
5444 F20101123_AADIEJ kanjanarat_p_Page_015thm.jpg
e6728a3db79927d9a41930efe21996f5
3cd687e6fd5393401e16f9c036a19db847ca4926
47766 F20101123_AADHZD kanjanarat_p_Page_149.pro
fcaaacaf6dbce7aeeb1d1521e7f7201d
3e3ec9eb38fefbf214b932bada010114b4121b8d
2222 F20101124_AAAANC kanjanarat_p_Page_170.txt
3e8f377161c907ab611f83a5cc103d31
9658284e4ebe27573bb5d473a6a43f67281d66c7
579 F20101124_AAAAMO kanjanarat_p_Page_156.txt
d07bbfd01cd57bbeb4329968d4351fe2
935a94f7d69db5c2d00da144475bd7ae6cde72e9
F20101124_AAAALZ kanjanarat_p_Page_138.txt
4b3ea91635413a36388ea15cadd2557b
0438e488849c4e33fe0884d585c74e89514a4f9e
16496 F20101123_AADIDW kanjanarat_p_Page_078.QC.jpg
418153b78641655ba380ed714f20b270
0a20e69e82a6f522e08019d76e37c25f1da1f641
19247 F20101123_AADHYQ kanjanarat_p_Page_023.QC.jpg
e3c87cfeda39e6166cd6fc5eeadf29cc
c83f67df55a28d44745172af0230288765fd8dd8
5976 F20101123_AADIEK kanjanarat_p_Page_070thm.jpg
6edbf3a73315b82f15bd7be6eccfb733
707ad09a5a9591c5d4f0f4931aff25b4d6797acc
25331 F20101123_AADHZE kanjanarat_p_Page_043.pro
0f0307d1ef70df554cdfe6259bffc666
4839e198703b7673900f42f6091d907018d06b55
3330 F20101124_AAAAND kanjanarat_p_Page_171.txt
ed8fedcb8e53dfb6046f8512d76f59d8
5b5e1bc7cd964b70e65df229a080e815e4406f44
2315 F20101124_AAAAMP kanjanarat_p_Page_157.txt
323228592a46931ad96cecbd5db7681a
2bcfa98146893c5a4ccff47d2caa1a43f1aedcfe
2821 F20101123_AADIDX kanjanarat_p_Page_045.txt
3f2076fbe3bb6f58a365d006e5900d1a
c4b7b18a6a8c6b7df2d9d4b228f20f4ef54b5965
51026 F20101123_AADHYR kanjanarat_p_Page_161.jpg
cb7a446dc644e38a52a719698b4e56af
1764ea56fbf67c0a7cabce89ccaf9af1afb65ae5
22265 F20101123_AADIEL kanjanarat_p_Page_062.QC.jpg
f133544f88d0959e5a26d816f59b46e5
a3d6f49ba165e884d6e51671c74b2b68ae33d444
23449 F20101123_AADHZF kanjanarat_p_Page_059.QC.jpg
8069cb6dd05c866f78312abd28f995ac
ca9f08651ea2ca63a4437b4fbcb3b43e88435297
2853 F20101124_AAAANE kanjanarat_p_Page_172.txt
3dff8f3a9ae7d1c068974de392726f4c
06b2c4e3a8d98d52f562564b799e11b72082b6f6
F20101124_AAAAMQ kanjanarat_p_Page_158.txt
1ff20ac0e7a984f2c54842ed6f70a64c
5a435fee7e688a638d31246122f38702356edd2a
102789 F20101123_AADIDY kanjanarat_p_Page_136.jp2
805d7c6a0af16710d1b8e509f674b954
ad7959afe1e81f5cdd700997b2a26e4944d8c5f8
496727 F20101123_AADHYS kanjanarat_p_Page_043.jp2
c43689336987ad83a08ad75b9eb4be80
5b19822a1591c6aff28844efd0ef31681ae3c9fc
22814 F20101123_AADIEM kanjanarat_p_Page_009.pro
bfe9a0275fa5d3d0874324b532a7255e
d829ca2c47b0800b070aaf11eced2cf6a93450fa
32053 F20101123_AADHZG kanjanarat_p_Page_113.pro
ba3e50370fb892b716c3a467afbced60
f2f79d6769b777f9e3636a2bb6519332c2a7ad7a
2430 F20101124_AAAANF kanjanarat_p_Page_174.txt
559bdf5614f020ab94290af60cdcb0f2
5fd3b4226f3f41e30587fb60cdf5c579ebff7bee
819 F20101124_AAAAMR kanjanarat_p_Page_159.txt
891b5dc18c7c71a14d995da3ffea06e8
0588b3a86fe35c47dfef7ea64ca52047659017e1
50464 F20101123_AADIDZ kanjanarat_p_Page_079.jpg
a63782cf79563ce9b9ddf74d0ef26bd6
990e26645fa2ad12c42982b8a176a8d80b9a8ae9
22991 F20101123_AADHYT kanjanarat_p_Page_044.QC.jpg
d6c16649d054515eb4eea17cdd8d5155
0fa96430fcdf36e30176fc4f278cf5ad2ad27f50
23459 F20101123_AADIFC kanjanarat_p_Page_001.jpg
98b0d254863c972047954bfc0e66bb91
bf52e8c734ab4d4f212faa8b1258d9e46d7944f9
71955 F20101123_AADIEN kanjanarat_p_Page_019.jpg
551e3a647f069863bf0a1dc2ebafd7c5
b4fcd9456be2286017832260f1a8e14aee979e51
2117 F20101123_AADHZH kanjanarat_p_Page_183.txt
d7d3671e39640bc1734b59d17ae6fa77
8b8a4a4dd70137e8226c29b5ed2dd77e296737df



PAGE 1

IMPACT OF THE 30-BAHT HEALTH INSURANCE POLICY ON HOSPITAL DRUG UTILIZATION IN THAILAND By PENKARN KANJANARAT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Penkarn Kanjanarat

PAGE 3

To my father, mother, brother, sister, ni ece, and my husband for their great love and selfless support

PAGE 4

ACKNOWLEDGMENTS I would like to express my gratitude to my major advisor, Dr. Almut Winterstein, for her guidance, encouragement, support, and friendship during my pursuit of the Ph.D. This study would not have been completed without her guidance and dedication. I would like to thank Dr. Earlene Lipowski who provided a broad perspective of healthcare systems and guided discussions about the applicability of this study. Dr. Lipowski also provided input from her working experiences in Thailand to refine my study. Her encouragement and respect are greatly appreciated. I would like to thank Dr. Abraham Hartzema, who shaped my research idea. His vision in pharmacoepidemiology and international research experiences also helped guide my research. I deeply appreciate his challenging questions. I greatly appreciate Dr. Lili Tian for her suggestions for the analysis of my study. With her help, I have greatly extended my knowledge in Time Series Analysis in health service research, which will greatly benefit me in my future research. I could not express my gratitude enough to the Department of Pharmacy Health Care Administration and the College of Pharmacy for their support. I appreciate the infinite support of Dr. Richard Segal, Dr. Carole Kimberlin, and Dr. Donna Berardo. I would like to express my gratitude to the College of Pharmacy, Chiang Mai University, for their support of my Ph.D. study. iv

PAGE 5

I would like to extend my gratitude to Dr. Jeffery Crane for his helps, encouragement, and friendship. I could not have achieved this goal without his indefinite support. This study is funded by a P.A. Foote Small Research Grant from the Perry A. Foote Health Outcomes and Pharmacoeconomics, University of Florida. v

PAGE 6

TABLE OF CONTENTS Page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...............................................................................................................x LIST OF FIGURES..........................................................................................................xii ABSTRACT.....................................................................................................................xiv CHAPTER 1 INTRODUCTION........................................................................................................1 Background...................................................................................................................1 Need for Study..............................................................................................................2 Purposes of Study.........................................................................................................4 Study Objectives...........................................................................................................5 Research Questions.......................................................................................................5 Significance..................................................................................................................6 2 REVIEW OF LITERATURE.......................................................................................7 Healthcare Care System in Thailand............................................................................7 Health-Seeking Behavior and Healthcare Utilization...........................................7 Access to Drugs.....................................................................................................8 Healthcare Financing.............................................................................................9 The civil servant medical benefit scheme....................................................10 The 30-Baht health insurance policy............................................................11 Evaluating Health Policy............................................................................................12 Methodology in Evaluating Health Policy..........................................................13 Time series in observational design.............................................................14 Nonconcurrent time series design................................................................15 Concurrent time series design......................................................................16 Quantitative and Qualitative Measurement of Drug Use...........................................20 Quantitative Approach to Measuring Drug Utilization.......................................21 Measurement Units of Drug Utilization..............................................................22 The defined daily dose (DDD).....................................................................22 Strengths of the DDD...................................................................................23 Limitation of the DDD.................................................................................24 vi

PAGE 7

Defined daily dose in drug utilization research............................................24 Alternatives measurements to the DDD.......................................................26 Qualitative Approaches in Measuring Drug Utilization......................................29 Concept of quality of care and its measurements.........................................29 Approaches in measuring appropriateness of drug use................................33 Measuring quality of drug utilization using explicit criteria........................34 Measuring quality of drug use by quality indicators (QIs)..........................35 Challenges of Measuring Quality of Drug Use...................................................38 Validity of the measures...............................................................................38 Reliability of the measures...........................................................................39 Sensitivity and specificity............................................................................39 Validating Computerized Administrative Databases..................................................39 Data Validation Methods.....................................................................................40 External data validation................................................................................41 Internal validation methods..........................................................................46 3 VALIDATING DATA QUALITY.............................................................................49 Interview of Hospital Directors..................................................................................52 Interviews of Hospital Database Managers................................................................52 Quantitative Assessment of Data Quality...................................................................52 Database Characteristics......................................................................................53 Missing Data and Outliers...................................................................................53 Face Validity (Plausibility of the Data)...............................................................54 Data Coherence...................................................................................................54 Validating disease diagnosis codes (ICD-10)..............................................56 Drug data......................................................................................................57 4 DATA VALIDATION RESULTS.............................................................................58 Hospitals Demographics.............................................................................................58 Database Characteristics.............................................................................................61 Face Validity...............................................................................................................63 Missing Data...............................................................................................................66 Data Coherence...........................................................................................................69 Disease Diagnosis and Gender............................................................................69 Disease Diagnosis and Drugs..............................................................................69 Diabetes and antidiabetic drugs....................................................................69 Hypertension and antihypertensive drugs....................................................72 Bacterial pneumonia and antibiotics............................................................74 Drugs and Disease Diagnosis..............................................................................76 Antidiabetic drugs and diabetes...................................................................76 Results from the Expert Interviews............................................................................77 5 METHODS.................................................................................................................80 Hypotheses..................................................................................................................81 vii

PAGE 8

Patient Selection.........................................................................................................85 Data Source.................................................................................................................87 Measures.....................................................................................................................87 Drug Utilization Rate (DR).................................................................................88 Percentage of Appropriate Prescribing (Drug Use Quality, DQ)........................89 Hospital Visit Rates (HR)....................................................................................90 Hospital Admission Rates (AR)..........................................................................90 Statistical Analyses.....................................................................................................92 Analysis of the Effects of the 30-Baht HI Policy on the Study Measures..........92 Changes of Drug Disparity Associated with the 30-Baht HI Policy...................93 6 RESULTS...................................................................................................................95 Patient Demographics.................................................................................................95 Diabetes......................................................................................................................96 Drug Utilization Rates.........................................................................................96 Hospital Visit Rates.............................................................................................97 Impact of the 30-Baht HI Policy on the Study Measures....................................99 Hypertension.............................................................................................................101 Drug Utilization Rates.......................................................................................101 Hospital Visit Rates...........................................................................................101 Impact of the 30-Baht HI Policy on the Study Measures..................................103 Infectious Diarrhea...................................................................................................105 Drug Utilization Rates.......................................................................................105 Percent Appropriate Prescribing........................................................................106 Hospital Visit and Admission Rates..................................................................106 Impact of 30-Baht HI Policy on the Study Measures........................................108 Bacterial Pneumonia.................................................................................................112 Drug Utilization Rates.......................................................................................112 Hospital Visit and Admission rates...................................................................113 Impact of the 30 Bath HI Policy on the Study Measures..................................114 Disparity of Drug Utilization Rates..........................................................................118 Disparity of the Percent Appropriate Prescribing for Infectious Diarrhea...............118 7 DISCUSSION...........................................................................................................119 The Effect of the 30-Baht HI Policy on Drug and Hospital Service Utilization......119 Drug Utilization and Access to Care.................................................................119 Effectiveness of the Implementation of the 30-Baht HI Policy........................120 Suggestions to Improve the Evaluation of 30-Baht HI Policy on Drug Utilization and Hospital Service Utilization...........................................................................124 Time Series Design............................................................................................126 The Defined Daily Dose....................................................................................126 Quality of the Data in the HI database......................................................................127 Study Limitations......................................................................................................130 Recommendations for Future Research....................................................................133 Contribution to Scientific Society............................................................................134 viii

PAGE 9

Conclusions...............................................................................................................135 APPENDIX A TERMINOLOGY AND DEFINITIONS.................................................................136 B HEALTH INSURANCE CODES............................................................................138 C RECORD LINKAGE...............................................................................................139 D DRUG LISTS FOR CALCULATING DRUG UTILIZATION RATES.................143 E EXPERT INTERVIEWS..........................................................................................155 LIST OF REFERENCES.................................................................................................163 BIOGRAPHICAL SKETCH...........................................................................................177 ix

PAGE 10

LIST OF TABLES Table page 2-1 Alternative units for measuring drug utilization......................................................27 2-2 Types, definitions, and policy purposes of access to healthcare services................30 4-1 Hospital information, numbers of hospital visits, hospital admissions, and prescriptions from 2000 to 2003..............................................................................59 4-2 Five most prevalent dispensed drugs for inpatient and outpatient use of the eight included hospitals (A-H) during from 2000 to 2003................................................64 4-3 Provincial Disease Statistics (2002) on the 10 most common causes of morbidity for outpatient services, Ubonratchatani province, Thailand....................66 4-4 Missing data on patient demographics Hospital Missing data.................................68 4-5 Missing data on disease diagnosis codes for inpatient and outpatient data of eight studied hospitals (2000-2003).........................................................................68 4-6 Missing data on prescribed drugs of eight studied hospitals over four years (2000-2003)..............................................................................................................68 6-1 Expected populations eligible for the 30-Baht HI policy and the CSMBS in 8 selected government community hospitals, Ubonratchatani province, Thailand (2003).......................................................................................................................95 6-2 Patient demographics...............................................................................................96 6-3 Regression parameters for drug utilization and hospital visit rates of diabetes.......99 6-4 Rates of drug utilization and hospital visits related to diabetes of the 30-Baht health insurance beneficiaries................................................................................100 6-5 Regression parameters for drug utilization and hospital visit rates of hypertension...........................................................................................................103 6-6 Monthly rates of drug utilization and hospital visit related to hypertension of the 30-Baht health insurance beneficiaries..................................................................104 x

PAGE 11

6-7 Regression paramerters for drug utilization, prescribing quality, hospital visit, and hospital admission rates of infectious diarrhea...............................................109 6-8 Monthly rates of drug utilization, hospital visit, hospital admission, and the percent prescribing appropriateness of the 30-Baht group.....................................110 6-9 Regression parameters for drug utilization, prescribing quality, hospital visit, and hospital admission rates of bacterial pneumonia.............................................115 6-10 Monthly rates of drug utilization, hospital visit, hospital admission related to bacterial pneumonia...............................................................................................116 B-1 Codes for health insurance status for the Health Insurance database.....................138 D-1 Antidiabetic drugs for calculating drug utilization rates........................................143 D-2 Antihypertensive drugs for calculating drug utilization rates................................145 D-3 Antibiotics for calculating drug utilization rates....................................................148 xi

PAGE 12

LIST OF FIGURES Figure page 2-1 Health seeking behavior, Thailand 1999 (23)............................................................8 2-2 Nonconcurrent control group in observational design (27).....................................16 2-3 Concurrent control group in observational design (27)...........................................17 2-4 Formulas for calculating sensitivity, specificity......................................................42 4-1 Geographic locations of the included eight community government hospitals in Ubonratchatani province, Thailand..........................................................................59 4-2 Percent of patients with two or more disease diagnosis codes assigned in the HI database (Hospital F)................................................................................................60 4-3 Linkage among the selected patient data in the HI database AN: scrambled hospital admission number.......................................................................................62 4-4 Percent data coherence between disease diagnosis of diabetes and antidiabetic drugs of outpatient data among eight hospitals, 2000-2003.....................................71 4-5 Percent data coherence between disease diagnosis of diabetes and antidiabetic drugs of inpatient data among eight hospitals, 2000-2003.......................................71 4-6 Percent data coherence between disease diagnosis of hypertension and antihypertensive drugs of outpatient data among eight hospitals, 2000-2003.........73 4-7 Percent data coherence between disease diagnosis of hypertension and antihypertensive drugs of inpatient data among eight hospitals, 2000-2003...........73 4-8 Percent data coherence between disease diagnosis of bacterial pneumonia and antibiotics of outpatient data among eight hospitals, 2000-2003.............................75 4-9 Percent data coherence between disease diagnosis of bacterial pneumonia and antibiotics of inpatient data among eight hospitals, 2000-2003...............................75 4-10 Percent data coherence between antidiabetic drugs and disease diagnosis of diabetes among eight hospitals, 2000-2003.............................................................76 xii

PAGE 13

5-1 Algorithm to assess quality of drug utilization (DQ) using community-acquired pneumonia as an example.(169;170)........................................................................91 5-2 Interrupted time series analysis of a drug utilization rate in the 30-Baht HI group and the control group before and after the policy....................................................93 6-1 Monthly drug utilization rates for diabetes in the 30-Baht HI population and of the CSMBS population from January 2000 to December 2003...............................97 6-2 Monthly hospital visit rates for diabetes in the 30-Baht health insurance population and of the CSMBS population from January 2000 to December 2003..98 6-3 Drug utilization rates for hypertension of the 30-Baht health insurance benefit group from January 2000 to December 2003.........................................................102 6-4 Monthly hospital visit rates for hypertension of the 30-Baht health insurance benefit group from January 2000 to December 2003.............................................102 6-5 Monthly drug utilization rates for infectious diarrhea of the 30-Baht HI from January 2000 to December 2003............................................................................105 6-6 Percentages of prescribing appropriateness for infectious diarrhea of the 30-Baht HI from January 2000 to December 2003..............................................................106 6-7 Monthly hospital visit rates for infectious diarrhea of the 30-Baht HI from January 2000 to December 2003............................................................................107 6-8 Monthly hospital admission rates for infectious diarrhea of the 30-Baht HI group January 2000 to December 2003............................................................................107 6-9 Monthly drug utilization rates for bacterial pneumonia of the 30-Baht HI from January 2000 to December 2003............................................................................112 6-10 Monthly hospital visit rates for bacterial pneumonia of the 30-Baht HI from January 2000 to December 2003............................................................................113 6-11 Monthly hospital admission rates for bacterial pneumonia of the 30-Baht HI from January 2000 to December 2003...................................................................114 C-1 Two tables identify patients who visited outpatient clinic during the study period139 C-2 Two tables identify patients admitted to the hospital for inpatient services during the study period, and length of stay........................................................................140 C-3 Patient demographic table identifies patients age, gender, marital status, types of health insurance.................................................................................................141 C-4 Two tables contain drug data.................................................................................142 xiii

PAGE 14

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy IMPACT OF THE 30-BAHT HEALTH INSURANCE POLICY ON HOSPITAL DRUG UTILIZATION IN THAILAND By PENKARN KANJANARAT August 2005 Chair: Almut G. Winterstein Major Department: Pharmacy Health Care Administration Thailand implemented the national 30-Baht health insurance policy in June 2001 to provide healthcare coverage (including drugs) for the uninsured population. We retrospectively assessed the impact of this policy on drug utilization in community government hospitals using computerized patient-specific data from 8 hospitals in Ubonratchatani province. Cross-sectional data from January 2000 to December 2003 included all inpatients and outpatients with bacterial pneumonia (CAP), gastrointestinal (GI) infections, diabetes (DM), and hypertension (HTN). We confirmed the internal validity of the database and conducted interviews of hospital personnel. The primary measure was drug utilization rate (DUR), measured as the Defined Daily Dose (DDD)/10,000 population/month. The secondary measures were percent prescribing appropriateness, hospital visit and admission rates. Monthly observations (N=48) were analyzed using segmented time series regression analysis (SARIMA model). xiv

PAGE 15

Prior to the policy, the average monthly DUR of the 30-Baht group for DM, HTN, Infectious diarrhea, and CAP were 2,727.8, 1,414.0, 40.3, and 7.8 DDDs/10,000 beneficiaries, hospital visit rates were 86.7, 25.8, 25.6, and 1.9 visits/10,000 beneficiaries, and hospital admission rates were 2.5 and 0.3 admissions/10,000 beneficiaries, respectively. After the policy, average monthly DUR for DM, HTN, Infectious diarrhea, and CAP were 3,105.6, 2,983.5, 52.8 and 12.4 DDDs/10,000 beneficiaries, monthly hospital visit rates were 87.4, 36.2, 26.5, and 1.9 visits/10,000 beneficiaries, while admission rates for Infectious diarrhea and CAP were 3.0 and 0.6 admissions/10,000 beneficiaries, respectively. Appropriate antibiotics were prescribed to less than one half of the patients with Infectious diarrhea (43.4% before and 47.5% after the policy). Analysis revealed no immediate or trend effect in DUR, hospital visit/admission rates for DM, HTN, infectious diarrhea, or CAP after the 30-Baht HI policy was implemented (p>.05 for all measures). There was no significant change on the percent appropriate antibiotic prescribing for Infectious diarrhea, p>.05. The study did not detect a change in drug and hospital service utilization associated with the 30-Bath HI policy, although there were positive trends in rates of drug utilization, hospital visit, and admission after the policy. Computerized hospital database prove valid and useful resource for research. xv

PAGE 16

CHAPTER 1 INTRODUCTION Background The Institute of Medicine (IOM) Report 2002 (1) addressed significant health problems of the uninsured. The report stated that approximately 30 million Americans with no health insurance had a higher risk of poor health and shorter life-expectancy than people with health insurance. Poorer health in the uninsured was the consequence of limited access to care (1). Until the year 2000, approximately 30% of the Thai population had no health insurance coverage. In 2001, the Thai government, to improve access to care and reduce healthcare disparity between the insured and uninsured Thais, implemented the 30-Baht Health Insurance (HI) Policy for the uninsured Thai population. The uninsured were the poor, self-employed, and children 12 to 18 years of age, who did not have any benefits from the three existing governmental health insurance plans: the Civil Servant Medical Benefit Scheme (CSMBS), the Medical Welfare Scheme (MWS), and the Social Security Scheme (SSS). The 30-Baht HI policy offers coverage for most healthcare services and prescription drugs provided by government hospitals. The National Health Insurance Office evaluates eligibility based on the uninsured status of an individual. The 30-Baht HI beneficiaries are then registered with the government community hospital in their district as a main healthcare provider. The HI benefits are not extended to care received by other providers except when the patients are referred to that facility. Similarly, 1

PAGE 17

2 prescription drugs are covered by the policy only if they are dispensed from the hospital pharmacy at the designated hospital. Our study aims to assess the impact of the 30-Baht HI policy on drug utilization because drug therapy is a core component of medical care. Although previous literature suggests that lack of drug benefits is associated with underutilization of prescription drugs (2-7) and that an increase of drug utilization was expected, our study evaluated any changes of drug utilization associated with this policy. Need for Study After this national policy was implemented in October 1, 2001, the government authorities, including the National Health Security Office (NHSO) and the National Health System Research Institute (HSRI), conducted various health policy evaluation studies to evaluate the effects of the policy on various aspects. However, most of the research from the NHSO and the HSRI focused only on administrative and economic issues, and patients and providers satisfaction with the policy (8-10). Few studies have yet examined the impact of the policy on drug utilization (8). Evaluating the effect of the new policy on drug utilization required identifying the effect of the policy on changes in the disparity of drug between subpopulations. While national healthcare expenditures in Thailand increased greatly over the last 2 decades (3.82% of GDP in 1978 to 6.21% in 1998) the distribution of this spending is skewed toward the insured. The increase mostly occurred in the insured population. Only a small proportion occurred in the uninsured, which indicates healthcare utilization disparity or unequal opportunity in healthcare service access and utilization among uninsured versus insured populations. The disparity was associated with various factors that inhibit access to care, including financial status (11).

PAGE 18

3 Assessments of national health policies in developing countries are often compromised by weak study designs that fail to establish a causal relationship between the policy, and health indicators and health outcomes (12). For example, various studies use cross-sectional designs, post-only intervention designs, or pre-post comparisons with no control group (13-19). Moreover, the selection of reliable and clinically significant outcomes measures is often compromised by data unavailability or inappropriate data format for the analysis (e.g., healthcare data was collected in paper format, with no record linkage among databases). Our study applied a sound study design (interrupted time series of impact analysis) to evaluate the causal association between the 30-Baht HI policy and drug use by the targeted population. Currently, Thailand has no standardized data source that contains drug utilization data at a national level. Available drug utilization data were derived from quarterly government hospital purchasing and inventory reports, which lack patient-level detail. Estimates of drug utilization can also be derived from drug importation and/or manufacturing reports put forward by pharmaceutical companies on an annual basis. No drug utilization data are available at a patient level, which would allow assessments of drug utilization in a certain disease state or across subpopulations. The opportunity for our study was offered by a novel computerized administrative database, Health Insurance (HI) database, used in every hospital in Ubonratchatani province since 1997. This database includes electronic medical records and drug dispensing data for every patient who received healthcare services (outpatient and inpatient) from the hospitals. Although never used in research, the database offers a

PAGE 19

4 unique opportunity for health outcomes and drug utilization studies because of its level of detail and its uniform structure in all community hospitals of the province. Evaluating of the effect of the health policy on drug utilization would contribute to an objective assessment of healthcare services established by the Thai government. It would also offer suggestions to stakeholders for further improving access to drug utilization and eliminating disparities in healthcare. Purposes of Study We proposed to quantitatively evaluate the impact of the 30-Baht HI policy on drug utilization in terms of drug utilization rates and prescribing quality, using drug data from the HI database. Drug utilization rates are measured using the Defined Daily Dose (DDD) and prescribing quality is assessed based on the appropriateness of prescribing for a given disease state. Effects of the 30-Baht HI policy on access to hospital services were also measured as hospital outpatient visit rates and hospital admission rates. Four disease states (diabetes, hypertension, bacterial pneumonia, and infectious diarrhea) are studied. We selected the listed disease states for two reasons. First, they are prevalent and associated with high mortality rates in Thailand. Second, they represent acute and chronic conditions and allow a more comprehensive assessment of the policy impact on healthcare. From the computerized patient-specific healthcare database (HI database) of the government community hospitals in Ubonratchatani province, Thailand, we chose subjects who were eligible for the 30-Baht HI policy during 2000 and 2003 as a study group. The Civil Servant Medical Benefit Scheme (CSMBS) was selected as a control population because of its comprehensive healthcare and drug benefits. For the control group, access to care and choices of treatment were not limited by the patients ability to

PAGE 20

5 pay for services. We analyzed the effect of the 30-Baht HI policy on drug utilization and access to hospital services using an interrupted time series analysis. By using a control group, we adjusted for the natural trend and other factors that might have an effect on drug utilization rather than the 30-Baht HI policy. In addition, we compared drug use disparity (difference of drug utilization between the 30-Baht HI beneficiaries and the CSMBS group) at one year before and after the implementation of the policy. Definitions and terminology used in our study are presented in Appendix A. This study was approved by the Institutional Review Board, University of Florida. Study Objectives To test the validity of the patient-specific healthcare data in the HI database of the community government hospitals in Ubonratchatani province, Thailand To assess the impact of the 30-Baht HI policy on drug utilization rates, the percent appropriate prescribing, hospital visit and hospital admission rates in the 30-Baht HI population controlling for longitudinal changes in the CSMBS group To evaluate the change of drug utilization disparity between the 30-Baht HI population and the CSMBS at one year before and after the policy was implemented. Research Questions Research question 1: Did drug utilization rates in the 30-Baht beneficiaries for the selected disease states change after the 30-Baht HI policy was implemented, controlling for trends of the drug utilization rates of the previously uninsured group and those of the CSMBS group? Research question 2: Did the prescribing quality in the 30-Baht beneficiaries change after the 30-Baht HI policy was implemented, controlling for the drug utilization quality of the previously uninsured group and those of the CSMBS group? Research question 3: Did hospital admission rates related to pneumonia and gastro intestinal infections for inpatient services in the 30-Baht beneficiaries change after the 30-Baht HI policy was implemented, controlling for the hospital admission rates of the previously uninsured before the policy and those of the CSMBS group? Research question 4: Did hospital outpatient visit rates related to diabetes and hypertension in 30-Baht beneficiaries change after the 30-Baht HI policy was implemented, controlling for the hospital visit rates of the previously uninsured before the policy and those of the CSMBS group?

PAGE 21

6 Research question 5: Did the extent of drug utilization disparity between the 30-Baht HI policy beneficiaries and the control group change at one year after the policy compared to the extent at one year before the policy? Significance We evaluates whether the 30-Baht HI policy improved drug utilization rates and quality of drug utilization as provided in governmental insurance schemes. Comparisons of drug utilization rates and quality between this population and the control population also addressed its effect on drug utilization disparity. Our results will complement other epidemiologic and/or economic evaluations of this policy and suggest areas of improvement of the health insurance benefit program to reach the goal of comprehensive national healthcare coverage. We used comprehensive electronic patient medical records from the HI databases that have never been validated or used for research purposes. Although this issue posed an additional challenge for main study objectives, we hope the validation of the HI databases facilitates future health service research and quality improvement activities.

PAGE 22

CHAPTER 2 REVIEW OF LITERATURE Healthcare Care System in Thailand Most healthcare services in Thailand are provided to the population by the public sector. Most of these facilities are affiliated with the Ministry of Public Health, followed by the Ministry of Defense (veteran hospitals) and the Ministry of Internal Affairs (municipal hospitals). Healthcare services provided by the private sector include private hospitals, private outpatient clinics, and independent pharmacies. Additionally, many patients seek cures from traditional treatments, although modern medicine is widely available. Health-Seeking Behavior and Healthcare Utilization Even though self-medication (44.2% to 86.3%) (20;21) is commonly practiced in Thailand, it decreased from 54.1 to 17.62% (more than 36%) from 1970 to 1996, while healthcare utilization from public health services increased from 15.5-44.0% (almost 30%) (22). Socio-economic factors, the healthcare infrastructure, and type of health problems appear to affect health seeking behaviors. For example, the 1999 Health and Welfare Survey indicated that people living in rural areas mainly seek healthcare services from public settings, while people who live in urban areas are more likely to seek healthcare services from private clinics or hospitals than from the public facilities (23) (Figure 2-1). A study of health-seeking behavior of the villagers in rural areas of Thailand using health diary by Osaka et al.(24) found that most patients with chronic diseases were more 7

PAGE 23

8 likely to receive treatment and medications from local health centers or community hospitals, while patients with severe and acute symptoms (e.g., high fever or abdominal pain) sought urgent care from private clinics. For minor illness, most villagers rarely used healthcare services from a local health center or a community hospital. 0 10 20 30 40 50 60 % of population Folk docto r Others N o treatment Self-treated Public facility Private clinic/hos p ital Urban Rural Figure 2-1. Health seeking behavior, Thailand 1999 (23). Access to Drugs Regarding the current drug delivery systems, almost all classes of drugs are accessible for Thais, from prescription drugs to herbal medications. Drug distribution channels include hospitals, private clinics, pharmacies, health centers, public health centers, and even grocery stores. Prescription drugs are sold in drug stores with pharmacists and without pharmacists. Some pharmacies are open for service with no pharmacists on duty, thus patients receive drug products from the store owner or a store clerk. In 1999, nearly half of the drug stores (5,351 of 12,548 stores) were authorized to sell prescription and controlled-substances (e.g., corticosteroids, benzodiazepines, and

PAGE 24

9 narcotics). The remaining drugstores were authorized to sell most of the commonly prescribed drugs, including antibiotics, with or without a pharmacist. Moreover, over 400,000 small grocery stores in villages around the country have prescription drugs available. Drug regulations and enforcement do not control this type of retail drugstores in Thailand (25). Not only pharmacists can dispense prescription drugs and controlled substances: Thai drug regulations allow physicians to dispense medications as well. It is commonly acknowledged that drugs are the major source of income of healthcare facilities. So every hospital has a hospital pharmacy department where most prescriptions are dispensed. While drugs can be obtained through various channels only government hospitals can dispensed drugs under the 30-Baht HI policy. Thus, the study databases can assess whether reimbursement for drugs from hospital improved, but it is not possible to assess whether overall drug access increased as a results of the 30-Baht HI policy. Healthcare Financing To understand how healthcare services are provided to Thai people, it is important to be familiar with existing healthcare financing schemes before the 30-Baht HI policy was implemented. The 30-Baht HI policy was created based on the four established health insurance schemes offered by the government, and targeted to fill the gap between the insured and uninsured. The four health insurance schemes offered by the Thai government included the Civil Servant Medical Benefit Scheme (CSMBS), the Social Security Scheme (SSS), the Medical Welfare Scheme (MWS), and the Health Card Scheme (HCS). Private Health Insurance (PHI) is also available for people who can afford it.

PAGE 25

10 The next section provides some background information on the CSMBS in terms of its benefits, program financing and related problems. This group was used as a control group to analyze the impact of the 30-Baht HI policy on drug utilization. We will not discuss details of the other health benefit schemes. The civil servant medical benefit scheme The CSMBS is the health benefit program provided by the government for civil servants who currently work for government entities and respective retirees. The benefits extend to immediate family members, including a spouse, parents, and children younger than 20 years of age. In 2000, there were approximately seven million beneficiaries. However, the exact number of beneficiaries is unknown due to lack of a registry database. The benefits include outpatient and inpatient, medical and surgical services, emergency services and drug expenses. The benefits exclude a small number of services such as cosmetic surgery and preventive services (vaccination and contraceptive medication), except for an annual health check-up. The beneficiaries have access to care from government hospitals or private hospitals that have registered with the plan. This program uses a retrospective reimbursement method to pay for healthcare based on a fee-for-service system. This type of payment system has minimal control on healthcare expenditure. Consequently, healthcare expenditures in the CSMBS population have increased dramatically over time both for outpatient and inpatient services and are expected to rise continuously. It has been reported that the expenditures per person were twice as much as those for patients in other healthcare benefit plans or those with no health insurance(11).

PAGE 26

11 As part of the Healthcare Reform in 1998, the benefit package of the CSMBS has been reviewed. Two major changes were proposed to control healthcare costs and improve the quality of care. First, the previous focus on treatment has shifted to health promotion and disease prevention (disease screening and vaccination plans are now included in the plan). Second, a certain level of healthcare quality is assured by the requirement that the providers must register with the Ministry of Public Health in order to get reimbursement for inpatient services for this population. These changes affect both drug utilization rates and quality, and the sources of drug utilization data of this population. The 30-Baht health insurance policy The concept of universal health insurance, which was ratified in the new national constitution amendment in 1998, states that All Thai people have an equal right to access quality health services Despite a quite substantial number of existing schemes offered by the government and by private insurance companies, only 70% of the population was covered by these plans. To expand the previous insurance schemes towards the goal of universal health coverage, the Thai government introduced the -Baht HI Policy for the remaining uninsured population. The 30-Baht HI policy was introduced by the new government elected in 2000. This policy is a product that responds to the 1998 Thai Health Care Reform which focuses on the improvement of access to care and reduction of health disparities. The pilot phase of this policy was implemented in 4 provinces (out of a total of 75 provinces in Thailand) in February and expanded to another 15 provinces, including Ubonratchatani, in June 2001. After successful completion of the pilot phase, the policy was implemented nationwide in October 2001.

PAGE 27

12 The policy covers every person who did not have any healthcare benefits from the existing healthcare benefit schemes mentioned earlier. Each person can register with the primary care center, or a community government hospital in their local district. The eligibility to participate in the plan is verified by the Office of National Health Insurance. Registered persons receive a gold card that has to be presented when healthcare services are sought. The beneficiaries are required to pay 30-Baht (75 US cents) per episode of care. The benefits of the policy include medical treatment and disease prevention, disease screening and diagnosis, and rehabilitation as necessary. They also include Thai traditional medicine and alternative medicine under the provision of a medical professional. Regarding prescription drugs, the 30-Baht HI policy covers only drugs listed in the National Drug Formulary that include most essential drugs recommended by the World Health Organization. These drugs are required by the Ministry of Public Health to be included in a hospital drug formulary of government hospitals. More recent and expensive drugs can be added to a hospital drug formulary, but are typically restricted to patients who are willing to pay for extra costs (e.g., the CSMBS and SSS population). Evaluating Health Policy In the past two decades, quality improvement (QI) has been a major focus in health care services worldwide. Among other QI initiatives, an increasing number of health-related policy interventions has been implemented to improve quality, and reduce and/or contain costs. As a consequence, the number of health policy evaluation studies and the use of automated healthcare databases for this purpose has increased dramatically in the public (e.g., Medicaid and Medicare populations) and private sectors such as Health

PAGE 28

13 Maintenance Organizations, such as the Group Health Cooperative of Puget Sound, Kaiser Permanente. Most health policies that have been implemented are either regulation or deregulation, both, with the intent to affect health service utilization and health outcomes. Examples of regulation policies are the state antibiotics vigilance in the Medicaid population, antibiotics restriction in hospitals (26), or co-payments and drug reference pricing (27). The evaluation of the impact of health policies is difficult because they are implemented in non-experimental conditions: the policy typically affects the entire target population and thus, control groups are difficult to establish. In addition, the policy is implemented in an already dynamic healthcare environment, and other factors that may affect the outcome of interest may not be excluded. A review of drug policy evaluations in developing countries found that most policy evaluation studies used weak study designs and were based on post-intervention measures only. Thus, the studied results were not conclusive (28). From a measurement perspective, evaluations of health policies are often restricted to administrative databases, which raise problems such as poor data quality, inability to ascertain the outcomes of interest, and incomplete patient information data. Methodology in Evaluating Health Policy Various methods have been used to evaluate the effects of health policy on healthcare utilization and health outcomes of the population. Among more advanced study design are randomized controlled trial, uncontrolled time series, and controlled time series. Only time series design, its methods, strengths, weaknesses, and its applications are discussed in this section.

PAGE 29

14 Time series in observational design Time series analysis in a longitudinal study design has been used in the fields of economics, political science, and engineering research for several decades. It has also been an increasing number of research studies using time series in the field of drug utilization, health policy evaluations, and pharmacoepidemiology. Several time series studies in healthcare research focused on the assessment of the impact of health policy on prescribing behavior and drug utilization (26;29-31;31-34). There are three major purposes of applying time series analysis: 1) identifying patterns of the series of the data, 2) evaluating the effect of an intervention, and 3) forecasting future values of the data. Our study intends to use the analysis of time series data to evaluate the effect of the policy (considered an intervention), thus, only the literature on the evaluation of interventions is included in this section. Segmented regression analysis requires a series of data that was collected regularly and is equally spaced over time, e.g., monthly, seasonally, yearly. There are several data sources that can be used to measure the effects of health policy on drug use. Data are derived either from primary data collection (e.g., direct follow-up of patients to ascertain information on health outcomes), or secondary data collection (e.g., medical records, insurance claim data, pharmacy dispensing data, or hospital discharge data). The benefit of using healthcare data that have been routinely collected for the purpose to document clinical interventions or reimbursement is the availability of the data longitudinally in large populations (35). Outcomes of interest can be healthcare or drug utilization, clinical measures, or costs. The outcome measures can be applied as averages, proportions or rates. Drug utilization measures that are often used in policy evaluation research are number of drugs

PAGE 30

15 prescribed per patient, average prescription cost, percent of beneficiaries receiving a particular drug, or percent of patients treated according to treatment guidelines (35). Lengths of stay and hospital admission rates are widely used to determine the effect of the intervention on access to healthcare services. For clinical measures, surrogate markers such as blood pressure can be used. Time series is composed of a set of observations that are measured longitudinally (36). The desired number of observations is usually more than 50 to yield sufficient sample size to precisely model the data (37). The observations can be measured from a single case (e.g., weight of a person over 10 years) but more often are seen in aggregate data from several cases, e.g., average drug utilization rates in diabetes patients over 10 years. Nonconcurrent time series design The time series design that has no concurrent controls is useful to identify immediate effects of a policy intervention; however it always encounters internal validity problems due to other causes that may affect the outcome of interests. Since it is impossible to control other factors that might affect the outcomes at the same time as the policy did, it is important to base the conclusion on the assumption that there is a close temporal relationship between the policy and the outcome, and the extrapolation of the baseline trend must be estimated as if the intervention was not implemented. Figure 2-2 shows the dotted line indicating the measure before policy and an extrapolation after the policy was implemented under the assumption that the policy has no effect on this measure. The full line represents the observed values of the measure if the policy has a negative effect on the measure. The time series analysis of non-concurrent data becomes more accurate when there is large number of observations before and after the

PAGE 31

16 intervention because the prediction of the trend before the intervention is more accurate (38). Time series analysis requires an adequate number of observations (some researchers have recommended more than 200 to produce accurate results) (37) before and after the intervention to be able to control for trends or seasonal effects. Intervention Expected trend without intervention Intervention group Time Unite Figure 2-2. Nonconcurrent control group in observational design (27) Concurrent time series design To overcome the validity problem due to other factors that might affect the outcome rather than the intervention, a concurrent control group time series design should be used if a comparison group exists. The concurrent control group can help adjusting for the effect of other factors that might oppose or distort the effect of the studied intervention. However, the conclusion of the analysis must be based on the assumption that the control trend is equal to the trend in the intervention group if there were no effect of the intervention. The dotted line in Figure 2-3 demonstrates the observations in a control population, which has similar trend before the policy (straight

PAGE 32

17 line). After the policy, trend in the control group remains the same as the trend before the policy. In the intervention group, there is a change in the trend after the intervention implementation, which suggests a negative effect in decreasing the value of the measures. The difference in the slope can indicate the effect of the intervention. Control group Intervention group Intervention Unit Figure 2-3. Concurrent control group in observational design (27) In this study, the researcher chose to apply a time series design using the CSMBS population as a control group to evaluate the effect of the 30-Baht HI policy on hospital drug utilization over a 4year period. This method is appropriate for aggregate data, e.g., drug utilization, hospital visit or hospital admission rates. Using a concurrent control group, the design provides strong evidence of a causal relationship between the intervention and outcome measures. The following section presents some examples of published studies of health/drug policy evaluations using the analysis of time series data. Some studies used concurrent control groups and some did not. The methods are presented for each study.

PAGE 33

18 A study in Manitoba, Canada, evaluated the drug benefit policy change from a fixed deductible and co-payment system to an income-based deductible system on receipt of prescriptions for inhaled-corticosteroids in children with asthma. This study compared a cohort before and after the drug benefit policy was changed. No concurrent control group was included in the study. Receipt of prescription and number of inhaled corticosteroid doses were compared by visual observation and odds ratios (39). Blais et al. conducted a study in Quebec to assess the unintended effects of a cost sharing drug insurance plan on drug utilization among individuals receiving social assistance. Drug utilization in three classes of medications was studied: inhaled corticosteroids, neuroleptics, and anticonvulsants. A control group was applied in this study. The Seasonal Autoregressive Integrated Moving Average (SARIMA) model was implemented to adjust seasonal effects on drug utilization(40). In 1997, the reference pricing policy for angiotensin-converting enzyme inhibitors (ACEIs) was implemented in British Columbia, Canada. Schneeweiss and his colleagues conducted a study to evaluate the effects of this policy on quantity and timing of drug utilization. Autoregressive time series of prescription drug claims data for 3 years was used to identify the effects of the policy. The authors conclude that analysis of time series data is able to provide detailed information for the policy makers regarding the extent and duration of the effects of the drug policy. In 1989, the State of New York implemented a Triplicate Prescription Program (TPP) as a drug prescribing surveillance system. After the TPP was implemented, there was 55% reduction in the monthly number of benzodiazepine recipients in the Medicaid cohort (41). Wagner and his research team further studied the intended and unintended

PAGE 34

19 effects on new post-hospitalization benzodiazepine use(42). The effects of the TPP were evaluated by an interrupted time-series of post-hospitalization benzodiazepine dispensing rates and the substitute medications in a Medicaid population. The control group in this study was the Medicaid population in New Jersey, which adds strength to this study in controlling for extraneous factors that might affect benzodiazepine prescribing behavior. In 2000, hospitals in Columbia implemented an educational intervention to improve appropriateness of antibiotic prescribing practices. Interrupted time series were conducted on three antibiotic groups (aminoglycosides, cephradine/cephalothin, and cefazidime/cefotaxime), comparing the hospital weekly rates of incorrect prescriptions, and prophylactic antibiotic use in elective surgery to assess the effects of the intervention. From a statistical method perspective, this study applied the Autoregressive Integrated Moving Average (ARIMA) model of time series data because the series of the data are not linear. The study did not have a reference group because of inability to find a comparable setting (31). Another study by Ansari et al. in Tayside, Scotland, applied segmented regression of interrupted time series analysis to identify the effect of an Alert Antibiotic Policy on improving use of appropriate antibiotics (26). The authors commented that segmented regression analysis of pharmacy data is a feasible method for the assessment of the effects of the policy (26). A study of drug prescribing in primary care practices was conducted in Boston, Massachusetts after all non-formulary drugs were removed from the formulary and generic samples stocked. Segmented linear regression analysis was applied to estimate changes in levels or trends in formulary compliance (43).

PAGE 35

20 In addition to quasi-experimental designs, time series is useful in identifying the impact of natural changes or influences. A study by Tu et al. intended to evaluate the effect of the finding from the Heart Outcomes Prevention Evaluation (HOPE) study that ACE inhibitor (ramipril) are effective in the prevention of secondary cardiovascular disease. By applying segmented regression of time series of prescribing data, the study was able to identify the effect of each publication type on ACE inhibitors prescribing (44). The above studies have applied time series design and analysis using a segmented regression model to evaluate the effect of the interventions or policies. Some studies allowed lag times in the ARIMA process in order to incorporate delayed effects of the intervention. Very few studies utilized a concurrent control group; however those studies had sufficient historical observation points to ensure that any change after the policy was not associated with the factors that affected the measure before the policy was implemented. For non-linear data (e.g., the percentage of prescribing appropriateness) the studies applied various transformation methods (e.g., logits) to establish more linear series. Seasonality that might cause non-linearity of the data was sometimes incorporated by using a SARIMA model. Quantitative and Qualitative Measurement of Drug Use Quantitative drug utilization studies usually refer to the measurement of numbers and rates of drug consumption or exposure in the population of interest in a certain period of time (e.g., number of drug doses/10,000 population/year, and proportion of patients receiving a certain drug per 10,000 population per year). A qualitative approach of measuring drug utilization is the evaluation of appropriateness of utilization of a certain drug or a class of drugs in a specific population with a certain disease states. Since the

PAGE 36

21 definition of the appropriateness of drug use theoretically incorporates every aspect of drug use, it is complicated to conduct a conclusive evaluation of the process of care and patient outcomes. In the section on qualitative measurements of drug utilization, definitions of quality, interpretation, and measurement methods, including explicit criteria and quality indicators are discussed. Sources of the data that have been used in assessing appropriateness of drug use are also discussed regarding their applications and the limitations. Quantitative Approach to Measuring Drug Utilization According to the WHO, drug utilization is defined as marketing, distribution, prescription and use of drugs in a society and economic consequences (45). A drug utilization study refers to the assessment of medical, sociological-behavioral and economic factors influencing drug utilization, including the effects of drug utilization at all levels. Lunde and Baksaas (46) describe general objectives of drug utilization studies as problem identification and problem analysis in relation to importance, causes, and consequences, and an establishment of a weighted basis for decisions on problem solution; assessment of the effects of the action taken. These objectives are relevant to problems and decision making throughout the drug and healthcare systems. The approaches may vary according to the purpose and the needs of the users. Those include the health authorities, the drug manufacturers, the academic and clinical health professionals, social scientists, and economists as well as the media and the consumers (46).

PAGE 37

22 Measurement Units of Drug Utilization Quantitative drug utilization studies involve drug utilization statistics (e.g., utilization rates) of the population by age, gender, disease stage, geographical area, or across time. Drug utilization rates can be used to determine access to care or areas that are underor over-utilized. These statistics are also useful to plan for drug importation, production, and distribution by pharmaceutical industries and government authorities. Drug utilization rates have been used as crude estimates of disease prevalence, morbidity and mortality rate, e.g., digitalis utilization for congestive heart failure. In addition, drug utilization rates allow monitoring of drug use in specific therapeutic groups that are associated with drug therapy problems (e.g., NSAIDs, narcotic analgesics, hypnotics and sedatives) (47). In addition, drug utilization rates can be used to monitor the effects of health-related policy and activities over time and across populations. The defined daily dose (DDD) There are several widely accepted measurements of drug utilization in research depending upon the purpose of the study, research community and institution, country, or the type of available database. In early 1970s when drug databases were not fully established and detached from population data, the researchers in drug utilization studies predominantly used costs of drug use and volume of prescription as an index of drug use. However, there are several limitations of using cost data to represent actual drug use in the population. Costs of drugs were drastically varied over time, depending on pricing policy, country, currency exchange rate, and quantity of drug purchasing. Additionally, cost does not have any correlation to the population denominator, such as, demographic factors, disease conditions, prescribers, or health of the population; thus, it is difficult to interpret this cost data in drug utilization studies.

PAGE 38

23 Another measurement unit of drug use is the number of prescriptions per patient or per population. This unit has been used in drug utilization review studies for several decades since prescription records, were recorded in the computer systems. The number of prescriptions could be used to indicate the number of patients who received drugs; however, it still has some limitations as various quantities may be prescribed per prescription or the gap between drug prescribed and dispensed. The number of prescription units remains insufficient in evaluation associations between drug use and the indication (disease) and patient health outcomes. This unit may not be an appropriate measure if numbers of drugs and quantity per prescription vary largely. In late 1980s, the World Health Organization introduced a standardized measurement unit of drug utilization, that is, the Defined Daily Dose (DDD). This effort aimed to provide a unit for comparison across countries and to enable drug utilization monitoring over time. The DDD is a technical unit of drug utilization that provides an estimate of the number of patients within a community who receive a drug(s) of 1 maintenance dose. It is calculated based on the assumed average dose per day for a drug product used for its major indication in actual practice (48). The identified dose is suggested by the medical literature and presumed to be the average maintenance dose when used for the major indication (with full patient compliance). We used Equation 2-1 to estimate utilization of a drug in a population of 10,000 per month, expressed in DDDs. DDDs/10,000 people/month = amount of drug (mg) sold in 1 year x 10,000 (2-1) DDD (mg) x 12 month x number of people Strengths of the DDD Measurement of drug use is independent of brand name, package size, and sales price

PAGE 39

24 Measuring drug utilization based on the same DDD allows comparison across settings, regions, and countries DDD could be used to monitor drug use overtime if the main indication and dosage of the drug remains the same DDD could be used to estimate drug utilization of the population using aggregate data, e.g., sales data DDD has been adopted widely for more than two decades in research and health policy and the WHO, which offer advantages to current researchers in understanding the application of the DDD and then apply to their research. Limitation of the DDD DDD is defined based on Scandinavian data using the therapeutic maintenance dose for an adult person of 70 kg bodyweight and normal organ functions. Generalizability to other countries or patient populations, such as children or patients with renal failure, is limited. DDD is only defined for the main indication of the drug for a disease and does not refer to other uses, e.g., prophylaxis. Moreover, several drugs are used with no indication (off-labeled). DDDs are not defined for a preparation for topical use, sera and vaccines, antineoplastic drugs, anesthetics, or contrast media. Number of patients receiving DDD based on population-based drug utilization estimates is a rough estimate under the assumption that patients have full compliance, and may not represent actual drug use. Defined daily dose in drug utilization research A good example of the applications of the DDD is demonstrated by the study of trend in drug consumption of calcium channel blockers in the Czech Republic (49). This study focused on the drug utilization pattern overtime and comparison across selected countries. This study applied the DDDs to identify effects of the intervention (publication of the adverse drug events of short acting nifedipine) on drug use. Drug utilization was measure by the DDD system using wholesale data from the General Health Insurance Company of each country. The DDD unit was used in a retrospective

PAGE 40

25 study of antipsychotic agent utilization in Spain, which indicates the benefit of this unit to compare pattern of drug use nationally and longitudinally (50). Defined daily dose in health policy and healthcare delivery A study by Damiani et al. in 2002, identifying the effect on non-hospital prescribing of intramuscular administered cephalosporins (IACs) in Italy, showed an application of the DDD in measuring drug overtime. Since drug expenditure for parenteral cephalosporins was a major part of the countrys healthcare drug budget, there was a community-level regulation to restrict use of IACs called CUF55, which limit the use of these antibiotics to specific infections that are resistant to other common antibiotics, or in patients with immunodeficiencies. CUF55 was then modified to deregulate second-generation cephalosporins, cefonicid and cefmetazole. The research applied DDD to measure changes of the targeted drugs (cefonicid and cefmetazole) and the drugs that might be used as a supplement (3rd generation or 1st generation cephalosporins) associated with the regulation (51). This study shows that DDD is sensitive to a measure change of drug use overtime and offer the detection of the effect of the health policy on drug utilization of the country using sales data. Defined Daily Dose in drug utilization review study (clinical analysis) Problems of drug use addressed by the IOM in the report in 1998 are well-known as overuse, underuse, and misuse (52). Can DDD be applied to find the evidence regarding problems of drug use? The selected examples using the DDD system below addressed appropriateness of drug use in different countries. The first example is the study of misuse of sumatriptan by measuring number of DDD prescribed from a prescription database in Denmark by Gaist et al.(53). Second

PAGE 41

26 example is a study of appropriateness of DM drug use between settings, GPs and specialists in two different geographical areas in Sweden (54). This method applied actual prescribed dose compared the DDD to detect change over time and identify the differences of prescribing pr actice in between the two towns. Another example is a comparison of quinolone use in general populatio n in long-term care facilities, and within a single institution in the Netherlands. This is a good example of the DDD application for exploratory research on appropriateness of drug use and prescribing behavior(55). Alternatives measurements to the DDD There are several units of drug utilization measurement available and have been used in drug utilization studies. However, at this time the alternatives do not prove much advantage over the DDDs. Five alternative measures of drug use are presented below and offer some strength and weaknesses in different issues.

PAGE 42

27 Table 2-1. Alternative units for measuring drug utilization Alternative Advantage Disadvantage Prescribed Daily Dose (PDD) (56) Description: A defined actual prescribed dosage by prescribers in a selected geographical area or setting Useful for a study of prescribing behavior and therapeutic traditions Useful for a study of morbidity prevalence Reflect actual practice than can be compared to the DDD Non-standardized measure make it difficult to compare across settings, regions, countries, and overtime Minimum marketed dose (MMD)(57) Description: A minimum dose that will produce a desired therapeutic effect, which is the minimum dose marketed by the manufacturer Same as the DDD Only reflect how many people received a minimum dose Minimum dose is subject to change depends on the market, thus it is not consistent for longitudinal comparison Variations of minimum dose produced by different manufacturers who produced the same drug make it difficult to decide which one to use Comparisons of drug use between countries are limited, because of the unavailability of the minimum dosage of a drug in some countries Because of the limitations there is only little information available in the literature to compare the results of a new study with. Therapeutic Course (TC)(58) Description: Measurement of drug dose for the whole course of treatment. This parameter is accounted for time of drug therapy. DDD combined with TC offer a more reliable information about exposure to drug use Provide meaningful drug utilization measure of how many people receive full course of therapy (appropriate care),e.g., antibiotics Length of some therapies may vary, e.g., infection becomes chronic with complications that the set length of TC may not apply.

PAGE 43

28 Table 2-1. Continued Alternative Advantage Disadvantage Equipotential dose (ED)(59) Description: ED is created by two doctors in a Danish research group for treatment of hypertension. It is defined as the amount of efficacious substance in relation to a given amount of a given drug, the amounts having the same potential effect on the blood pressure. Not obvious advantage over the DDD, because the DDD is calculated based on equipotency assumption. Defined by only two doctors, thus generalizability is questioned. Indicate treatment effort in terms of lowering blood pressure, but does not indicate how many patients received the drugs. Average Daily Dose (ADD), (German Drug Index)(60;61) Description: Defined by pharmacologists in the German Scientific Institute of General Health Insurance according to German drug use situation. Same concept as the DDD, but specific to German drug therapy situation No advantage over the DDD

PAGE 44

29 Qualitative Approaches in Measuring Drug Utilization To discuss the measurement methods of drug utilization quality, it is necessary to understand how quality is defined from the level of quality of healthcare services to quality of drug use in the literature. Then, the measurement tools will be discussed for this application in quality improvement and drug safety studies. Concept of quality of care and its measurements In the US, since 1965 when the government offered Medicaid and Medicare programs to the poor and older populations, the need to monitor and improve quality of care has required most healthcare institutions to set a framework to assess and improve quality of care. Quality of care is a multifaceted concept. The most widely referred concept of quality of care was introduced by Donabedian in 1978 (62), which suggests that quality of care can be measured at three levels; health-related structure, process, and outcomes. Health-related structure is described as factors related to healthcare systems and health policy, such as the healthcare delivery system, population needs, or an economic situation. Heath-related process is a measure of ability to access care financially and non-financially, and a level of health risks. Lastly, the ultimate goal of healthcare quality is to produce efficient and equitable of health status (health outcomes) (62-64). There are six types of access to healthcare: potential, realized, equitable, inequitable, effective, and efficient (65). The definitions of access to care and the types of health policy that target each level of access are described in Table 2-2. Literature studying access to care has emphasized that access to care is a relative term based on healthcare needs (66;67). In 1993, the IOM redefined access as the timely use of personal health services to achieve the best possible health outcomes(68). From this

PAGE 45

30 definition, the interpretation of access to care is expanded to include patients needs and the consequent health outcomes. In this study, three levels of access, potential access, realized access, and inequitable access, are studied. This study intends to evaluate the effects of the health policy that offer health benefits to increase healthcare service utilization (potential access). Drug utilization rates are measured to determine realized access as a result of the 30-Baht HI policy. Prescribing appropriateness is the measure of effectiveness access to drugs. Equitable access is measured by comparing drug utilization measures (rates and quality) between two populations. Gaps of drug utilization rates and quality suggest inequity (disparity) of access to drug therapy. Table 2-2. Types, definitions, and policy purposes of access to healthcare services Type Definition Policy Purpose 1. Potential access Healthcare system characteristics and enabling resources that influence use of health services To increase or decrease health service use 2. Realized access Use of health services To monitor and evaluate policies to influence health service use 3. Equitable access Use of health services is determined by demographic characteristics and need To ensure health services distribution is determined by need 4. Inequitable access Use of health services is determined by social characteristics and enabling resources To reduce the influence of social characteristics and enabling resources on health services distribution 5. Effective access Use of health services improve health status or satisfaction To improve the outcomes (health status, satisfaction) from health services use 6. Efficient access Minimizes the cost of health services use and maximizes health status or satisfaction To minimize the costs of improving outcomes from health services use Source: Adapted from Measuring access and trend. Introduction to health services, 5ed.(65) Donabedian (63;63) further discussed the concept and the measurement of quality of care in the subsequent publications and suggested that quality is composed of seven attributes: 1) efficacy: the ability of care to improve health; 2) effectiveness: the actual health improvement resulted from care; 3) efficiency: the greatest health improvement with the lowest cost; 4) optimality: marginal health improvement compared with cost; 5)

PAGE 46

31 acceptability: access to care, patient-practitioner relationship, convenience of care, patient preference to effectiveness and cost of care; 6) legitimacy: value of care to the society; and 7) equity: distribution of access and quality of care that is fair based on needs (63). The evaluation of quality of drug therapy, healthcare product and service, must be conducted based on these attributes. However, in defining quality of drug therapy in this study the researcher focuses on only two attributes: effectiveness and equity, more than the rest of the attributes, because they are feasible to operationalize and measure with the hospital data sources, which the researcher selected for the analyses of this study. Further more, other attributes are less likely to be affected by the implementation of the 30-Baht health insurance policy. Definitions of efficacy, effectiveness, and equity of drug therapy are described below. Effectiveness Effectiveness of drug therapy is the ability to improve health outcomes in patients with a specific disease under the real world of everyday practice. The real world situation may include heterogeneous patients characteristics comparing to the patients in the clinical trials (e.g., age, race, compromised renal function), low patient adherence (e.g., patients stop taking the drugs, or miss doses), and inappropriate prescribing by physicians resulting from certain conditions (e.g., limited choice of drug in the formulary, limited resources, or lack of knowledge). The evidence of drug effectiveness is available in randomized clinical trials and for a number of diseases. This evidence is summarized as clinical treatment guidelines that can be used as a gold standard to evaluate the effectiveness of drug use in the population of interest (63). However, there are some

PAGE 47

32 limitations of using clinical treatment guidelines to support drug effectiveness (e.g., some patients fail to respond the recommended drugs or dosage). Equity Equity is fair distribution of drug utilization and quality based on individual needs. Individual needs include patients disease condition, co-morbidities, and demographics (63). Equity is a relative measure that can be measured by comparing drug utilization rates and quality with those of other population(s). This means that patients with the same disease state, and with similar co-morbidities and demographics, should receive similar drug therapy. The Global Health Equity Initiative (GHEI), a collaborative research network of twenty countries addressing the increasing inequities in health, has defined equity as fairness that healthcare services should be available and accessible as needed. This definition could imply that needy (sicker or more vulnerable) groups within a society require access to care at a higher level of resource consumption than those people who have better health status. Then, the gap of equity of care will be reduced (69). In Minnesota, Minnesota Health Care Commission applied the concept of equity in providing universal health care coverage to the residents that set the goal as affordable, accessible, and accountable healthcare for everyone. It addresses financial barriers and non-financial barriers to access to care related to geography, culture, language, race, transportation, and a shortage of providers (70). The provision of the 30-Baht HI policy is different from the universal healthcare in Minnesota, as it did not include non-financial factors. Thus, these factors remain influence on drug utilization.

PAGE 48

33 Approaches in measuring appropriateness of drug use Quality of drug use has been evaluated and studied by a method called drug utilization review or drug utilization evaluation. Drug utilization review is defined as an authorized, structured process that reviews, analyzes, and interprets the pattern of drug use in a given health care delivery system in relation to explicitly predetermined criteria, guidelines or standards(71;72). In general, quality is used interchangeably with appropriateness in drug utilization review literature. Schmader and his colleagues have defined the term appropriateness as the selection of a medication and instructions for use that agrees with accepted medical standards to provide safe, effective care(73). Assessment of quality is judged by appropriateness of drug use, including choice, dosage, duration of therapy, drug administration, drug monitoring, and patient compliance (74). In many cases, practitioners use their expert judgment for appropriateness. On the other hand, researchers seek to apply methods that are more objective and based on the evidence published in the literature. Some researchers used categorical measures to evaluate the appropriateness in terms of whether a drug was present or not (e.g., the percentage of patients who received antihypertensive medications to control blood pressure). However, there are several reviews criticizing that it provides insufficient information addressing the complexity of medication use (75-77). As stated earlier, quality of drug use is defined as quality of prescribing in this study because drug use quality is based on what physician prescribed in the hospital environment. Three components of quality of prescribing are selected for the assessment in this study: choices of drug, dosage, and duration of therapy, as these components are addressed by the WHO regarding quality of prescribing. The WHO has defined quality of prescribing as prescribing the right drug, with the correct dose, duration, and drug

PAGE 49

34 form for the right indication, with adequate information and instruction to the right person in accordance with co-morbidity and other medication used. In evaluating the quality of prescribing, various data should be available, e.g., patients demographics, disease diagnosis, and detailed information of prescribed drugs (dose regimen, duration of therapy), and reimbursement methods (78). In this study, these required data are available in the administrative database of the selected hospitals. Measuring quality of drug utilization using explicit criteria In drug utilization studies, explicit criteria have been introduced to evaluate prescribing appropriateness at a patient-level, mostly in inpatient populations, where a usual amount of information related to drug use is available. These explicit criteria were established based on current clinical knowledge, practicality, and the available data sources. Extensive literature on drug appropriateness using explicit criteria exists in elderly populations because of their vulnerability and complexity of care. An example of an explicit criterion for elderly populations is an indicator that evaluates the inappropriateness of benzodiazepine use. Talerico (79) conducted a critique of six measures (Beers criteria, Avorn index of potentially inappropriate drug use, Medication Appropriateness Index, the Defined Daily Dose, Panel Assessment for drug regimen, and Swedish medical product agency guidelines) for their utility in assessing inappropriate psychoactive drug utilization in the elderly. This critique applied six criteria to evaluate the measures validity and reliability: 1) indication for drug therapy, 2) the effectiveness of drug therapy, 3) correct dosage adjusted for pharmacodynamic and pharmacokinetic changes in elderly, 4) appropriate duration of therapy, 5) duplication of drug therapy, and 6) the risk of adverse events.(79) The results from the review confirmed that most measures based the

PAGE 50

35 appropriateness of drug use on the effectiveness or choice, dosage, and duration of therapy. This finding indicates that the three criteria are important components in assessing quality of drug use. Other criteria are also important, however, difficult to measure without complete drug information and patient outcomes. Thus, other criteria that are not used in this study where not discussed in this chapter. A study by Owen et al.(80) estimated the sensitivity and specificity of explicit criteria in assessing quality of antipsychotic drug use compared with implicit criteria as a gold standard. The explicit criteria produced high sensitivity (84.6%), however, relatively low specificity (71.7%). Using the explicit dose criterion may result in a systematic overestimation of inappropriate dosing. Medicaid claims data has been used to measure inappropriateness of psychotropic drug use. However, the researchers suggest that the finding has a potential overestimation of inappropriateness of psychotropic drug, because many filled prescriptions are never taken, and the adherence with the prescribed drugs is often not verified (81). Thus, prevalence of (in) appropriateness of drugs in the population based on prescription refill data should be interpreted carefully. Measuring quality of drug use by quality indicators (QIs) Quality indicators have been introduced with the concept of quality improvement for more than a decade(82). Almost every healthcare institution has adopted the concept of quality improvement to be able to achieve standards set by accreditation organizations and to compete in the healthcare market. To be able to measure quality of healthcare services, quality indicators must be established. The Agency of Healthcare Research and Quality (AHRQ), and Health Care Financing Administration (HCFA), the Joint Commission on Accreditation of Healthcare Organization (JCAHO), and the Center of

PAGE 51

36 Medicaid and Medicare Services (CMS) are responsible to provide guidelines for quality for each type of services, healthcare institutions, and populations. The AHRQ is responsible for creating tools and resources that help healthcare providers and their institutions to provide high quality of care that is safe, accessible and affordable to the patients. These tools include surveys of medical expenditure, and healthcare utilization, questionnaires to assess patients experience in healthcare services, and quality measurement tools to assess clinical performance, including the AHRQ Quality Indicators (QIs). The AHRQ developed three sets of QIs to identify specific areas that are potentially problematic in hospital services: prevention, inpatient care, and patient safety. These QIs are specifically developed to allow the assessment of quality using hospital inpatient administrative data. The following discussion focuses only on inpatient care quality indicators that are applied to this current study. Process of developing the indicators and the criteria of the selection of quality indicators are described. The AHRQ inpatient QIs have been developed based on the Healthcare Cost and Utilization Project (HCUP) QIs, which are consisted of 33 clinical performance measures. The HCUP QIs covers three dimensions of care: 1. Potentially avoidable adverse hospital outcomes, e.g., measurement of mortality rates among low-risk patients receiving common procedures, complication rates during hospitalizations, e.g. UTIs 2. Potentially inappropriate utilization of hospital procedures, e.g., overuse or underuse of a certain healthcare services (cesarean section) 3. Potentially avoidable hospital admissions, e.g., immunization rates to prevent pneumonia in elderly population However, there are several limitations of measuring quality of care using the HCUP QIs. These indicators do not include considerations of severity of risk adjustment; the

PAGE 52

37 denominators of the measurement were based on hospital discharges rather than populations; and the indicators mainly focus on surgical procedures and did not represent chronic diseases or pediatric illnesses (83). The revised Inpatient QIs by the AHRQ include the following areas: volume of healthcare utilization (esophageal resection volume), mortality indicators for inpatient procedures, and mortality indicators for inpatient conditions, utilization rates (healthcare services overuse, underuse or misuse, e.g., cesarean delivery rate) (84). The measurements of quality related to drug use mandated to report to the Congress, are the percentages of persons with outpatient visit and the percentage of persons with prescription drugs. It is important to understand what criteria are used to establish those indicators to be able to justify the selection of the QIs in measuring the quality or appropriateness of drug utilization in this study. Reliability and validity criteria for QI selection Quality indicators to measure quality should be valid and reliable. The AHRQ has used six criteria (84) to evaluate the validity and reliability of the established quality indicators as follows: 1. Face validity: the QIs must be established based on sound clinical evidence from the literature and must be able to indicate quality aspects that apply to providers or the healthcare system. 2. Precision: the QIs should be able to capture the quality across populations (e.g., types of providers or health plans) 3. Minimum bias: the QIs should take into account of the disease severity and patient co-morbidity. The differences of quality must not be biased because of different patients disease-related characteristics. 4. Construct validity: the QIs should be consistent with other QIs measures that evaluate the same aspect of quality. An example for construct validity is that prescribing of

PAGE 53

38 beta-blocker after myocardial infarction should correlate to reduction of cardiovascular mortality. 5. Fosters real quality improvement: the QIs should facilitate the implementation of quality improvement programs. 6. Application: the QIs should be able to implement with other indicators and together provide broader picture of the quality. Proposed measures of the three targeted quality criteria: efficacy, effectiveness, and equity will be described, and the operationalization discussed. Challenges of Measuring Quality of Drug Use Validity of the measures In our study, effectiveness of drug therapy will be measured based on only three health-related-process components: drug choice, dosage, and duration of the therapy. Other levels of health-related processes are not measured, e.g., severity of the diseases that affects choices, dosage, and duration of therapy. In addition, the measurement of quality neither includes health-related structure (e.g., availability of drugs in the hospital formulary), nor health-related outcomes (e.g., lengths of stay in the hospital, adverse drug events, and mortality). Secondly, this unit of measurement of quality of drug use in the unit of proportions of patients receiving effective drugs based on the US clinical treatment guidelines from reputable sources may not have some limitations, in the situation that physician compliance to the guidelines is low. Standard treatment guidelines are not well-established or well-adopted from the practitioners in primary care hospital level in Thailand. It seems unfair to use standard dose, duration, and choice of therapy of some institutions to measure quality of care before introducing them prior to the measurement of the quality in their settings. However, modern practice of medicine came from

PAGE 54

39 evidence-based that the researcher would assume that the practitioners learn this concept from medical school, and should update their practice based on current published evidence. From this reason, variation of effectiveness of drug therapy may be observed from different hospitals. Reliability of the measures The measurement of quality of drug use that is reliable should allow the comparison of drug use overtime. The measurement unit is not able to capture the change of effectiveness of drug use over time, if the standard of quality changes (e.g., new drugs are approved and recommended by the guideline, or dosage recommendation is changed). It is important to clarify the standard of effectiveness of drug use when comparison over time is conducted. Sensitivity and specificity Since effectiveness of a drug will be concluded only if all three components (choice, dosage, and duration of therapy) are correct based on the guideline the measure have high specificity, but low sensitivity. By this, it means the measure has ability to capture drug ineffectiveness better than drug effectiveness (i.e., patient received a correct choice of antibiotic and dosage, but the duration was one day shorter to the recommendation, it will be categorized as ineffective, however patients might recover as well as receiving full course of therapy). Validating Computerized Administrative Databases Administrative databases have been collected primarily for reimbursement purposes. Additionally, clinical data in these databases are collected in the electronic format that facilitates data extraction and statistical analyses. For these reasons, these databases have become one of the main sources of drug and disease data for

PAGE 55

40 pharmacoepidemiologic research and health policy evaluations(85). The characteristics of the databases differ depending on the size of the data, timeliness of the data (e.g., how recent the data is available in the databases), number of variables collected in the databases, and types of institution that provide healthcare services (e.g., hospital databases, community pharmacy databases, and claims databases). Most administrative databases contain patient demographics (e.g., age, gender, marital status, disease diagnosis, drugs, hospital charges, and health insurance status), which are necessary for the charge reimbursement and suffice to identify disease condition and drug utilization for health policy evaluations. Van Eijk et al. suggested in his article about data requirement for research in 2001 that the data should be accurate and in computerized and standardized format. In addition, record linkage should be unique and easy to link the data on patients characteristics, medical, and prescription drug data (78). Examples of administrative databases that have been used in pharmacoepidemiologic and drug utilization research and quality assurance purposes are: Medicaid Management Information System (38;42;72;76;85-128), the Group Health Cooperative of Puget Sound(42;129), the Manitoba Health Service Commission(39;130-142), and the Medicare database (94;143-148). Data Validation Methods Data validation methods widely used to validate healthcare data are the external and internal data validation. External data validation means the data is compared with other data sources that contain the same data. The comparison data source for external data validation should be accurate and, thus, is considered the gold standard database. However, when a gold standard or a validated database is not available, internal validation methods are used for evaluation. Internal validation method is basically a

PAGE 56

41 cross-check within the same database, or longitudinal comparisons to identify any inconsistency of the data that are unexpected (e.g., downward peak of prescription drug claims in a particular month). Methods of data validation, measurements of data quality, and information from the literature regarding data validation from administrative databases are discussed below. External data validation West et al. describes three major quantitative methods of measurement errors of the data in a database that is compared with one or more other data sources: 1) reliability, 2) validity, and 3) agreement (149). The reliability measure is used when the same source of the data is used more than once for the same information on the same person. An example of reliability is the consistency of blood pressures in repeated measurements of an individual. Validity or accuracy of the data can be assessed by comparing specific data from one database to a superior source(s) that is considered the gold standard. There are two measures for validity testing, sensitivity and specificity. Sensitivity or completeness is used to represent the extent to which the studied database correctly identifies individuals who have the characteristics of interest (e.g., benzodiazepine users, or patients with diabetes) compared with the true occurrence in the gold standard source. Specificity measures the extent to which the studied database correctly identifies individuals who do not have the characteristics of interest (e.g., elderly who never used benzodiazepines, or patients without diabetes) (149). (Figure 2-5) In general, data that has high sensitivity are more likely to have low specificity (149). West et al. suggests that absolute values of these measures can be falsely interpreted. For example, if the true prevalence of insulin use is 5%, and by using claims

PAGE 57

42 data with a specificity of 95% (and 100% sensitivity) the prevalence will double to 10% (149). For these reasons, the benchmark (or expected values) of the sensitivity and specificity should be carefully set and interpreted. A True positive B False positive C False negative D True negative Gold standard Present Absent Present Absent Database m1 m2 n1 n2 N Sensitivity = A/n1 Specificity = D/n2 Positive predictive value = A/m1 Negative predictive value = D/m2 Figure 2-4. Formulas for calculating sensitivity, specificity Adapted from Pharmacoepidemiology, 3rd ed. Strom BL, 2000.(45) Other terms that have been used in research to identify the performance of a database are positive predictive values (PPV) and negative predictive values (NPV). Validity cannot be concluded with these measurements because they are calculated using the denominator from the studied database (not from the gold standard). (Figure 2-4) The PPV and the NPV rely on sensitivity and specificity (or validity) of the data source and the prevalence of the exposure or the outcome of interest. For example, even if the database has the same validity of drug exposure in two populations, the PPV and NPV

PAGE 58

43 will be different if the prevalence of drug exposure in the two populations is not similar(149). In this study, percent agreement will be used to identify validity of the diagnosis and prescription drug data. An example of assessing validity of diagnosis data for pneumonia from the HI database is to compare the ICD-10 code of J13 and J15 with prescribed antibiotics (in this database bacterial culture results are not available for comparison). Using the same method, the reliability of prescribed drug data can be assessed using the diagnosis codes for any bacterial infections that are sensitive to the prescribed antibiotics. Disease diagnosis code Diagnostic codes that are found in databases are either the International Code of Diagnosis (ICD) version 9 or 10 (150). The first concern in using the ICD Codes to identify patients with the disease is whether the ICD Codes are accurately and completely assigned to the corresponding clinical conditions (151;152). The ICD code system does not provide any disease description (i.e., standardized diagnostic criteria), which might produce misclassification of the disease. Diagnosis coding accuracy varies depending on various factors, e.g., institutional policy of coding for reimbursement purposes, whether the disease is easy to diagnose or not, incentive for coding, and how well the data coding personnel are trained (150). Examples of disease diagnosis coding bias were found in a study by Hsia et al. in 1988, when data accuracy was tested by comparing the claim data with the diagnosis from original medical records in Medicare patients in 239 hospitals The results show that 20.8% of diagnosis codes was not accurate, and 61.7% of these discrepancies financially

PAGE 59

44 favored the hospitals (153). The study was repeated in 1988 (154), the results showed the same problem that 14.7% of records contained errors that altered the diagnosis-related groups; 50.7% of these errors financially benefited hospitals. The validity of the ICD coding varies across disease states. Payne et al. addresses that the ICD-9-CM system is not able to capture several clinical problems in outpatient settings and important functional, socioeconomic, and psychosocial factors. For example, Alzheimers disease and related dementias (ADRD) were under-coded in Medicaid, Medicare and managed care populations (155). Generally, administrative databases are designed to contain a limited number of slots for assignment of ICD codes, which might be sufficient for uncomplicated cases, but inadequate for patient with multiple complications. However, the studies of validity of disease diagnosis codes in Medicare population (156;157) suggest an increase in the number of slots for diagnosis and procedure coding might not assure the improvement in completeness of the data. Medicare claims data has been validated for different purposes, e.g., measurement of drug utilization, disease prevalence, and adverse drug event. Buchmueller validated the Medicare data for measuring tumor stage by comparing with SEER data as a gold standard. For the inpatient population, the diagnosis (ICD-9-CM) (158) was obtained from the Medicare Provider Analysis and Review (MEDPAR) files, while the data for outpatients were extracted from Medicare claim database. Sensitivity was higher in the inpatient than the outpatient population. Sensitivity was lower in regional disease than distant disease. Positive predictive values varied with the affected organs (38.3%-84.8%). The sensitivity and PPV values were never simultaneously 80% within one

PAGE 60

45 stage of a specific cancer. The findings suggest that Medicare claims data have limited utility for the study of cancer stage due to the high rate of misclassification (143). Another data source that has been used in research is the Group Health Cooperative of Puget Sound (GHC) database. The automated data of GHC was validated to identify complications and co-morbidities of diabetes patients by comparing the ICD-9-CM and Current Procedural Terminology codes with medical chart data (gold standard) (129). The overall sensitivity of diabetes complications was high, but varied by each complication (79.2% for ostomyelitis to 95.2% for myocardial infarction). The PPV was low (mean 46.3%, [8.6%-88.5%], except for amputation (82.9%). Even though, sensitivity was acceptable to detect the complications and co-morbidities of diabetes, the overall PPV was low. Prescription drug data Completeness (or accuracy) of prescription drug data in administrative databases varies depending upon various factors (e.g., whether it is voluntary data collection or mandatory) and the complication of data submission. A study by Kozyrskyj et al. (138), the Drug Programs Information Network (DPIN) electronic prescription claims database in Manitoba, Canada, confirmed that voluntary data collection is prone to data incompleteness. The completeness was assessed by comparing the prescription number from the DPIN with the original pharmacy records. The results showed that prescriptions reimbursed from the provinces drug befit plan, Pharmacare, were 93% (98%CI 92.4% to 93.6%) with the original records. The completeness of prescription were lower for the treaty status Indians (Indian Affairs) and social assistant recipients (Manitoba Family Services/City of Winnipeg Social Services)

PAGE 61

46 compared to the Pharmacare group (79.7%, 98% CI 78.0% to 81.4% and 90.1%, 98% CI 88.8% to 91.4%, respectively). The study found that the completeness of the data was varied by type of drug benefit plans, where the pharmacists receive drug reimbursement from Pharmacare based solely on the data entered in the DPIN, while the drug data for Indian Affairs and the social assistant recipients were voluntarily collected for drug utilization evaluation purposes (138). McKenzie et al. validated the accuracy of Medicaid pharmacy claims for estimating drug use among elderly nursing home residents in Oregon by comparing drug data with medical charts found that the percent agreement and PPV were above 85% for antipsychotics, antidepressants, and anxiolytics (kappa = 0.81, 0.63, and .0.52, respectively) (107). The findings suggest that Medicaid pharmacy claim data are accurate (107). Since the computerized healthcare database in this study is used in place of paper patient records, external data validation (comparing the data from the HI database with other data sources) is not possible. Thus, internal validation methods will be used. Internal validation methods Internal validation of the data has been used widely for disease incidence(159), diagnosis confirmation(151), and adverse drug event measurement(97;160-162). The researchers validated specific data, e.g., disease diagnosis (ICD-9-CM) codes by comparing with other data that represent the disease such as laboratory results or prescription drug data. For example, the ICD-10 for community-acquired pneumonia can be validated with positive results of the blood culture for Streptococcus pneumoniae. Gerstman et al. used prescription drugs (e.g., anticoagulants) with the disease diagnosis (deep vein thrombosis or pulmonary embolism) to identify patients with venous

PAGE 62

47 thromboemobolism (163). Movig et al. validated the ICD-9 codes for hyponatremia (276.1) from the Dutch PHARMO record linkage database by comparing them to laboratory serum sodium (Na+) measurements. The study found that sensitivity was low and only 30% of the cases were found using the ICD-9 compared to the Na+ measurements. However, specificity was high (>99%). (151) Hennessy and his colleagues conducted a macro-level data quality assessment of Medicaid data, the Computerized On-line Medical Pharmaceutical Analysis and Surveillance System (COMPASS) (164) from six states. The authors used descriptive explanation of the data quality based on missing data, unusual presence of the disease in patients with specific characteristics, and the data inconsistency over time periods. The study classified potential data errors into 4 types: incomplete claims for certain time periods; absences of an accurate indicator to identify patient hospitalizations; incomplete hospitalization data for the beneficiaries; and diagnosis codes in demographic groups that would be expected to have a low frequency of the diseases (99). The study first examined missing data by comparing the number of enrollees and the prescription claims over time periods. Secondly, the incompleteness of the data in a suspect group (in this study elderly patients because Medicare data may be more complete) compare with other age groups. Thirdly, the researchers examined the validity of diagnosis and demographic data. The disease diagnoses were cross-checked with patient characteristics: 1) the diseases that are specific to a certain age or gender, e.g., breast cancer in female, childbirth and pregnancy complications in patients younger than 60, and lung cancer in patients older than 40.

PAGE 63

48 Data quality was identified by using graphs of the number of prescription drug claims plotted longitudinally to indicate unusual data (upward or downward peaks). Number of hospitalizations per enrollee was observed by age group. Other analyses were to assess the accuracy of diagnosis code with patients demographics. The level of the graph and its consistency overtime indicate whether the data is valid and reliable, or not. The macro-level data quality assessment methods used in this study suggests that despite the comparison data is not available; it is possible to check internal data validity using various data cross-checking and longitudinal comparisons.

PAGE 64

CHAPTER 3 VALIDATING DATA QUALITY The HI databases contain electronic medical records, pharmacy dispensing data, and administrative information. The databases were designed by a physician to facilitate a report of information required by the Ministry of Public Health, to generate disease statistics, to monitor of healthcare utilization, financing, and lastly for quality improvement purposes. Accordingly, these databases have comprehensive clinical and administrative data. The electronic medical records included patient demographics, type of health insurance, disease diagnosis (ICD-10), and details of dispensed medications, some physical examination information and laboratory test values (e.g., blood pressure, virology and serology test results). However, laboratory data were incomplete and not usable for this study. Drug names, drug classes, dosage, amount of drug supplies, and the instruction for use were completely available in the pharmacy dispensing database. The clinical information appears on the computer screen when healthcare staff renders hospital services. Physicians are able to view past medical history, previous diagnosis and treatment, including the medications that have been dispensed. Any new clinical information obtained during a patient encounter is immediately updated in the database system. Once the physicians prescribe medications to the patients, the order is sent electronically to the pharmacy department of the hospital and, as a result, most prescriptions are dispensed at the hospital pharmacy. Moreover, only prescriptions that are dispensed at the hospital pharmacy are covered by most health insurance plans, 49

PAGE 65

50 except the CSMBS group while provide retrospective reimbursement for the prescriptions dispensed at retail pharmacies. Most CSMBS patients fill prescriptions at the hospital pharmacy, if the drugs were available in the hospital drug formulary. Because dispensing activities are directly limited to the electronic record, we assumed that drug dispensing data from these hospital databases represent hospital drug utilization for both inpatients and outpatients. Since the HI database system was implemented in the hospitals, paper medical records are no longer used. Healthcare providers at the hospitals rely solely on these electronic medical records for clinical intervention and administrative purposes. For this reason, data validation could not be performed by comparing the electronic medical records and pharmacy database with the medical chart in paper format. The databases were selected for this study because they are able to capture most aspects of care provided by the community government hospitals, while are the main provider of primary and secondary care to the 30 HI beneficiaries. By applying the patient-specific healthcare data from the 19 government hospitals in Ubonratchatani province, the researcher was able to measure drug utilization covered by the 30-Baht HI policy at the provincial level. The researcher used two steps to obtain the datasets from the target hospitals: 1) an official letter for data request to the Ministry of Public Health, Thailand, who has authority to suggest the hospitals participate in the research project and to allow the data to be used for research purposes, 2) interviews wtih hospitals directors and database managers regarding data quality. The request was approved by the Minister of Public Health, and the approval communicated to the 19 targeted hospitals in December 2003.

PAGE 66

51 Fourteen of 19 hospitals agreed to provide the data. Two hospitals did not agree to provide the data, two hospitals did not use the HI database, and one is a tertiary care hospital that use different database system. One tertiary care hospital in Ubonratchatani was not included in this study because it was not the main provider of primary care of the patients covered by 30-Baht HI policy. Officially, services by tertiary care hospitals are only covered under the 30-Baht HI policy if patients are referred from the community government hospitals. The HI Database System was used only in the community government hospitals in Ubonratchatani province. Healthcare data extracted for data validation analysis in this study were: 1) patient demographics, 2) disease diagnoses and, 3) dispensed drugs. These data were the core elements to this study in identifying the effect of the 30-Baht HI policy on hospital drug utilization in the affected population and in the control group. The data in the HI database were linked with a unique patient identification number (HN) assigned when patients first visited the hospital. For demographic data, the researcher included age, gender, marital status, and job based on the data availability. Disease diagnoses were extracted separately for inpatient and outpatient services using discharge diagnosis codes. To be able to understand, interpret and generalize the results of this study, the researcher interviewed experts that were involved with the HI database regarding data quality, and the healthcare providers from the hospitals who used the database and were experienced in patient care. The interview had two major purposes: 1) to explore the possible errors from data collection, data coding, and data entry, and 2) to identify extraneous factors that could compromise internal validity of the effect of the policy on drug use. Details of each expert interview were described as follows:

PAGE 67

52 Interview of Hospital Directors Interviews with hospital directors of 8 community hospitals were conducted to gather information of how well the data in the databases represent the actual clinical practice, clinical interventions, and the treatment. The researcher interviewed one hospital director of each hospital. If the hospital director was not available, another knowledgeable practitioner in the hospital was interviewed. The following issues were discussed: Completeness and accuracy of patient-specific data in the database Factors affecting prescribing behaviors that might have an effect on drug utilization rates rather than or in addition to the 30-Baht HI policy (choice of drugs, amount of medical supplies, hospital admission decision);disease severity, type of patients, (health insurance benefit), drug price, or physical factors (physical factors, i.e., clinic schedule, census) Interviews of Hospital Database Managers To explore the validity of the data, interviews of database managers were conducted to determine the coding system, relationship of the databases (i.e., links between medical database and pharmacy database), limitations of the software, similarity of the database between hospitals, and known problems reported from the current users. Completeness of the data, potential information bias (e.g., from coding errors), frequent missing data, and other possible factors that lead to lower quality of the data was also gathered. Finally, eight hospital databases from 19 community government hospitals were selected for data validation based on data availability and the qualitative information provided by the hospital directors and the database managers. Quantitative Assessment of Data Quality The assessment of the quality of the hospital data was extensively conducted because the data had not been used for drug utilization research purposes. Quantitative

PAGE 68

53 assessment of the data was performed in regards to four major characteristics: descriptive analysis of the HI database and data characteristics, face validity, missing data, and coherence of the data. If the data quality did not met the criteria specified in each validation process, the data from that hospital was excluded from the study. Details of how to handle the data problems and the data exclusion criteria are discussed below. Database Characteristics Initially, the researcher explored the data structure and located patient-specific data on demographics, disease diagnoses, types of health insurance plan, and dispensed drugs. Next, the linkage was established in the HI database among patient demographics, disease diagnoses, types of health insurance, and dispensed drugs. In this study, the database was designed to link patient demographics and diagnosis data by medical record number and hospital admission number. However, the diagnosis and prescription drug data from the pharmacy database was linked by prescription number. The differences of the record linkage posed some difficulties in combining the data. Missing Data and Outliers Descriptive statistics were generated to present types and magnitude of missing data in the databases. Missing data from the following variables were examined: patient demographics, types of health insurance, disease diagnosis codes (ICD-10), lists of prescription drugs, volume of drug dispensed, and directions of drug use. For ICD-10, the cases that contained missing diagnosis codes were excluded from the data analysis. Unusual and outlier values resulting from data entry error of patient demographics, disease diagnosis codes, and prescription drugs and doses were identified based on the

PAGE 69

54 clinical knowledge and experience of the researcher. If unusual values appeared to occur at random and did not exceed 5% of the total number of values for a given variable, they were deleted from the dataset (107). If the unusual values exceeded 5%, the whole database of the hospital would have been excluded from the study, but this was not necessary. Face Validity (Plausibility of the Data) If data were similar among the studied hospitals and the studied years the researcher assumed that the data were valid and could be combined among the hospitals for further analysis of the policy. The assessment of the plausibility of the data focused only on the five most prevalent drugs that were prescribed in the studied period. The five most prevalent drugs were compared among hospitals, and studied years. Secondly, the five most prevalent drugs were compared with disease prevalence/incidence provided by the provincial disease statistics. If the five most dispensed drugs were consistent with the diseases epidemiology, it was assumed that the hospital drug data was valid. Data Coherence The researcher applied the internal data validation method to assess validity of the data recommended by Hennessy et al., when external validation was not possible. Hennessy and his colleagues conducted a macro-level data quality assessment of Medicaid data, the Computerized On-line Medical Pharmaceutical Analysis and Surveillance System (COMPASS) (164) from six states. The authors used descriptive explanation of the data quality based on missing data, unusual presence of the disease in patients with specific characteristics, and the data inconsistency over time periods. The study classified potential data errors into 4 types: incomplete claims for certain time periods; absences of an accurate indicator to identify patient hospitalizations; incomplete

PAGE 70

55 hospitalization data for the beneficiaries; and diagnosis codes in demographic groups that would be expected to have a low frequency of the diseases. (99) The study first examined missing data by comparing the number of enrollees and the prescription claims over time periods. Secondly, the incompleteness of the data in a suspect group (in this study elderly patients because U.S. Medicare data may be more complete) was compared with other age groups. Thirdly, the researchers examined the validity of diagnosis and demographic data. The disease diagnoses were cross-checked with patient characteristics: 1) the diseases that are specific to a certain age or gender, e.g., breast cancer in female, childbirth and pregnancy complications in patients younger than 60, and lung cancer in patients older than 40. Data quality was identified by using graphs of the number of prescription drug claims plotted longitudinally to indicate unusual data (upward or downward peaks). Number of hospitalizations per enrollee was observed by age group. Other analyses concerned the accuracy of diagnosis code with patients demographics. The macro-level data quality assessment methods used in this study suggests that even though the comparison data was not available it was possible to check data validity using various internal data cross-checking and longitudinal comparisons. The data was considered valid if the percent coherence was equal to or greater than 80% (113). Coherence was measured as a percent agreement between two related variables, e.g., a particular disease diagnosis and drugs dispensed to treat the disease, or a gender-specific disease and the corresponding gender.

PAGE 71

56 Validating disease diagnosis codes (ICD-10) Coherence between ICD-10 and patient demographics The percent coherence was observed across the studied hospitals and among the studied years. Graphic plots of the percent coherence of each hospital by study year were generated and compared with the plots from other hospitals. Coherence between a gender-specific disease and gender was checked for the following conditions: Pregnancy x female Prostate cancer x male Coherence between ICD 10 and dispensed drugs Three disease diagnosis codes were cross-checked against the drugs usually used to treat these diseases, including diabetes and antidiabetic drugs, hypertension and antihypertensive drugs, and bacterial pneumonia and antibiotics. Gastrointestinal tract infections were not validated because there are varieties of treatment methods depending on the pathogens, severity and the onset of type symptoms. Coherence level was cautiously interpreted because it was not only associated with reliability of the data, but also with the appropriateness of drug prescribing. Even though, the measure was not an exact measure of validity, it provided valuable information about reliability, in the absence of a gold standard. Diabetes ((ICD10=E1xx) x all antidiabetes drugs in the hospital drug formulary Hypertension (ICD10 = I1xx) x all antihypertensive drugs in the hospital drug formulary Bacterial pneumonia (ICD10= J13x or J15x or J16x or J17x or J18x) x all prescribed antibiotics that were effective for the treatment of pneumonia

PAGE 72

57 Drug data To identify coherence of drug data, we cross-checked the antidiabetic drugs with the diagnosis codes of diabetes. The researcher calculated the percent of unmatched data between drugs and the diabetes diagnosis codes, i.e., the percentage of patients who were prescribed insulin without a diagnosis code of diabetes. The percent coherence and the percent mismatched was summarized in Chapter 4.

PAGE 73

CHAPTER 4 DATA VALIDATION RESULTS Hospitals Demographics The HI database system is used in 16 out of 18 community government hospitals in Ubonratchatani province (2004).(Table 4-1) The HI databases were installed in the hospitals at different times depending on the financial and staff readiness to implement this database system. The completeness and accuracy of the data appeared to be associated with the length of time the databases have been used. Of 18 community government community hospitals in Ubonratchatani province, Thailand, eight hospitals had complete data for disease diagnosis, dispensed drugs, and types of health insurance available for the entire study period, six hospitals had incomplete data in 2000, two hospitals did not use the HI database system, and two hospitals refused to provide the data for our study. Locations of the eight included hospitals suggest good geographic representation of the government community hospitals in Ubonratchatani province. (Figure 4-1) Descriptive information of hospital service utilization, including the number of medical records in the studied period, hospital visits, hospital admissions, and prescription drugs is shown in Table 4-1. The capacity of the government community hospitals varied from 30 to 90 beds. The medical record data or censuses ranged from 36,572 to 179,986 per hospital. This information corresponds with the size of the hospitals. The numbers of hospital visits, hospital admissions, and prescriptions during the study period also appeared to correlate with the size of the hospitals. 58

PAGE 74

59 Figure 4-1. Geographic locations of the included eight community government hospitals in Ubonratchatani province, Thailand Table 4-1. Hospital information, numbers of hospital visits, hospital admissions, and prescriptions from 2000 to 2003 Hospital Number of beds Start HI database (year) Number of medical records Number of hospital visits Number of hospital admissions Number of prescriptions A 30 1999 50,840 264,097 12,832 564,058 B 60 1998 108,131 494,615 31,294 1,147,042 C 30 1999 62,990 254,670 11,766 579,203 D 30 1998 45,075 227,583 11,296 573,720 E 90 1999 179,986 711,957 46,418 1,770,982 F 60 1997 166,116 640,928 34,188 1,373,993 G 60 1998 36,572 568,611 31,839 1,368,213 H 30 1999 63,788 410,585 24,152 902,064

PAGE 75

60 Healthcare providers were able to enter up to 16 diagnosis codes in the HI database. However in most instances, one to two disease diagnosis codes were used at each visit/admission. There were higher percentages of patients assigned with two or more diagnosis codes in inpatient than outpatient settings. For outpatient services, the percent of patients assigned two or more diagnosis codes were 8.3% in 2000 and increased to 19.6% in 2003, while the percentage increased in inpatients from 10.03% to 49.3%. (Figure 4-2) The hospital directors commented in the interviews that at the beginning of the implementation of the HI database system, there was a dual (paper and electronic) system to keep medical records until the systems were considered stable. Because of limitations in numbers of data entry personnel, only data that were important for patient care and billing were entered. In addition, physicians and the data entry personnel were not familiar with the ICD-10 lists in the databases, thus, it required more time to enter more diagnosis codes into the database. An increased number of diagnosis codes does not suggest an increase in co-morbidity, but rather increasing comprehensiveness in coding as personnel became more familiar with the system. 0102030405060% patients with two or more ICD10 2000200120022003Year outpatient inpatient Figure 4-2. Percent of patients with two or more disease diagnosis codes assigned in the HI database (Hospital F)

PAGE 76

61 Database Characteristics Healthcare data used for our study were stored in four main databases: 1) TBL containing table files, e.g., data coding sheets, and patient demographics, 2) IPD containing data related to inpatient services, 3) OPD containing data related to outpatient services, and 4) PHM containing dispensed drug data. Patient demographics, types of health insurance, disease diagnosis, and dispensed drug data were linked using unique identification numbers (in our study, the identification numbers were deidentified by the hospitals). (Figure 4-3) TBL folder containing the following tables: o PT table containing patient demographic data o Code book tables for the data in HI database IPD folder: o IPTyear containing admission date and time, types of health insurance, and length of hospital stay o IPTDXyear: hospital admission date and time, and disease diagnoses OPD folder: o OVSTyear: hospital visit date and time, types of health insurance o OVSTDX: hospital visit date and time, and disease diagnoses PHM folder: o PRSCyear: unique patient identifier, prescription number o PRSCDTyear: prescription number and drug data The HI database has been continuously updated, e.g., changing disease diagnosis coding system from ICD-9 CM to ICD-10, adding drug codes for new drugs in the hospital drug formulary. Additionally, the system was supplementedwith drug information tools such as drug interaction database to improve quality of service. Based

PAGE 77

62 on its elements, the HI database appears to be a comprehensive source of computerized patient-specific healthcare data, including drug utilization. Figure 4-3. Linkage among the selected patient data in the HI database AN: scrambled hospital admission number; DRUGITEM: drug item code; HN: hospital number (scrambled medical record number); ICD10: disease diagnosis code; IPD: inpatient service folder; IPT: inpatient admission folder; IPTDX: inpatient diagnosis folder; OPD: outpatient service folder; OVST: outpatient visit folder; OVSTDX: outpatient diagnosis folder; PRSCNO: prescription number; PHM: pharmacy folder; PRSCD: prescription and patient information; PRSCDT: dispensed drugs; PTTYPE: type of health insurance; REGISTERDATE: hospital admission date; REGISTERTIME: hospital admission time; ; TBL: table folder; VISITDATE: hospital visit date; and VISITTIME: hospital visit time; TBL PT HN PTTYPE GENDER OCCUPATION MARITAL STATUS BIRTHDATE IPD OPD PHM PRSC PRSCNO HN PRSCDT PRSCNO DRUGITEM OVST HN VISITDATE VISITTIME PTTYPE OVSTDX HN VISITDATE VISITTIME ICD10 IPT HN AN REGISTERDATE REGISTERTIME PTTYPE IPTDX AN ICD 10

PAGE 78

63 Face Validity Comparisons of dispensing frequencies s uggested that the most frequent by dispensed therapeutic categories were simila r across hospitals.(Tabl e 4-2) Analgesics, antidiabetic drugs, antianxiet y drugs, antibiotics, vitamin supplements, antihypertensive drugs, and drugs for gastrointestinal disorder s were most prevalent among the hospitals and across the studied period. Analgesic dr ugs including acetaminophen in both tablet and solution preparation, and its combination with muscle relaxants, e.g., orphenadrine. NSAIDs (ibuprofen and indomethacin) we re commonly found. The most commonly dispensed antidiabetic drug was glibenclamide, while metformin was prevalent in only some hospitals (6 out of 8 hospitals). Antianxiety drugs included diazepam in various strengths (2-10 mg), furazepam and amitrip tyline. In the antibiotic class, only amoxicillin was found within the top five prescribed drugs. Vitamin B complex, multivitamin, and ferrous sulfate were the most prevalent vitamin supplements. For antihypertensive drugs, hydrochl orothiazide and enalapril were prescribed frequently. Lastly, antacid and H2-receptor antagonist (r anitidine) were found within the five most prevalent dispensed drugs across hospitals and over time.

PAGE 79

64 Table 4-2. Five most prevalent dispensed drugs for inpatient and outpatient use of the eight included hospitals (A-H) during from 2000 to 2003 Hospital 2000 2001 2002 2003 A Para* 500 Mg Tab Para 500 Mg Tab Para 500 Mg Tab Para 500 Mg Tab Para Syr. 120 Mg/5ml Para Syr. 120 Mg/5ml Para Syr. 120 Mg/5ml Para Syr. 120 Mg/5ml Diazepam 2 Mg Ibuprofen 400 Mg Tab. Diazepam 2 Mg Diazepam 2 Mg Ibuprofen 400 Mg Tab. Diazepam 2 Mg Mtv Coated Tab. Mtv Coated Tab. Glyceryl Guiacolate Syr Ammonium Carbonate Mixture Para 325 Mg Tab. Para 325 Mg Tab. B 2000 2001 2002 2003 Paracetamol Tab. 500 Mg Paracetamol Tab. 500 Mg Paracetamol Tab. 500 Mg Paracetamol Tab. 500 Mg Paracetamol Syr.120 Mg/5ml Ibuprofen Tab. 400 Mg Ibuprofen Tab. 400 Mg Ibuprofen Tab. 400 Mg Para 450 +Orphenadrine 25 Mg Para 450 +Orphenadrine 25 Mg Para 450 +Orphenadrine 25 Mg Para 450 +Orphenadrine 25 Mg Bromhexine Tab. 8 Mg Paracetamol Syr.120 Mg/5ml Paracetamol Syr.120 Mg/5ml Paracetamol Syr.120 Mg/5ml Ascorbic Acid Tab. 100 Mg Amoxycillin 125mg/5ml Amoxycillin Cap. 500 Mg Ranitidine 150 Mg Tab C 2000 2001 2002 2003 Paracetamol Tab. 500 Mg Paracetamol Tab. 500 Mg Paracetamol Tab. 500 Mg Paracetamol Tab.500 Mg Paracetamol Syr.120 Mg/5ml Paracetamol Syr.120 Mg/5ml Paracetamol Syr.120 Mg/5ml Paracetamol Syr.120 Mg/5ml Diazepam Tab. 2mg Vitamin B Complex Tab. Vitamin B Complex Tab. Vitamin B Complex Tab. Antacid Suspension 240 Ml Diazepam Tab. 2mg Diazepam Tab. 2mg Glibenclamide Tab. 5 Mg Indomethacin Cap. 25 Mg Antacid Suspension Antacid Suspension 240 Ml Diazepam Tab. 2mg D 2000 2001 2002 2003 Para 500 Mg Tab. Para 500 Mg Tab. Para 500 Mg Tab. Para 500 Mg Tab. Diazepam 2 Mg Para Syr.120 Mg/5ml Para Syr.120 Mg/5ml Para Syr.120 Mg/5ml Multivitamin Tab Diazepam 2 Mg Ibuprofen 200 Mg Tab. Norgesic Para Syr.120 Mg/5ml Multivitamin Tab Balm Balm Diazepam 5 Mg Diazepam 5 Mg Ammonium Carbonate Ibuprofen 400 Mg

PAGE 80

65 Table 4-2. Continued Hospital 2000 2001 2002 2003 E Paracetamol Tab. 500 Mg Paracetam ol Tab. 500 Mg Paracetamol Tab. 500 Mg Paracetamol Tab. 500 Mg Alumina And Magnesia Susp. Di azepam Tab. 5mg Diazepam Tab. 5mg Alumina And Magnesia Susp. Diazepam Tab. 2mg Diazepam Tab. 2mg Diazepam Tab. 2mg Diazepam Tab. 5mg Diazepam Tab. 5mg Alumina And Magnesia Alumina And Ma gnesia Diazepam Tab. 2mg Ibuprofen Tab. 400 Mg Ibuprofen Tab. 400 Mg Ibuprofen Tab. 400 Mg Paracet.500 +Orphenadrine F 2000 2001 2002 2003 Para 500 Mg Tab. Para 500 Mg Tab. Para 500 Mg Tab. Para 500 Mg Tab. Glibenclamide 5 Mg Para Syr.120 Mg/5ml 60 Ml Para Syr.120 Mg/5ml 60 Ml Para Syr.120 Mg/5ml 60 Ml Hctz Tab. 50 Mg Glibenclam ide 5 Mg Amoxy 250 Mg Balm Para Syr.120 Mg/5ml 60 Ml Ammonium Carbonate Mixture Balm Ranitidine 150mg Ammonium Carbonate Hctz Tab. 50 Mg Ammonium Carbonate Amoxy 250 Mg G 2000 2001 2002 2003 Para 500 Mg Tab. Para 500 Mg Tab. Para 500 Mg Tab. Para 500 Mg Tab. Diazepam 2 Mg Diazepam 2 Mg Diazepam 2 Mg Diazepam 2 Mg Para Syr.120 Mg/5ml 60 Ml Ranitidine 150 Mg. Tab. Para Syr.120 Mg/5ml 60 Ml Para Syr.120 Mg/5ml Ranitidine 150 Mg. Tab. Para Syr.120 Mg/5ml Furapam Tab. 500 Mg Ranitidine 150 Mg. Tab Furapam Tab. 500 Mg Furapam Tab. 500 Mg Ranitidine 150 Mg. Tab. Furapam Tab. 500 Mg H 2000 2001 2002 2003 Paracetamol Tab. 500 Mg Paracetam ol Tab. 500 Mg Paracetamol Tab. 500 Mg Paracetamol Tab. 500 Mg Paracetamol Syr.120 Mg/5ml Paracetamol Syr.120 Mg/5ml Paracetamol Syr.120 Mg/5ml Paracetamol Syr.120 Mg/5ml Paracetamol+Orphenadine Diazepam Tab. 2mg Antacid Suspension 240 Ml Ranitidin e 150 Mg Tab. Antacid Suspension 240 Ml Multivitamin Tab. Multivitamin Tab. Ibuprofen Tab. 400 Mg Multivitamin Tab. Antacid Suspension Ibuprofen Tab. 400 Mg Antacid Suspension *PARA: paracetamol or acetaminophen

PAGE 81

66 Drug utilization pattern were consistent with the data from Provincial Disease Statistics Report in 2002 (165). (Table 4-3) The 10 most common causes of morbidity for outpatient servies include respiratory tract and gastrointestinal tract disorders. musculoskeletal disorders, and infectious diseases. Table 4-3. Provincial Disease Statistics (2002) on the 10 most common causes of morbidity for outpatient services, Ubonratchatani province, Thailand Cause of morbidity Number of cases Incidence/1,000 population Respiratory tract disorders 857,007 483.22 Gastrointestinal tract disorders 572,429 322.76 Musculoskeletal system disorders 301,827 170.18 Infectious diseases 249,033 140.42 Endocrine disorders 230,662 130.06 Unindentified causes 213,488 120.37 Skin and connective tissue disorders 195,440 110.20 Urogenital disorders 168,024 94.74 Metal and behavioral disorders 161,917 91.30 Cardiovascular diseases 147,528 83.18 Source: Annual Epidemiological Surveillance Report 2002. Ministry of Public Health, Thailand (165) The results suggest that the patterns of drug prescribing were not different among the studied hospitals. Similar patterns of dispensed drugs across the studied years indicate that validity of drug data was maintained over time. The congruence between the most frequently dispensed drugs and disease incidences suggests that the data were accurate. Missing Data Small numbers of missing data were observed in patient demographics, types of health insurance, disease diagnosis codes for outpatient and inpatient service (defined as presence of at least 1 disease diagnosis code), and prescription drugs among hospitals and across the studied period. The data in Table 4-4 shows that age (date of birth) contains

PAGE 82

67 0.02-0.22% missing data, while gender, marital status, and occupation have 0% missing, with an exception of hospital A, C and D, wh ich have a small percentage of missing data in one or two characteristics. This suggests the demographi c data were sufficient for subgroup analyses. There was no missing data on disease di agnosis of inpatient data in most hospitals, except hospital F (0.02%) (Table 4-5). However, it is unknown whether all disease diagnoses related to a patients conditions were recorded in the database. There were higher percentages of missi ng data in the outpatient da taset with a median of 1.83% [0.05, 3.13] compared with the i npatient data. This result mi ght suggest that missing data was not specific to a particular hospital, but rather the process of data entry in outpatient and inpatient care. Pharmacy databases of the HI databa se systems contained small amounts of missing data with a median of 0.04% [0.01, 0.08] for drug codes assigned to dispensed drugs. The percent missing data on drug qua ntity were higher (0.25% vs. 0.04%). Hospital A and G had complete data on drug qua ntity. (Table 4-6) Again, we could not verify whether all dispensed drugs were entered into the database. Overall, patient demographics, disease diagnosis, and drug data contained small percentage of missing data, which was acceptable by common standards (<5%). The results from the analysis of missing data sugge st that the HI database had sufficient data quality and could be used for the analyses in our study.

PAGE 83

68 Table 4-4. Missing data on patient demographics Hospital Missing data Data/Hospital A B C D E F G H Numbers of patients N=50,840 N=108,117 N=62,974 N=4,900 N= 179,986 N= 166,116 N=36,572 N=63,788 Age 10 (0.02) 240 (0.22) 11 (0.02) 0 40 (0.02) 63 (0.04) 24 (0.06) 80 (0.12) Gender 0 0 0 0 1 (<0.01) 0 0 0 Marital status 0 0 154 (0.24) 0 0 0 0 0 Occupation 2 (<0.01) 0 32 (0.05) 0 0 0 0 0 Table 4-5. Missing data on diseas e diagnosis codes for inpatient and outpatient data of ei ght studied hospitals (2000-2003) Missingdata(%) Hospital A B C D E F G H Inpatient ICD-10 0/16,253 (0) 0/44,779 (0) 0/16,647 (0) 1/16043 (0) 2/63923 (0) 7/34461 (0.02) 0/41756 (0) 0/34323 (0) Outpatient ICD-10 140/264,086 (0.05) 11,940/443,989 (2.69) 7585/242,558 (3.13) 5874/305,427 (1.92) 13447/792,365 (1.70) 5143/470,923 (1.09) 9497/556,875 (1.70) 8130/407,121 (2.00) Table 4-6. Missing data on prescribed drugs of eight studied hospitals over four years (2000-2003) Missingdata(%) Hospital A N= 564,058 B N= 960,965 C N= 429,641 D N= 573,720 E N= 1,770,976 F N= 1,030,151 G N= 1,368,213 H N= 902,064 Drug name 279 (0.05) 255 (0.03) 152 (0.04) 164 (0.03) 652 (0.01) 601 (0.06) 1081 (0.08) 411 (0.05) Prescribed quantity 0 2573 (0.27) 2066 (0.48) 1433 (0.25) 3919 (0.22) 12248 (1.19) 2 (0) 2479 (0.03)

PAGE 84

69 Data Coherence Disease Diagnosis and Gender The percent coherence between gender-specific diagnosis and gender was high for pregnancy and female among eight hospitals over the studied period with a median of 98.89% [95.42, 99.86]. Inpatient data tended to have a higher percent coherence than outpatient data (median 99.18% [96.63, 99.86] vs. 98.68% [94.74, 99.83]). The accuracy remained high from the beginning of the study period to the end. There was no significant change of the percent coherence over time. Disease diagnosis with benign prostate hyperplasia and prostate conditions validated with male gender yielded 100% accuracy in every hospital, except hospital E in the year of 2000 (87.5%). There were two hospitals (Hospital A and D) that did not have any cases of the above conditions in 2001 and 2002, thus the percent coherence could not be calculated. Disease Diagnosis and Drugs The researcher validated the accuracy of the disease diagnosis codes by calculating the percent of persons who were diagnosed with the disease and received medications. The diagnosis codes for three disease states, diabetes, hypertension, and pneumonia, were included in the validation of the accuracy of the diagnosis code. The disease diagnosis codes for gastrointestinal tract infections were not validated because there were a variety of drug therapy regimens for these conditions, and which in tern are not specific enough to identify the percent coherence between the disease diagnosis codes and the drugs. Diabetes and antidiabetic drugs The type of antidiabetic drugs available in the hospital formulary varied among the hospitals. However, every hospital included oral antidiabetics and insulin. Most of

PAGE 85

70 the hospitals contained only glibenclamide in the sulfonylurea class. Chlorpropamide were found in only 6 hospitals (except A and H). Glipizide and glyburide were found in only 2 hospitals (Hospital B and G). In the biguanide group, only metformin was included in the hospital drug formulary in all hospitals. Regular and intermediate-acting insulin (NPH) were included in the drug formulary of every hospital. Disease diagnosis codes and antidiabetic drugs showed a median percent coherence of 89.81% [74.56, 97.71]. (Figure 4-4) The percent coherence of the inpatient population was lower than the findings for the outpatient populations, 80.44% [65.20, 89.47] (Figure 4-5). The findings for both inpatients and outpatients appear reasonable as a small proportion of diabetes patients is typically managed with dietary restrictions only. In our study, the researcher used only ICD-10 codes to select patients with the disease for the measurement of drug utilization and hospital visits/admission. Thus, accuracy of the diseases diagnosis is crucial for patient identification of the disease. Using disease diagnosis with low accuracy might underestimate drug use in a population and impose misclassification bias.

PAGE 86

71 0.0010.0020.0030.0040.0050.0060.0070.0080.0090.00100.002000200120022003YearData coherence (%) A B C D E F G H Figure 4-4. Percent data coherence between disease diagnosis of diabetes and antidiabetic drugs of outpatient data among eight hospitals, 2000-2003 0.0010.0020.0030.0040.0050.0060.0070.0080.0090.00100.002000200120022003YearData coherence (%) A B C D E F G H Figure 4-5. Percent data coherence between disease diagnosis of diabetes and antidiabetic drugs of inpatient data among eight hospitals, 2000-2003.

PAGE 87

72 Hypertension and antihypertensive drugs Antihypertensive drugs included in every hospital formulary were diuretics, beta-blockers, alpha-blockers, and ACEIs. Most of the drugs were diuretics, including hydrochlorothiazide (HCTZ), furosemide, amiloride and the combination of the two drugs (amiloride and HCTZ). Calcium channel blockers (verapamil and felodipine) were included in only three hospitals (B, E and H). There were limited choices of beta-blockers (propranolol and atenolol) and ACEIs (enalapril and lisinopril). However, they were available in every hospital. The median percent coherence between the diagnosis of hypertension (I1xx) and the above antihypertensive medication was 89.53% [80.81, 95.56] for the outpatient population. Similar coherence levels were found across hospitals and studied years. (Figure 4-6) For the inpatient population, the percent coherence varied with a median of 80.24% [58.46, 100] (Figure 4-7). Four hospitals showed lower coherence in 2000 and 2001 than in the following years. A median of 10-20% of the patients who had a diagnosis of hypertension did not receive any antihypertensive drugs, which appears reasonable as some patients may try behavioral changes prior to drug treatment. Alternatively, patients may not be able to afford drugs, or patients may have obtained antihypertensive drugs from other sources. Therefore, it is not appropriate to conclude misclassification for diagnosis for diabetes or hypertension. While the coherence levels appear realistic no final decision about diagnosis accuracy can be made. Coherence levels varied more for outpatient data when individual hospitals were compared, suggesting that either diagnosis coding accuracy or the completeness of drug records may vary.

PAGE 88

73 0.0010.0020.0030.0040.0050.0060.0070.0080.0090.00100.002000200120022003YearData coherence (%) A B C D E F G H Figure 4-6. Percent data coherence between disease diagnosis of hypertension and antihypertensive drugs of outpatient data among eight hospitals, 2000-2003 0.0010.0020.0030.0040.0050.0060.0070.0080.0090.00100.002000200120022003YearData coherence (%) A B C D E F G H Figure 4-7. Percent data coherence between disease diagnosis of hypertension and antihypertensive drugs of inpatient data among eight hospitals, 2000-2003

PAGE 89

74 Bacterial pneumonia and antibiotics Since there was no official treatment guideline for the treatment of bacterial pneumonia implemented in the community hospitals, we selected all antibiotics that had an indication for the treatment of bacterial pneumonia based on Clinical Pharmacology Online 2005 drug information database (166). These include beta-lactams (e.g., amoxicillin, penicillin, and cloxacillin), cepalosporins (e.g., cephazolin, ceftriazone), macrolides (e.g., erythromycin, tetracycline, doxycycline, and roxithromycin), fluoroquinolone (e.g., norfloxacin, ciprofloxacin), aminoglycosides (e.g., gentamicin), sulfamethoxazole and trimethoprim, and lincomycin. The selected antibiotics had indications for multiple infectious diseases and thus, were non-specific for the treatment of bacterial pneumonia. Bacterial cultures were not commonly ordered in the community hospitals, i.e., empirical treatment was common. Most of the hospitals similar coherence levels between disease diagnosis of bacterial pneumonia and antibiotics with a median of 92.52% [64.25, 100] for outpatient service and 89.20%, [42.67, 100] for inpatient services. Hospital E had distinctively lower coherence compared with other hospitals 72.08%, [64.25, 72.47] and 53.95%, [42.67, 59.33] for outpatient and inpatient service, respectively. All patients who were diagnosed with bacterial pneumonia should be given an antibiotic. From this analysis, 7% of the outpatient population did not receive any antibiotics. The number was higher in the inpatient population. These discrepancies might be the results of coding errors. Thus, using the ICD-10 codes in identifying patients with bacterial pneumonia approximately 10% of the patients may be falsify classified as having pneumonia, or may have missing medication records.

PAGE 90

75 0.0010.0020.0030.0040.0050.0060.0070.0080.0090.00100.002000200120022003YearData coherence (%) A B C D E F G H Figure 4-8. Percent data coherence between disease diagnosis of bacterial pneumonia and antibiotics of outpatient data among eight hospitals, 2000-2003 0.0010.0020.0030.0040.0050.0060.0070.0080.0090.00100.002000200120022003YearData coherence (%) A B C D E F G H Figure 4-9. Percent data coherence between disease diagnosis of bacterial pneumonia and antibiotics of inpatient data among eight hospitals, 2000-2003

PAGE 91

76 Drugs and Disease Diagnosis Antidiabetic drugs and diabetes The median percent data coherence of antidiabetic drugs with the diagnosis of diabetes had a median of 99.16% [95.09, 100], indicating good data accuracy. There were small variations of the percent coherence of antidiabetic drugs and disease diagnosis codes, however they were not significant. The percent coherence of Hospital E decreased slightly over the studied period (98.76%-95.46%). (Figure 4-10) Since antidiabetic drugs have few non-diabetes indications, only small discrepancies between codes for antidiabetic drugs and diagnosis codes of diabetes should be expected. 92.0093.0094.0095.0096.0097.0098.0099.00100.002000200120022003YearData coherence (%) A B C D E F G H Figure 4-10. Percent data coherence between antidiabetic drugs and disease diagnosis of diabetes among eight hospitals, 2000-2003 No comparisons between drugs and diagnosis were conducted for antihypertensive drugs and hypertension, and bacterial pneumonia and antibiotics, because both drug groups have multiple indications and thus low coherence levels would be expected.

PAGE 92

77 The results from data validation analysis using data coherence method suggest that disease diagnosis code had sufficient accuracy to be used to identify patients with diseases. The disease diagnosis codes captured approximately 90% of the patient who received corresponding drug therapy. Dispensed drugs in this database were a good indicator only for patients with diabetes. At least 10% of diabetes patients were not captured by antidiabetic drugs. In conclusion, the results of the validation of data quality, including the observation of database characteristics, the assessment of missing data, face validity of drug data, and the assessment of data coherence, suggest that the data contains sufficient quality for the analysis of the effect of the 30-Baht HI policy on hospital drug utilization. As stated earlier, the assessment lacked comparison with a gold standard, thus, it is recommended to revalidate data quality before using the database for other purpose. Results from the Expert Interviews A summary of the interviews with representatives of 8 study hospitals is summarized below. Details of the interviews are presented in Appendix E. Regarding major health problems in the community, diabetes, hypertension, acute illness related to seasons, e.g., cold or muscle soreness in rainy season related to working in the rice field, and infectious diseases, e.g., tuberculosis, Infectious diarrhea, urinary tract infections, were common. HIV/AIDS was one of the prevalent conditions in this area. Most hospital directors commented that the hospitals did not have any specific action plan to accommodate increases in census due to the 30-Baht HI policy, except in one hospital where the P&T committee had set up plans for patient referral and drug dispensing (allowed local public health office to dispense some antibiotics (e.g., amoxicillin, tetracycline). Regarding the impact of the 30-Baht HI policy, the hospital

PAGE 93

78 directors found there was a natural increase in numbers of patients visiting the hospitals, but no significant change in the numbers of hospital visits by diabetes patients. There was no difference in drug choice amongst different health insurance plans. Prescribing patterns were reported as beingpatient centered that focused on the patients needs. In addition, prescribing choice was mostly restricted by the hospital drug formulary. One hospital reported that CSMBS beneficiaries received more drug supply per visit. Another hospital reduced the duration of supplied drugs (e.g., dispensing of antidiabetic drugs was changed from 3-month to 2-month supply after the 30-Baht HI policy was implemented. Regarding other factors that might affect drug utilization rates, every hospital director referred to the Hospital Accreditation program that helps foster hospitals to set standards and assure equity of care regardless of patients health insurance status. One hospital initiated patient education programs for patients with chronic diseases (diabetes and hypertension). In addition, this hospital adapted nine clinical practice guidelines from the Royal Medical School, Thailand for the treatment of common illnesses (e.g., infectious diseases, and asthma). Experts from three from eight interviewed hospitals reported no data available in 2000. Of the five hospitals that had data during the study period, the experts commented that patient data in the HI database were accurate and useful for providing information on hospital services and administrative tasks. However, several hospital directors commented that physicians might not know all the available disease diagnosis codes in the database. One hospital instituted a data validation team to randomly check completeness and accuracy of the data (only for inpatient service by comparing the computerized patient data with the medication administration record (MAR).

PAGE 94

79 In summary, the interviewees felt that an increasing numbers of patients eligible for 30-Baht policy were encountered after implementation of the 30-Baht policy. There was no different in the patterns of drug prescribing between 30-Baht beneficiaries and those with others health insurance plans. Most hospitals did not have any plan to support the implementation of the 30-Baht HI policy. Accuracy and completeness of the data in the electronic database was rated high, however, formal quantitative validation was broadly missing.

PAGE 95

CHAPTER 5 METHODS Our study applies a cross-sectional observational design using segmented regression analysis of time series data with a concurrent control group to evaluate the impact the 30-Baht HI policy on hospital drug utilization in Ubonratchatani province, Thailand. In other words, only observations of patients who visited or were admitted to the hospital at a given month were measured; no cohort of the patients was followed during the study period. To identify the effect of this policy, we selected drug utilization rates and drug utilization quality as primary measures, and hospital visit rates or hospital admission rates as secondary measures. These measures were calculated based upon data extracted for the 30-Baht HI beneficiaries (considered the intervention group) and the CSMBS group (control group) from the patient-specific healthcare data from the HI database from the eight community government hospitals. The impact of the 30-Baht HI policy was evaluated in four of the most prevalent diseases that affect the Thai population. The four diseases states included community-acquired pneumonia, gastrointestinal tract infections, hypertension, and diabetes. Pneumonia and gastrointestinal tract infections were selected to identify whether the policy had an effect on acute conditions, while hypertension and diabetes were selected to identify the effect of the policy on chronic conditions. The measurement of drug utilization included all drugs that were available in the hospital drug formulary and had an indication to treat the selected diseases based on official labeling by the US Food and Drug Administration. Drug utilization rates of each 80

PAGE 96

81 selected disease were measured using the DDD unit per population. Drug utilization quality was assessed based on the percent appropriateness of prescribing. Hospital visit/admission rates were estimated based on the eligible population registered for each of the study hospitals. Linear regression analysis of time series data was used to analyze the effect of the policy on drug utilization rates, drug utilization quality, hospital visit rates or hospital admission rates. By using a time series technique with a concurrent control group, any immediate change (level) and a gradual change (trend) of the measures associated with the 30 Bath HI policy could be identified. The concurrent control group was included in the regression model by differencing the study measures between the intervention and the control group. Including a concurrent control group, we were able to control for external factors that might influence the measures at the same time when the 30-Baht HI policy was implemented. In addition to the above measures, drug use disparity (i.e., differences of drug utilization between the two populations) was calculated based on the assumption that the previously uninsured population had lower rates of drug utilization. Changes of drug use disparity related to the 30-Baht HI policy was identified by comparing the differences of drug utilization rate and quality between the two groups one year before and one year after the implementation of the policy using T-test and Chi-Square test at an alpha level of .05. Hypotheses There were six hypotheses for statistical tests for research questions 1 to 6. We chose an Seasonal Autoregressive Integrated Moving average (SARIMA) linear regression model because it incorporated associations (a trend) of the measures during the pre-policy period and intervention (input) variables, including a pulse function variable to

PAGE 97

82 identify an abrupt effect, and a step function variable to identify trend change. The model was set up to integrate seasonality or the association of the measures in a repeated interval (e.g., annually). If trend is taken into account in the analysis, history and maturation biases on the measures can be reduced. A constant in the regression model was not included by centering the data with the mean of the measures of the whole series to simplify the model. We allowed a delay in the effect of the 30-Baht Hi policy, if visual observations of the plots suggested changes of the study measures after the implementation of the policy. However to simplify the model, we did not include a decay parameter (the rate that the policy took effect) into the model. Equation 5-1, 5-2, and 5-3 were chosen for testing the hypothesis 1 to 4. Hypothesis 5 and 6 were tested with t-test and Chi-Square, respectively. All hypotheses were tested at the significance level of 0.05. Equation 5-1 is the full regression model that includes historical trends, immediate impact variable, and trend impact variables. Equation 5-2 illustrates the regression model for autoregressive process, while Equation 5-3 illustrates the moving average process. We used logit in the regression analysis for the measure of appropriate prescribing. Yt = 1Tt + 2 Pt + 3 St +et (5-1) Autoregressive process Yt = 1yt-1 + 2yt-2 + 3yt-3+ nyt-n + et (5-2) Moving Average Yt = et 1e t-1 2e t-2 3e t-3 ne t-n ... (5-3) Yt : A difference of the studied measure between the 30-Baht HI beneficiaries and the CSMBS group after the policy in month t

PAGE 98

83 Tt: A continuous variable indicating time in months from the beginning of the observation period to the month before the implementation of the policy (prepolicy period from January 2000 to May 2001) Pt: A nominal variable for time t indicating the presence of the policy only at month 18. t < 18, Pt = 0; t = 18, Pt =1; t >18, Pt = 0 St: A nominal variable for time t indicating the presence of the policy at month 18 and after. t < 18, St = 0; t >=18, St =1 1: A regression parameter of the slope of the measure in prepolicy period (prepolicy trend) 2: A regression parameter estimating the level change in the measure immediately after the policy (abrupt change) 3: A regression parameter estimating the trend change in the measure after the policy t: Autoregressive coefficient t: Moving Average coefficient n: numbers of observation For infectious diarrhea, bacterial pneumonia, diabetes and hypertension: Research question 1: Did drug utilization rates of the 30-Baht beneficiaries (DRi) change after the 30-Baht HI policy was implemented, controlling for the DRi before the policy and the drug utilization rates of the control group (DRc)? Hypothesis1: DRi changed after the 30-Baht HI policy was implemented, controlling for the DRi before the policy and the DRc H0: 2 =0; 3 = 0 H1: 2 0; 3 0 Research question 2: Did the percent prescribing appropriateness of the 30-Baht beneficiaries (DQi) change after the 30-Baht HI policy was implemented, controlling for the DQi before the policy and those of the control group (DQc)? Hypothesis2: DQi changed after the 30-Baht HI policy was implemented, controlling for the DQi before the policy and the DQc

PAGE 99

84 H0: 2 =0; 3 = 0 H1: 2 0; 3 0 Research question 3: Did hospital admission rates for inpatient services for bacterial pneumonia and infectious diarrhea change in the 30-Baht beneficiaries (ARi) change after the 30-Baht HI policy was implemented, controlling for admission rates of the control group (ARc)? Hypothesis 3: ARi change after the 30-Baht HI policy was implemented, controlling for the ARi before the policy and ARc H0: 2 =0; 3 = 0 H1: 2 0; 3 0 Research question 4: Did hospital visit rates for outpatient services for diabetes and hypertension of the 30-Baht beneficiaries (HRi) change after the 30-Baht HI policy was implemented, controlling for the HRi before the policy and the rates of the control group (HRc)? Hypothesis 4: HRi change after the 30-Baht HI policy was implemented, controlling for the HRi before the policy and HRc H0: 2 =0; 3 = 0 H1: 2 0; 3 0 Research question 5: Did the difference of drug utilization rates (dDA) between the 30 Bath HI and the CSMBS groups change after the 30-Baht HI policy? Hypothesis 5: dDA change after the 30-Baht HI policy H0: pbefore = pafter H1: pbefore pafter

PAGE 100

85 pbefore: dDA before the 30-Baht HI policy pbefore: dDA after the 30-Baht HI policy Research question 6: Did the difference of the percentages of prescribing appropriateness (dDQ) between the 30 Bath HI and the CSMBS groups change after the 30-Baht HI policy? Hypothesis 6: dDQ change after the 30-Baht HI policy H0: pbefore = pafter H1: pbefore pafter pbefore: dDQ before the 30-Baht HI policy pbefore: dDQ after the 30-Baht HI policy Patient Selection The primary and secondary measures were extracted for patients 18 and older, who visited the hospitals in Ubonratchatani for outpatient services and/or were admitted to the hospitals for inpatient services from January 1, 2000 to December 31, 2003. Patients who were admitted to the hospital for inpatient services were excluded if they were referred to other healthcare facilities with less than 1 day from the admission time. Patients who visited outpatient clinics, but did not have any assigned disease diagnosis code were not included, because these patients might have self discharged prior to seeing the healthcare providers for examination, diagnosis, or treatment. We included patients who were diagnosed with at least one of the four most prevalent diseases/conditions in Thailand, according to The Thai National Disease Statistics 2002 (110;165). The selected diseases defined by the International Statistical Classification of Diseases and Related Health Problems, Version 10 (ICD10) (167) were: 1) bacterial

PAGE 101

86 pneumonia (ICD-10:J13, J15), gastrointestinal tract infections (A00-A09), hypertension with no complications (I1xx), and diabetes (E1xx). As these disease states largely affect health of the Thai population, they were considered suitable indicators of improvements in healthcare quality. In addition, treatment of these diseases relies heavily on drug therapy and thus, measures the effect of the 30-Baht HI policy on drug use as intended in our study. Because our study did not measure a direct effect on health outcomes, it was important to select disease states where drug therapy (process) was closely related to the health outcomes (e.g., cure rates, or mortality). Patients were identified using the dates and time of visits or admissions, ICD-10, and the type of health insurance at each episode of care to classify the patients into intervention and control groups. Patients who were uninsured before June 1, 2001 (the introduction of the 30-Baht HI policy) were coded as cash or health card or no health insurance. Patients status and codes are listed in Appendix B. We chose the CSMBS group as a control in our study because the 30-Bath HI policy coverage did not extend to this group, thus the CSMBS was used to control external factors that might affect drug utilization rather than the 30-Baht HI policy. The CSMBS group lives in the same geographical, so we assumed similar disease exposure. No other control groups were available within this database. More importantly, considering equity of drug use, the CSMBS group was considered a gold standard (110). This population received full healthcare coverage from the Thai government with the most comprehensive prescription drug benefits with no co-payment, except non-formulary drugs. Thus, using the CSMBS as a control, the issues regarding economic barriers to receive prescription drugs in the 30-Baht HI population could be addressed.

PAGE 102

87 From the above reasons, the CSMBS population seemed to be an appropriate control population for our study. Given their type of employment, there were some differences in population demographics, socio-economic status and likely, overall health between the two populations. In order to utilize the CSMBS group in this study, the following assumptions had to be made: the CSMBS group is not different from the 30-Baht HI population, and the exposures to other factors that might affect the study measures are similar in both groups. Data Source The measures were extracted from the HI databases from 8 community government hospitals in Ubonratchatani province, Thailand, which had been validated for the purpose of our study. Since our study intended to measure access to drugs and hospital services, we chose to relate hospital-specific utilization data to the expected numbers of the people who were eligible for the 30-Baht HI policy. For the control group, the expected numbers of people in the CSMBS group was used as the denominators as well. Both denominators were extracted from the health insurance registry database of the National Health Security Office, Thailand (168). Several studies of health service access have presented their measures based on the whole population to demonstrate the effect of the interventions on the large scale (27;29;51). Comparing rates of drug utilization based on the population who accessed the hospital can only identify changes within this hospital but it does not incorporate possible changes in the total population assigned to a specific hospital. Measures Primary measures of the study are monthly drug utilization rates and the percentage of appropriate prescribing of each month. The assessment of prescribing

PAGE 103

88 appropriateness was conducted only for infectious diarrhea and bacterial pneumonia. Appropriateness prescribing for antidiabetic and antihypertensive drugs was not conducted because it required more clinical data that were unavailable in the HI database. Secondary measures were monthly hospital visit rates for diabetes and hypertension, and monthly hospital admission rates for gastrointestinal infection and bacterial pneumonia. The measures were calculated for each selected disease for the 30-Baht HI group and the CSMBS group. Since we intended to identify the effect of the 30-Baht HI policy not only on utilization, but also on access to drug therapy and hospital services of the target population, the expected number of beneficiaries for the 8 selected hospitals for each health insurance plan was used as a denominator for the measurement of drug utilization and hospital visit/admission rates. Drug Utilization Rate (DR) Drug utilization rates of patients with diabetes and hypertension were calculated only for outpatient visits. DRs of bacterial pneumonia and Infectious diarrhea were calculated for outpatient and inpatient services because approximately half of the patients were treated in outpatient clinic and the other half in hospital wards. Drug utilization rates were calculated by using the total number of the DDDs of all drugs that were prescribed for patients with the disease, divided by the expected number of beneficiaries. Lists of the drugs used for the calculation of drug utilization rates are presented in Appendix D. DR = Number of the DDDs of prescribed for outpatient with the disease The expected number of beneficiaries

PAGE 104

89 Percentage of Appropriate Prescribing (Drug Use Quality, DQ) The assessment of prescribing appropriateness in our study included three major components: 1) selection of appropriate drugs with the FDA-approved medication for the disease; 2) appropriate dosage for the specified disease state; and 3) appropriate duration of therapy. Criteria of the appropriateness of drug therapy were set based on the current evidence in recognized clinical practice guidelines published in Thailand and in the US. Quality of drug use was expressed as binary measure (YES: correct choice, dosage, and duration; or NO: incorrect for at least one of the three components). For example, drug therapy was appropriate for patients with community-acquired pneumonia (CAP) patients, if at least one macrolide (erythromycin or clarithromycin), doxycycline, beta-lactam antibiotic (amoxicillin, amoxicillin-clavulanate, cefotaxime, ceftriaxone), or fluoroquinolone (ciprofloxacin, levofloxacin) (169) was prescribed. In addition, the dose and duration had to be correct, for examples amoxicillin was prescribed, the appropriate dosage had to range between 1.5-2.0 gram/day, and the appropriate duration of therapy had to be 10-14 days(170). DQ = Number of visits/admissions with appropriate prescribing Total number of visits/admissions for the selected disease The percentage of appropriate prescribing was an aggregate of the assessment of drug(s) prescribed for a patient at each visit/admission. If two drugs were prescribed at one visit, we assessed the appropriateness for both drugs, and the visit was assigned 1, if at least one drug was appropriate. The denominator was the total number of visits/admissions of each month. The appropriate prescribing algorithm is illustrated in Figure 5-1.

PAGE 105

90 Hospital Visit Rates (HR) Hospital visit rates were calculated by the monthly number of outpatient visits divided by total number of expected beneficiaries. HR = Number of outpatient visits Total number of expected beneficiaries Hospital Admission Rates (AR) Hospital admission rates were calculated by the number of hospital admissions divided by the total number of expected beneficiaries. HR = Number of hospital admissions Total number of expected beneficiaries Drug Disparity (DP) Drug disparity was calculated separately for utilization rates and the percentage of appropriate prescribing at one year before and one year after the implementation of 30-Baht HI policy. Disparity of drug utilization rate: DPdr = DR 30-Baht DRCSMBS Disparity of appropriate prescribing: DPdq = DQ 30-Baht DQCSMBS In our study, drug utilization rates, the percentage of appropriate prescribing, hospital visit rates, and hospital admission rates were calculated monthly from January 1, 2000 to December 31, 2003. This totaled 48 observations: 17 before and 31 observations after the implementation of the 30-Baht HI policy. The segment from month 1 to 17 was called the pre-policy period, and month 18 to 48 was called post-policy period. The number of the observations was considered sufficient for the statistical analysis using segmented linear regression of time series data (48 observations) and with 4 year-observation period, we were able to be adjusted for any seasonal factor that might affect the measures.

PAGE 106

91 Patients with communityacquired pneumonia (ICD10 code = E10-E14) 1. Choice of drug Prescribed drug is in the following class: Macrolide Doxycycline Beta-lactam Fluoroquinolone Cephalosporin The combination of the above dru g classes 2. Dosage range The drug is used in the appropriate dosage range as recommended by guidelines for adults with the condition 3. Duration of therapy YES YES Appropriate drug use Quality of drug utilization (DQ) = 1 YES NO Inappropriate drug use Quality of drug utilization (DQ) = 0 NO NO Figure 5-1. Algorithm to assess quality of drug utilization (DQ) using community-acquired pneumonia as an example.(169;170)

PAGE 107

92 Statistical Analyses Analysis of the Effects of the 30-Baht HI Policy on the Study Measures First, we used visual observation of the plots between the measure and time to observe sudden change or change of the slope after the implementation of the policy. Then, segmented regression analysis of interrupted time-series was used to test for a statistical significant change of the measures, controlling for historical trends and external factors using the control group. Any abrupt or sudden change at month 18 or later and gradual change (slope change) could be identified using the selected regression model. (Figure 5-2) Box-Jenkins Autoregressive Integrated Moving Average (ARIMA (p,d,q)) was applied to identify autocorrelation and to transform non-stationary data into stationary data, and to identify the order of the model for chronic diseases (diabetes and hypertension). Seasonal Autoregressive Integrated Moving Average (SARIMA (p,d,q) x (P,D,Q)) models (Season cycle is 12 months) were used for acute conditions (Infectious diarrhea and bacterial pneumonia). Seasonality was adjusted (decomposed) in the autoregressive process, because drug utilization for acute conditions varied by season, e.g., high antibiotics use for infectious diarrhea in the summer months. The level of significance for all statistical tests was 0.05 two-sided. For statistical analysis, SAS version 8 (171) was used to conduct regression analysis of time series data. The direction of the change of the measure was not predetermined because it was not certain that the policy had negative or positive effect on drug utilization rates, prescribing quality, or hospital visit/admission rates.

PAGE 108

93 CSMBS 30-Baht HI Jun 01 Jan 00 Dec 03 Total numbers of DDDs/population 1 2 3 Figure 5-2. Interrupted time series analysis of a drug utilization rate in the 30-Baht HI group and the control group before and after the policy Changes of Drug Disparity Associated with the 30-Baht HI Policy The assessment of changes of disparity levels of DR and DQ associated with the policy was conducted by comparing levels of disparity at one year before and at one year after the policy implementation. The differences of the disparity levels of DR were identified using unpaired T-test at two-sided significance level of .05. Statistically significant changes of disparity levels of DQ (%) were identified using Chi-Square test at a significance level of .05. We expected that the 30-Baht HI group had a lower drug utilization rate and a lower percentage of appropriate prescribing than those of the CSMBS group during the pre-policy period associated with its uninsured status. The difference of the measure between the two populations should be smaller if the policy has a positive effect in improving drug utilization rates and quality.

PAGE 109

94 Patient-specific healthcare data 8 Community government hospitals Adult patients admitted or visited during January 2000 to December 2003 with the selected diseases* Data quality assessment Qualitative assessment: expert interview of hospital directors and database managers Quantitative tests: data structure, missing data, plausibility, coherence Test an effect of 30-Baht HI policy: Segmented linear regression of time series monthly data 48 observations (17 before and 31 after the 30 Baht policy implementation) Controlled for the extraneous effects using control group Acute condition (inpatient and outpatient service) Pneumonia Gastro in testin al tract Chronic condition (outpatient service) Hypertension Diabetes Measures of the 30-Baht HI group and the control group Drug utilization rates % of appropriate prescribing Hospital visit rates or admission rates or Figure 5-3. Study framework demonstrating sources of the data used in the study, inclusion criteria for data extraction, measurements, and the analyses *The selected diseases are: diabetes, hypertension, pneumonia, and gastrointestinal tract infection Civil Servants Medical Benefits Scheme

PAGE 110

CHAPTER 6 RESULTS Patient Demographics The expected eligible population was 324,931 for the 30-Baht HI policy and 38,109 for the CSMBS group in 2003. These numbers were used for the calculation throughout the study period because only minor changes occurred during 2000-2003. These numbers represent the population with health benefits provided by the 8 selected government community hospitals in the eight districts of Ubonratchatani province, Thailand. (Table 6-1) Table 6-1. Expected populations eligible for the 30-Baht HI policy and the CSMBS in 8 selected government community hospitals, Ubonratchatani province, Thailand (2003) Hospital Total number of 30-Baht HIa Policy Total number of CSMBSb A 18,201 855 B 43,575 3,794 C 18,928 929 D 16,179 1,389 E 78,198 5,498 F 67,240 14,237 G 45,759 4,074 H 36,852 1,389 Total 324,931 38,109 a 30-Baht health insurance policy. b The Civil Service Medical Benefit Scheme On average, the 30-Baht HI patients were younger than the CSMBS patients. (Table 6-2) There was a higher proportion of female in the 30-Baht HI population than the CSMBS. Most patients in both groups were married. Most patients in the 30-Baht group were farmers, while only 42% in the CSMBS group were reported as farmers. It appears that the two groups were different, particularly, in age and occupation. 95

PAGE 111

96 Table 6-2. Patient demographics Patient demographics 30-Baht HI group CSMBS group Total number of patient 132,397 10,148 Age, year (mean SD) 29.79.4 35.013.1 Gender male (%) 60,771(45.9) 5,068 (50.1) Marital status (%) Married Single Widowed Others 123,262 (93.1) 4,672 (3.5) 3,341 (2.5) 1,122 (0.8) 9,120 (89.9) 388 (3.8) 552 (5.4) 88 (0.9) Occupation (%) Farmer Worker Trade Government officer No full-time job/student Others 110,176(83.2) 6,617 (5.0) 4,114 (3.1) 4,013 (3.0) 4,009 (3.0) 3,468 (2.7) 4,299 (42.5) 503 0 0 4,099 (40.5) 495 (4.9) 1,255 (12.1) Diabetes Drug Utilization Rates The mean of the drug utilization rates for the 30-Baht HI population was approximately ten times greater than the rates of the CSMBS group in the pre-policy period (2,727.79 vs. 348.84/10,000 population/month). (Figure 6-1) After the policy was implemented, the rate in the 30 Bath HI population slowly increased about 14 % (from 2,727.79 to 3,105.61), while the rate in the CSMBS continuously increased 4.5-fold from month 25 until the end of the observation period (from 348.84 to 1,892.76). At the end of the 2003, the rate of the CSMBS was as high as the rate of the 30-Baht HI population. Comparing these figures with the diabetes prevalence of 8.6% (165), each diabetic with 30-Baht HI benefits used approximately a daily maintenance dose of a antidiabetic drug for 3 days per month. In other words, about 10% of diabetics received a diabetic drug every day. The utilization rates of antidiabetic drugs of the 30-Baht group showed a positive trend during pre-policy period. At month 18, the rate slightly increased, then remained

PAGE 112

97 stable. The drug utilization rates in the CSMBS group were steady until month 18, and then increased significantly and continuously until the end of the observation period. 050010001500200025003000350040001357911131517192123252729313335373941434547Month 30.00 CSMBS Figure 6-1. Monthly drug utilization rates for diabetes in the 30-Baht HI population and of the CSMBS population from January 2000 to December 2003 Hospital Visit Rates Hospital visit rates in the 30-Baht HI population were steady over the study period (86.66 and 87.35 visits/10,000 beneficiaries before and after the policy). Comparisons of these rates with the diabetic prevalence in 2002 suggest that only 10% of the diabetics in both groups used hospital services. In the CSMBS group, the rates was significantly lower than those of the 30-Baht population (25.14 vs. 86.66) in the pre-policy period, but increased drastically in month 25 and remained high through the end of the observation period (from 25.14 to 73.66).(Figure 6-2) Since the drug utilization and hospital visit rates in the CSMBS group started to increase at 7 months after the policy was implemented, it suggests that the factors associated with these increase were not likely to affect changes associated with the3030-Baht policy

PAGE 113

98 Baht HI policy at month 18 in the 30-Baht HI population. For this reason, we decided to analyze the effect of the policy on drug utilization and hospital visit rates of the 30-Baht HI population without the control group. If the control group were used, differences in rates would show a large decrease after month 25 due to a significant increase of the rates in the CSMBS group, which were neither associated with the 30-Baht HI policy nor could they be explained by every other phenomenon. The drastic increase in drug utilization rates and hospital visit rates was also observed from the plots for hypertension, Infectious diarrhea, and bacterial pneumonia. Therefore, the control group was excluded from the analyses of the impact of the policy on drug utilization and hospital visit rates of all selected disease states. 0204060801001201357911131517192123252729313335373941434547Month 30.00 CSMBS Figure 6-2. Monthly hospital visit rates for diabetes in the 30-Baht health insurance population and of the CSMBS population from January 2000 to December 2003 30-Baht policy

PAGE 114

99 Impact of the 30-Baht HI Policy on the Study Measures The results from the segmented time series analysis show no significant sudden or trend change of the drug utilization rates at month 18 after the policy was implemented (p= 0.7484 and p=.5907, respectively) using autoregressive order of 3. (Table 6-3) Table 6-3. Regression parameters for drug utilization and hospital visit rates of diabetes Measure Ordera Delayb Parameter (SEc) T-ratio p AICd A. Drug utilization rate of the 30-Baht HI group AR AR AR Level change Trend change 1 2 3 0 -0.003 (0.121) 0.288 (0.116) 0.557 (0.130) 76.667 (239.0) 146.960 (273.237) -0.02 2.45 4.29 0.32 0.54 0.9825 0.0141 <.0001 0.7484 0.5907 687.38 B. Hospital visit rate of the 30 Bath HI group AR AR AR Level change Trend change 1 2 3 0 -0.094 (0.144) 0.286 (0.139) 0.362 (0.146) 4.005 (6.849) 0.663 (4.062) -0.65 2.06 2.47 0.58 0.16 0.5137 0.0394 0.0135 0.5587 0.8704 330.61 a ARIMA order of the linear regression model. b Expected time delay (month) of the effect of the policy. c Standard error. d Akaike Information Criteria The parameter for level change suggests that at month 18 the drug utilization rate was 76.67 DDDs above the mean of the data series. With a very large standard error (239.0) it is evident that the increase of the drug utilization rate at month 18 was not statistically significant. The trend parameter suggests that during the post-policy period the drug utilization rates increased by 146.96 from the mean before the policy, however with a large standard error no statistically significant difference were observed. Visual observation from Figure 6-1 suggests that no delay effect should be included in the model. The analysis of the hospital visit rates in the 30-Baht population proposes that the 30-Baht HI policy did not have an immediate or trend impact on hospital visits related to diabetes (p =0.5587 and p=.9704, respectively). (Table 6-3) The monthly numbers of

PAGE 115

100 DDDs and visits, the expected number of beneficiaries, and the rates of drug utilization and hospital visit are presented in Table 6-4. Table 6-4. Rates of drug utilization and hospital visits related to diabetes of the 30-Baht health insurance beneficiaries Month DDDsa Visits Population Drug utilization rate Hospital visit rate Time Pulseb Stepc Jan-2000 65,752.19 2,502 324,931 2023.57 77.00 1 0 0 Feb-2000 78,679.51 2,556 324,931 2421.42 78.66 2 0 0 Mar-2000 86,707.72 2,932 324,931 2668.50 90.23 3 0 0 Apr-2000 74,255.62 2,560 324,931 2285.27 78.79 4 0 0 May-2000 92,560.67 3,048 324,931 2848.63 93.80 5 0 0 Jun-2000 80,677.08 2,764 324,931 2482.90 85.06 6 0 0 Jul-2000 73,517.54 2,614 324,931 2262.56 80.45 7 0 0 Aug-2000 94,979.10 3,177 324,931 2923.05 97.77 8 0 0 Sep-2000 80,139.11 2,581 324,931 2466.34 79.43 9 0 0 Oct-2000 94,293.16 2,740 324,931 2901.94 84.33 10 0 0 Nov-2000 89,880.15 2,662 324,931 2766.13 81.93 11 0 0 Dec-2000 68,766.66 2,407 324,931 2116.35 74.08 12 0 0 Jan-2001 107,414.80 3,084 324,931 3305.77 94.91 13 0 0 Feb-2001 100,049.81 2,858 324,931 3079.11 87.96 14 0 0 Mar-2001 102,295.92 3,056 324,931 3148.24 94.05 15 0 0 Apr-2001 96,868.78 2,851 324,931 2981.21 87.74 16 0 0 May-2001 119,946.78 3,476 324,931 3691.45 106.98 17 0 0 Jun-2001 104,063.41 3,021 324,931 3202.63 92.97 18 1 1 Jul-2001 105,184.19 3,131 324,931 3237.12 96.36 19 1 0 Aug-2001 109,844.40 3,226 324,931 3380.55 99.28 20 1 0 Sep-2001 96,947.37 2,861 324,931 2983.63 88.05 21 1 0 Oct-2001 117,656.64 3,386 324,931 3620.97 104.21 22 1 0 Nov-2001 107,435.19 3,077 324,931 3306.40 94.70 23 1 0 Dec-2001 99,096.62 2,827 324,931 3049.77 87.00 24 1 0 Jan-2002 109,111.14 3,032 324,931 3357.98 93.31 25 1 0 Feb-2002 90,985.12 2,482 324,931 2800.14 76.39 26 1 0 Mar-2002 101,378.25 2,834 324,931 3119.99 87.22 27 1 0 Apr-2002 106,743.65 2,890 324,931 3285.12 88.94 28 1 0 May-2002 105,869.21 2,953 324,931 3258.21 90.88 29 1 0 Jun-2002 90,921.33 2,584 324,931 2798.17 79.52 30 1 0 Jul-2002 104,388.42 2,809 324,931 3212.63 86.45 31 1 0 Aug-2002 95,356.83 2,730 324,931 2934.68 84.02 32 1 0 Sep-2002 90,126.05 2,633 324,931 2773.70 81.03 33 1 0 Oct-2002 103,947.83 2,989 324,931 3199.07 91.99 34 1 0 Nov-2002 90,154.76 2,580 324,931 2774.58 79.40 35 1 0 Dec-2002 95,759.91 2,779 324,931 2947.08 85.53 36 1 0 Jan-2003 95,917.96 2,560 324,931 2951.95 78.79 37 1 0 Feb-2003 88,155.86 2,422 324,931 2713.06 74.54 38 1 0 Mar-2003 89,734.46 2,559 324,931 2761.65 78.76 39 1 0 Apr-2003 110,590.91 3,015 324,931 3403.52 92.79 40 1 0 May-2003 96,581.02 2,791 324,931 2972.35 85.90 41 1 0 Jun-2003 91,140.80 2,590 324,931 2804.93 79.71 42 1 0 Jul-2003 113,902.26 3,144 324,931 3505.43 96.76 43 1 0 Aug-2003 91,859.71 2,549 324,931 2827.05 78.45 44 1 0 Sep-2003 104,044.30 2,929 324,931 3202.04 90.14 45 1 0

PAGE 116

101 Table 6-4. Continued Month DDDsa Visits Population Drug utilization rate Hospital visit rate Time Pulseb Stepc Oct-2003 112,809.16 3,058 324,931 3471.79 94.11 46 1 0 Nov-2003 108,528.65 2,794 324,931 3340.05 85.99 47 1 0 Dec-2003 100,001.90 2,754 324,931 3077.63 84.76 48 1 0 a Numbers of the define daily dose; b Pulse variable for the test of immediate or sudden change; c Step variable for the test of long-term or trend change Hypertension Drug Utilization Rates The drug utilization rates of the 30-Baht HI population remained constant in the first 11 months of the study. The rate abruptly increased in December 2000 and remained steady until the end of 2003.(Figure 6-3) The average drug utilization rate of the 30-Baht HI population was 1,413.99 DDDs/10,000 beneficiaries before the implementation of the policy. After the policy, the average drug utilization rate of the 30-Baht population was 2,172.00 DDDs/10,000 beneficiaries/month. Comparing with the prevalence of hypertension of 9.5% before the policy (165), the results suggested that every hypertensive patient in the 30-Baht HI population used 1.5 doses of antihypertensive drug per month or approximately 5% of the patients received at least one dose every day. After the policy, the proportion of hypertensive patients in the 30-Baht group who received a drug at one dose for the whole month increased by approximately 5%. Hospital Visit Rates The average hospital visit rate was 25.84 visits/10,000 beneficiaries/month during pre-policy period. Compared with the disease prevalence of 9.5%, this means only 2.6% of hypertensive patients in the 30-Baht HI group had one visit at the hospitals per month. After the policy, the rates increased by about one third (36.20 visits/10,000 beneficiaries/month) ( Figure 6-4). An additional 1% of the 30-Baht patients with

PAGE 117

102 hypertension had one visit to the hospitals per month after the policy compared to the rates during the pre-policy period. 05001000150020002500300035001357911131517192123252729313335373941434547MonthDDDs/10,000/month Figure 6-3. Drug utilization rates for hypertension of the 30-Baht health insurance benefit group from January 2000 to December 2003 01020304050601357911131517192123252729313335373941434547Monthvisits/10,000/month 30-Baht policy 30-Baht policy Figure 6-4. Monthly hospital visit rates for hypertension of the 30-Baht health insurance benefit group from January 2000 to December 2003

PAGE 118

103 Impact of the 30-Baht HI Policy on the Study Measures The analysis of the linear regression model of the drug utilization rate for the 30-Baht HI population showed no statistically significant difference between the 30-Baht HI policy and the level or trend changes in the drug utilization rates (p= 0.2461, and p= 0.6626, respectively) with autoregressive of 2. The results from the analysis of hospital visit rates suggest that the 30-Baht policy did not have any sudden or gradual effects on hospital visit rates with autoregressive of 4 (p=0.3742, and p=0.2473, respectively). (Table 6-5) The monthly numbers of DDDs and visits, the expected number of beneficiaries, and the rates of drug utilization and hospital visit are presented in Table 6-6. Table 6-5. Regression parameters for drug utilization and hospital visit rates of hypertension Measure Ordera Delayb Parameter (SEc) T-ratio p AICd A. Drug utilization rate of 30-Baht HI group AR AR Level change Trend change 1 2 0 0.439 (0.129) 0.494 (0.132) 257.589(222.066) 106.702 (244.571) 3.40 3.74 -1.16 0.44 0.0007 0.0002 0.2461 0.6626 661.27 B. Hospital visit rate of the 30-Baht HI group AR AR AR AR Level change Trend change 1 2 3 4 0 0.300(0.153) 0.357 (0.137) 0.571 (0.154) -0.253 (0.178) -0.170 (2.441) 0.984 (2.450) 1.96 2.60 3.71 -1.42 -.070 0.40 0.0499 0.0093 0.0002 0.1560 0.4864 0.6879 243.53 a ARIMA order of the linear regression model. b Expected time delay (month) of the effect of the policy. c Standard error. d Akaike Information Criteria

PAGE 119

104 Table 6-6. Monthly rates of drug utilization and hospital visit related to hypertension of the 30-Baht health insurance beneficiaries Month DDDsa Visits Population Drug utilization rate Hospital visit rate Time Pulseb Stepc Jan-2000 37,356.31 852 324,931 1149.67 26.22 1 0 0 Feb-2000 40,851.25 829 324,931 1257.23 25.51 2 0 0 Mar-2000 48,051.00 949 324,931 1478.81 29.21 3 0 0 Apr-2000 35,313.31 789 324,931 1086.79 24.28 4 0 0 May-2000 38,527.69 887 324,931 1185.72 27.30 5 0 0 Jun-2000 36,476.88 829 324,931 1122.60 25.51 6 0 0 Jul-2000 35,513.44 857 324,931 1092.95 26.37 7 0 0 Aug-2000 40,218.75 949 324,931 1237.76 29.21 8 0 0 Sep-2000 39,981.69 893 324,931 1230.47 27.48 9 0 0 Oct-2000 36,897.63 785 324,931 1135.55 24.16 10 0 0 Nov-2000 41,080.50 861 324,931 1264.28 26.50 11 0 0 Dec-2000 35,945.75 782 324,931 1106.26 24.07 12 0 0 Jan-2001 61,391.69 805 324,931 1889.38 24.77 13 0 0 Feb-2001 54,421.00 738 324,931 1674.85 22.71 14 0 0 Mar-2001 67,884.19 826 324,931 2089.19 25.42 15 0 0 Apr-2001 60,680.88 767 324,931 1867.50 23.61 16 0 0 May-2001 70,474.31 876 324,931 2168.90 26.96 17 0 0 Jun-2001 58,257.38 772 324,931 1792.92 23.76 18 1 1 Jul-2001 68,688.63 871 324,931 2113.95 26.81 19 1 0 Aug-2001 64,937.50 882 324,931 1998.50 27.14 20 1 0 Sep-2001 58,157.75 793 324,931 1789.85 24.41 21 1 0 Oct-2001 74,871.06 973 324,931 2304.21 29.94 22 1 0 Nov-2001 66,568.56 924 324,931 2048.70 28.44 23 1 0 Dec-2001 68,646.63 886 324,931 2112.65 27.27 24 1 0 Jan-2002 62,564.90 1,073 324,931 1925.48 33.02 25 1 0 Feb-2002 57,700.94 959 324,931 1775.79 29.51 26 1 0 Mar-2002 62,509.75 1,019 324,931 1923.79 31.36 27 1 0 Apr-2002 67,827.44 1,023 324,931 2087.44 31.48 28 1 0 May-2002 62,687.77 1,094 324,931 1929.26 33.67 29 1 0 Jun-2002 68,610.81 1,078 324,931 2111.55 33.18 30 1 0 Jul-2002 66,052.75 1,152 324,931 2032.82 35.45 31 1 0 Aug-2002 67,086.63 1,192 324,931 2064.64 36.68 32 1 0 Sep-2002 70,208.00 1,221 324,931 2160.70 37.58 33 1 0 Oct-2002 78,892.96 1,378 324,931 2427.99 42.41 34 1 0 Nov-2002 68,856.21 1,144 324,931 2119.10 35.21 35 1 0 Dec-2002 71,971.33 1,296 324,931 2214.97 39.89 36 1 0 Jan-2003 78,021.43 1,340 324,931 2401.17 41.24 37 1 0 Feb-2003 67,967.30 1,231 324,931 2091.75 37.88 38 1 0 Mar-2003 71,429.07 1,294 324,931 2198.28 39.82 39 1 0 Apr-2003 71,609.37 1,363 324,931 2203.83 41.95 40 1 0 May-2003 75,833.56 1,334 324,931 2333.84 41.05 41 1 0 Jun-2003 68,263.11 1,277 324,931 2100.85 39.30 42 1 0 Jul-2003 82,619.45 1,491 324,931 2542.68 45.89 43 1 0 Aug-2003 70,606.46 1,317 324,931 2172.97 40.53 44 1 0 Sep-2003 77,727.86 1,448 324,931 2392.13 44.56 45 1 0 Oct-2003 94,676.13 1,747 324,931 2913.73 53.77 46 1 0 Nov-2003 82,450.57 1,436 324,931 2537.48 44.19 47 1 0 Dec-2003 81,524.27 1,455 324,931 2508.97 44.78 48 1 0 a Numbers of the define daily dose. b Pulse variable for the test of immediate or sudden change. c Step variable for the test of long-term or trend change

PAGE 120

105 Infectious Diarrhea Drug Utilization Rates Before the policy was implemented, the average drug utilization rates related to Infectious diarrhea was 40.28 DDDs/10,000 beneficiaries/month in the 30-Baht HI group. These numbers can be interpreted as of ten thousand 30-Baht HI beneficiaries; there were 40 patients with Infectious diarrhea receiving one maintenance dose of an antibiotic per month. Comparisons of the rates with the incidence reported by the national disease incidence 2002 (165) of 67.9/10,000/year or approximately 14 cases/10,000 population/month suggests that every 30-Baht HI patient with GI infection received an average defined daily dose of an antibiotic for 3 days. After the policy, the drug utilization rate in the 30-Baht population slightly increased (40.28 to 52.75). (Figure 6-5) In comparison with the disease incidence, this means every 30-Baht HI patient with Infectious diarrhea received an antibiotic drug approximately 4 days for the 30-Baht HI population. 01020304050607080123456789101112131415161718192021222324252627282930313233343536373839404142434445464748Monthddds/10,000/month Figure 6-5. Monthly drug utilization rates for infectious diarrhea of the 30-Baht HI from January 2000 to December 2003 30-Baht policy

PAGE 121

106 Percent Appropriate Prescribing Approximately 43% of the hospital visits an appropriate antibiotic was prescribed for 30-Baht patients with Infectious diarrhea before the policy was implemented. After the policy, the percent appropriate prescribing increased slightly from 43.41% to 47.47%. (Figure 6-6) 010203040506070801357911131517192123252729313335373941434547MonthPercent antibiotic prescribing appropriateness 30 Baht CSMBS Figure 6-6. Percentages of prescribing appropriateness for infectious diarrhea of the 30-Baht HI from January 2000 to December 2003 Hospital Visit and Admission Rates The average hospital visit rate of the 30-Baht HI population was 25.65 visits/10,000 beneficiaries/month) during the pre-policy period. Comparing with the disease incidence rate of 14 cases/10,000/month, the 30-Baht patients with Infectious diarrhea visited outpatient clinics twice. After the policy, there was a modest change of the rate from 25.65 to 26.49 visits/10,000/month. (Figure 6-7). In contrast, the average hospital admission rate was 2.52 admissions/10,000 beneficiaries/month before the policy. Then the rate increased about 20% (2.52 to 3.02) after the policy. (Figure 6-8) 30-Baht policy

PAGE 122

107 0510152025303540123456789101112131415161718192021222324252627282930313233343536373839404142434445464748Monthvisits/10,000/month Figure 6-7. Monthly hospital visit rates for infectious diarrhea of the 30-Baht HI from January 2000 to December 2003 0112233445123456789101112131415161718192021222324252627282930313233343536373839404142434445464748Monthadmissions/10,000/month 30-Baht policy 30-Baht policy Figure 6-8. Monthly hospital admission rates for infectious diarrhea of the 30-Baht HI group January 2000 to December 2003

PAGE 123

108 Impact of 30-Baht HI Policy on the Study Measures The results from the segmented regression analyses show that there was no statistically significant change in the level or the trend of the drug utilization rates in the 30-Baht HI population (p=.6677 and .0915, respectively) using first order autoregressive process, Autoregressive of 1, without the seasonal trend in the regression model. Autoregressive of 1 indicated that the drug utilization rate at a certain month was correlated to the rate of the prior month. (Table 6-7) Similarly, the policy was not associated with any change in the percent appropriate prescribing either at a sudden or gradual change (p=0.6799 and p= 0.0634, respectively). The results were similar for the hospital visit rates in that the policy had no statistically significant effect on hospital visit rates either on a sudden or a trend change in the 30-Baht HI population (p= 0.9868 and 0.1994, respectively). The results are consistent with the plot in Figure 6-7 in that the hospital visit rates of the 30 Bath HI group remained steady over time. The analysis of hospital admission rates show no statistical significance in a sudden (p=0.1033) or a gradual change (p=0.3672). The 30-Baht HI policy had no association on the change of drug utilization, hospital visits, and hospital admission rates, and the percent appropriateness of prescribing. However, natural positive trends of drug utilization and hospital visit rates, and the percent of appropriate prescribing and natural negative trend for hospital admission rates were observed. The monthly numbers of DDDs, the percent appropriate prescribing, hospital visits and admissions, the expected number of beneficiaries, and the rates of drug utilization and hospital visit are presented in Table 6-8.

PAGE 124

109 Table 6-7. Regression paramert ers for drug utilization, prescribing quality, hospital visit, and hospital admission rates of infectious diarrhea Measure Ordera Delayb Parameter (SEc) T-ratio p AICd A Drug utilization rate for 30 Baht HI population AR Level change Trend change 1 0 0.622 (0.129) -3.330(7.758) 10.409(6.168) 4.83 -0.43 1.69 <.0001 0.6677 0.0915 345.05 B Prescribing quality of 30-Baht group AR AR Level change Trend change 1 10 0.390 (0.121) -0.436 (0.139) -0.028 (0.069) 0.054 (0.029) 3.21 -3.13 -0.41 1.86 0.0013 0.0017 0.6799 0.0634 -100.84 C Hospital visit rate for 30-Baht HI population AR AR Level change Trend change 1 12 0 0.693 (0.111) 0.864 (0.227) 0.040 (2.400) 2.816 (2.194) 6.24 3.81 0.02 1.28 <.0001 0.0001 0.9868 0.1994 250.01 D Hospital admission rate for 30Baht HI population AR Level change Trend change 1 5 0.963 (0.038) -5.521 (3.389) -4.280 (4.736) 25.33 -1.63 -0.90 <.0001 0.1033 0.3672 234.55 a ARIMA order of the linear regression model. b Expected time delay (month) of the effect of the policy. c Standard error. d Akaike Information Criteria

PAGE 125

110 Table 6-8. Monthly rates of drug utilization, hospital visit, hospital admission, and the percent prescribing appropriateness of the 30-Baht group DDDsa Visits Admis-sions Population Drugutilization rate Hospital visit rate Hospital admission rate Appropriate Prescribing (%) Log odds of appropriate prescribing Month Pulseb Stepc Jan-2000 1,438.94 690 66 324,931 44.28 21.24 2.03 52.69 0.047 1 0 0 Feb-2000 996.80 709 56 324,931 30.68 21.82 1.72 39.46 -0.186 2 0 0 Mar-2000 1,765.63 1,003 100 324,931 54.34 30.87 3.08 48.53 -0.026 3 0 0 Apr-2000 1,285.72 833 87 324,931 39.57 25.64 2.68 44.28 -0.100 4 0 0 May-2000 1,141.53 884 90 324,931 35.13 27.21 2.77 45.40 -0.080 5 0 0 Jun-2000 1,197.24 892 104 324,931 36.85 27.45 3.20 43.23 -0.118 6 0 0 Jul-2000 1,483.80 804 132 324,931 45.67 24.74 4.06 45.10 -0.085 7 0 0 Aug-2000 1,345.96 762 76 324,931 41.42 23.45 2.34 35.24 -0.264 8 0 0 Sep-2000 1,420.33 719 68 324,931 43.71 22.13 2.09 39.42 -0.187 9 0 0 Oct-2000 993.22 642 46 324,931 30.57 19.76 1.42 35.83 -0.253 10 0 0 Nov-2000 788.44 560 74 324,931 24.26 17.23 2.28 35.15 -0.266 11 0 0 Dec-2000 792.10 710 58 324,931 24.38 21.85 1.78 56.27 0.110 12 0 0 Jan-2001 1,424.70 987 78 324,931 43.85 30.38 2.40 46.05 -0.069 13 0 0 Feb-2001 1,283.49 852 74 324,931 39.50 26.22 2.28 43.77 -0.109 14 0 0 Mar-2001 1,417.84 1,005 81 324,931 43.64 30.93 2.49 39.62 -0.183 15 0 0 Apr-2001 1,666.74 1,069 98 324,931 51.30 32.90 3.02 44.37 -0.098 16 0 0 May-2001 1,809.51 1,046 105 324,931 55.69 32.19 3.23 43.59 -0.112 17 0 0 Jun-2001 1,983.93 1,135 118 324,931 61.06 34.93 3.63 49.44 -0.010 18 1 1 Jul-2001 2,133.56 1,153 132 324,931 65.66 35.48 4.06 52.70 0.047 19 0 1 Aug-2001 2,035.16 1,088 117 324,931 62.63 33.48 3.60 50.29 0.005 20 0 1 Sep-2001 1,684.57 863 84 324,931 51.84 26.56 2.59 47.12 -0.050 21 0 1 Oct-2001 1,554.86 892 99 324,931 47.85 27.45 3.05 42.55 -0.130 22 0 1 Nov-2001 1,272.54 668 62 324,931 39.16 20.56 1.91 42.39 -0.133 23 0 1 Dec-2001 1,704.23 823 77 324,931 52.45 25.33 2.37 50.36 0.006 24 0 1 Jan-2002 1,727.57 884 103 324,931 53.17 27.21 3.17 49.84 -0.003 25 0 1 Feb-2002 1,669.73 904 98 324,931 51.39 27.82 3.02 47.45 -0.044 26 0 1 Mar-2002 2,108.36 1,011 109 324,931 64.89 31.11 3.35 52.57 0.045 27 0 1 Apr-2002 2,010.40 1,012 124 324,931 61.87 31.15 3.82 49.93 -0.001 28 0 1 May-2002 2,232.76 1,029 126 324,931 68.71 31.67 3.88 48.63 -0.024 29 0 1 Jun-2002 1,945.68 964 126 324,931 59.88 29.67 3.88 46.02 -0.069 30 0 1

PAGE 126

111 Table 6-8. Continued DDDsa Visits Admissions Population Drug utilization rate Hospital visit rate Hospital admission rate Appropriate Prescribing (%) Log odds of appropriate prescribing Month Pulseb Stepc Jul-2002 2,055.90 872 116 324,931 63.27 26.84 3.57 54.47 0.078 31 0 1 Aug-2002 1,673.58 762 96 324,931 51.51 23.45 2.95 46.78 -0.056 32 0 1 Sep-2002 1,715.69 763 94 324,931 52.80 23.48 2.89 47.59 -0.042 33 0 1 Oct-2002 2,151.25 860 116 324,931 66.21 26.47 3.57 49.25 -0.013 34 0 1 Nov-2002 1,542.96 701 81 324,931 47.49 21.57 2.49 53.13 0.054 35 0 1 Dec-2002 1,881.91 807 86 324,931 57.92 24.84 2.65 52.35 0.041 36 0 1 Jan-2003 1,678.14 782 93 324,931 51.65 24.07 2.86 50.09 0.002 37 0 1 Feb-2003 1,688.84 815 96 324,931 51.98 25.08 2.95 51.40 0.024 38 0 1 Mar-2003 1,940.85 969 85 324,931 59.73 29.82 2.62 51.95 0.034 39 0 1 Apr-2003 2,130.03 1,022 135 324,931 65.55 31.45 4.15 54.14 0.072 40 0 1 May-2003 1,848.86 873 101 324,931 56.90 26.87 3.11 48.03 -0.034 41 0 1 Jun-2003 1,652.25 785 104 324,931 50.85 24.16 3.20 45.93 -0.071 42 0 1 Jul-2003 1,705.25 870 93 324,931 52.48 26.77 2.86 43.11 -0.120 43 0 1 Aug-2003 1,169.84 665 79 324,931 36.00 20.47 2.43 42.01 -0.140 44 0 1 Sep-2003 1,176.71 719 74 324,931 36.21 22.13 2.28 39.54 -0.184 45 0 1 Oct-2003 1,141.79 728 75 324,931 35.14 22.40 2.31 39.44 -0.186 46 0 1 Nov-2003 971.92 598 76 324,931 29.91 18.40 2.34 38.97 -0.195 47 0 1 Dec-2003 942.26 661 63 324,931 29.00 20.34 1.94 34.04 -0.287 48 0 1 a Numbers of the define daily dose. b Pulse variable for the test of immediate or sudden change. c Step variable for the test of long-term or trend change

PAGE 127

112 Bacterial Pneumonia Drug Utilization Rates The average drug utilization rate for bacterial pneumonia of the 30-Baht HI population was 7.73 DDDs/10,000 population/month during the pre-policy period. It could be interpreted as about 8 doses of an antibiotic were prescribed monthly for bacterial pneumonia in 10,000 of the 30-Baht HI population. Comparing with the disease incidence report in 2002(165) of 19.6/10,000 population/month, about 40% of the 30-Baht patients with bacterial pneumonia received one dose of antibiotic per month. This means 5% of the 30-Baht HI patients with bacterial pneumonia received an antibiotic for 7 days (full course of treatment). After the policy, the rate increased from 7.73 to 12.40 DDDs/10,000 beneficiaries/month. (Figure 6-9). 0.005.0010.0015.0020.0025.001357911131517192123252729313335373941434547MonthDDDs/10,000/month 30-Baht p olic y Figure 6-9. Monthly drug utilization rates for bacterial pneumonia of the 30-Baht HI from January 2000 to December 2003

PAGE 128

113 This means about 63% of the 30-Baht HI patients with bacterial pneumonia received an antibiotic for one day per month or 11% of the patients received a full course of antibiotic treatment per month after the policy. The drug utilization rates increased gradually over time, with an exception of an outlier in month 40 (30.93 DDDs/10,000/month) that were replaced with an average of the post-policy series for the statistical analysis (11.79 DDDs/10,000/month). Hospital Visit and Admission rates The average hospital visit rates of the 30-Baht population remained stable over the observation period (pre-policy of 1.88 and post-policy of 1.93 visits/10,000/month). (Figure 6-10) The average hospital admission rate increased about 75% from 0.33 to 0.58 admissions/10,000/month. (Figure 6-11) Again, compared with the disease incidence of bacterial pneumonia, these numbers could be interpreted as only about 20% of the patients with bacterial pneumonia visited or were admitted to the hospitals. 0.00.51.01.52.02.53.03.54.014710131619222528313437404346MonthVisits/10,000/month Figure 6-10. Monthly hospital visit rates for bacterial pneumonia of the 30-Baht HI from January 2000 to December 2003 30-Baht policy

PAGE 129

114 0.00.20.40.60.81.01.21.414710131619222528313437404346MonthAdmissions/10,000/mont h Figure 6-11. Monthly hospital admission rates for bacterial pneumonia of the 30-Baht HI from January 2000 to December 2003 Impact of the 30 Bath HI Policy on the Study Measures The analysis of the segmented regression model of AR (1) showed that the 30-Baht HI policy was not associated with any changes of the level or the trend of the drug utilization rates (p = 0.0911 and 0.0641, respectively) (Table 6-9). Autoregressive order of 1 indicates that there was a correlation between the current measure and the measure one month prior. The hospital visit rates showed a drop at month 18, however it was not statistically significant (p=0.7294). The slope of hospital visit rates post-policy was not different from the slope during the pre-policy period (p=0.8288). Similarly, the 30-Baht HI policy was not associated with any changes in the level and the trend of hospital admission rates related to bacterial pneumonia (p= 0.5565, and 0.8247, respectively. The monthly numbers of DDDs, hospital visits and admissions, the expected number of beneficiaries, and the rates of drug utilization, hospital visit, and hospital admission are presented in Table 6-10. 30-Baht policy

PAGE 130

115 Table 6-9. Regression parameters for drug uti lization, prescribing quality, hospital visit, and hospital admission rates of bacterial pneumonia Measure Ordera Delayb Parameter (SEc) T-ratio p AICd A. Drug utilization rate of the 30-Baht HI group AR Level change Trend change 1 0 0.406 (0.137) -5.095 (3.016) 4.421 (1.540) 2.97 -1.69 2.87 0.0030 0.0911 0.0641 249.36 B. Hospital visit rate for 30Baht HI group AR MA Level change Trend change 1 1 0 0.508 (0.214) 0.884 (0.107) -0.158 (0.455) -0.076 (0.353) 2.38 8.26 -0.35 -0.22 0.0174 <.0001 0.7294 0.8288 65.61 C. Hospital admission rate of the 30-Baht HI group AR AR Level change Trend change 1 2 0 3.064 (0.147) 3.348 (0.148) -0.270 (0.460) 0.076 (0.350) 2.48 2.35 -0.59 0.22 0.0132 0.0188 0.5565 0.8247 65.40 a ARIMA order of the linear regression model. b Expected time delay (month) of the effect of the policy. c Standard error. d Akaike Information Criteria In summary, the visual observations of th e rates of drug utiliza tion, hospital visit, hospital admission, and the percen tage of the appropriateness of antibiotic pr escribing in the acute conditions in the 30-Baht HI popul ation alone, there were no abrupt changes when the policy was implemented. The results from the analyses of segmented linear regression of the time series data propose that the 30-Baht HI policy had no effect on any change in hospital drug utiliz ation, the quality of prescribing by physicians, or hospital outpatient and inpatient services in the selected acute and chronic conditions.

PAGE 131

Table 6-10. Monthly rates of drug utilization, hospital visit, hospital admission related to bacterial pneumonia DDDsa Visits Admissions Population Drugutilization rate Hospital visit rate Hospital admission rate Month Pulseb Stepc Jan-2000 209.25 56 15 324,931 6.44 1.72 0.46 1 0 0 Feb-2000 369.75 63 13 324,931 11.38 1.94 0.40 2 0 0 Mar-2000 384.50 74 21 324,931 11.83 2.28 0.65 3 0 0 Apr-2000 315.50 61 8 324,931 9.71 1.88 0.25 4 0 0 May-2000 188.50 80 14 324,931 5.80 2.46 0.43 5 0 0 Jun-2000 200.00 69 15 324,931 6.16 2.12 0.46 6 0 0 Jul-2000 250.00 65 8 324,931 7.69 2.00 0.25 7 0 0 Aug-2000 160.25 54 4 324,931 4.93 1.66 0.12 8 0 0 Sep-2000 234.50 57 9 324,931 7.22 1.75 0.28 9 0 0 Oct-2000 326.00 50 8 324,931 10.03 1.54 0.25 10 0 0 Nov-2000 315.00 49 6 324,931 9.69 1.51 0.18 11 0 0 Dec-2000 184.50 71 11 324,931 5.68 2.19 0.34 12 0 0 Jan-2001 269.00 52 12 324,931 8.28 1.60 0.37 13 0 0 Feb-2001 205.00 55 7 324,931 6.31 1.69 0.22 14 0 0 Mar-2001 268.00 74 13 324,931 8.25 2.28 0.40 15 0 0 Apr-2001 241.50 71 9 324,931 7.43 2.19 0.28 16 0 0 May-2001 149.00 36 10 324,931 4.59 1.11 0.31 17 0 0 Jun-2001 186.38 43 14 324,931 5.74 1.32 0.43 18 1 1 Jul-2001 379.00 53 22 324,931 11.66 1.63 0.68 19 0 1 Aug-2001 251.75 35 13 324,931 7.75 1.08 0.40 20 0 1 Sep-2001 259.00 36 19 324,931 7.97 1.11 0.58 21 0 1 Oct-2001 304.55 46 18 324,931 9.37 1.42 0.55 22 0 1 Nov-2001 271.00 39 18 324,931 8.34 1.20 0.55 23 0 1 Dec-2001 95.50 22 5 324,931 2.94 0.68 0.15 24 0 1 Jan-2002 286.50 52 16 324,931 8.82 1.60 0.49 25 0 1 Feb-2002 337.50 45 13 324,931 10.39 1.38 0.40 26 0 1 Mar-2002 395.00 55 18 324,931 12.16 1.69 0.55 27 0 1 Apr-2002 219.50 43 11 324,931 6.76 1.32 0.34 28 0 1 May-2002 257.00 48 19 324,931 7.91 1.48 0.58 29 0 1 Jun-2002 364.00 45 14 324,931 11.20 1.38 0.43 30 0 1 Jul-2002 419.90 56 22 324,931 12.92 1.72 0.68 31 0 1 116

PAGE 132

117 Table 6-10. Continued DDDsa Visits Admissions Population Drugutilization rate Hospital visit rate Hospital admission rate Month Pulseb Stepc Aug-2002 370.75 59 14 324,931 11.41 1.82 0.43 32 0 1 Sep-2002 603.55 77 27 324,931 18.57 2.37 0.83 33 0 1 Oct-2002 692.42 86 22 324,931 21.31 2.65 0.68 34 0 1 Nov-2002 360.38 45 12 324,931 11.09 1.38 0.37 35 0 1 Dec-2002 568.50 69 25 324,931 17.50 2.12 0.77 36 0 1 Jan-2003 335.38 60 13 324,931 10.32 1.85 0.40 37 0 1 Feb-2003 501.00 74 13 324,931 15.42 2.28 0.40 38 0 1 Mar-2003 452.00 108 13 324,931 13.91 3.32 0.40 39 0 1 Apr-2003 1,004.90 81 27 324,931 30.93 2.49 0.83 40 0 1 May-2003 505.80 96 24 324,931 15.57 2.95 0.74 41 0 1 Jun-2003 448.75 78 20 324,931 13.81 2.40 0.62 42 0 1 Jul-2003 608.50 112 25 324,931 18.73 3.45 0.77 43 0 1 Aug-2003 457.50 81 32 324,931 14.08 2.49 0.98 44 0 1 Sep-2003 439.25 91 22 324,931 13.52 2.80 0.68 45 0 1 Oct-2003 477.00 82 38 324,931 14.68 2.52 1.17 46 0 1 Nov-2003 302.75 73 22 324,931 9.32 2.25 0.68 47 0 1 Dec-2003 338.00 53 9 324,931 10.40 1.63 0.28 48 0 1 a Numbers of the define daily dose. b Pulse variable for the test of immediate or sudden change. c Step variable for the test of long-term or trend change

PAGE 133

118 Disparity of Drug Utilization Rates The change of drug disparity associated with the 30 HI policy was not conducted as proposed. The evidence from the drug utilization and hospital visit rates of the CSMBS suggests that the data were not appropriate to use as a comparison to the measures in the 30-Baht HI group. Comparison of the rates using the number of the DDDs per eligible population to identify disparity would not be valid. Disparity of the Percent Appropriate Prescribing for Infectious Diarrhea The disparity of the percentage of appropriate prescribing was 16% at one year before the policy and 8% at one year after the policy. The results from Chi-square test of the difference of the median percentages of prescribing appropriateness between the two populations at one year before and one year after the policy showed that there was a statistical significant reduction of the disparity with Pearsons chi-square of 3.03, p= 0.0817. The direction of the disparity was however unexpected, in that the CMSBS patients had lower percentage of appropriate prescribing than the 30-Baht HI population.

PAGE 134

CHAPTER 7 DISCUSSION The Effect of the 30-Baht HI Policy on Drug and Hospital Service Utilization Drug Utilization and Access to Care Drug utilization and hospital visit rates for bacterial pneumonia, hypertension, and diabetes found in our study were considered extremely low in the 30-Baht HI population, comparing to the disease incidence/prevalence. This indicates that most of the patients with hypertension, diabetes, and bacterial pneumonia are either under-diagnosed or underutilized drug therapy and hospital services. Approximately 2.6% of individuals with hypertension utilized hospital services and of these patients 5% received antihypertensive drugs. In the US., the prevalence of hypertension was higher than the rate in Thailand 2002 (172) (28.6% vs. 9.5%), and approximately 12% used hospital services (173) and 84% of these patients received an antihypertensive drug (174). Higher access rates in the US were also reported for diabetes. Our study found that only 10% of patients with diabetes received care from the hospitals and of these 10% were prescribed an antidiabetic drug. The US diabetes prevalence was reported 4.9% in 2004, with 42% using hospital services (173) (the hospital visit rate was 7.9 per person per year (175;176)), and 60-85% of these patients received an antidiabetic drug. Overall, the access rates for hospital service and utilization of drugs in the selected chronic diseases in Thailand were incomparable to the rates in the US. The results suggested that the 30-Baht HI policy failed to improve access to drug and hospital services to the population who need it. With only slow a natural positive 119

PAGE 135

120 trend in drug utilization and hospital visit rates (approximately 2.5% per year for hypertension and 5% in diabetes), it is not likely that the health status of the population will improve in the near future. Effectiveness of the Implementation of the 30-Baht HI Policy Our study found no changes in the drug utilization rates, appropriateness of prescribing and/or hospital service utilization associated with the 30 Bath HI policy. Results were similar for both acute and chronic conditions. The drug utilization, hospital visit, and hospital admission rates highly fluctuated throughout the observation period, which might pose difficulty to identify statistical significant changes of the measure associated with the policy. However, there was a positive trend in drug utilization rates, quality, and/or hospital visit over time without the influence of the 30-Baht HI policy. The results did not support the hypothesis that the 30-Baht HI policy that offers healthcare with a considerably small co-payment changes drug utilization rates, drug utilization quality, and hospital visit/admission rates. If the 30-Baht HI policy had an effect on these measures, an increase or decrease of the measures at or shortly after implementation of the policy would had been observed. There are several factors that might explain these results. First, the provision of the 30-Baht HI policy has changed from the beginning of the implementation to the end of the study period in regards of program eligibility, benefits, geographical allocation of the beneficiaries, and reimbursement rates. While there were efforts in promoting the policy to the target population using health marketing techniques in several types of media, response to such promotions normally takes time to reach the poorest rural populations. Second, the healthcare providers at community government hospitals were required to provide services to larger numbers of patients and/or accommodate more patient visits.

PAGE 136

121 While the infrastructure may have slowly improved to support the policy, the incentives to the health providers to promote the policy did not change. Most community government hospitals had limited numbers of personnel, especially physicians and pharmacists. This policy was seen by local providers as having the potential to overload the current healthcare system. Third, the 30-Baht HI policy not only includes the coverage for the treatment of diseases, it also included disease prevention and health promotion (e.g., vaccination, prenatal care, and health education programs). The effect of disease prevention and health promotion could have had successful results in the prevention of acute conditions (bacterial pneumonia) and reduced the need for drug utilization and hospital visits/admissions over the course of our study. The effects of disease prevention and health promotion were not likely to reduce the onsets of chronic conditions such as diabetes and hypertension however, it might prevent emergency visits from disease complications related to the uncontrolled condition. Fourth, the study was limited to government hospitals that provide primary and secondary care. However, during the observation period, the government community hospitals delegated some tasks to a local primary care units (PCU), including dispensing of some antibiotics and the refill of medications for chronic diseases (e.g., diabetes and hypertension). These drug dispensing records were not kept in the HI database and could not be included in the analysis. PCUs play an important role in healthcare systems in Thailand as they are the gate keepers who provide basic primary care services to local population. However, only limited numbers of drugs could be dispensed by the healthcare

PAGE 137

122 workers at the PCU compared with the included drugs in our study. From this reason, the missing records seemed to have small influence to the findings of the study. Fifth, the co-payment might have been one of the economic barriers in the access to drug utilization and hospital care. Even though 30-Baht ($0.75) was considered a small amount of money in developed countries and in wealthier urban areas in Thailand, this amount is almost 50% of the average daily wage in the poor population in Thailand. The common generic drugs used to treat the studied conditions may be purchased from local drugstores for less than the 30-Baht co-payment. The implementation of co-pays for health services and pharmaceuticals are used in the US healthcare system to limit the number of unnecessary patient visits and prescriptions filled. With Thailands co-pay in the range of 50% of a workers daily pay, it is fair to assume that the co-pay effectively served as a barrier to needed health care access. The 30-Baht population was considered poor compared to the populations with other health insurance status based on their income levels as farmers. In addition, they lived in rural areas where access to care seemed to be cost of care, transportation, numbers of providers, and health information. Sixth, the 30-Baht ($0.75) had to be paid for every hospital encounter. In our study, the population who were eligible for the 30-Baht HI policy was those without health insurance benefits and those who had elected to purchase a government health card that covered the costs of hospital care for the whole family for a period of one year. Patients without a health card who had to pay all healthcare costs out-of-pocket, would be more likely to pay for the 30-Baht co-pay, so an increase in utilization should have been observed in that population. On the other hand, patients who used to purchase a health card for 500 Baht ($12.50) per year to cover the whole family before the policy was in

PAGE 138

123 place, now might consider the 30-Baht per episode of care expensive and thus, be less likely to use healthcare services. A negative impact of an increased co-payment on drug utilization has been reported the literature (40;109;177). The increase in the use of healthcare services by the previously uninsured and the decrease of healthcare services by the families that normally would have purchased the health card may have counteracted the effects of the policy observed in the entire population. Seventh, our study only focused on drug utilization in the four prevalent diseases that were associated with hospital service utilization. The 30-Baht policy could have had an impact on other diseases that were not studied such as upper-respiratory infections, musculoskeletal conditions, and mental health. Similarly, the drug utilization rates were calculated for every drug on the hospital drug formulary, with an indication for the studied diseases. Because all drugs were collapsed for a given disease state, any change in drug utilization rates of a specific drug could not be identified. Eight, for drug utilization quality, only drugs that were recommended in the selected standard practice guidelines were evaluated. The evaluation of these drugs might not cover all the drugs that were used in the actual practice in the community government hospitals. In other words, the studied drugs and the disease state might not be sensitive enough to capture the effect of the 30-Baht HI policy. Finally, while the 30-Baht policy to reduce the financial barriers to the low income populations access of healthcare was a major step, there are other access to care barriers that must be corrected (i.e. distance to a provider and ability to set timely appointments Access to primary care issues must be addressed holistically.

PAGE 139

124 Suggestions to Improve the Evaluation of 30-Baht HI Policy on Drug Utilization and Hospital Service Utilization Our study results suggest that the implementation of the policy alone does not abide by the spirit in which the legislation was approved; access to drugs for the uninsured has not improved. To improve the effectiveness of the 30-Baht HI policy in providing access to care and utilization of drugs in the target population, we recommends that policy makers should consider several strategies. First, the channel of drug distribution should be expanded. At this stage, only drugs received from the hospitals are free. Patients who received drugs from local community pharmacies or other doctors clinics cannot get reimbursement. Considering the 30-Baht HI population was poor, patients have limited ability to reach the main provider because of poor road conditions and/or no vehicle to travel. These patients might consider self-treatment from local pharmacies, traditional therapy, or even no treatment. To improve patient outcomes, the policy should include doctors clinics in the district to provide primary care services (e.g., disease screening and diagnosis and prescribe medications). Local community pharmacies should be included for dispensing drugs as prescribed. Both settings should be included as a referral point and at the same time a gate keeper to prevent patient overflow to the hospitals. In addition, these settings should be a center of health and drug therapy monitoring using the advantages of close proximity to the patients. Second, community outreach by healthcare team and health workers should be implemented to reach the target population, but who have never been identified by the hospital databases should be considered. These populations are more likely to fall through the screens of the healthcare service systems. Health promotions along with

PAGE 140

125 communication of the policy should be emphasized. In particular with chronic disease, education about diseases and drug, and disease screening should be promoted. Lastly, a supportive infrastructure should be established in terms of physical facility, healthcare personnel, policy promotion, and database systems. When the 30-Baht HI policy results in larger numbers of patients receiving more hospital services and drugs, quality of healthcare service must be monitored to ensure that is not compromised. Methods in Evaluating the Health Policy A certain level of misclassification of the patients based on the assigned health insurance codes existed. We selected patients who were eligible for the 30-Baht HI policy by the codes assigned for the patients who paid cash (when no reference to a health benefit was found), using a health card (considered poor), or who were labeled as having no health benefits. Some patients who had benefits with the CSMBS or other health plans had to pay cash when they did not present their identification cards. This misclassification of the 30 Bath HI group might have underestimated the effect of the 30-Baht HI policy on drug utilization and hospital access rates when compared to the CSMBS group or the others. The CSMBS group was planned in our study as a concurrent control group to improve internal validity by controlling for external factors that might affect the measures rather than the 30-Baht HI policy. Since the control group could not be used, external factors that might have an effect of the study measures were not controlled in our study. The selection of the control group was crucial to eliminate external factors that could affect the measures at the same time when the policy was implemented.

PAGE 141

126 Time Series Design Segmented regression analysis of time series data is an analysis technique for observational studies that have multiple longitudinal observations, which improves internal validity when a randomization study is not possible. For example, as health policies or drug policies are generally applied to large populations, random assignment of receivers and non-receivers of the policy is rarely possible. Although in our study the concurrent control group was not valid, the historical control should be sufficient to isolate the effect of the policy, if there was any. Even though segmented regression analysis of time series data has several strong characteristics, it has some limitations. The analysis in our study assumed a linear trend in the outcome measures in both pre-policy and post-policy periods. This assumption may be incorrect, for example disease incidence follows seasonal pattern. Advanced mathematical models (e.g., Box-Jenkins Tao model) may be employed to accommodate a non-linear trend; however it requires a large number of observations (at least 50) to achieve sufficient statistical power. Segmented regression analysis applied aggregate data of the population thus; individual data did not pertain anymore in the results. Therefore, individual adjustment for ones characteristics could not be conducted. As a consequent, the results of the analysis may not apply to an individual. The Defined Daily Dose The measurement of drug utilization using the defined daily dose provided simple and objective measures of prescribing in studies of utilization of a single drug or several drugs with the same indication to treat a particular disease. This unit of measurement was useful for comparisons of drug utilization among subpopulations or drug use

PAGE 142

127 statistics in most developing and European countries where the defined daily dose was widely used to measure drug utilization. Quality of the Data in the HI database The results of the database validation using internal data validation methods suggested that the quality of the data from the HI database was sufficient for policy evaluation in our study. The HI database contained a comprehensive backbone with several sub-databases (physician office, inpatient ward, pharmacy, laboratory, and billing) that were linkable, and was found sufficient to compliment paper medical records for both outpatient and inpatient services. The database contained patient-specific healthcare data, which allowed drug and healthcare service utilization monitoring at the patient level. Patient outcomes were available in this database as discharge status for inpatient records that can be included in a health policy evaluation. In summary, the well-designed structure of the database and the comprehensiveness of the data at the patient-level offers opportunities for health policy evaluation, drug utilization, patient outcome, or even economic research. The current limitation of the data for research associated with the database structure was the incomplete linkage of the databases. The assessment of data quality suggested that a minimal number of missing data was found throughout the observation period and across the studied hospitals. However, there was evidence of higher percentage of missing data in the year 2000, in which several hospitals began to use the HI database system. There was also a trend in decreasing missing data over time, especially for the number of disease diagnosis codes and health insurance status. This could be explained by improvements of the data entry as users become more familiar with the system. Outpatient data contained more missing data compared to the inpatient data, which might be related to a larger amount of data

PAGE 143

128 that had to be recorded daily or the short duration of patient-provider interactions. Data entry personnel may have focused on the necessary clinical and financial data needed for hospital service and billing transactions, while auxiliary data (e.g., vital sign, lab data, or disease severity) had lower priority (Personal communication, Hospital D & E Directors, July 20 & 22, 2004). The database missed all dictated reports containing clinical data that were necessary to justify the use of drugs in our study. This incomplete clinical data posed limitations to accurately identify patients with a disease and to evaluate appropriateness of prescribing. Several studies of the validity of disease diagnosis codes (mostly ICD-9) suggested that over-coding issues took place often to take advantage of higher reimbursement in capitation-based payment systems(153;178). On the other hand, under-coding of diseases such as mental health (155) and lung cancer (152) conditions have been described. In our study, most of the patients were assigned only one to two ICD-10 codes, while there were 16 available spaces for data entry in the HI database. Moreover, the proportion of patients with more than two disease diagnosis codes increased during the observation period. The increased number of diagnosis codes and the smaller number of missing data suggested that the data input quality has improved over time and was less likelihood of over-coding. The reduction of missing data over time due to an improvement of data entry could affect the analysis in our study. More patients with the disease were identified even though the disease incidence or hospital access did not change, thus resulting in an

PAGE 144

129 artificially increased drug utilization rate. The effect of increased data entry could bias the analysis of the effect of the 30-Baht HI policy on drug utilization rates. The assessment of the data validity for the disease diagnosis and drug coding suggested that the observed level of coding errors was minimal. However, it was determined to be in an acceptable range for the analysis. The results of the gender-specific disease diagnosis validation showed a very small percentage of mismatched cases. This analysis offered important information about coding errors, which suggested the gender-specific disease diagnosis had considerably high accuracy. However, our ability to validate diagnosis codes was limited as no paper records were available for the whole data set and internal cross checks were confined to demographic and drug data. Using drug therapy to validate disease coding was not useful because not every patient with the disease may have received drugs for the treatment. In contrast, validation of drug data using diagnosis codes was successful. Antidiabetic drugs should have prescribed only for diabetic patients. Thus, most discrepancies were related to coding errors. For hypertension, ACEIs, beta-blockers, alpha-adrenergic blockers, and diuretics had multiple indications to treat cardiovascular conditions. These drugs were not specific to hypertensive conditions. Non-specificity was more prominent in antibiotics, where each antibiotic could be used to treat any infectious conditions that were susceptible to the pathogens covered by the antibiotics. Even though, validation of drug data using disease diagnosis in our study was valid only for diabetes drugs, the information of the percentages of data coherence between disease diagnosis and drugs could be useful to identify coding errors when comparing these percentages across time and hospitals.

PAGE 145

130 Patient demographic, disease diagnosis, and drug data appeared to be valid for our study, however, the lack of external validation suggests that the results of our study using the data in the HI data should be carefully interpreted. External data validation for future drug utilization research is required. The results of the datas face validity in terms of data consistency across the selected hospitals and longitudinally had several implications for the validity of the study. Similar lists of dispensed drugs among the studied hospitals allowed the assumption that the case mix and drug use were similar and thus, the data from each hospital could be combined for analysis purposes of our study. Simultaneous analysis of the combined data in turn improved external validity. Study Limitations First, our study has some limitations related to the use of administrative healthcare databases. Despite large benefits of administrative healthcare data including saved time related to data entry, large numbers of records, ease of computerized statistical analysis, and problems related to unavailability and incomplete clinical data exist. Disease severity and laboratory results were not recorded in the hospital database, which limited the ability to adjust for healthcare needs and appropriateness of drug selection and dosage for the comparison of drug use between the 30-Baht HI patients and the CSMBS group. Second, because we calculated drug utilization rates from computerized pharmacy dispensing records, it represented only drugs that were prescribed and dispensed in the hospitals. A prescription that was dispensed outside the designated hospital or drug use from other sources (e.g., private hospitals or clinics, local health offices, community pharmacy, or self-treatment) was not measured in our study. Consequently, needs of healthcare services that might have been affected by other drug distribution channels,

PAGE 146

131 could not be adjusted. If the implementation of the 30-Baht HI policy was successful in improving access to services and drugs from hospitals, it is expected that drug utilization from other sources was steady if the policy improve utilization in the patients who had never used healthcare services, or drug utilization was reduced if the policy shifted patients from self-treatment to receive hospital services. Third, CSBMS patients were able to receive care from other hospitals, including private hospitals, provincial hospital and the tertiary hospital, that might have underestimated drug use through the government hospitals. On the other hand, 30-Baht beneficiaries must have received care from the government hospitals to use the benefit of the new policy. The unequal choice of care might affect patients use of healthcare services from the selected hospitals, which, unfortunately, could not be adjusted by using the data from the government hospitals. Since the CSMBS beneficiaries have wider choice of providers, the measurement of drug and healthcare services from the selected community government hospitals might underestimate actual access and utilization in this population. This may further explain why rates in the CSMBS group fluctuated heavily. Fourth, since the patient-specific data from the hospital databases has not been used for research purposes, there has not been any intensive investigation process of data quality by the hospitals. A comparison of the data in electronic with paper records was not possible for outpatients because the paper records were no longer available. As a consequence, data that were not entered to the database were not included. Without external validation, the magnitude of missing cases was unknown. Validity assessment was conducted only through the self-data check and subjective quality assessment from database users, and expert opinions.

PAGE 147

132 Fifth, our study focused heavily on the effect of the new policy on drug utilization rates and drug utilization quality as defined for the purpose of our study. Appropriateness of drug use or drug utilization review by other criteria other than choices, dosage, and duration of therapy was not evaluated in our study. Patient health outcomes as a result of drug therapy (e.g., drug effectiveness, mortality, patient satisfaction, or costs) were not studied. Sixth, the study hypothesized that there was a positive effect on drug use as a result of the 30-Baht HI policy by reducing economic barriers. However, other barriers to healthcare services (e.g., geographical, or personal barriers) were not studied. These barriers might hinder the effect of the 30-Baht HI policy on drug use in the target population. Seventh, the time series study design utilized aggregate data of drug utilization rates, drug quality, and hospital visit/admission rates of a population and thus, the results do not provide any information at individual level whether the 30-Baht HI policy had an impact on drug use or the quality of prescription received or frequency of hospital visit of an individual. Eighth, since the control group could not be included in the analysis because of the unidentified changes in all measures, other external factors that might have had an impact on drug utilization (e.g., changing in prescribing patterns over time, changing a group of physicians periodically due to shortage of practitioners, or any other health-related policies) could not be controlled, and that might confound the results of the study. Ninth, since the data were extracted from only 8 from 19 government community hospitals in Ubonratchatani province, the results could not be generalized to other

PAGE 148

133 hospitals in other districts, other types of hospitals, or other provinces. This was because the prevalence of the diseases, patterns of care, geographical barriers to receive care from community government hospitals, and other related-factors might have been different from the studied populations in the selected 8 hospitals. Lastly, despite the assumption of geographical similarity among the selected hospitals there were some differences between community hospitals and the hospitals at the same level in other provinces (e.g., patterns of diseases, variations of medical practice among physicians and hospital drug formularies). Thus, the finding from our study might not be generalized to drug use for other Thai provinces. Despite the study found no effect of the 30-Baht HI policy drug utilization and access to hospital services, the study was attempted to apply computerized patient-specific drug utilization data extracted from the main providers of the 30-Baht HI beneficiaries that allow diseases-based analysis. We utilized time series analysis that is the strongest study design for a policy evaluation where a randomized controlled trial is not possible and sufficiently powered statistical analysis to identify the effect of the health policy in our study. Recommendations for Future Research To study the effects of a health policy on drug utilization rates, researchers should consider including quality of care provided and patient outcomes (clinical and patient-reported outcomes). It is recommended that future researchers validate disease diagnosis codes and drug data in the HI database by comparing with electronic data for inpatients with progress notes that are used in the hospital wards to ascertain that the data are accurate and reliable for use in drug utilization studies and the assessment of prescribing appropriateness.

PAGE 149

134 Contribution to Scientific Society Our study utilized computerized patient-specific healthcare data that has never been used for health service research. Internal data validation methods were used to ensure adequate data quality, while external data validation was not possible. The initial data validation could lead to further assessment of the HI databases quality for use in research, and for other assessment purposes, such as quality improvement programs in the hospitals. From a public health perspective, this database was useful for generating meaningful disease and health statistics regarding service utilization of the population. While the availability of patient-specific data from hospital administrative databases in developing countries are typically limited, we demonstrated that this methodology produced useful estimates of drug and heath service utilization that may be more accurate than traditional techniques currently available. Developing countries should look for existing databases within their current systems and validate the data for possible use, as was done for our study. The low rates of drug utilization, hospital visit, and admission from our study suggested problems of the healthcare delivery systems. This information should alert governmental officials that the implementation of the 30-Baht HI policy is still ongoing that the development of a more supportive infrastructure in Thailand to improve access to care and drugs should be considered. The study applied the strongest statistical analysis for an observational study available to improve the validity of the findings. The segmented regression analysis of time series data with a concurrent control group is recommended for future policy evaluation and health service utilization monitoring.

PAGE 150

135 While the implementation of the 30-Baht HI policy to improve quantity of drug utilization was the major focus, its effects on quality should also be considered. Our study provided the impact of the policy on both the quantity and the quality of drug utilization for, which may be useful to facilitate decision making in the improvement and continuation of such a policy. Conclusions The findings of our study indicated that the 30-Baht HI policy had no impact on drug utilization quantitatively and qualitatively, neither in acute or chronic conditions in community government hospitals. Access to hospital services and drugs was extremely low in the target population and was not improved by the 30-Baht HI policy as intended. There was a natural positive trend in drug and hospital service utilization in the 30-Baht HI population; however the increase was not significant in improving health of the population. Effective intervention of the 30-Baht HI policy is needed.

PAGE 151

APPENDIX A TERMINOLOGY AND DEFINITIONS Drug Utilization Study A study of drug utilization focuses on drug consumption in a specific population, patterns of use cross-sectional and overtime, drug use in specific disease states, the detection of adverse outcomes, drug use as a surrogate measure for disease states, drug use for health policy evaluation tool, drug use across geographical areas, and the effect of demographic variables on drug use. Drug utilization studies are widely used for health policy evaluation worldwide (35). The 30-Baht Health Insurance Policy The 30-Baht HI policy is a strategy to expand the existing health insurance system in Thailand towards universal health coverage. The Thai government introduced the new policy nationwide in October 2001. This policy aims at reforming the health service system to provide health benefits to all Thai citizens for a small co-payment of 30-Baht per episode of care. The 30-Baht HI policy beneficiaries are able to receive all available healthcare services, including outpatient, inpatient, medication, and preventive care from any government hospitals in Thailand. 30-Baht beneficiaries are required to pre-register to a primary care government hospital in the district of their residence. The 30-Baht HI policy is a separate health insurance benefit from in the existing health insurance plans. There are five existing major forms of health insurance in Thailand: 1) the Low-income Card Scheme (LICS) for low-income individuals, 2) the Civil Servants Medical Benefits Scheme (CSMBS), which covers all government 136

PAGE 152

137 employees and pensioners, and their dependents, 3) the Voluntary Health Card Scheme (VCS) aimed at the near poor, 4) the Social Security Scheme (SSS) and the Workmans compensation Scheme (WCS), which covers workers using a co-payment method, and 5) Private Indemnity Insurance. Disparity of Drug Use In our study, disparity of drug use is defined as the differences of drug utilization rates and quality of drug use between 30-Baht HI beneficiaries and of CSMBS patients. Disparity can be explained by using the opposite term, equity, which is often used in healthcare policy and healthcare service utilization research. Equity is defined as an equal access to quality care based on needs, regardless of financial barriers(110). Equity of receiving quality care in this study is defined as presence of pharmacological similar drugs with equally adequate dosage range and duration of treatment among the two groups. However, equity of drug therapy may still be compromised by restricted hospital formularies and other costs control techniques or local treatment modalities if differently applied to groups.

PAGE 153

APPENDIX B HEALTH INSURANCE CODES Table B-1. Codes for health insurance status for the Health Insurance database Patient status or payment method Code 30-Baht health insurance* Cash 10 Low income 52 No benefit 60 30-Baht in the district 81 30-Baht outside of the district 82 30-Baht outside of the province 83 Civil Servant Medical Benefit Scheme Government/Enterprise 20 Civil service medical benefit 21 Governmental retirement benefit 22 MOPH Civil service 23 Patients under the age of 18 are covered by the 30-Baht HI policy but will not be included in this study because of the age criteria. 138

PAGE 154

APPENDIX C RECORD LINKAGE Record Linkage of Outpatient Data Figure C-1. Two tables identify patients who visited outpatient clinic during the study period. A) The OVST year table. B) The OVSTDX year table. HN = hospital number, a unique patient identifier. VSTDATE = visit date. VSTTIME = visit time. PTTYPE = type of patient by health insurance status. ICD10 = disease diagnosis code. 139

PAGE 155

140 Record Linkage of Inpatient Data Figure C-2. Two tables identify patients admitted to the hospital for inpatient services during the study period, and length of stay. A) The IPT year table. B) The IPTDX year table. HN = hospital number, a unique patient identifier. AN = admission numbers, restarted annually. RGTDATE = hospital admission date. RGTTIME = hospital admission time. PTTYPE = type of health insurance. PREDIAG = preliminary diagnosis. DCHDATE = discharge date. DAYCNT= length of stay. ITEMNO = number of diagnoses. ICD10 = disease diagnosis code

PAGE 156

141 Patient Demographic Table Figure C-3. Patient demographic table identifies patients age, gender, marital status, types of health insurance. HN = hospital number. BRTHDATE = date of birth. MALE = gender. OCCPTN = occupation. PTTYPE = type of health insurance. MRTLST = marital status.

PAGE 157

142 Record Linkage for Prescription Drug Link to OVSTyear, OVSTDXyear, or PT table Link to IPTyear, IPTDXyear Figure C-4. Two tables contain drug data. A) The PRSC year table. B) The PRSCDT year table. PRSCNO = prescription number. PRSCDATE = prescribing date. PRSCTIME = prescribing time. HN = hospital number, a unique patient identifier. AN = hospital admission number. PTTYPE = type of health insurance. MEDITEM = medication item number. QTY = quantity of dispensed drugs. MEDUSAGE = instruction of drug administration

PAGE 158

APPENDIX D DRUG LISTS FOR CALCULATING DRUG UTILIZATION RATES Table D-1. Antidiabetic drugs for calculating drug utilization rates Hospital Drug code Drug name A 245 Glibenclamide Tab. 5 Mg 350 Metformin Tab. 500 Mg. 288 Isophane Insulin 100 Iu/Ml 810 Human Insulin 70/30 485 Regular Insulin Inj. 40 Iu/Ml B 54 Chlorpropramide Tab 250 Mg 771 Glipizide Tab 5 Mg 29 Glybenclamide 5 Mg Tab 800 Humulin N Cartidge 300 U X5 288 Isophane Insulin 100 Iu/Ml 289 Isophane Insulin 100 Iu/Ml,Human 350 Metformin Tab. 500 Mg. 170 Ri Inj 100 Iu/Ml,Human C 103 Chlorpropamide Tab.250mg 245 Glibenclamide Tab. 5 Mg 288 Isophane Insulin 100 Iu/Ml 350 Metformin Tab. 500 Mg. 172 Nph Insulin 485 Ri Insulin Inj. 797 Ri Stat Dose 40 Iu D 103 Chlorpropamide Tab.250mg 245 Glibenclamide Tab. 5 Mg 350 Metformin Tab. 500 Mg. 9 Nph 485 Ri Insulin Inj. 40 Iu/Ml E 103 Chlorpropamide Tab.250mg 245 Glibenclamide Tab. 5 Mg 288 Isophane Insulin 100 Iu/Ml 350 Metformin Tab. 500 Mg. 810 Ri Insulin Inj 100 Iu/Ml 485 Ri Insulin Inj. 100 Iu/Ml 143

PAGE 159

144 Table D-1. Continued Hospital Drug code Drug name 103 Chlorpropamide Tab.250mg 248 Glybenclamide Tab. 5 Mg 288 Isophane Human Insulin 100 Iu/Ml 73 Humulin N 81 Biphasic30/70 Human Insulin 6031 Mixtard Penfill Insulin 300 U 350 Metformin Tab. 500 Mg. 6023 Penfill Nph Insulin 6035 Ri Insulin 1 Unit (Inj.At Er) 485 Ri Insulin Inj. 40 Iu/Ml G 248 Glibenclamide 5 Mg 245 Glybenclamide 5 Mg 350 Metformin 500 Mg 288 Nph Insulin 300 Iu/Ml 868 Nph 300 Unit/3ml (Penfill) 103 Diabenes 250 Mg 811 Diabenes 250 Mg 103 Chlorpropamide 250 Mg 840 Glipizide 5 Mg 18 Gliclazide 80 Mg. Tab. H 274 Glibenclamide Tab. 5 Mg 271 Insulin Regular 100 Iu/Ml 302 Isophane Insulin 100 Iu/Ml 545 Mixtard Insulin 450 Ri Insulin Inj. 40 Iu/Ml 343 Metformin Tab. 500 Mg.

PAGE 160

145 Table D-2. Antihypertensive drugs for calculating drug utilization rates Hospital Meditem Name A 25 Amiloride+Hctz Tab. 207 Enalapril 5 Mg Tab 233 Furosemide Inj. 10 Mg/Ml 235 Furosemide Tab. 40 Mg 262 Hctz Tab. 50 Mg 354 Methyldopa Tab. 250 Mg 475 Propanolol Hcl Tab. 40 Mg 474 Propanolol Hcl Tab. 10 Mg 804 Enalapril 20 Mg Tab 832 Furosemide Tab.500 Mg 828 Spironolactone 25mg B 25 Amiloride+Hctz Tab. 206 Enalapril Tab. 20 Mg 207 Enalapril Tab. 5 Mg 23 Felodipine Tab 5 Mg 115 Furosemide Hd Inj. 233 Furosemide Inj. 10 Mg/Ml 235 Furosemide Tab. 40 Mg 262 Hctz Tab. 50 Mg 373 Furosemide+Amiloride Tab. 353 Methyldopa Tab. 125 Mg 354 Methyldopa Tab. 250 Mg 457 Prazosin Tab.2 Mg 474 Propranolol Hcl Tab. 10 Mg 475 Propranolol Hcl Tab. 40 Mg 774 Verapamil 40 Mg Tab 775 Verapamil Inj 5mg/2ml C 106 Methyldopa Tab. 250 Mg 182 Prazosin 1mg 207 Enalapril Tab. 5 Mg 233 Furosemide Inj. 10 Mg/Ml 235 Furosemide Tab. 40 Mg 25 Amiloride+Hctz Tab. 262 Hctz Tab. 50 Mg 353 Methyldopa Tab. 125 Mg 474 Propanolol Hcl Tab. 10 Mg 475 Propanolol Hcl Tab. 40 Mg 524 Spironolactone Tab. 25 Mg 60 Moduretic

PAGE 161

146 Table D-2. Continued Hospital Meditem Name D 233 Furosemide Inj. 10 Mg/Ml 207 Enalapril Tab. 5 Mg 235 Furosemide Tab. 40 Mg 262 Hctz Tab. 50 Mg 354 Methyldopa Tab. 250 Mg 373 Moduretic Tab. 474 Propanolol Hcl Tab. 10 Mg 475 Propanolol Hcl Tab. 40 Mg 524 Spironolactone Tab. 25 Mg 628 Enalapril Maleate 20 Mg 759 Amiloride+Hctz 777 Furosemide 500 Mg Tab. 783 Prazosin 2 Mg Tab. 794 Furosemide 250 Mg/10ml 827 Felodipine 2.5 Mg E 186 Diltiazem 30 Mg. Tab. 152 Diltiazem Hcl 30 Mg. Tab. 183 Enalapril 20 Mg. Tab. 191 Felodipine 5 Mg. Tab. 767 Atenolol 100mg. 576 Verapamil Tab. 40 Mg 831 Diltiazem Hcl. 30 Mg. 207 Enalapril 5 Mg Tab. 233 Furosemide Inj. 10 Mg/Ml 234 Furosemide Inj. Hd 10mg/Ml 235 Furosemide Tab. 40 Mg 828 Furosemide Tab.500 Mg. 262 Hctz Tab. 50 Mg 353 Methyldopa Tab. 125 Mg 354 Methyldopa Tab. 250 Mg 373 Moduretic Tab. 474 Propanolol Hcl Tab. 10 Mg 475 Propanolol Hcl Tab. 40 Mg 524 Spironolactone Tab. 25 Mg 575 Verapamil Inj. 1 Mg 213 Verapramil Hcl 40 Mg.

PAGE 162

147 Table D-2. Continued Hospital Meditem Name F 207 Enalapril Tab. 5 Mg 6060 Enalapril 20 Mg 6007 Furosemide 500mg/Tab 233 Furosemide Inj. 20 Mg/Ml 235 Furosemide Tab. 40 Mg 25 Amiloride+Hctz Tab. 5990 Atenolol 100 Mg 262 Hctz Tab. 50 Mg 353 Methyldopa Tab. 125 Mg 354 Methyldopa Tab. 250 Mg 474 Propanolol Hcl. 10 Mg 62 Propranolol 40 Mg 524 Spironolactone Tab. 25 Mg G 280 Zestril (Lisinopril) 474 Inderal 10 Mg 262 Hctz 373, 806 Moduretic 474 Propanolol 10 Mg 37 Verapamil 40 Mg 614 Aldactone 29 Verapamil Inj 353 Aldomet 125 Mg 354 Aldomet 250 Mg 756 Atenolol 100 Mg 207 Enalapril 5 Mg 206 Enalapril 10 Mg 234 Furosemide Inj 10 Mg/Ml 794 Furosemide 500 Mg 233 Lasix Inj 10 Mg/Ml 235 Lasix Tab 40 Mg H 288 Hctz Tab. 50 Mg 106 Amiloride+Hctz Tab. 548 Atenolol 100 Mg. 536 Enalapril 5 Mg 7 Enarapril 20 Mg 268 Furosemide Inj. 10 Mg/Ml 269 Furosemide Tab. 40 Mg 438 Propranolol Hcl Tab. 10 Mg 543 Verapamil 40 Mg Tab. 148 Verapamil Hydrochloride 514 Verapamil Inj. 1 Mg

PAGE 163

148 Table D-3. Antibiotics for calculating drug utilization rates Hospital Drug Code Drug name A 35 Amoxycillin Cap. 250 Mg 36 Amoxycillin Cap. 500 Mg 38 Amoxycillin Syr.125mg/5ml 41 Ampicillin Inj. 1.0 Gm 42 Ampicillin Inj. 250 Mg 43 Ampicillin Inj. 500 Mg 108 Chloramphenicol Cap.250mg 111 Chloramphenicol Inj. 1 Gm 135 Cloxacillin Cap 250 Mg 138 Cloxacillin Sodium 1 Gm 139 Cloxacillin Syr. 125 Mg/Ml 150 Cotrimoxazole Susp. 151 Cotrimoxazole Tab. 184 Dicloxacillin Cap.250 Mg 200 Doxycycline Cap 100 Mg 219 Erythromycin Tab. 250 Mg (As Stearate) 220 Erythromycin Syr 125mg/5ml 238 Gentamicin Inj. 40 Mg/Ml 321 Lincomycin Inj. 300 Mg/Ml 400 Norfloxacin Tab. 100 Mg 401 Norfloxacin Tab. 200 Mg 402 Norfloxacin Tab. 400 Mg 424 Penicillin G 5 M Iu Inj. 425 Penicillin V Tab. 125 Mg 426 Penicillin V Syr.125mg/5ml 427 Penicillin V Tab. 250 Mg 543 Tetracycline Cap. 250 Mg 764 Lincomycin 300mg/Ml In 2 Ml 765 Streptomycin 811 Ceftriaxone 773 Roxithromycin 150 Mg 822 Ofloxacine 200 Mg B 108 Chloramphenicol Cap.250mg 111 Chloramphenicol Inj. 1 Gm 117 Ofloxacin Tab.100 Mg. 135 Cloxacillin Cap.250 Mg 137 Cloxacillin Cap.500 Mg 138 Cloxacillin Inj.1 Gm 150 Cotrimoxazole Susp. 151 Cotrimoxazole Tab. 200 Doxycycline Cap 100 Mg

PAGE 164

149 Table D-3. Continued Hospital Drug Code Drug name 219 Erythromycin Tab. 250 Mg 238 Gentamicin Inj. 40 Mg/Ml 3 Cloxacillin Dry Syrup 321 Lincomycin Inj.300 Mg/Ml 33 Roxithromycin 150 Mg Tab 35 Amoxycillin Cap. 250 Mg 36 Amoxycillin Cap. 500 Mg 38 Amoxycillin 125mg/5ml Syr 400 Norfloxacin Tab. 100 Mg 402 Norfloxacin Tab. 400 Mg 41 Ampicillin Inj. 1.0 Gm 42 Ampicillin Inj. 250 Mg 424 Penicillin G 5 M Iu Inj. 425 Penicillin V Tab. 125 Mg 426 Penicillin V Syr.125mg/5ml 427 Penicillin V Tab. 250 Mg 43 Ampicillin Inj. 500 Mg 526 Streptomycin Inj. 5 Gm 641 Ceftriazone 1g.Inj. 70 Benzathine Penicillin Inj 763 Dicloxacillin 250 Mg 787 Kanamycin Injection 1g/Vial 788 Ciprofloxacin Tablet 500 Mg 806 Amoxycillin 250mg/5ml Dry Syrup 808 Cefazolin Inj.1 Gm 826 Dicloxa Dry Syrup 92 Kanamycin 1 G.Inj C 136 Cloxacillin Cap. 250 Mg 138 Cloxacillin Sodium 1 Gm 139 Cloxacillin Syr. 25 Mg/Ml 150 Cotrimoxazole Susp. 151 Cotrimoxazole Tab. 184 Dicloxacillin Tab.250 Mg 189 Ceftriaxone 1 G 191 Roxithromycin 150 Mg 200 Doxycycline Cap 100 Mg 219 Erythromycin Tab. 250 Mg 220 Erythromycin Syr 125mg/5ml 238 Gentamicin Inj. 40 Mg/Ml

PAGE 165

150 Table D-3. Continued Hospital Drug Code Drug name 28 Norfloxacin Tab. 200 Mg 36 Amoxycillin Cap. 500 Mg 38 Amoxycillin Syr.125mg/5ml 402 Norfloxacin Tab. 400 Mg 41 Ampicillin Inj.1.0 Gm 42 Ampicillin Inj. 250 Mg 424 Penicillin G 5 M Iu Inj. 425 Penicillin V Tab. 125 Mg 426 Penicillin V Syr.125mg/5ml 427 Penicillin V Tab. 250 Mg 43 Ampicillin Inj. 500 Mg 525 Streptomycin Inj. 1 Gm 526 Streptomycin Inj. 5 Gm 543 Tetracycline Cap. 250 Mg 760 Lincomycin 300 Mg/Ml 10 Ml D 108 Chloramphenicol Cap.250mg 111 Chloramphenicol Inj. 1 Gm 136 Cloxacillin Cap. 250 Mg 137 Cloxacillin Cap. 500 Mg 138 Cloxacillin Sodium 1 Gm 139 Cloxacillin Syr. 25 Mg/Ml 150 Cotrimoxazole Susp. 151 Cotrimoxazole Tab. 200 Doxycycline Cap 100 Mg 219 Erythromycin Tab. 250 Mg 220 Erythromycin Syr 125mg/5ml 238 Gentamicin Inj. 40 Mg/Ml 321 Lincomycin Inj. 300 Mg/Ml 35 Amoxycillin Cap. 250 Mg 37 Amoxycillin Cap. 500 Mg 38 Amoxycillin Syr.125mg/5ml 41 Ampicillin Inj. 1.0 Gm 42 Ampicillin Inj. 250 Mg 424 Penicillin G 5 M Iu Inj. 425 Penicillin V Tab. 125 Mg 426 Penicillin V Syr.125mg/5ml 427 Penicillin V Tab. 250 Mg 43 Ampicillin Inj. 500 Mg 525 Streptomycin Inj. 1 Gm

PAGE 166

151 Table D-3. Continued Hospital Drug Code Drug name 526 Streptomycin Inj. 5 Gm 606 Norfloxacin 200 Mg. 776 Cotrimoxazole 960 Mg 784 Penicillin G Benzathine 790 Norflox 400 Mg 791 Dicloxa 250 Mg 805 Ceftriazone 813 Ceftriazone 826 Norfloxacin 100 Mg E 170 Amoxy 250 + Clavulanic Acid 125 Mg. Tab. 39 Ampicillin Cap. 250 Mg 40 Ampicillin Cap. 500 Mg 44 Ampicillin Syr. 125mg/5ml 761 Cloxacillin 250 Mg. Cap. 760 Cloxacillin 250mg.Cap. 35 Amoxycillin Cap. 250 Mg 36 Amoxycillin Cap. 500 Mg 38 Amoxycillin Syr.125mg/5ml 41 Ampicillin Inj. 1.0 Gm 42 Ampicillin Inj. 250 Mg 43 Ampicillin Inj. 500 Mg 70 Benzathine Penicillin Inj 97 Cefazolin Inj 1 Gm. 641 Ceftriazone 1 Gm. Inj. 111 Chloramphenicol Inj. 1 Gm 762 Cloxacillin 250 Mg Cap. 138 Cloxacillin Sodium 1 Gm 150 Cotrimoxazole Susp. 151 Cotrimoxazole Tab. 850 Dicloxacillin 62.5 Mg./Ml.Syr. 135 Dicloxacillin Cap 250 Mg 137 Dicloxacillin Cap. 500 Mg 139 Dicloxacillin Syr.125 Mg/Ml 200 Doxycycline Cap 100 Mg 220 Erythromycin Syr 125mg/5ml 219 Erythromycin Tab. 250 Mg 238 Gentamicin Inj. 40 Mg/Ml 402 Norflox. 400 Mg Tab 404 Norflox. 400 Mg Tab 400 Norfloxacin Tab. 100 Mg

PAGE 167

152 Table D-3. Continued Hospital Drug Code Drug name 403 Norfloxacin Tab. 200 Mg 425 Penicillin V Tab. 125 Mg 427 Penicillin V Tab. 250 Mg 206 Roxithomicin 150 Mg. 142 Streptomycin 1 Gm Inj. 543 Tetracycline Cap. 250 Mg 321 Lincomycin Inj. 300 Mg/Ml 322 Lincomycin Inj. 300 Mg/Ml F 5997 Amoxy+Clavulanic Acid Susp. 6061 Amoxycillin Cap 500 Mg 35 Amoxycillin Cap. 250 Mg 38 Amoxycillin Syr.125mg/5ml 641 Ceftriazone 1 G. Inj. 135 Cloxacillin Cap 250 Mg 137 Cloxacillin Cap. 500 Mg 138 Cloxacillin Sodium 1 Gm 139 Cloxacillin Syr. 25 Mg/Ml 6009 Cotrimoxazole Inj. 150 Cotrimoxazole Susp. 151 Cotrimoxazole Tab. 200 Doxycycline Cap 100 Mg 220 Erythromycin Syr 125mg/5ml 219 Erythromycin Tab. 250 Mg 238 Gentamicin Inj. 40 Mg/Ml 606 Norfloxacin 200 Mg. 402 Norfloxacin Tab. 400 Mg 5988 Roxithromycin 150 Mg 321 Lincomycin Inj. 300 Mg/Ml G 35 Amoxycillin Cap. 250 Mg 36 Amoxycillin Cap. 500 Mg 38 Amoxycillin Syr.125mg/5ml 41 Ampicillin Inj. 1.0 Gm 42 Ampicillin Inj. 250 Mg 43 Ampicillin Inj. 500 Mg 70 Benzathine Penicillin Inj 108 Chloramphenicol Cap.250mg 111 Chloramphenicol Inj. 1 Gm 135 Cloxacillin Cap 250 Mg 138 Cloxacillin Sodium 1 Gm 139 Cloxacillin Syr. 25 Mg/Ml

PAGE 168

153 Table D-3. Continued Hospital Drug Code Drug name 150 Cotrimoxazole Susp. 200 Doxycycline Cap 100 Mg 219 Erythromycin Tab. 250 Mg 220 Erythromycin Syr 125mg/5ml 238 Gentamicin Inj. 40 Mg/Ml 292 Kanamycin Inj. 1 Gm 321 Lincomycin Inj.300mg/Ml 10 Ml. 322 Lincomycin Inj.300 Mg/Ml 2ml. 402 Norfloxacin Tab. 400 Mg 424 Penicillin G 5 M Iu Inj. 425 Penicillin V Tab. 125 Mg 426 Penicillin V Syr.125mg/5ml 427 Penicillin V Tab. 250 Mg 543 Tetracycline Cap. 250 Mg 641 Ceftriaxone 250 Mg. Inj. 768 Ceftriaxone 1 Gm. Inj. 783 Roxithromucin 150 Mg. 835 Cephalexin Syr. 843 Amoxicillin250+Clavulanic Acid Tab. 845 Cotrimoxazole Inj. 860 Clindamycin 150 Mg. Cap. 879 Ceftazidime 1 G. Inj. H 113 Amoxycillin Cap. 250 Mg 114 Amoxycillin Cap. 500 Mg 115 Amoxycillin Syr.125mg/5ml 8 Ampicillin 1 Gm Inj. 533 Ampicillin 250 Mg Inj. 118 Ampicillin Inj. 500 Mg 140 Benzathine Penicillin Inj 158 Cefazolin Inj 500 Mg 541 Ceftriaxone 1 G Inj. 540 Ceftriaxone 1 G Inj. 291 Cetrizine 10mg Tab 169 Chloramphenicol Inj. 1 Gm 189 Cloxacillin Sodium 1 Gm 190 Cloxacillin Syr. 25 Mg/Ml 194 Cotrimoxazole Susp. 195 Cotrimoxazole Tab. 305 Dicloxacillin 250 Mg 535 Dicloxacillin 500 Mg

PAGE 169

154 Table D-3. Continued Hospital Drug Code Drug name 242 Doxycycline Cap 100 Mg 256 Erythromycin Tab. 250 Mg 270 Gentamicin Inj. 40 Mg/Ml 322 Lincomycin Inj. 300 Mg/Ml 380 Norfloxacin Tab. 200 Mg 381 Norfloxacin Tab. 400 Mg 397 Penicillin G 5 M Iu Inj. 398 Penicillin V Syr.125mg/5ml 399 Penicillin V Tab. 125 Mg 400 Penicillin V Tab. 250 Mg 494 Tetracycline Cap. 250 Mg

PAGE 170

APPENDIX E EXPERT INTERVIEWS Details of the interviews of hospital directors and database managers of eight community government hospitals in Ubonratchatani province, Thailand are presented in this appendix. However, the data from only five hospitals out of the eight hospitals were included in this study due to incomplete data. Then, the detail of two additional interviews with director of the National Health Security Office and chief of Division of Health Insurance, Provincial Public Health Office, Ubonratchatani province, Thailand are provided. Five Included Hospitals Hospital C 1. Hospital Demographic Information: Date of 30-Baht policy was implemented: June 2001 Major health problems: chronic disease, musculoskeletal systems 2. Hospital Administrative plans in serving 30-Baht policy No specific plan was implemented. 3. Impact of the 30-Baht Policy: Number of patients: slightly increased. Type of patients: N/A Prescribing pattern: depend of physicians and interns. Mostly, physicians prescribed drugs without consideration of the types of health insurance. Drug utilization: Drug utilization rate might increase because the number of patients increased. Other health policies that might have an impact on pattern of care: Hospital Accreditation. Patient Referral policy: severe cases, e.g., stroke, surgery Economic impact: considering efficiency of drug prescribing, e.g., for diabetic patients. 4. Hospital Database: good The database is stable. Hospital D 1. Hospital Demographic Information: 30-bed community hospital Date of 30-Baht policy was implemented: June 2001 Major health problems: DM (E11), Dispepsia (K30), Musculoskeletal system (M791), Common cold (J00), Pharyngitis (J03) 2. Hospital Administrative plans in serving 30-Baht policy Hospital P&T committee and health office committee met regarding the implementation of 30-Baht HI policy to set up plans for patient referral system 155

PAGE 171

156 and drug dispensing. Local health offices are allowed to dispense some medications including some antibiotics (amoxicillin, tetracycline). 3. Impact of the 30-Baht Policy: Number of patients: Significantly increasing number of patients visiting the hospital after the policy implementation Type of patients: N/A Prescribing pattern: Depend on clinical condition, disease severity, and efficiency of drug use. There is no difference in dispensed drugs between patient groups, which might be influenced by limited number of drugs available in hospital formulary. There are some changes in pattern of drug use. An example of the changes in prescribing pattern is the change in the duration of the supply; DM patients used to receive 3-month supply of DM medications after the policy implementation it was changed to 2-month supply. Dr. Vilpong gave the rational to the change that this way MDs and other healthcare professionals could closely monitor DM patients because of the low compliance. Drug utilization: Drug utilization rate is increased because there is increasing number of patients. Other health policies that might have an impact on pattern of care: Hospital Accreditation program is currently the major policy for Sirinthorn hospital. There are several programs initiated to serve this policy and might have impacts on healthcare services and patient outcomes. There are 9 clinical practice guidelines in prevalent conditions/diseases (e.g., asthma). Drug Utilization Management Team monitors number and efficiency of drug use through local health offices affiliated to the hospital. Drug Prescribing review by other healthcare professionals. Death case conference was set up for quality improvement. The hospital provided training to health officers in dispensing antibiotics to the patients. Patient Referral policy: based on patient clinical conditions and hospital health service ability. Costs are the least priority. Economic impact: drug prescribing and dispensing is more efficient, e.g., number of DM drugs is decreased because of the closer monitoring and less supply/visit. There are quite large numbers of patients register with this hospital, which benefit for their financial management. After 30-Baht HI policy was implemented there were questions whether patients could go to other hospitals near their house without registration, will the same drugs dispensed as other patients, and will they receive less supply. 5. Hospital Database: accuracy good to very good. However, at the beginning only important clinical information was entered, e.g., primary diagnosis, patients health insurance, and drug data. When the users are more familiar with the database system, more data were recorded. 6. Data validation team works with database manager to randomly check completeness and correction of the data. The validation is conducted only for IPD data by comparing data in paper-based with computer-based. For OPD, there is only electronic data thus no validation with other sources.

PAGE 172

157 7. There are sets of guideline to assign diagnosis codes, e.g., MD has to give at least 2 digit code, for the case that no clear evidence was found (fever) MD has to make a note and then change to a proper code. Hospital E 1. Hospital Demographic Information 90-bed hospital with 11 physicians Date of 30-Baht policy was implemented: June 2001 Major health problems: chronic diseases, e.g., DM, HTN, upper-respiratory infections The hospital covers patients from several nearby subdistricts. 2. Hospital Administrative plans in serving 30-Baht policy This hospital provided services for the eligible patients from nearby districts because of the better facilities. 3. Impact of the 30-Baht Policy: Number of patients: no effect on hospital admission rates Type of patients: more 30-Baht HI patients Prescribing pattern: there is no different of types of prescribed drugs for patients with different health insurance plan. Prescribing pattern based on hospital drug formulary with cost conscience. Focus on patients needs. Drug utilization: drug utilization rate has been increasing but it is more likely caused by having specialists in the hospital. Drugs that are significantly increased in use are ceftriazone, insulin, antihypertensive drugs. Other health policies that might have an impact on pattern of care: Hospital Accreditation program drives the hospital to set a standard in providing care to all patients at the same quality. Patient Referral policy: stroke cases are referred because there is no appropriate medical equipment (CAT scan), operations that require gas anesthesia. There is no effect from 30-Baht policy on patient referral system. Economic impact: Since there is an increasing in drug utilization rate in the hospital (even though it is not related to 30-Baht policy) hospital has an increasing in economic burden. Patients moved from other hospitals to this hospital because the requirement that if the patient registers with one hospital by geographical area patient must go to that hospital first. These are patients who registered with this hospital but usually used health care services from other hospitals that are easily accessed. 4. Hospital Database Accuracy of the data in the database: moderate. The program had some problems at the beginning in 1999, but quite stable now. Some data, e.g., lab results, disease severity were not complete because of limited numbers of data entry personnel. OPD data is available in electronic format only since 1997. IPD data is available both in MAR and electronic format.

PAGE 173

158 Hospital F 1. Hospital Demographic Information Date of 30-Baht policy was implemented: June 2001 Major health problems: chronic diseases, e.g., DM, HTN, TB, AIDs 2. Hospital Administrative plans in serving 30-Baht policy No specific plans for 30-Baht policy at hospital level 3. Impact of the 30-Baht Policy: Number of patients: increasing in a natural rate Type of patients: DM patients are the primary groups with no significant increasing in number or frequency of hospital visit. Prescribing pattern: there is no different of types of prescribed drugs for patients with different health insurance plan. However, Civil Servant group received more home medication supplies. In other words, civil servant patients will come for hospital visit less often than other groups. The major influence on choice of drugs is restricted hospital drug formulary. Most medications have only one brand each. Drug utilization: drug utilization rate has been increasing but it is more likely caused by having specialists in the hospital. Drugs that are significantly increased in use are ceftriazone, insulin, antihypertensive drugs. Other health policies that might have an impact on pattern of care: Hospital Accreditation program drives the hospital to set a standard in providing care to all patients at the same quality. Patient Referral policy: stroke cases are referred because there is no appropriate medical equipment (CAT scan), operations that require gas anesthesia. There is no effect from 30-Baht policy on patient referral system. Economic impact: Since there is an increasing in drug utilization rate in the hospital (even though it is not related to 30-Baht policy) hospital has an increasing in economic burden. Patients moved from other hospitals to this hospital because the requirement that if the patient registers with one hospital by geographical area patient must go to that hospital first. These are patients who registered with this hospital but usually used health care services from other hospitals that are easily accessed. 4. Hospital Database Accuracy of the data in the database: accurate OPD data is available in electronic format only since 1997. IPD data is available both in MAR and electronic format. Hospital G 1. Hospital Demographic Information 60-bed hospital with surgery care service with 8 physicians Date of 30-Baht policy was implemented: June 2001 Major health problems: chronic diseases, e.g., DM, HTN, TB, AIDs 2. Hospital Administrative plans in serving 30-Baht policy No specific plans for 30-Baht policy at hospital level

PAGE 174

159 3. Impact of the 30-Baht Policy: Number of patients: did not increase, depending on the season. In rainy season, numbers of patients decreased because the road conditions are not good. Only severe case visits the hospital. Type of patients: diabetic and hypertensive patients. The numbers increased because of the disease screening promotion supported by the Ministry of Public Health Prescribing pattern: depends on physicians and types of health insurance The hospital used standard practice guidelines for treatment of diabetes and hypertension. Drug utilization: drug utilization rates might increase from having specialists and patients requested more per visit. Patient Referral policy: other hospitals referred patients to this hospital for surgery. 4. Hospital Database Accuracy of the data in the database: accurate Database system is stable with small user-related errors Hospital H 1. Hospital Demographic Information Date of 30-Baht policy was implemented: June 2001 30-Bed hospital with 5 physicians Major health problems: infectious diseases, diabetes, hypertension, mental health 2. Hospital Administrative plans in serving 30-Baht policy No specific plans for 30-Baht policy at hospital level 3. Impact of the 30-Baht Policy: Number of patients: increasing in a natural rate Type of patients: patient with chronic diseases Prescribing pattern: there was no differentiation on drug prescribing based on health insurance. However, if the patients concerned that 30-Baht group received low quality drug. Drug utilization: drug utilization rate has been increasing naturally. Other health policies that might have an impact on pattern of care: Hospital Accreditation program Patient Referral policy: severe cases, e.g., pneumonia, or surgery cases were referred to the tertiary care hospital. Economic impact: hospital has to better manage the budget of the hospital to cover the 30-Baht patients 4. Hospital Database Accuracy of the data in the database: quite accurate. The hospital has random data validity check with progress notes for the data in 2003 (approximately 95% accuracy) The database system is stable.

PAGE 175

160 Three Excluded Hospitals Hospital I (Khong Jiam) 1. Hospital Demographic Information: 30-bed community hospital Date of 30-Baht policy was implemented: June 2001 Major health problems: Infectious diarrhea, URT infections, intoxication cases, acute illnesses by seasons, e.g., rainy season: cold, musculoskeletal disorders (backache) depending on activities in that area such as farming. 2. Hospital Administrative plans in serving 30-Baht policy No specific plans for 30-Baht policy at hospital level because there is a small number of patients registering with this hospital thus are difficult for financial management. 3. Impact of the 30-Baht Policy: Number of patients: no significant differences in number of hospital visits since 30-Baht policy implementation. Type of patients: N/A Prescribing pattern: there is no different of types of prescribed drugs for patients with different health insurance plan. However, there is a rotation of physicians working at this hospital thus; changing pattern of drug use is most likely caused by physician preference from their experience rather than an effect from the 30-Baht health policy. Drug utilization: Other health policies that might have an impact on pattern of care: Hospital Accreditation program drives the hospital to set a standard in providing care to all patients at the same quality. Patient education programs are also set up for patients who have chronic illness, e.g., DM, HTP. Patient Referral policy: Reimbursement from 30-Baht policy is one of the factors that influence the inter-hospital referral decision. However, it is not a restriction to refer patients who need more advanced care from other hospitals. Stroke cases are referred because there is no appropriate medical equipment (CAT scan), operations that require gas anesthesia. Economic impact: Since there is an increasing in drug utilization rate in the hospital (even though it is not related to 30-Baht policy) hospital has an increasing in economic burden. This hospital is at the border to Laos; patients from Laos also came to receive care from this hospital. They received the same quality care and medication as other patients. From this situation, hospital has to cover these costs. This hospital has 10 health offices under their supervision, thus, drug utilization depends on the rates of drug utilization from these health offices. It appears that there is an increasing in drug use from them. 4. Hospital Database Hospital implemented the HI system in October 2000. Thus, the data are incomplete.

PAGE 176

161 Hospital J (Tansum) 1. Hospital Demographic Information: 30-bed community hospital MD 3, pharmacist (3), nurses (15), dentist (1), and other healthcare professionals Major health problems: Respiratory system, GI, musculoskeletal, infectious diseases, endocrine. Chronic health problems: HTN, DM, thyroid, heart diseases, mental health 2. Hospital Administrative plans in serving 30-Baht policy No specific policy set up 3. Impact of the 30-Baht Policy: Number of patients: increasing number of patients eligible for 30-Baht HI. Type of patients: 30-Baht group Prescribing pattern: There is no difference of dispensed drugs between groups of patients. No clinical treatment guidelines set up in the hospital for prescribing. Hospital has a policy on patient center that consider patients need. Drug utilization: Drug utilization rate is increased because there is increasing number of patients. Other health policies that might have an impact on pattern of care: patient education programs, e.g., illicit drugs, exercise, elderly health, food safety, extended OPD service at the local primary care unit, and collecting family health information Patient Referral policy: based on patient clinical conditions and hospital health service ability regardless of patient health insurance benefit. MD tries to treat as much as the facilities allowed but severe cases and stroke cases are mostly referred. Economic impact: increased debt because there are small numbers of patients registered with this hospitals, which make it difficult to manage the budget. 4. Hospital Database: Electronic data: since Oct 2000 No electronic data for dental care Some ICD 10 codes were assigned with insufficient detail. MD might not know all the available codes. No data validation process Hospital K (Tung Sri Udon) 1. Hospital Demographic Information: 10-bed community hospital Major health problems: Respiratory system, Infectious diarrhea. Chronic health problems: HTN, DM. 2. Hospital Administrative plans in serving 30-Baht policy No specific policy set up 3. Impact of the 30-Baht Policy: Number of patients: did not see much increase in the number of patients eligible for 30-Baht HI. Type of patients: 30-Baht group and the UC with no copayment

PAGE 177

162 Prescribing pattern: There is no difference of dispensed drugs between groups of patients. No clinical treatment guidelines set up in the hospital for prescribing. Drug utilization: Drug utilization rate is the same Other health policies that might have an impact on pattern of care: no other policies Patient Referral policy: severe cases are referred to the tertiary care hospital 4. Hospital Database: Electronic data: since 2001 No permanent database system manager The LAN systems had failed several times during the first year and the second year. Sometimes the computer system could not be used for a few weeks. Paper medical records were used during that time. Data were not updated from the paper medical records after the computer system was No data validation process In summary, increasing number of patients who are eligible for 30-Baht policy was observed after policy implementation. There is no different in pattern of drug prescribed between 30-Baht beneficiaries and other groups. The given explanations are the limited number of drugs in hospital formulary and ethics in medical practices that eliminate variation of practices from economic influences. Most hospitals did not have any specific supportive plan for 30-Baht HI policy. Hospital Accreditation policy has significant impact on every aspects of healthcare services in the hospitals, including quality of healthcare services and drug use. Accuracy and completeness of the data in the electronic database is rated high, however, quantitative validation is needed.

PAGE 178

LIST OF REFERENCES 1. Board on Health Care Service, IOM. Care Without Coverage: Too little, Too late. 1-194. 2002. Washington, D.C., National Academy Press. 2. Piette JD, Heisler M. Problems due to medication costs among VA and non-VA patients with chronic illnesses. Am J Manag Care 2004; 10(11 Pt 2):861-868. 3. Lurk JT, DeJong DJ, Woods TM, Knell ME, Carroll CA. Effects of changes in patient cost sharing and drug sample policies on prescription drug costs and utilization in a safety-net-provider setting. Am J Health Syst Pharm 2004; 61(3):267-272. 4. Poisal JA, Chulis GS. Medicare beneficiaries and drug coverage. Health Aff (Millwood ) 2000; 19(2):248-256. 5. Dong H, Bogg L, Rehnberg C, Diwan V. Association between health insurance and antibiotics prescribing in four counties in rural China. Health Policy 1999; 48(1):29-45. 6. Dong H, Bogg L, Wang K, Rehnberg C, Diwan V. A description of outpatient drug use in rural China: evidence of differences due to insurance coverage. Int J Health Plann Manage 1999; 14(1):41-56. 7. Aday LA, Lee ES, Spears B, Chung CW, Youssef A, Bloom B. Health insurance and utilization of medical care for children with special health care needs. Med Care 1993; 31(11):1013-1026. 8. Na Ranong V. Monitoring and assessing universal health coverage in Chiangmai, Pissanuloke, Nakornratchasrima, Srisaket, Pathumtani, Saraburi and Phuket with description of not using facilities registrated plus overview of care quality: assessment in 21 provinces under pilot project of the UC Scheme in 2001-2002. HSRI000588. 2002. Nontaburi, Thailand, Health Systems Research Institute. 9. Laphying P. Oral health financing in universal coverage system in Thailand. HSRI000453. 2002. Nontaburi, Thailand, Health Systems Research Institute. 10. Tangcharoensathien V. Overview of health insurance systems. HSRI000503. 2001. Nontaburi, Thailand, Health Systems Research Institute. 11. Ministry of Public Health T. Thailand Health Profile 1997-1999. 2004. Bangkok, Thailand, Ministry of Public Health. 163

PAGE 179

164 12. Ratanawijitrasin S, Soumerai SB, Weerasuriya K. Do national medicinal drug policies and essential drug programs improve drug use?: a review of experiences in developing countries. Soc Sci Med 2001; 53(7):831-844. 13. Udomthavornsuk B, Patjanasoontorn B, Bhuripanyo K. Intervention of inappropriate antibiotic use at the university teaching hospital. J Med Assoc Thai 1991; 74(10):429-436. 14. Asawapokee N, Vaithayapichet S, Komoltri C. The failure of a preprinted order form to alter physicians' antimicrobial prescribing pattern. J Med Assoc Thai 1991; 25(4):223-230. 15. Limpanasithikul W, Wangsaturaka D, Nantawan P, Itthipanichapong C, Thamaree S, Wittayalertpanya S et al. Effects of managerial intervention on drug utilization pattern at King Chulalongkorn Memorial Hospital. J Med Assoc Thai 2002; 85 Suppl 1:S336-S343. 16. Swaddiwudhipong W, Nguntra P, Chaovakiratipong C, Koonchote S, Lerdlukanavonge P, Chandoun C. Effect of health education and condom promotion on behavioral change among low socioeconomic prostitutes in Mae Sot, Tak, Thailand. Southeast Asian J Trop Med Public Health 1990; 21(3):453-457. 17. Suwangool P, Moola-Or P, Waiwatana A, Sitthi-Amorn C, Israsena S, Hanvanich M. Effect of a selective restriction policy on antibiotic expenditure and use: an institutional model. J Med Assoc Thai 1991; 74(7):272-275. 18. Thamlikitkul V, Danchaivijitr S, Kongpattanakul S, Ckokloikaew S. Impact of an educational program on antibiotic use in a tertiary care hospital in a developing country. J Clin Epidemiol 1998; 51(9):773-778. 19. Ratanawijitrasin S, Soumerai SB, Weerasuriya K. Do national medicinal drug policies and essential drug programs improve drug use?: a review of experiences in developing countries. Soc Sci Med 2001; 53(7):831-844. 20. Ministry of Public Health T. A survey of health status, health-seeking behavior and relationship between health service provider and health service user in local area. 1979. Bangkok, Thailand, Health Planning Division. 21. Grande Le A, Sri-ngernyuang L, Streefland PH. Enhancing appropriate drug use: the contribution of herbal medicine promotion. Soc Sci Med 1993; 36:1023-1035. 22. Donaldson D, Pannarunothai S, Tangcharoensathien V. Health financing in Thailand. 1999. Nonthaburi, Thailand, Health System Research Institute. Thailand: Health Financing and Management Study Project, ADB#2997 THA. 23. Office of the National Statistics OotPM. Report of the 1999 household socio-economic survey. 20-22. 2000. Bangkok, Thailand, Text and Journal Publication.

PAGE 180

165 24. Osaka R, Nanakorn S. Health care of villagers in northeast Thailand--a health diary study. Kurume Med J 1996; 43(1):49-54. 25. Thamrongwarangkul A, Wibulpolprasert S, Ruamtham N. Improvement of the avialability and rational use of essential drugs through village grocery stores in rural Thailand. 2000. Nonthaburi, Thailand, Ministry of Public Health. 26. Ansari F, Gray K, Nathwani D, Phillips G, Ogston S, Ramsay C et al. Outcomes of an intervention to improve hospital antibiotic prescribing: interrupted time series with segmented regression analysis. J Antimicrob Chemother 2003; 52(5):842-848. 27. Schneeweiss S, Maclure M, Soumerai SB. Prescription duration after drug copay changes in older people: methodological aspects. J Am Geriatr Soc 2002; 50(3):521-525. 28. Ratanawijitrasin S, Soumerai SB, Weerasuriya K. Do national medicinal drug policies and essential drug programs improve drug use?: a review of experiences in developing countries. Soc Sci Med 2001; 53(7):831-844. 29. Schneeweiss S, Maclure M, Carleton B, Glynn RJ, Avorn J. Clinical and economic consequences of a reimbursement restriction of nebulised respiratory therapy in adults: direct comparison of randomised and observational evaluations. BMJ 2004; 328(7439):560. 30. Vander Stichele RH, Peys F, Van Tielen R, Van Eeckhout H, van Essche O, Seys B. A decade of growth in public and private pharmaceutical expenditures: the case of Belgium 1990-1999. Acta Clin Belg 2003; 58(5):279-289. 31. Perez A, Dennis RJ, Rodriguez B, Castro AY, Delgado V, Lozano JM et al. An interrupted time series analysis of parenteral antibiotic use in Colombia. J Clin Epidemiol 2003; 56(10):1013-1020. 32. Gillings D, Makuc D, Siegel E. Analysis of interrupted time series mortality trends: an example to evaluate regionalized perinatal care. Am J Public Health 1981; 71(1):38-46. 33. Martin BC, Miller LS, Kotzan JA. Antipsychotic prescription use and costs for persons with schizophrenia in the 1990s: current trends and five year time series forecasts. Schizophr Res 2001; 47(2-3):281-292. 34. Lopez-Lozano JM, Monnet DL, Yague A, Burgos A, Gonzalo N, Campillos P et al. Modelling and forecasting antimicrobial resistance and its dynamic relationship to antimicrobial use: a time series analysis. Int J Antimicrob Agents 2000; 14(1):21-31. 35. Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regression analysis of interrupted time series studies in medication use research. Journal of Clinical Pharmacy and Therapeutics 2002;(27):299-309.

PAGE 181

166 36. Granger CWJ, Newbold P. Forecasting Econometric Time Series. New York: Academic Press, 1986. 37. Yaffee R, McGee M. Introduction to Time Series Analysis and Forecasting with Applications of SAS and SPSS. New York: Academic Press, 2000. 38. Soumerai SB, Ross-Degnan D, Fortess EE, Abelson J. A critical analysis of studies of state drug reimbursement policies: research in need of discipline. Milbank Q 1993; 71(2):217-252. 39. Kozyrskyj AL, Mustard CA, Cheang MS, Simons FE. Income-based drug benefit policy: impact on receipt of inhaled corticosteroid prescriptions by Manitoba children with asthma. CMAJ 2001; 165(7):897-902. 40. Blais L, Couture J, Rahme E, LeLorier J. Impact of a cost sharing drug insurance plan on drug utilization among individuals receiving social assistance. Health Policy 2003; 64(2):163-172. 41. Ross-Degnan D, Simoni-Wastila L, Brown JS, Gao X, Mah C, Cosler LE et al. A controlled study of the effects of state surveillance on indicators of problematic and non-problematic benzodiazepine use in a Medicaid population. Int J Psychiatry Med 2004; 34(2):103-123. 42. Wagner AK, Soumerai SB, Zhang F, Mah C, Simoni-Wastila L, Cosler L et al. Effects of state surveillance on new post-hospitalization benzodiazepine use. Int J Qual Health Care 2003; 15(5):423-431. 43. Mukamal KJ, Markson LJ, Flier SR, Calabrese D. Restocking the sample closet: results of a trial to alter medication prescribing. J Am Board Fam Pract 2002; 15(4):285-289. 44. Tu K, Mamdani MM, Jacka RM, Forde NJ, Rothwell DM, Tu JV. The striking effect of the Heart Outcomes Prevention Evaluation (HOPE) on ramipril prescribing in Ontario. CMAJ 2003; 168(5):553-557. 45. Pharmacoepidemiology. 3 ed. New York: Jonh Wiley & Sons, Ltd., 2000. 46. Lunde PK, Baksaas I. Epidemiology of drug utilization--basic concepts and methodology. Acta Med Scand Suppl 1988; 721:7-11. 47. Martin BC, Miller LS, Kotzan JA. Antipsychotic prescription use and costs for persons with schizophrenia in the 1990s: current trends and five year time series forecasts. Schizophr Res 2001; 47(2-3):281-292. 48. WHO Collaboring Center for Drug Statistics Methodology. About the ATC/DDD system. the World Health Organization 2004.

PAGE 182

167 49. Dolezal T, Nemecek K, Krsiak M. Trends in consumption of calcium channel blockers in the Czech Republic during 1993-1999 and comparison with selected countries: drug utilization study. Eur J Clin Pharmacol 2002; 58(7):477-478. 50. Santamaria B, Perez M, Montero D, Madurga M, de Abajo FJ. Use of antipsychotic agents in Spain through 1985-2000. Eur Psychiatry 2002; 17(8):471-476. 51. Damiani G, Federico B, Oradei M, Bellocco R, Vanini GC. Impact of drug policy on the use of parenteral cephalosporins in Italy. Pharmacy World & Science 2002; 24(5):169-171. 52. Board on Health Care Service IoM. Statement on quality of care. Institute of Medicine, 1-26. 1-1-1998. 53. Gaist D, Hallas J, Sindrup SH, Gram LF. Is overuse of sumatriptan a problem? Eur J Clin Pharmacol 1996; 50:161-165. 54. Olsson J, Lindberg G, Gottsater M, Lindwall K, Sjostrand A, Tisell A et al. Differences in pharmacotherapy and in glucose control of type 2 diabetes patients in two neighbouring towns: a longitudinal population-based study. Diabetes, Obesity and Metabolism 2001; 3:249-253. 55. Natsch S, Hekster YA, de Jong R, Heerdink ER, Herings RM, van der Meer JW. Application of the ATC/DDD methodology to monitor antibiotic drug use. Eur J Clin Microbiol Infect Dis 1998; 17(1):20-24. 56. Capella D. Descriptive tools and analysis. Dukes MNG. 1993. Copenhagen, WHO Regional Office for Europe. Methods and uses. 57. Ried LD, Johnson RE. Evaluation of the reliability and validity of a measure of anxiolytic drug-use intensity for pharmacoepidemiologic studies. Ann Pharmacother 1992; 26(11):1441-1446. 58. Resi D, Castelvetri C, Vaccheri A, Montanaro N. The therapeutic course as a measure complementary to defined daily doses when studying exposure to antibacterial agents. Eur J Clin Pharmacol 2001; 57(2):177-180. 59. Christiansen T, Pederson K, Harvald B, Rasmussen K, Jorgensen J, Svader C. A health economic investigation of the effect of treatment for hypertension. 1987. Odense, Odense University Department of Health Economics and Public Health. Research Report No. 6/1987. 60. Linden M, Gothe H. Benzodiazepine substitution in medical practice. Analysis of pharmacoepidemiologic data based on expert interviews. Pharmacopsychiatry 1993; 26(4):107-113.

PAGE 183

168 61. Linden M, Gothe H. Benzodiazepines in medical teaching. A survey of university departments of pharmacology, psychiatry, and internal medicine in Germany. Pharmacopsychiatry 1993; 26(1):6-10. 62. Donabedian A. The quality of medical care. Science 1978; 200(4344):856-864. 63. Donabedian A. The seven pillars of quality. Arch Pathol Lab med 1990; 114:1115-1118. 64. Donabedian A. The quality of care. How can it be assessed? JAMA 1988; 260(12):1743-1748. 65. Measuring access and trend. In: Williams SJ and Torrens PR, editor. Introduction to health services. Albany, NY: Delmar Publishers, 1993: 91-93. 66. Andersen R, Aday LA. Access to medical care in the US: realized and potential. Med Care 1978; 16(7):533-546. 67. Andersen RM, McCutcheon LA, Aday LA, Chiu GY, Bell R. Exploring dimensions of access to medical care. Health Serv Res 1983; 18(1):49-74. 68. IOM. Access to Health Care in America. 1993. Washington DC, National Academy Press. 69. Global Health Equity Initiative. The equity gauge: an approach to monitoring equity in health and health care in develoing countries. 2003. 70. Nelms RC. Providing univeral health care coverage. Minnesota Medicine 1994; 77:51. 71. Stolar M. Drug use review: Operational definitions. Am J Hosp Pharm 1978; 33:225-230. 72. Soumerai SB, Lipton HL. Computer-based drug-utilization review--risk, benefit, or boondoggle? N Engl J Med 1995; 332(24):1641-1645. 73. Schmader K, Hanlon JT, Weinberger M, Landsman PB, Samsa GP, Lewis I et al. Appropriateness of medication prescribing in ambulatory elderly patients. J Am Geriatr Soc 1994; 42(12):1241-1247. 74. Hanlon JT, Schmader KE, Samsa GP, Weinberger M, Uttech KM, Lewis IK et al. A method for assessing drug therapy appropriateness. J Clin Epidemiol 1992; 45(10):1045-1051. 75. Garrard J, Chen V, Dowd B. The impact of the 1987 federal regulations on the use of psychotropic drugs in Minnesota nursing homes. Am J Public Health 1995; 85(6):771-776.

PAGE 184

169 76. Shorr RI, Fought RL, Ray WA. Changes in antipsychotic drug use in nursing homes during implementation of the OBRA-87 regulations. JAMA 1994; 271(5):358-362. 77. Lantz MS, Giambanco V, Buchalter EN. A ten-year review of the effect of OBRA-87 on psychotropic prescribing practices in an academic nursing home. Psychiatr Serv 1996; 47(9):951-955. 78. van Eijk ME, Krist LF, Avorn J, Porsius A, de Boer A. Do the research goal and databases match? A checklist for a systematic approach. Health Policy 2001; 58(3):263-274. 79. Talerico KA. A critique of research measures used to assess inappropriate psychoactive drug use in older adults. J Am Geriatr Soc 2002; 50(2):374-377. 80. Owen RR, Thrush CR, Hudson TJ, Mallory SR, Fischer EP, Clardy JA et al. Using an explicit guideline-based criterion and implicit review to assess antipsychotic dosing performance for schizophrenia. Int J Qual Health Care 2002; 14(3):199-206. 81. Gurwitz JH, Soumerai SB, Avorn J. Improving medication prescribing and utilization in the nursing home. J Am Geriatr Soc 1990; 38(5):542-552. 82. Chassin MR. Quality of health care. Part 3: improving the quality of care. N Engl J Med 1996; 335(14):1060-1063. 83. Agency for Healthcare Research and Quality. Achive: Healthcare Costs and Utilization Project Quality indicators (HCUP QIs). Agency for Healthcare Research and Quality 2003. 84. Agency for Healthcare Research and Quality. AHRQ Quality Indicators. http://www.qualityindicators.ahrq.gov/hcup_archive.htm#1 2001. 85. Gerstman BB, Lundin FE, Stadel BV, Faich GA. A method of pharmacoepidemiologic analysis that uses computerized Medicaid. J Clin Epidemiol 1990; 43(12):1387-1393. 86. Birkholz G, Viens DC. Medicaid claims data comparisons for nurse practitioners, physician assistants, and primary care physicians in New Mexico. J Am Acad Nurse Pract 1999; 11(1):3-10. 87. Briesacher BA, Stuart B, Peluso R. Drug use and prescribing problems in the community-dwelling elderly: a study of three state Medicaid programs. Clin Ther 1999; 21(12):2156-2172. 88. Bright RA, Avorn J, Everitt DE. Medicaid data as a resource for epidemiologic studies: strengths and limitations. J Clin Epidemiol 1989; 42(10):937-945.

PAGE 185

170 89. Buescher PA, Jones-Vessey K. Using Medicaid data to estimate stateand county-level prevalence of asthma among low-income children. Matern Child Health J 1999; 3(4):211-216. 90. Coleman CI, Reddy P, Laster-Bradley NM, Dorval S, Munagala B, White CM. Effect of practitioner education on adherence to asthma treatment guidelines. Ann Pharmacother 2003; 37(7-8):956-961. 91. Cosler LE, Lambrinos J. Zidovudine's impact on resource use by patients with symptomatic HIV illness: a large sample analysis. Inquiry 1992; 29(3):345-355. 92. Dubois RW, Chawla AJ, Neslusan CA, Smith MW, Wade S. Explaining drug spending trends: does perception match reality? Health Aff (Millwood ) 2000; 19(2):231-239. 93. Every NR, Frederick PD, Robinson M, Sugarman J, Bowlby L, Barron HV. A comparison of the national registry of myocardial infarction 2 with the cooperative cardiovascular project. J Am Coll Cardiol 1999; 33(7):1886-1894. 94. Fleming C, Fisher ES, Chang CH, Bubolz TA, Malenka DJ. Studying outcomes and hospital utilization in the elderly. The advantages of a merged data base for Medicare and Veterans Affairs hospitals. Med Care 1992; 30(5):377-391. 95. Fortess EE, Soumerai SB, McLaughlin TJ, Ross-Degnan D. Utilization of essential medications by vulnerable older people after a drug benefit cap: importance of mental disorders, chronic pain, and practice setting. J Am Geriatr Soc 2001; 49(6):793-797. 96. Giaquinta D. Health care resource evaluation. Manag Care Interface 2001; 14(1):47-48. 97. Griffin MR, Ray WA, Schaffner W. Nonsteroidal anti-inflammatory drug use and death from peptic ulcer in elderly persons. Ann Intern Med 1988; 109(5):359-363. 98. Hawes C, Morris JN, Phillips CD, Mor V, Fries BE, Nonemaker S. Reliability estimates for the Minimum Data Set for nursing home resident assessment and care screening (MDS). Gerontologist 1995; 35(2):172-178. 99. Hennessy S, Bilker WB, Weber A. Descriptive analyses of the integrity of a US Medicaid claims database. Pharmacoepidemiology and Drug Safety 2003; 12:103-111. 100. Irwin RG. How OASIS data quality affects you! Home Healthc Nurse 1999; 17(9):558-563. 101. Joish VN, Cady PS. Effect of sumatriptan on health care resource use among patients with migraine. Manag Care Interface 2001; 14(1):68-72.

PAGE 186

171 102. Kozma CM, Reeder CE, Lingle EW. Expanding Medicaid drug formulary coverage. Effects on utilization of related services. Med Care 1990; 28(10):963-977. 103. Lipton HL, Byrns PJ, Soumerai SB, Chrischilles EA. Pharmacists as agents of change for rational drug therapy. Int J Technol Assess Health Care 1995; 11(3):485-508. 104. Martin BC, McMillan JA. The impact of implementing a more restrictive prescription limit on Medicaid recipients. Effects on cost, therapy, and out-of-pocket expenditures. Med Care 1996; 34(7):686-701. 105. McCombs JS, Shi L, Stimmel GL, Croghan TW. A retrospective analysis of the revocation of prior authorization restrictions and the use of antidepressant medications for treating major depressive disorder. Clin Ther 2002; 24(11):1939-1959. 106. McCombs JS, Nichol MB, Stimmel GL, Shi J, Smith RR. Use patterns for antipsychotic medications in medicaid patients with schizophrenia. J Clin Psychiatry 1999; 60 Suppl 19:5-11. 107. McKenzie DA, Semradek J, McFarland BH, Mullooly JP, McCamant LE. The validity of medicaid pharmacy claims for estimating drug use among elderly nursing home residents: The Oregon experience. J Clin Epidemiol 2000; 53(12):1248-1257. 108. Nash DB, Goldstein P, Kron S. Quality assurance and utilization review in a health insuring organization. The HealthPASS example. Qual Assur Util Rev 1989; 4(2):32-38. 109. Nelson AA, Jr., Reeder CE, Dickson WM. The effect of a Medicaid drug copayment program on the utilization and cost of prescription services. Med Care 1984; 22(8):724-736. 110. Pai CW, Ozcan YA, Jiang HJ. Regional variation in physician practice pattern: an examination of technical and cost efficiency for treating sinusitis. J Med Syst 2000; 24(2):103-117. 111. Rascati KL, Okano GJ, Burch C. Evaluation of physician intervention letters. Med Care 1996; 34(8):760-766. 112. Ray WA. Policy and program analysis using administrative databases. Ann Intern Med 1997; 127(8 Pt 2):712-718. 113. Ray WA, Griffin MR. Use of Medicaid data for pharmacoepidemiology. Am J Epidemiol 1989; 129(4):837-849.

PAGE 187

172 114. Rushton JL, Whitmire JT. Pediatric stimulant and selective serotonin reuptake inhibitor prescription trends: 1992 to 1998. Arch Pediatr Adolesc Med 2001; 155(5):560-565. 115. Sambamoorthi U, Moynihan PJ, McSpiritt E, Crystal S. Use of protease inhibitors and non-nucleoside reverse transcriptase inhibitors among Medicaid beneficiaries with AIDS. Am J Public Health 2001; 91(9):1474-1481. 116. Sambamoorthi U, Collins S, Crystal S. Dually eligible individuals with AIDS: characteristics and health services use. J Health Soc Policy 2001; 14(1):19-35. 117. Schorr B. How HCFA will track your script. Physicians Manage 1988; 28(11):140, 142, 155. 118. Sclar DA, Robison LM, Skaer TL, Dickson WM, Kozma CM, Reeder CE. Sulfonylurea pharmacotherapy regimen adherence in a Medicaid population: influence of age, gender, and race. Diabetes Educ 1999; 25(4):531-538. 119. Sena MM, Pashko S. Drug utilization review using a Medicaid claims database. Clin Ther 1993; 15(5):900-904. 120. Soumerai SB, Avorn J, Ross-Degnan D, Gortmaker S. Payment restrictions for prescription drugs under Medicaid. Effects on therapy, cost, and equity. N Engl J Med 1987; 317(9):550-556. 121. Soumerai SB, Ross-Degnan D, Gortmaker S, Avorn J. Withdrawing payment for nonscientific drug therapy. Intended and unexpected effects of a large-scale natural experiment. JAMA 1990; 263(6):831-839. 122. Stuart B, Briesacher BA, Ahern F, Kidder D, Zacker C, Erwin G et al. Drug use and prescribing problems in four state Medicaid programs. Health Care Financ Rev 1999; 20(3):63-78. 123. Stuart B, Lago D. Prescription drug coverage and medical indigence among the elderly. J Aging Health 1989; 1(4):451-469. 124. Swan JH, Benjamin AE. IV drug use, dementia, and nursing home care for PWAs (persons with AIDS). J Health Soc Policy 1993; 4(3):79-91. 125. Thamer M, Richard C, Ray NF, Greer JW, Cotter DJ, Pearson BC. The effect of insurance status on use of recombinant erythropoietin therapy among end-stage renal disease patients in three states. Am J Kidney Dis 1996; 28(2):235-249. 126. Xiao H, Sorofman B, Manasse HR, Jr. The association between pharmacy closures and prescription drug use: a retrospective analysis of Medicaid prescription claims in Iowa. Am J Manag Care 2000; 6(3):366-372.

PAGE 188

173 127. Zechnich AD, Greenlick M, Haxby D, Mullooly J. Elimination of over-the-counter medication coverage in the Oregon Medicaid population: the impact on program costs and drug use. Med Care 1998; 36(8):1283-1294. 128. Zulkowski K. MDS+ RAP items associated with pressure ulcer prevalence in newly institutionalized elderly: study I. Ostomy Wound Manage 1998; 44(11):40-8, 50, passim. 129. Newton KM, Wagner EH, Ramsey SD, McCulloch D, Evans R, Sandhu N et al. The use of automated data to identify complications and comorbidities of diabetes: a validation study. J Clin Epidemiol 1999; 52(3):199-207. 130. Brownell M, Mayer T, Martens PJ, Kozyrskyj A, Fergusson P, Bodnarchuk J et al. Using a population-based health information system to study child health. Can J Public Health 2002; 93 Suppl 2:S9-14. 131. Kozyrskyj AL, Mustard CA, Simons FE. Inhaled corticosteroids in childhood asthma: Income differences in use. Pediatr Pulmonol 2003; 36(3):241-247. 132. Kozyrskyj AL, Mustard CA, Becker AB. Childhood wheezing syndromes and healthcare data. Pediatr Pulmonol 2003; 36(2):131-136. 133. Kozyrskyj AL, Fergusson P, Bodnarchuk J, Brownell M, Burchill C, Mayer T. Community resources and determinants of the future health of Manitobans. Can J Public Health 2002; 93 Suppl 2:S70-S76. 134. Kozyrskyj AL. Prescription medications in Manitoba children: are there regional differences? Can J Public Health 2002; 93 Suppl 2:S63-S69. 135. Kozyrskyj AL, Hildes-Ripstein GE. Assessing health status in Manitoba children: acute and chronic conditions. Can J Public Health 2002; 93 Suppl 2:S44-S49. 136. Kozyrskyj AL, Mustard CA, Simons FE. Socioeconomic status, drug insurance benefits, and new prescriptions for inhaled corticosteroids in schoolchildren with asthma. Arch Pediatr Adolesc Med 2001; 155(11):1219-1224. 137. Kozyrskyj AL, Mustard CA, Cheang MS, Simons FE. Income-based drug benefit policy: impact on receipt of inhaled corticosteroid prescriptions by Manitoba children with asthma. CMAJ 2001; 165(7):897-902. 138. Kozyrskyj AL, Mustard CA. Validation of an electronic, population-based prescription database. Ann Pharmacother 1998; 32(11):1152-1157. 139. Martens PJ, Brownel MD, Kozyrskyj A. The virtual classroom: a summary of child health indicators. Can J Public Health 2002; 93 Suppl 2:S77-S80. 140. Metge C, Black C, Peterson S, Kozyrskyj AL. The population's use of pharmaceuticals. Med Care 1999; 37(6 Suppl):JS42-JS59.

PAGE 189

174 141. Mustard CA, Kaufert P, Kozyrskyj A, Mayer T. Sex differences in the use of health care services. N Engl J Med 1998; 338(23):1678-1683. 142. Nair C. Trends in cesarean section deliveries in Canada. Health Rep 1991; 3(3):203-219. 143. Cooper GS, Yuan Z, Stange KC, Amini SB, Dennis LK, Rimm AA. The utility of Medicare claims data for measuring cancer stage. Med Care 1999; 37(7):706-711. 144. Cox ER, Jernigan C, Coons SJ, Draugalis JL. Medicare beneficiaries' management of capped prescription benefits. Med Care 2001; 39(3):296-301. 145. Fisher ES, Baron JA, Malenka DJ, Barrett JA, Kniffin WD, Whaley FS et al. Hip fracture incidence and mortality in New England. Epidemiology 1991; 2(2):116-122. 146. Jencks SF, Cuerdon T, Burwen DR. Qaulity of medical care delivered to Medicare beneficiaries. JAMA 2000; 284(13):1670-1676. 147. Johnson RE, Goodman MJ, Hornbrook MC, Eldredge MB. The impact of increasing patient prescription drug cost sharing on therapeutic classes of drugs received and on the health status of elderly HMO members. Health Serv Res 1997; 32(1):103-122. 148. Johnson RE, Goodman MJ, Hornbrook MC, Eldredge MB. The effect of increased prescription drug cost-sharing on medical care utilization and expenses of elderly health maintenance organization members. Med Care 1997; 35(11):1119-1131. 149. West SL, Strom BL, Poole C. Validity of pharmacoepidemiology drug and diagnosis data. In: Strom BL, editor. Pharmacoepidemiology. New York, NY: John Wiley & Sons, Ltd., 2000: 661-705. 150. Iezzoni LI. Using administrative diagnostic data to assess the quality of hospital care. Pitfalls and potential of ICD-9-CM. Int J Technol Assess Health Care 1990; 6(2):272-281. 151. Movig KL, Leufkens HG, Lenderink AW, Egberts AC. Validity of hospital discharge International Classification of Diseases (ICD) codes for identifying patients with hyponatremia. J Clin Epidemiol 2003; 56(6):530-535. 152. Thomas SK, Brooks SE, Mullins CD, Baquet CR, Merchant S. Use of ICD-9 coding as a proxy for stage of disease in lung cancer. Pharmacoepidemiol Drug Saf 2002; 11(8):709-713. 153. Hsia DC, Krushat WM, Fagan AB, Tebbutt JA, Kusserow RP. Accuracy of diagnostic coding for Medicare patients under the prospective-payment system. N Engl J Med 1988; 318(6):352-355.

PAGE 190

175 154. Hsia DC, Ahern CA, Ritchie BP, Moscoe LM, Krushat WM. Medicare reimbursement accuracy under the prospective payment system. JAMA 1992; 268:896-899. 155. Fillit H, Geldmacher DS, Welter RT, Maslow K, Fraser M. Optimizing coding and reimbursement to improve management of Alzheimer's disease and related dementias. J Am Geriatr Soc 2002; 50(11):1871-1878. 156. Iezzoni LI, Foley SM, Daley J, Hughes J, Fisher ES, Heeren T. Comorbidities, complications, and coding bias. Does the number of diagnosis codes matter in prediting in-hosptal mortality? JAMA 1992; 267:2197-2203. 157. Romano PS, Mark DH. Bias in the coding of hospital discharge data and its implications for quality assessment. Med Care 1994; 32(81):90. 158. International classification of diseases. 9th Rev. Clinical Modification. New York: McGraw Hill, 1994. 159. Wang PS, Walker AM, Tsuang MT. Finding incident breast cancer cases through US claims data and a state cancer registry. Cancer Cases and Control 2001; 12:257-265. 160. Ray WA, Griffin MR, Downey W, Melton LJ, III. Long-term use of thiazide diuretics and risk of hip fracture. Lancet 1989; 1(8640):687-690. 161. Ray WA, Griffin MR, Schaffner W, Baugh DK, Melton LJ, III. Psychotropic drug use and the risk of hip fracture. N Engl J Med 1987; 316(7):363-369. 162. Lichtenstein MJ, Griffin MR, Cornell JE, Malcolm E, Ray WA. Risk factors for hip fractures occurring in the hospital. Am J Epidemiol 1994; 140(9):830-838. 163. Gerstman BB, Freiman JP, Hine LK. Use of subsequent anticoagulants to increase the predictive value of Medicaid deep venous thromboembolism diagnoses. Epidemiology 1990; 1(2):122-127. 164. Carson JL, Strom BL. Medicaid Databases. In: Strom BL, editor. Pharmacoepidemiology. New York, NY: John Wiley & Sons, Ltd., 2000: 307-324. 165. Ministry of Public Health TDoDC. Annual Epidemiological Surveillance Database 2002. http://epid.moph.go.th/ 2004. 166. Gold Standard. Clinical Pharmacology. Gold Standard. 2005. 2-10-2005. 167. International Statistical Classification of Diseases and Related Health Problems, version10. World Health Organization. 2003. 168. National Health Security Office. Health Insurance Registry and Statistics. National Health Security Office, Thailand. 4-20-2005. 1-20-2005.

PAGE 191

176 169. Mandell LA, Bartlett JG, Dowell SF, File TM, Jr., Musher DM, Whitney C. Update of practice guidelines for the management of community-acquired pneumonia in immunocompetent adults. Clin Infect Dis 2003; 37(11):1405-1433. 170. Clinical Pharmacology. Clinical Pharmacology: Drug Information Monograph. http://cpip.gsm.com/ 2004. 171. SAS Version 8. SAS Institute. 2001. 172. Centers for Disease Control and Prevention. Racial/Ethnic Disparities in Prevalence, Treatment, and Control of Hypertension --United States, 1999--2002. Morbidity and Mortality Weekly Report 7-9-2005. 173. Agency of Healthcare Research and Quality. Healthcare Cost and Utilization Project. Agency of Healthcare Research and Quality. 2005. 174. Pittrow D, Kirch W, Bramlage P, Lehnert H, Hofler M, Unger T et al. Patterns of antihypertensive drug utilization in primary care. Eur J Clin Pharmacol 2004; 60(2):135-142. 175. National Center for Chronic Disease Prevention and Health Promotion. National Diabetes Surveillance System. Centers for Disease Control and Prevention. 7-9-2005. 176. Boccuzzi SJ, Wogen J, Fox J, Sung JC, Shah AB, Kim J. Utilization of oral hypoglycemic agents in a drug-insured U.S. population. Diabetes Care 2001; 24(8):1411-1415. 177. Reeder CE, Nelson AA. The differential impact of copayment on drug use in a Medicaid population. Inquiry 1985; 22(4):396-403. 178. Assaf AR, Lapane KL, McKenny JL. Possible influence of the prospective payment system on the assignment of discharge diagnoses for coronary heart disease. N Engl J Med 1993; 329(13):931-935.

PAGE 192

BIOGRAPHICAL SKETCH Penkarn Kanjanarat received a Bachelor of Pharmacy degree from Chiangmai University, Thailand in 1992. She continued her education with a masters degree in Hospital Pharmacy from Mahidol University, Thailand in 1994. Her thesis title was "The effect of generic drug name use in hospitals on pharmacy inventory management." During her masters study, she worked as a part-time hospital pharmacist at two private hospitals. Penkarn served as a faculty in the Department of Pharmaceutical Care, College of Pharmacy, Chiangmai University for 4 years. During that time, she taught pharmacy management, pharmacy law and ethics, and a hospital pharmacy course. She gained experience in practicing pharmacy and managing a community pharmacy while working at her own community pharmacy for 2 years before she began her doctoral program in Pharmacy Health Care Administration at the University of Florida, College of Pharmacy. 177


Permanent Link: http://ufdc.ufl.edu/UFE0011544/00001

Material Information

Title: Impact of the 30-Baht Health Insurance Policy on Hospital Drug Utilization in Thailand
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0011544:00001

Permanent Link: http://ufdc.ufl.edu/UFE0011544/00001

Material Information

Title: Impact of the 30-Baht Health Insurance Policy on Hospital Drug Utilization in Thailand
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0011544:00001


This item has the following downloads:


Full Text












IMPACT OF THE 30-BAHT HEALTH INSURANCE POLICY ON HOSPITAL
DRUG UTILIZATION IN THAILAND















By

PENKARN KANJANARAT


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Penkarn Kanjanarat

































To my father, mother, brother, sister, niece, and my husband for their great love and
selfless support















ACKNOWLEDGMENTS

I would like to express my gratitude to my major advisor, Dr. Almut Winterstein,

for her guidance, encouragement, support, and friendship during my pursuit of the Ph.D.

This study would not have been completed without her guidance and dedication.

I would like to thank Dr. Earlene Lipowski who provided a broad perspective of

healthcare systems and guided discussions about the applicability of this study. Dr.

Lipowski also provided input from her working experiences in Thailand to refine my

study. Her encouragement and respect are greatly appreciated. I would like to thank Dr.

Abraham Hartzema, who shaped my research idea. His vision in pharmacoepidemiology

and international research experiences also helped guide my research. I deeply

appreciate his challenging questions. I greatly appreciate Dr. Lili Tian for her

suggestions for the analysis of my study. With her help, I have greatly extended my

knowledge in Time Series Analysis in health service research, which will greatly benefit

me in my future research.

I could not express my gratitude enough to the Department of Pharmacy Health

Care Administration and the College of Pharmacy for their support. I appreciate the

infinite support of Dr. Richard Segal, Dr. Carole Kimberlin, and Dr. Donna Berardo. I

would like to express my gratitude to the College of Pharmacy, Chiang Mai University,

for their support of my Ph.D. study.









I would like to extend my gratitude to Dr. Jeffery Crane for his helps,

encouragement, and friendship. I could not have achieved this goal without his indefinite

support.

This study is funded by a P.A. Foote Small Research Grant from the Perry A. Foote

Health Outcomes and Pharmacoeconomics, University of Florida.
















TABLE OF CONTENTS

Page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES .................................................. .................. ..... ....

L IST O F FIG U R E S .... ...... ...................... ........................ .. ....... .............. xii

ABSTRACT .............. .......................................... xiv

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

B a c k g ro u n d ................................................................................................... 1
N eed for Study .................................................................. ............................. . 2
Purposes of Study ................................................................ 4
Study Objectives.............................. ........5
R research Q uestions.............................................. 5
S ig n ifican ce ....................................................... 6

2 REVIEW OF LITERATURE .............................................................. .............. 7

Healthcare Care System in Thailand ........................................................
Health-Seeking Behavior and Healthcare Utilization ........................................7
A cc e ss to D ru g s ................................................................................. 8
H ealthcare Financing ................ ............................................. ............... 9
The civil servant medical benefit scheme ..................................... 10
The 30-Baht health insurance policy .....................................................11
Evaluating H health Policy ...................................... ........ ................................ 12
Methodology in Evaluating Health Policy ................. ................. ..........13
Time series in observational design ......................................................14
N onconcurrent tim e series design ........................................ ............... 15
Concurrent tim e series design ........................................ ......... ......16
Quantitative and Qualitative Measurement of Drug Use ............................20
Quantitative Approach to Measuring Drug Utilization ..................... ..........21
Measurement Units of Drug Utilization ............ ............................ ..... 22
The defined daily dose (DDD) ............................................... 22
Strengths of the D D D ........................................................ ............... 23
Lim station of the D D D ....................................................... .... ........... 24









Defined daily dose in drug utilization research.......................................24
Alternatives measurements to the DDD ........... .....................................26
Qualitative Approaches in Measuring Drug Utilization............. ............... 29
Concept of quality of care and its measurements........................................29
Approaches in measuring appropriateness of drug use.............................33
Measuring quality of drug utilization using explicit criteria........................34
Measuring quality of drug use by quality indicators (QIs) ........................35
Challenges of Measuring Quality of Drug Use ................................................ 38
V alidity of the m easures..................................... ......................... .. ......... 38
Reliability of the m easures................................................. ............... 39
Sensitivity and specificity ................................................. ............... 39
Validating Computerized Administrative Databases.......................................39
D ata V alidation M ethods.......................................................... ............... 40
External data validation............. .............................. 41
Internal validation m ethods ....................................................................... 46

3 VALIDATING DATA QUALITY............................... ......................... 49

Interview of H hospital D directors ........................................................ ............... 52
Interview s of Hospital Database M managers ..............................................................52
Quantitative Assessment of Data Quality .............. ...........................................52
Database Characteristics.................... ......... ............................. 53
M missing D ata and O utliers .................................... ........................ ............... 53
Face Validity (Plausibility of the Data)............... ............................................ 54
Data Coherence ............................................ .................54
Validating disease diagnosis codes (ICD-10) ........................... ........56
D ru g d ata .............................. ......... .................................. .............. 5 7

4 DATA VALIDATION RESULTS ........................................ ......................... 58

H hospitals D em graphics ......... ................. ............................................................ 58
D database C characteristics ......... ................. .................................... ............... 61
F a c e V a lid ity ......................................................................................................... 6 3
M missing D ata ................................................................................................66
Data Coherence............................................ 69
Disease Diagnosis and Gender ...... ..................... ......... 69
Disease Diagnosis and Drugs .................................. ............... 69
Diabetes and antidiabetic drugs................ ....... ................................. 69
Hypertension and antihypertensive drugs ..................................... 72
Bacterial pneumonia and antibiotics ................. ................. ............74
Drugs and Disease Diagnosis .................................. ............... 76
Antidiabetic drugs and diabetes ........................... ....... ............... 76
Results from the Expert Interviews ................................ ............... 77

5 M E T H O D S ........................................................................................................... 8 0

H y p o th e se s ............................................................................................................ 8 1









Patient Selection ................................... .. .. ........ .. ............85
Data Source.............................................. 87
M e a su re s ............................................................................................................... 8 7
D rug U utilization R ate (D R ) .................................................................................88
Percentage of Appropriate Prescribing (Drug Use Quality, DQ) ........................89
H hospital V isit R ates (H R )......................................................... ............... 90
H hospital A dm mission Rates (AR) ........................................ ....... ............... 90
Statistical A naly ses ....................................................... ............. .. .. .. ............. 92
Analysis of the Effects of the 30-Baht HI Policy on the Study Measures ..........92
Changes of Drug Disparity Associated with the 30-Baht HI Policy .................93

6 R E S U L T S .......................................................................... 9 5

Patient D em graphics .............................................. .. ...... ................. 95
D iab ete s ..............................................................................9 6
D rug U tilization R ates ........................................................... ............... 96
H hospital V isit Rates ............................................................ .... .... ... ....... 97
Impact of the 30-Baht HI Policy on the Study Measures............................... 99
H y p erten sio n .................. .................................................. ............... 10 1
D rug U tilization R ates ......................................................... .............. 101
H hospital V isit R ates ............................................................ .... ....... ....... 101
Impact of the 30-Baht HI Policy on the Study Measures..............................103
Infectious D diarrhea ......................... ........ ......................................105
D rug U tilization R ates ......................................................... .............. 105
Percent A appropriate Prescribing.................................... ........................ 106
Hospital Visit and Admission Rates......................................................106
Impact of 30-Baht HI Policy on the Study Measures ............. ... .................108
B bacterial Pneum onia .......................................................... .. ........ .. 112
D rug U tilization R ates .................................................................. ............ 112
H hospital V isit and A dm mission rates ................................................................113
Impact of the 30 Bath HI Policy on the Study Measures ...............................1.14
D isparity of D rug U tilization R ates..................................................... ..................118
Disparity of the Percent Appropriate Prescribing for Infectious Diarrhea .............18

7 D IS C U S S IO N ...... .. .................. .. .. ........ .... .............................................. 1 19

The Effect of the 30-Baht HI Policy on Drug and Hospital Service Utilization ......119
Drug Utilization and Access to Care................................... ............... 119
Effectiveness of the Implementation of the 30-Baht HI Policy ........................120
Suggestions to Improve the Evaluation of 30-Baht HI Policy on Drug Utilization
and H hospital Service U tilization ........................................ ....................... 124
Tim e Series D esign .................. ........................... .. ......... ........ .... 126
T he D efined D aily D ose ........................................................................ ... ... 126
Quality of the Data in the HI database............................................. ...............127
Study L im itations............... ..... ........................ .. .... .... ........... ....... 130
Recom m endations for Future Research................................................................ 133
Contribution to Scientific Society ........................................ ........................ 134









C o n clu sio n s.................................................... ................ 13 5

APPENDIX

A TERMINOLOGY AND DEFINITIONS .............................................. ...............136

B HEALTH INSURANCE CODES ........................................ ......................... 138

C R E C O R D L IN K A G E ...................................................................... .................... 139

D DRUG LISTS FOR CALCULATING DRUG UTILIZATION RATES................. 143

E EXPERT IN TERV IEW S .............................................. ....... ........................ 155

L IST O F R E FE R E N C E S ......................................................................... ................... 163

BIOGRAPH ICAL SKETCH .............................................................. ............... 177
















LIST OF TABLES


Table page

2-1 Alternative units for measuring drug utilization ............................................... 27

2-2 Types, definitions, and policy purposes of access to healthcare services................30

4-1 Hospital information, numbers of hospital visits, hospital admissions, and
prescriptions from 2000 to 2003 ........................................ ......................... 59

4-2 Five most prevalent dispensed drugs for inpatient and outpatient use of the eight
included hospitals (A-H) during from 2000 to 2003 ............................................. 64

4-3 Provincial Disease Statistics (2002) on the 10 most common causes of
morbidity for outpatient services, Ubonratchatani province, Thailand.................66

4-4 Missing data on patient demographics Hospital Missing data..............................68

4-5 Missing data on disease diagnosis codes for inpatient and outpatient data of
eight studied hospitals (2000-2003) .............................................. ............... 68

4-6 Missing data on prescribed drugs of eight studied hospitals over four years
(2000-2003) .................................................... ................. 68

6-1 Expected populations eligible for the 30-Baht HI policy and the CSMBS in 8
selected government community hospitals, Ubonratchatani province, Thailand
(2 0 0 3 ) ............................................................................ 9 5

6-2 P patient dem graphics ....................................................................... ..................96

6-3 Regression parameters for drug utilization and hospital visit rates of diabetes .......99

6-4 Rates of drug utilization and hospital visits related to diabetes of the 30-Baht
health insurance beneficiaries ........................................... ......................... 100

6-5 Regression parameters for drug utilization and hospital visit rates of
hyperten sion ........................................................................ 103

6-6 Monthly rates of drug utilization and hospital visit related to hypertension of the
30-Baht health insurance beneficiaries ........................... ............... 104









6-7 Regression paramerters for drug utilization, prescribing quality, hospital visit,
and hospital admission rates of infectious diarrhea ..................... .....................109

6-8 Monthly rates of drug utilization, hospital visit, hospital admission, and the
percent prescribing appropriateness of the 30-Baht group...................................110

6-9 Regression parameters for drug utilization, prescribing quality, hospital visit,
and hospital admission rates of bacterial pneumonia............................................ 115

6-10 Monthly rates of drug utilization, hospital visit, hospital admission related to
bacterial pneum onia .............................. ........... ........ .................. .. .. 116

B-1 Codes for health insurance status for the Health Insurance database...................138

D-1 Antidiabetic drugs for calculating drug utilization rates ............. ... ............ 143

D-2 Antihypertensive drugs for calculating drug utilization rates ..............................145

D-3 Antibiotics for calculating drug utilization rates...........................................148















LIST OF FIGURES


Figure pge

2-1 Health seeking behavior, Thailand 1999 (23). ........................................ ...............8

2-2 Nonconcurrent control group in observational design (27) ...................................16

2-3 Concurrent control group in observational design (27) .......................................17

2-4 Formulas for calculating sensitivity, specificity .............................................. 42

4-1 Geographic locations of the included eight community government hospitals in
Ubonratchatani province, Thailand ................ ........ ..... ........................... 59

4-2 Percent of patients with two or more disease diagnosis codes assigned in the HI
database (H hospital F )............. .............................. ................. .. ..... 60

4-3 Linkage among the selected patient data in the HI database AN: scrambled
hospital adm mission num ber............................................. .............................. 62

4-4 Percent data coherence between disease diagnosis of diabetes and antidiabetic
drugs of outpatient data among eight hospitals, 2000-2003...............................71

4-5 Percent data coherence between disease diagnosis of diabetes and antidiabetic
drugs of inpatient data among eight hospitals, 2000-2003 ........... ...............71

4-6 Percent data coherence between disease diagnosis of hypertension and
antihypertensive drugs of outpatient data among eight hospitals, 2000-2003 .........73

4-7 Percent data coherence between disease diagnosis of hypertension and
antihypertensive drugs of inpatient data among eight hospitals, 2000-2003 ..........73

4-8 Percent data coherence between disease diagnosis of bacterial pneumonia and
antibiotics of outpatient data among eight hospitals, 2000-2003...........................75

4-9 Percent data coherence between disease diagnosis of bacterial pneumonia and
antibiotics of inpatient data among eight hospitals, 2000-2003.............................75

4-10 Percent data coherence between antidiabetic drugs and disease diagnosis of
diabetes am ong eight hospitals, 2000-2003 .................................. ............... ..76









5-1 Algorithm to assess quality of drug utilization (DQ) using community-acquired
pneum onia as an example.(169; 170)......................................................... ......... 91

5-2 Interrupted time series analysis of a drug utilization rate in the 30-Baht HI group
and the control group before and after the policy ........................ ................. 93

6-1 Monthly drug utilization rates for diabetes in the 30-Baht HI population and of
the CSMBS population from January 2000 to December 2003............................. 97

6-2 Monthly hospital visit rates for diabetes in the 30-Baht health insurance
population and of the CSMBS population from January 2000 to December 2003..98

6-3 Drug utilization rates for hypertension of the 30-Baht health insurance benefit
group from January 2000 to December 2003....................................................... 102

6-4 Monthly hospital visit rates for hypertension of the 30-Baht health insurance
benefit group from January 2000 to December 2003...........................................102

6-5 Monthly drug utilization rates for infectious diarrhea of the 30-Baht HI from
January 2000 to D ecem ber 2003 ........................................ ........................ 105

6-6 Percentages of prescribing appropriateness for infectious diarrhea of the 30-Baht
HI from January 2000 to December 2003 ....................................................... 106

6-7 Monthly hospital visit rates for infectious diarrhea of the 30-Baht HI from
January 2000 to D ecem ber 2003 ........................................ ........................ 107

6-8 Monthly hospital admission rates for infectious diarrhea of the 30-Baht HI group
January 2000 to D ecem ber 2003 ........................................ ........................ 107

6-9 Monthly drug utilization rates for bacterial pneumonia of the 30-Baht HI from
January 2000 to D ecem ber 2003 ................................... .............. ......... ...... 112

6-10 Monthly hospital visit rates for bacterial pneumonia of the 30-Baht HI from
January 2000 to D ecem ber 2003 ..................................... ............ ....... ........ 113

6-11 Monthly hospital admission rates for bacterial pneumonia of the 30-Baht HI
from January 2000 to December 2003 ........... ........................... ...... .......... 114

C-l Two tables identify patients who visited outpatient clinic during the study periodl39

C-2 Two tables identify patients admitted to the hospital for inpatient services during
the study period, and length of stay................................................140

C-3 Patient demographic table identifies patient's age, gender, marital status, types
of health insurance. ........ ........ .... ............................. ... ............ ........141

C-4 Tw o tables contain drug data ...........................................................................142














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

IMPACT OF THE 30-BAHT HEALTH INSURANCE POLICY ON HOSPITAL
DRUG UTILIZATION IN THAILAND


By

PENKARN KANJANARAT

August 2005

Chair: Almut G. Winterstein
Major Department: Pharmacy Health Care Administration

Thailand implemented the national 30-Baht health insurance policy in June 2001 to

provide healthcare coverage (including drugs) for the uninsured population. We

retrospectively assessed the impact of this policy on drug utilization in community

government hospitals using computerized patient-specific data from 8 hospitals in

Ubonratchatani province. Cross-sectional data from January 2000 to December 2003

included all inpatients and outpatients with bacterial pneumonia (CAP), gastrointestinal

(GI) infections, diabetes (DM), and hypertension (HTN). We confirmed the internal

validity of the database and conducted interviews of hospital personnel. The primary

measure was drug utilization rate (DUR), measured as the Defined Daily Dose

(DDD)/10,000 population/month. The secondary measures were percent prescribing

appropriateness, hospital visit and admission rates. Monthly observations (N=48) were

analyzed using segmented time series regression analysis (SARIMA model).









Prior to the policy, the average monthly DUR of the 30-Baht group for DM, HTN,

Infectious diarrhea, and CAP were 2,727.8, 1,414.0, 40.3, and 7.8 DDDs/10,000

beneficiaries, hospital visit rates were 86.7, 25.8, 25.6, and 1.9 visits/10,000

beneficiaries, and hospital admission rates were 2.5 and 0.3 admissions/10,000

beneficiaries, respectively. After the policy, average monthly DUR for DM, HTN,

Infectious diarrhea, and CAP were 3,105.6, 2,983.5, 52.8 and 12.4 DDDs/10,000

beneficiaries, monthly hospital visit rates were 87.4, 36.2, 26.5, and 1.9 visits/10,000

beneficiaries, while admission rates for Infectious diarrhea and CAP were 3.0 and 0.6

admissions/10,000 beneficiaries, respectively. Appropriate antibiotics were prescribed to

less than one half of the patients with Infectious diarrhea (43.4% before and 47.5% after

the policy).

Analysis revealed no immediate or trend effect in DUR, hospital visit/admission

rates for DM, HTN, infectious diarrhea, or CAP after the 30-Baht HI policy was

implemented (p>.05 for all measures). There was no significant change on the percent

appropriate antibiotic prescribing for Infectious diarrhea, p>.05.

The study did not detect a change in drug and hospital service utilization associated

with the 30-Bath HI policy, although there were positive trends in rates of drug

utilization, hospital visit, and admission after the policy. Computerized hospital database

prove valid and useful resource for research.














CHAPTER 1
INTRODUCTION

Background

The Institute of Medicine (IOM) Report 2002 (1) addressed significant health

problems of the uninsured. The report stated that approximately 30 million Americans

with no health insurance had a higher risk of poor health and shorter life-expectancy than

people with health insurance. Poorer health in the uninsured was the consequence of

limited access to care (1).

Until the year 2000, approximately 30% of the Thai population had no health

insurance coverage. In 2001, the Thai government, to improve access to care and reduce

healthcare disparity between the insured and uninsured Thais, implemented the 30-Baht

Health Insurance (HI) Policy for the uninsured Thai population. The uninsured were the

poor, self-employed, and children 12 to 18 years of age, who did not have any benefits

from the three existing governmental health insurance plans: the Civil Servant Medical

Benefit Scheme (CSMBS), the Medical Welfare Scheme (MWS), and the Social Security

Scheme (SSS).

The 30-Baht HI policy offers coverage for most healthcare services and

prescription drugs provided by government hospitals. The National Health Insurance

Office evaluates eligibility based on the uninsured status of an individual. The 30-Baht

HI beneficiaries are then registered with the government community hospital in their

district as a main healthcare provider. The HI benefits are not extended to care received

by other providers except when the patients are referred to that facility. Similarly,









prescription drugs are covered by the policy only if they are dispensed from the hospital

pharmacy at the designated hospital.

Our study aims to assess the impact of the 30-Baht HI policy on drug utilization

because drug therapy is a core component of medical care. Although previous literature

suggests that lack of drug benefits is associated with underutilization of prescription

drugs (2-7) and that an increase of drug utilization was expected, our study evaluated any

changes of drug utilization associated with this policy.

Need for Study

After this national policy was implemented in October 1, 2001, the government

authorities, including the National Health Security Office (NHSO) and the National

Health System Research Institute (HSRI), conducted various health policy evaluation

studies to evaluate the effects of the policy on various aspects. However, most of the

research from the NHSO and the HSRI focused only on administrative and economic

issues, and patients' and providers' satisfaction with the policy (8-10). Few studies have

yet examined the impact of the policy on drug utilization (8).

Evaluating the effect of the new policy on drug utilization required identifying the

effect of the policy on changes in the disparity of drug between subpopulations. While

national healthcare expenditures in Thailand increased greatly over the last 2 decades

(3.82% of GDP in 1978 to 6.21% in 1998) the distribution of this spending is skewed

toward the insured. The increase mostly occurred in the insured population. Only a

small proportion occurred in the uninsured, which indicates healthcare utilization

disparity or unequal opportunity in healthcare service access and utilization among

uninsured versus insured populations. The disparity was associated with various factors

that inhibit access to care, including financial status (11).









Assessments of national health policies in developing countries are often

compromised by weak study designs that fail to establish a causal relationship between

the policy, and health indicators and health outcomes (12). For example, various studies

use cross-sectional designs, post-only intervention designs, or pre-post comparisons with

no control group (13-19). Moreover, the selection of reliable and clinically significant

outcomes measures is often compromised by data unavailability or inappropriate data

format for the analysis (e.g., healthcare data was collected in paper format, with no record

linkage among databases). Our study applied a sound study design (interrupted time

series of impact analysis) to evaluate the causal association between the 30-Baht HI

policy and drug use by the targeted population.

Currently, Thailand has no standardized data source that contains drug utilization

data at a national level. Available drug utilization data were derived from quarterly

government hospital purchasing and inventory reports, which lack patient-level detail.

Estimates of drug utilization can also be derived from drug importation and/or

manufacturing reports put forward by pharmaceutical companies on an annual basis. No

drug utilization data are available at a patient level, which would allow assessments of

drug utilization in a certain disease state or across subpopulations.

The opportunity for our study was offered by a novel computerized administrative

database, "Health Insurance (HI) database", used in every hospital in Ubonratchatani

province since 1997. This database includes electronic medical records and drug

dispensing data for every patient who received healthcare services (outpatient and

inpatient) from the hospitals. Although never used in research, the database offers a









unique opportunity for health outcomes and drug utilization studies because of its level of

detail and its uniform structure in all community hospitals of the province.

Evaluating of the effect of the health policy on drug utilization would contribute to

an objective assessment of healthcare services established by the Thai government. It

would also offer suggestions to stakeholders for further improving access to drug

utilization and eliminating disparities in healthcare.

Purposes of Study

We proposed to quantitatively evaluate the impact of the 30-Baht HI policy on drug

utilization in terms of drug utilization rates and prescribing quality, using drug data from

the HI database. Drug utilization rates are measured using the Defined Daily Dose

(DDD) and prescribing quality is assessed based on the appropriateness of prescribing for

a given disease state. Effects of the 30-Baht HI policy on access to hospital services were

also measured as hospital outpatient visit rates and hospital admission rates. Four disease

states (diabetes, hypertension, bacterial pneumonia, and infectious diarrhea) are studied.

We selected the listed disease states for two reasons. First, they are prevalent and

associated with high mortality rates in Thailand. Second, they represent acute and

chronic conditions and allow a more comprehensive assessment of the policy impact on

healthcare.

From the computerized patient-specific healthcare database (HI database) of the

government community hospitals in Ubonratchatani province, Thailand, we chose

subjects who were eligible for the 30-Baht HI policy during 2000 and 2003 as a study

group. The Civil Servant Medical Benefit Scheme (CSMBS) was selected as a control

population because of its comprehensive healthcare and drug benefits. For the control

group, access to care and choices of treatment were not limited by the patients' ability to









pay for services. We analyzed the effect of the 30-Baht HI policy on drug utilization and

access to hospital services using an interrupted time series analysis. By using a control

group, we adjusted for the natural trend and other factors that might have an effect on

drug utilization rather than the 30-Baht HI policy. In addition, we compared drug use

disparity (difference of drug utilization between the 30-Baht HI beneficiaries and the

CSMBS group) at one year before and after the implementation of the policy.

Definitions and terminology used in our study are presented in Appendix A.

This study was approved by the Institutional Review Board, University of Florida.

Study Objectives

* To test the validity of the patient-specific healthcare data in the HI database of the
community government hospitals in Ubonratchatani province, Thailand

* To assess the impact of the 30-Baht HI policy on drug utilization rates, the percent
appropriate prescribing, hospital visit and hospital admission rates in the 30-Baht HI
population controlling for longitudinal changes in the CSMBS group

* To evaluate the change of drug utilization disparity between the 30-Baht HI
population and the CSMBS at one year before and after the policy was implemented.

Research Questions

* Research question 1: Did drug utilization rates in the 30-Baht beneficiaries for the
selected disease states change after the 30-Baht HI policy was implemented,
controlling for trends of the drug utilization rates of the previously uninsured group
and those of the CSMBS group?

* Research question 2: Did the prescribing quality in the 30-Baht beneficiaries change
after the 30-Baht HI policy was implemented, controlling for the drug utilization
quality of the previously uninsured group and those of the CSMBS group?

* Research question 3: Did hospital admission rates related to pneumonia and gastro
intestinal infections for inpatient services in the 30-Baht beneficiaries change after the
30-Baht HI policy was implemented, controlling for the hospital admission rates of
the previously uninsured before the policy and those of the CSMBS group?

* Research question 4: Did hospital outpatient visit rates related to diabetes and
hypertension in 30-Baht beneficiaries change after the 30-Baht HI policy was
implemented, controlling for the hospital visit rates of the previously uninsured
before the policy and those of the CSMBS group?









* Research question 5: Did the extent of drug utilization disparity between the 30-
Baht HI policy beneficiaries and the control group change at one year after the policy
compared to the extent at one year before the policy?

Significance

We evaluates whether the 30-Baht HI policy improved drug utilization rates and

quality of drug utilization as provided in governmental insurance schemes. Comparisons

of drug utilization rates and quality between this population and the control population

also addressed its effect on drug utilization disparity. Our results will complement other

epidemiologic and/or economic evaluations of this policy and suggest areas of

improvement of the health insurance benefit program to reach the goal of comprehensive

national healthcare coverage.

We used comprehensive electronic patient medical records from the HI databases

that have never been validated or used for research purposes. Although this issue posed

an additional challenge for main study objectives, we hope the validation of the HI

databases facilitates future health service research and quality improvement activities.














CHAPTER 2
REVIEW OF LITERATURE

Healthcare Care System in Thailand

Most healthcare services in Thailand are provided to the population by the public

sector. Most of these facilities are affiliated with the Ministry of Public Health, followed

by the Ministry of Defense (veteran hospitals) and the Ministry of Internal Affairs

(municipal hospitals). Healthcare services provided by the private sector include private

hospitals, private outpatient clinics, and independent pharmacies. Additionally, many

patients seek cures from traditional treatments, although modern medicine is widely

available.

Health-Seeking Behavior and Healthcare Utilization

Even though self-medication (44.2% to 86.3%) (20;21) is commonly practiced in

Thailand, it decreased from 54.1 to 17.62% (more than 36%) from 1970 to 1996, while

healthcare utilization from public health services increased from 15.5-44.0% (almost

30%) (22). Socio-economic factors, the healthcare infrastructure, and type of health

problems appear to affect health seeking behaviors. For example, the 1999 Health and

Welfare Survey indicated that people living in rural areas mainly seek healthcare services

from public settings, while people who live in urban areas are more likely to seek

healthcare services from private clinics or hospitals than from the public facilities (23)

(Figure 2-1).

A study of health-seeking behavior of the villagers in rural areas of Thailand using

health diary by Osaka et al.(24) found that most patients with chronic diseases were more









likely to receive treatment and medications from local health centers or community

hospitals, while patients with severe and acute symptoms (e.g., high fever or abdominal

pain) sought urgent care from private clinics. For minor illness, most villagers rarely

used healthcare services from a local health center or a community hospital.


60

50C










o
o20 [Rural






0 4 ~2--
10

o co

Figure 2-1. Health seeking behavior, Thailand 1999 (23).

Access to Drugs

Regarding the current drug delivery systems, almost all classes of drugs are

accessible for Thais, from prescription drugs to herbal medications. Drug distribution

channels include hospitals, private clinics, pharmacies, health centers, public health

centers, and even grocery stores. Prescription drugs are sold in drug stores with

pharmacists and without pharmacists. Some pharmacies are open for service with no

pharmacists on duty, thus patients receive drug products from the store owner or a store

clerk.

In 1999, nearly half of the drug stores (5,351 of 12,548 stores) were authorized to

sell prescription and controlled-substances (e.g., corticosteroids, benzodiazepines, and









narcotics). The remaining drugstores were authorized to sell most of the commonly

prescribed drugs, including antibiotics, with or without a pharmacist. Moreover, over

400,000 small grocery stores in villages around the country have prescription drugs

available. Drug regulations and enforcement do not control this type of retail drugstores

in Thailand (25).

Not only pharmacists can dispense prescription drugs and controlled substances:

Thai drug regulations allow physicians to dispense medications as well. It is commonly

acknowledged that drugs are the major source of income of healthcare facilities. So

every hospital has a hospital pharmacy department where most prescriptions are

dispensed.

While drugs can be obtained through various channels only government hospitals

can dispensed drugs under the 30-Baht HI policy. Thus, the study databases can assess

whether reimbursement for drugs from hospital improved, but it is not possible to assess

whether overall drug access increased as a results of the 30-Baht HI policy.

Healthcare Financing

To understand how healthcare services are provided to Thai people, it is important

to be familiar with existing healthcare financing schemes before the 30-Baht HI policy

was implemented. The 30-Baht HI policy was created based on the four established

health insurance schemes offered by the government, and targeted to fill the gap between

the insured and uninsured. The four health insurance schemes offered by the Thai

government included the Civil Servant Medical Benefit Scheme (CSMB S), the Social

Security Scheme (SSS), the Medical Welfare Scheme (MWS), and the Health Card

Scheme (HCS). Private Health Insurance (PHI) is also available for people who can

afford it.









The next section provides some background information on the CSMBS in terms of

its benefits, program financing and related problems. This group was used as a control

group to analyze the impact of the 30-Baht HI policy on drug utilization. We will not

discuss details of the other health benefit schemes.

The civil servant medical benefit scheme

The CSMBS is the health benefit program provided by the government for civil

servants who currently work for government entities and respective retirees. The benefits

extend to immediate family members, including a spouse, parents, and children younger

than 20 years of age. In 2000, there were approximately seven million beneficiaries.

However, the exact number of beneficiaries is unknown due to lack of a registry

database.

The benefits include outpatient and inpatient, medical and surgical services,

emergency services and drug expenses. The benefits exclude a small number of services

such as cosmetic surgery and preventive services (vaccination and contraceptive

medication), except for an annual health check-up. The beneficiaries have access to care

from government hospitals or private hospitals that have registered with the plan.

This program uses a retrospective reimbursement method to pay for healthcare

based on a fee-for-service system. This type of payment system has minimal control on

healthcare expenditure. Consequently, healthcare expenditures in the CSMBS population

have increased dramatically over time both for outpatient and inpatient services and are

expected to rise continuously. It has been reported that the expenditures per person were

twice as much as those for patients in other healthcare benefit plans or those with no

health insurance(1 1).









As part of the Healthcare Reform in 1998, the benefit package of the CSMBS has

been reviewed. Two major changes were proposed to control healthcare costs and

improve the quality of care. First, the previous focus on treatment has shifted to health

promotion and disease prevention (disease screening and vaccination plans are now

included in the plan). Second, a certain level of healthcare quality is assured by the

requirement that the providers must register with the Ministry of Public Health in order to

get reimbursement for inpatient services for this population. These changes affect both

drug utilization rates and quality, and the sources of drug utilization data of this

population.

The 30-Baht health insurance policy

The concept of universal health insurance, which was ratified in the new national

constitution amendment in 1998, states that "All Thai people have an equal right to

access quality health services..." Despite a quite substantial number of existing schemes

offered by the government and by private insurance companies, only 70% of the

population was covered by these plans. To expand the previous insurance schemes

towards the goal of "universal health coverage", the Thai government introduced the "30-

Baht HI Policy" for the remaining uninsured population.

The 30-Baht HI policy was introduced by the new government elected in 2000.

This policy is a product that responds to the 1998 Thai Health Care Reform which

focuses on the improvement of access to care and reduction of health disparities. The

pilot phase of this policy was implemented in 4 provinces (out of a total of 75 provinces

in Thailand) in February and expanded to another 15 provinces, including

Ubonratchatani, in June 2001. After successful completion of the pilot phase, the policy

was implemented nationwide in October 2001.









The policy covers every person who did not have any healthcare benefits from the

existing healthcare benefit schemes mentioned earlier. Each person can register with the

primary care center, or a community government hospital in their local district. The

eligibility to participate in the plan is verified by the Office of National Health Insurance.

Registered persons receive a "gold card" that has to be presented when healthcare

services are sought. The beneficiaries are required to pay 30-Baht (75 US cents) per

episode of care.

The benefits of the policy include medical treatment and disease prevention,

disease screening and diagnosis, and rehabilitation as necessary. They also include Thai

traditional medicine and alternative medicine under the provision of a medical

professional. Regarding prescription drugs, the 30-Baht HI policy covers only drugs

listed in the National Drug Formulary that include most essential drugs recommended by

the World Health Organization. These drugs are required by the Ministry of Public

Health to be included in a hospital drug formulary of government hospitals. More recent

and expensive drugs can be added to a hospital drug formulary, but are typically

restricted to patients who are willing to pay for extra costs (e.g., the CSMBS and SSS

population).

Evaluating Health Policy

In the past two decades, quality improvement (QI) has been a major focus in health

care services worldwide. Among other QI initiatives, an increasing number of health-

related policy interventions has been implemented to improve quality, and reduce and/or

contain costs. As a consequence, the number of health policy evaluation studies and the

use of automated healthcare databases for this purpose has increased dramatically in the

public (e.g., Medicaid and Medicare populations) and private sectors such as Health









Maintenance Organizations, such as the Group Health Cooperative of Puget Sound,

Kaiser Permanente.

Most health policies that have been implemented are either "regulation" or

"deregulation", both, with the intent to affect health service utilization and health

outcomes. Examples of regulation policies are the state antibiotics vigilance in the

Medicaid population, antibiotics restriction in hospitals (26), or co-payments and drug

reference pricing (27).

The evaluation of the impact of health policies is difficult because they are

implemented in non-experimental conditions: the policy typically affects the entire target

population and thus, control groups are difficult to establish. In addition, the policy is

implemented in an already dynamic healthcare environment, and other factors that may

affect the outcome of interest may not be excluded. A review of drug policy evaluations

in developing countries found that most policy evaluation studies used weak study

designs and were based on post-intervention measures only. Thus, the studied results

were not conclusive (28). From a measurement perspective, evaluations of health

policies are often restricted to administrative databases, which raise problems such as

poor data quality, inability to ascertain the outcomes of interest, and incomplete patient

information data.

Methodology in Evaluating Health Policy

Various methods have been used to evaluate the effects of health policy on

healthcare utilization and health outcomes of the population. Among more advanced

study design are randomized controlled trial, uncontrolled time series, and controlled

time series. Only time series design, its methods, strengths, weaknesses, and its

applications are discussed in this section.









Time series in observational design

Time series analysis in a longitudinal study design has been used in the fields of

economics, political science, and engineering research for several decades. It has also

been an increasing number of research studies using time series in the field of drug

utilization, health policy evaluations, and pharmacoepidemiology. Several time series

studies in healthcare research focused on the assessment of the impact of health policy on

prescribing behavior and drug utilization (26;29-31;31-34).

There are three major purposes of applying time series analysis: 1) identifying

patterns of the series of the data, 2) evaluating the effect of an intervention, and 3)

forecasting future values of the data. Our study intends to use the analysis of time series

data to evaluate the effect of the policy (considered an intervention), thus, only the

literature on the evaluation of interventions is included in this section.

Segmented regression analysis requires a series of data that was collected regularly

and is equally spaced over time, e.g., monthly, seasonally, yearly. There are several data

sources that can be used to measure the effects of health policy on drug use. Data are

derived either from primary data collection (e.g., direct follow-up of patients to ascertain

information on health outcomes), or secondary data collection (e.g., medical records,

insurance claim data, pharmacy dispensing data, or hospital discharge data). The benefit

of using healthcare data that have been routinely collected for the purpose to document

clinical interventions or reimbursement is the availability of the data longitudinally in

large populations (35).

Outcomes of interest can be healthcare or drug utilization, clinical measures, or

costs. The outcome measures can be applied as averages, proportions or rates. Drug

utilization measures that are often used in policy evaluation research are number of drugs









prescribed per patient, average prescription cost, percent of beneficiaries receiving a

particular drug, or percent of patients treated according to treatment guidelines (35).

Lengths of stay and hospital admission rates are widely used to determine the effect of

the intervention on access to healthcare services. For clinical measures, surrogate

markers such as blood pressure can be used.

Time series is composed of a set of observations that are measured longitudinally

(36). The desired number of observations is usually more than 50 to yield sufficient

sample size to precisely model the data (37). The observations can be measured from a

single case (e.g., weight of a person over 10 years) but more often are seen in aggregate

data from several cases, e.g., average drug utilization rates in diabetes patients over 10

years.

Nonconcurrent time series design

The time series design that has no concurrent controls is useful to identify

immediate effects of a policy intervention; however it always encounters internal validity

problems due to other causes that may affect the outcome of interests. Since it is

impossible to control other factors that might affect the outcomes at the same time as the

policy did, it is important to base the conclusion on the assumption that there is a close

temporal relationship between the policy and the outcome, and the extrapolation of the

baseline trend must be estimated as if the intervention was not implemented. Figure 2-2

shows the dotted line indicating the measure before policy and an extrapolation after the

policy was implemented under the assumption that the policy has no effect on this

measure. The full line represents the observed values of the measure if the policy has a

negative effect on the measure. The time series analysis of non-concurrent data becomes

more accurate when there is large number of observations before and after the










intervention because the prediction of the trend before the intervention is more accurate

(38). Time series analysis requires an adequate number of observations (some

researchers have recommended more than 200 to produce accurate results) (37) before

and after the intervention to be able to control for trends or seasonal effects.



Intervention
Expected
trend without
intervention

S- .- Intervention
S. "group









Time

Figure 2-2. Nonconcurrent control group in observational design (27)

Concurrent time series design

To overcome the validity problem due to other factors that might affect the

outcome rather than the intervention, a concurrent control group time series design

should be used if a comparison group exists. The concurrent control group can help

adjusting for the effect of other factors that might oppose or distort the effect of the

studied intervention. However, the conclusion of the analysis must be based on the

assumption that the control trend is equal to the trend in the intervention group if there

were no effect of the intervention. The dotted line in Figure 2-3 demonstrates the

observations in a control population, which has similar trend before the policy (straight









line). After the policy, trend in the control group remains the same as the trend before the

policy. In the intervention group, there is a change in the trend after the intervention

implementation, which suggests a negative effect in decreasing the value of the measures.

The difference in the slope can indicate the effect of the intervention.


Intervention
Control
group





S- -- --------------- Intervention
I -group










Figure 2-3. Concurrent control group in observational design (27)

In this study, the researcher chose to apply a time series design using the CSMBS

population as a control group to evaluate the effect of the 30-Baht HI policy on hospital

drug utilization over a 4- year period. This method is appropriate for aggregate data, e.g.,

drug utilization, hospital visit or hospital admission rates. Using a concurrent control

group, the design provides strong evidence of a causal relationship between the

intervention and outcome measures.

The following section presents some examples of published studies of health/drug

policy evaluations using the analysis of time series data. Some studies used concurrent

control groups and some did not. The methods are presented for each study.









A study in Manitoba, Canada, evaluated the drug benefit policy change from a

fixed deductible and co-payment system to an income-based deductible system on receipt

of prescriptions for inhaled-corticosteroids in children with asthma. This study compared

a cohort before and after the drug benefit policy was changed. No concurrent control

group was included in the study. Receipt of prescription and number of inhaled

corticosteroid doses were compared by visual observation and odds ratios (39).

Blais et al. conducted a study in Quebec to assess the unintended effects of a cost

sharing drug insurance plan on drug utilization among individuals receiving social

assistance. Drug utilization in three classes of medications was studied: inhaled

corticosteroids, neuroleptics, and anticonvulsants. A control group was applied in this

study. The Seasonal Autoregressive Integrated Moving Average (SARIMA) model was

implemented to adjust seasonal effects on drug utilization(40).

In 1997, the reference pricing policy for angiotensin-converting enzyme inhibitors

(ACEIs) was implemented in British Columbia, Canada. Schneeweiss and his colleagues

conducted a study to evaluate the effects of this policy on quantity and timing of drug

utilization. Autoregressive time series of prescription drug claims data for 3 years was

used to identify the effects of the policy. The authors conclude that analysis of time

series data is able to provide detailed information for the policy makers regarding the

extent and duration of the effects of the drug policy.

In 1989, the State of New York implemented a Triplicate Prescription Program

(TPP) as a drug prescribing surveillance system. After the TPP was implemented, there

was 55% reduction in the monthly number of benzodiazepine recipients in the Medicaid

cohort (41). Wagner and his research team further studied the intended and unintended









effects on new post-hospitalization benzodiazepine use(42). The effects of the TPP were

evaluated by an interrupted time-series of post-hospitalization benzodiazepine dispensing

rates and the substitute medications in a Medicaid population. The control group in this

study was the Medicaid population in New Jersey, which adds strength to this study in

controlling for extraneous factors that might affect benzodiazepine prescribing behavior.

In 2000, hospitals in Columbia implemented an educational intervention to improve

appropriateness of antibiotic prescribing practices. Interrupted time series were

conducted on three antibiotic groups (aminoglycosides, cephradine/cephalothin, and

cefazidime/cefotaxime), comparing the hospital weekly rates of incorrect prescriptions,

and prophylactic antibiotic use in elective surgery to assess the effects of the intervention.

From a statistical method perspective, this study applied the Autoregressive Integrated

Moving Average (ARIMA) model of time series data because the series of the data are

not linear. The study did not have a reference group because of inability to find a

comparable setting (31).

Another study by Ansari et al. in Tayside, Scotland, applied segmented regression

of interrupted time series analysis to identify the effect of an Alert Antibiotic Policy on

improving use of appropriate antibiotics (26). The authors commented that segmented

regression analysis of pharmacy data is a feasible method for the assessment of the

effects of the policy (26).

A study of drug prescribing in primary care practices was conducted in Boston,

Massachusetts after all non-formulary drugs were removed from the formulary and

generic samples stocked. Segmented linear regression analysis was applied to estimate

changes in levels or trends in formulary compliance (43).









In addition to quasi-experimental designs, time series is useful in identifying the

impact of natural changes or influences. A study by Tu et al. intended to evaluate the

effect of the finding from the Heart Outcomes Prevention Evaluation (HOPE) study that

ACE inhibitor (ramipril) are effective in the prevention of secondary cardiovascular

disease. By applying segmented regression of time series of prescribing data, the study

was able to identify the effect of each publication type on ACE inhibitors prescribing

(44).

The above studies have applied time series design and analysis using a segmented

regression model to evaluate the effect of the interventions or policies. Some studies

allowed lag times in the ARIMA process in order to incorporate delayed effects of the

intervention. Very few studies utilized a concurrent control group; however those studies

had sufficient historical observation points to ensure that any change after the policy was

not associated with the factors that affected the measure before the policy was

implemented. For non-linear data (e.g., the percentage of prescribing appropriateness)

the studies applied various transformation methods (e.g., logits) to establish more linear

series. Seasonality that might cause non-linearity of the data was sometimes incorporated

by using a SARIMA model.

Quantitative and Qualitative Measurement of Drug Use

Quantitative drug utilization studies usually refer to the measurement of numbers

and rates of drug consumption or exposure in the population of interest in a certain period

of time (e.g., number of drug doses/10,000 population/year, and proportion of patients

receiving a certain drug per 10,000 population per year). A qualitative approach of

measuring drug utilization is the evaluation of "appropriateness" of utilization of a certain

drug or a class of drugs in a specific population with a certain disease states. Since the









definition of the appropriateness of drug use theoretically incorporates every aspect of

drug use, it is complicated to conduct a conclusive evaluation of the process of care and

patient outcomes.

In the section on qualitative measurements of drug utilization, definitions of

quality, interpretation, and measurement methods, including explicit criteria and quality

indicators are discussed. Sources of the data that have been used in assessing

appropriateness of drug use are also discussed regarding their applications and the

limitations.

Quantitative Approach to Measuring Drug Utilization

According to the WHO, drug utilization is defined as marketing, distribution,

prescription and use of drugs in a society and economic consequences (45). A drug

utilization study refers to the assessment of medical, sociological-behavioral and

economic factors influencing drug utilization, including the effects of drug utilization at

all levels.

Lunde and Baksaas (46) describe general objectives of drug utilization studies as

"problem identification and problem analysis in relation to importance, causes, and

consequences, and an establishment of a weighted basis for decisions on problem

solution; assessment of the effects of the action taken. These objectives are relevant to

problems and decision making throughout the drug and healthcare systems. The

approaches may vary according to the purpose and the needs of the users. Those include

the health authorities, the drug manufacturers, the academic and clinical health

professionals, social scientists, and economists as well as the media and the consumers

(46)".









Measurement Units of Drug Utilization

Quantitative drug utilization studies involve drug utilization statistics (e.g.,

utilization rates) of the population by age, gender, disease stage, geographical area, or

across time. Drug utilization rates can be used to determine access to care or areas that

are under- or over-utilized. These statistics are also useful to plan for drug importation,

production, and distribution by pharmaceutical industries and government authorities.

Drug utilization rates have been used as crude estimates of disease prevalence, morbidity

and mortality rate, e.g., digitalis utilization for congestive heart failure. In addition, drug

utilization rates allow monitoring of drug use in specific therapeutic groups that are

associated with drug therapy problems (e.g., NSAIDs, narcotic analgesics, hypnotics and

sedatives) (47). In addition, drug utilization rates can be used to monitor the effects of

health-related policy and activities over time and across populations.

The defined daily dose (DDD)

There are several widely accepted measurements of drug utilization in research

depending upon the purpose of the study, research community and institution, country, or

the type of available database. In early 1970's when drug databases were not fully

established and detached from population data, the researchers in drug utilization studies

predominantly used costs of drug use and volume of prescription as an index of drug use.

However, there are several limitations of using cost data to represent actual drug

use in the population. Costs of drugs were drastically varied over time, depending on

pricing policy, country, currency exchange rate, and quantity of drug purchasing.

Additionally, cost does not have any correlation to the population denominator, such as,

demographic factors, disease conditions, prescribers, or health of the population; thus, it

is difficult to interpret this cost data in drug utilization studies.









Another measurement unit of drug use is the number of prescriptions per patient

or per population. This unit has been used in drug utilization review studies for several

decades since prescription records, were recorded in the computer systems. The number

of prescriptions could be used to indicate the number of patients who received drugs;

however, it still has some limitations as various quantities may be prescribed per

prescription or the gap between drug prescribed and dispensed. The number of

prescription units remains insufficient in evaluation associations between drug use and

the indication (disease) and patient health outcomes. This unit may not be an appropriate

measure if numbers of drugs and quantity per prescription vary largely.

In late 1980's, the World Health Organization introduced a standardized

measurement unit of drug utilization, that is, the Defined Daily Dose (DDD). This effort

aimed to provide a unit for comparison across countries and to enable drug utilization

monitoring over time. The DDD is a technical unit of drug utilization that provides an

estimate of the number of patients within a community who receive a drugs) of 1

maintenance dose. It is calculated based on the assumed average dose per day for a drug

product used for its major indication in actual practice (48). The identified dose is

suggested by the medical literature and presumed to be the average maintenance dose

when used for the major indication (with full patient compliance). We used Equation 2-1

to estimate utilization of a drug in a population of 10,000 per month, expressed in DDDs.

DDDs/10,000 people/month = amount of drug (mg) sold in 1 year x 10,000 (2-1)
DDD (mg) x 12 month x number of people


Strengths of the DDD

* Measurement of drug use is independent of brand name, package size, and sales price









* Measuring drug utilization based on the same DDD allows comparison across
settings, regions, and countries

* DDD could be used to monitor drug use overtime if the main indication and dosage of
the drug remains the same

* DDD could be used to estimate drug utilization of the population using aggregate
data, e.g., sales data

* DDD has been adopted widely for more than two decades in research and health
policy and the WHO, which offer advantages to current researchers in understanding
the application of the DDD and then apply to their research.

Limitation of the DDD

* DDD is defined based on Scandinavian data using the therapeutic maintenance dose
for an adult person of 70 kg bodyweight and normal organ functions.
Generalizability to other countries or patient populations, such as children or patients
with renal failure, is limited.

* DDD is only defined for the main indication of the drug for a disease and does not
refer to other uses, e.g., prophylaxis. Moreover, several drugs are used with no
indication (off-labeled).

* DDDs are not defined for a preparation for topical use, sera and vaccines,
antineoplastic drugs, anesthetics, or contrast media.

* Number of patients receiving DDD based on population-based drug utilization
estimates is a rough estimate under the assumption that patients have full compliance,
and may not represent actual drug use.

Defined daily dose in drug utilization research

A good example of the applications of the DDD is demonstrated by the study of

trend in drug consumption of calcium channel blockers in the Czech Republic (49). This

study focused on the drug utilization pattern overtime and comparison across selected

countries. This study applied the DDDs to identify effects of the intervention

(publication of the adverse drug events of short acting nifedipine) on drug use. Drug

utilization was measure by the DDD system using wholesale data from the General

Health Insurance Company of each country. The DDD unit was used in a retrospective









study of antipsychotic agent utilization in Spain, which indicates the benefit of this unit to

compare pattern of drug use nationally and longitudinally (50).

Defined daily dose in health policy and healthcare delivery

A study by Damiani et al. in 2002, identifying the effect on non-hospital

prescribing of intramuscular administered cephalosporins (IACs) in Italy, showed an

application of the DDD in measuring drug overtime. Since drug expenditure for

parenteral cephalosporins was a major part of the country's healthcare drug budget, there

was a community-level regulation to restrict use of IACs called CUF55, which limit the

use of these antibiotics to specific infections that are resistant to other common

antibiotics, or in patients with immunodeficiencies. CUF55 was then modified to

deregulate second-generation cephalosporins, cefonicid and cefmetazole. The research

applied DDD to measure changes of the targeted drugs (cefonicid and cefmetazole) and

the drugs that might be used as a supplement (3rd generation or 1st generation

cephalosporins) associated with the regulation (51). This study shows that DDD is

sensitive to a measure change of drug use overtime and offer the detection of the effect of

the health policy on drug utilization of the country using sales data.

Defined Daily Dose in drug utilization review study (clinical analysis)

Problems of drug use addressed by the IOM in the report in 1998 are well-known

as overuse, underuse, and misuse (52). Can DDD be applied to find the evidence

regarding problems of drug use? The selected examples using the DDD system below

addressed appropriateness of drug use in different countries.

The first example is the study of misuse of sumatriptan by measuring number of

DDD prescribed from a prescription database in Denmark by Gaist et al.(53). Second









example is a study of appropriateness of DM drug use between settings, GPs and

specialists in two different geographical areas in Sweden (54). This method applied

actual prescribed dose compared the DDD to detect change over time and identify the

differences of prescribing practice in between the two towns. Another example is a

comparison of quinolone use in general population in long-term care facilities, and within

a single institution in the Netherlands. This is a good example of the DDD application

for exploratory research on appropriateness of drug use and prescribing behavior(55).

Alternatives measurements to the DDD

There are several units of drug utilization measurement available and have been

used in drug utilization studies. However, at this time the alternatives do not prove much

advantage over the DDDs. Five alternative measures of drug use are presented below

and offer some strength and weaknesses in different issues.














Table 2-1. Alternative units for measuring drug utilization


Alternative
Prescribed Daily Dose (PDD) (56)
Description:
A defined actual prescribed dosage by
prescribers in a selected geographical area or
setting
Minimum marketed dose (MMD)(57)
Description:
A minimum dose that will produce a desired
therapeutic effect, which is the minimum dose
marketed by the manufacturer








Therapeutic Course
(TC)(58)
Description:
Measurement of drug dose for the whole
course of treatment. This parameter is
accounted for time of drug therapy.


Advantage
Useful for a study of prescribing behavior
and therapeutic traditions
Useful for a study of morbidity prevalence
Reflect actual practice than can be compared
to the DDD
Same as the DDD












DDD combined with TC offer a more
reliable information about exposure to drug
use
Provide meaningful drug utilization measure
of how many people receive full course of
therapy (appropriate care),e.g., antibiotics


Disadvantage
Non-standardized measure make it difficult to compare across
settings, regions, countries, and overtime



Only reflect how many people received a minimum dose
Minimum dose is subject to change depends on the market, thus
it is not consistent for longitudinal comparison
Variations of minimum dose produced by different
manufacturers who produced the same drug make it difficult to
decide which one to use
Comparisons of drug use between countries are limited, because
of the unavailability of the minimum dosage of a drug in some
countries
Because of the limitations there is only little information
available in the literature to compare the results of a new study
with.
Length of some therapies may vary, e.g., infection becomes
chronic with complications that the set length of TC may not
apply.


_ __ __ __














Table 2-1. Continued
Alternative
Equipotential dose (ED)(59)
Description:
ED is created by two doctors in a Danish
research group for treatment of hypertension.
It is defined as "the amount of efficacious
substance in relation to a given amount of a
given drug, the amounts having the same
potential effect on the blood pressure".
Average Daily Dose (ADD), (German Drug
Index)(60;61)
Description:
Defined by pharmacologists in the German
Scientific Institute of General Health
Insurance according to German drug use
situation.


Advantage
Not obvious advantage over the DDD,
because the DDD is calculated based on
equipotency assumption.





Same concept as the DDD, but specific to
German drug therapy situation


Disadvantage
Defined by only two doctors, thus generalizability is
questioned.
Indicate treatment effort in terms of lowering blood pressure,
but does not indicate how many patients received the drugs.




No advantage over the DDD









Qualitative Approaches in Measuring Drug Utilization

To discuss the measurement methods of drug utilization quality, it is necessary to

understand how quality is defined from the level of quality of healthcare services to

quality of drug use in the literature. Then, the measurement tools will be discussed for

this application in quality improvement and drug safety studies.

Concept of quality of care and its measurements

In the US, since 1965 when the government offered Medicaid and Medicare

programs to the poor and older populations, the need to monitor and improve quality of

care has required most healthcare institutions to set a framework to assess and improve

quality of care.

Quality of care is a multifaceted concept. The most widely referred concept of

quality of care was introduced by Donabedian in 1978 (62), which suggests that quality

of care can be measured at three levels; health-related structure, process, and outcomes.

Health-related structure is described as factors related to healthcare systems and health

policy, such as the healthcare delivery system, population needs, or an economic

situation. Heath-related process is a measure of ability to access care financially and

non-financially, and a level of health risks. Lastly, the ultimate goal of healthcare quality

is to produce efficient and equitable of health status (health outcomes) (62-64).

There are six types of access to healthcare: potential, realized, equitable,

inequitable, effective, and efficient (65). The definitions of access to care and the types

of health policy that target each level of access are described in Table 2-2. Literature

studying access to care has emphasized that access to care is a relative term based on

healthcare needs (66;67). In 1993, the IOM redefined access as the timely use of

personal health services to achieve the best possible health outcomes(68). From this










definition, the interpretation of access to care is expanded to include patient's needs and

the consequent health outcomes.

In this study, three levels of access, potential access, realized access, and

inequitable access, are studied. This study intends to evaluate the effects of the health

policy that offer health benefits to increase healthcare service utilization (potential

access). Drug utilization rates are measured to determine realized access as a result of the

30-Baht HI policy. Prescribing appropriateness is the measure of effectiveness access to

drugs. Equitable access is measured by comparing drug utilization measures (rates and

quality) between two populations. Gaps of drug utilization rates and quality suggest

inequity (disparity) of access to drug therapy.


Table 2-2. Types, definitions, and policy purposes of access to healthcare services
Type Definition Policy Purpose
1. Potential access Healthcare system characteristics and To increase or decrease health service
enabling resources that influence use of use
health services
2. Realized access Use of health services To monitor and evaluate policies to
influence health service use
3. Equitable access Use of health services is determined by To ensure health services distribution is
demographic characteristics and need determined by need
4. Inequitable access Use of health services is determined by To reduce the influence of social
social characteristics and enabling characteristics and enabling resources
resources on health services distribution
5. Effective access Use of health services improve health To improve the outcomes (health status,
status or satisfaction satisfaction) from health services use
6. Efficient access Minimizes the cost of health services To minimize the costs of improving
use and maximizes health status or outcomes from health services use
satisfaction
Source: Adapted from Measuring access and trend. Introduction to health services, 5ed.(65)

Donabedian (63;63) further discussed the concept and the measurement of quality

of care in the subsequent publications and suggested that quality is composed of seven

attributes: 1) efficacy: the ability of care to improve health; 2) effectiveness: the actual

health improvement resulted from care; 3) efficiency: the greatest health improvement

with the lowest cost; 4) optimality: marginal health improvement compared with cost; 5)









acceptability: access to care, patient-practitioner relationship, convenience of care, patient

preference to effectiveness and cost of care; 6) legitimacy: value of care to the society;

and 7) equity: distribution of access and quality of care that is fair based on needs (63).

The evaluation of quality of drug therapy, healthcare product and service, must be

conducted based on these attributes.

However, in defining quality of drug therapy in this study the researcher focuses on

only two attributes: effectiveness and equity, more than the rest of the attributes, because

they are feasible to operationalize and measure with the hospital data sources, which the

researcher selected for the analyses of this study. Further more, other attributes are less

likely to be affected by the implementation of the 30-Baht health insurance policy.

Definitions of efficacy, effectiveness, and equity of drug therapy are described below.

Effectiveness

Effectiveness of drug therapy is the ability to improve health outcomes in patients

with a specific disease under the "real world" of everyday practice. The real world

situation may include heterogeneous patient's characteristics comparing to the patients in

the clinical trials (e.g., age, race, compromised renal function), low patient adherence

(e.g., patients stop taking the drugs, or miss doses), and inappropriate prescribing by

physicians resulting from certain conditions (e.g., limited choice of drug in the formulary,

limited resources, or lack of knowledge). The evidence of drug effectiveness is available

in randomized clinical trials and for a number of diseases. This evidence is summarized

as clinical treatment guidelines that can be used as a gold standard to evaluate the

effectiveness of drug use in the population of interest (63). However, there are some









limitations of using clinical treatment guidelines to support drug effectiveness (e.g., some

patients fail to respond the recommended drugs or dosage).

Equity

Equity is fair distribution of drug utilization and quality based on individual needs.

Individual needs include patient's disease condition, co-morbidities, and demographics

(63). Equity is a relative measure that can be measured by comparing drug utilization

rates and quality with those of other populationss. This means that patients with the

same disease state, and with similar co-morbidities and demographics, should receive

similar drug therapy.

The Global Health Equity Initiative (GHEI), a collaborative research network of

twenty countries addressing the increasing inequities in health, has defined equity as

"fairness" that healthcare services should be available and accessible as needed. This

definition could imply that needy (sicker or more vulnerable) groups within a society

require access to care at a higher level of resource consumption than those people who

have better health status. Then, the gap of equity of care will be reduced (69).

In Minnesota, Minnesota Health Care Commission applied the concept of equity in

providing universal health care coverage to the residents that set the goal as affordable,

accessible, and accountable healthcare for everyone. It addresses financial barriers and

non-financial barriers to access to care related to geography, culture, language, race,

transportation, and a shortage of providers (70). The provision of the 30-Baht HI policy

is different from the universal healthcare in Minnesota, as it did not include non-financial

factors. Thus, these factors remain influence on drug utilization.









Approaches in measuring appropriateness of drug use

Quality of drug use has been evaluated and studied by a method called drug

utilization review or drug utilization evaluation. Drug utilization review is defined as "an

authorized, structured process that reviews, analyzes, and interprets the pattern of drug

use in a given health care delivery system in relation to explicitly predetermined criteria,

guidelines or standards(71;72)." In general, quality is used interchangeably with

"appropriateness" in drug utilization review literature. Schmader and his colleagues have

defined the term "appropriateness" as "the selection of a medication and instructions for

use that agrees with accepted medical standards to provide safe, effective care(73)."

Assessment of quality is judged by appropriateness of drug use, including choice,

dosage, duration of therapy, drug administration, drug monitoring, and patient

compliance (74). In many cases, practitioners use their expert judgment for

appropriateness. On the other hand, researchers seek to apply methods that are more

objective and based on the evidence published in the literature. Some researchers used

categorical measures to evaluate the appropriateness in terms of whether a drug was

present or not (e.g., the percentage of patients who received antihypertensive medications

to control blood pressure). However, there are several reviews criticizing that it provides

insufficient information addressing the complexity of medication use (75-77).

As stated earlier, quality of drug use is defined as quality of prescribing in this

study because drug use quality is based on what physician prescribed in the hospital

environment. Three components of quality of prescribing are selected for the assessment

in this study: choices of drug, dosage, and duration of therapy, as these components are

addressed by the WHO regarding quality of prescribing. The WHO has defined quality

of prescribing as "prescribing the right drug, with the correct dose, duration, and drug









form for the right indication, with adequate information and instruction to the right

person in accordance with co-morbidity and other medication used". In evaluating the

quality of prescribing, various data should be available, e.g., patients' demographics,

disease diagnosis, and detailed information of prescribed drugs (dose regimen, duration

of therapy), and reimbursement methods (78). In this study, these required data are

available in the administrative database of the selected hospitals.

Measuring quality of drug utilization using explicit criteria

In drug utilization studies, explicit criteria have been introduced to evaluate

prescribing appropriateness at a patient-level, mostly in inpatient populations, where a

usual amount of information related to drug use is available. These explicit criteria were

established based on current clinical knowledge, practicality, and the available data

sources. Extensive literature on drug appropriateness using explicit criteria exists in

elderly populations because of their vulnerability and complexity of care. An example of

an explicit criterion for elderly populations is an indicator that evaluates the

inappropriateness of benzodiazepine use.

Talerico (79) conducted a critique of six measures (Beer's criteria, Avorn index of

potentially inappropriate drug use, Medication Appropriateness Index, the Defined Daily

Dose, Panel Assessment for drug regimen, and Swedish medical product agency

guidelines) for their utility in assessing inappropriate psychoactive drug utilization in the

elderly. This critique applied six criteria to evaluate the measures' validity and

reliability: 1) indication for drug therapy, 2) the effectiveness of drug therapy, 3) correct

dosage adjusted for pharmacodynamic and pharmacokinetic changes in elderly, 4)

appropriate duration of therapy, 5) duplication of drug therapy, and 6) the risk of adverse

events.(79) The results from the review confirmed that most measures based the









appropriateness of drug use on the effectiveness or choice, dosage, and duration of

therapy. This finding indicates that the three criteria are important components in

assessing quality of drug use. Other criteria are also important, however, difficult to

measure without complete drug information and patient outcomes. Thus, other criteria

that are not used in this study where not discussed in this chapter.

A study by Owen et al.(80) estimated the sensitivity and specificity of explicit

criteria in assessing quality of antipsychotic drug use compared with implicit criteria as a

gold standard. The explicit criteria produced high sensitivity (84.6%), however,

relatively low specificity (71.7%). Using the explicit dose criterion may result in a

systematic overestimation of inappropriate dosing.

Medicaid claims data has been used to measure inappropriateness of psychotropic

drug use. However, the researchers suggest that the finding has a potential

overestimation of inappropriateness of psychotropic drug, because many filled

prescriptions are never taken, and the adherence with the prescribed drugs is often not

verified (81). Thus, prevalence of (in) appropriateness of drugs in the population based

on prescription refill data should be interpreted carefully.

Measuring quality of drug use by quality indicators (QIs)

Quality indicators have been introduced with the concept of quality improvement

for more than a decade(82). Almost every healthcare institution has adopted the concept

of quality improvement to be able to achieve standards set by accreditation organizations

and to compete in the healthcare market. To be able to measure quality of healthcare

services, quality indicators must be established. The Agency of Healthcare Research and

Quality (AHRQ), and Health Care Financing Administration (HCFA), the Joint

Commission on Accreditation of Healthcare Organization (JCAHO), and the Center of









Medicaid and Medicare Services (CMS) are responsible to provide guidelines for quality

for each type of services, healthcare institutions, and populations.

The AHRQ is responsible for creating tools and resources that help healthcare

providers and their institutions to provide high quality of care that is safe, accessible and

affordable to the patients. These tools include surveys of medical expenditure, and

healthcare utilization, questionnaires to assess patients' experience in healthcare services,

and quality measurement tools to assess clinical performance, including the AHRQ

Quality Indicators (QIs).

The AHRQ developed three sets of QIs to identify specific areas that are potentially

problematic in hospital services: prevention, inpatient care, and patient safety. These QIs

are specifically developed to allow the assessment of quality using hospital inpatient

administrative data. The following discussion focuses only on inpatient care quality

indicators that are applied to this current study. Process of developing the indicators and

the criteria of the selection of quality indicators are described.

The AHRQ inpatient QIs have been developed based on the Healthcare Cost and

Utilization Project (HCUP) QIs, which are consisted of 33 clinical performance

measures. The HCUP QIs covers three dimensions of care:

1. Potentially avoidable adverse hospital outcomes, e.g., measurement of mortality rates
among low-risk patients receiving common procedures, complication rates during
hospitalizations, e.g. UTIs

2. Potentially inappropriate utilization of hospital procedures, e.g., overuse or underuse
of a certain healthcare services (cesarean section)

3. Potentially avoidable hospital admissions, e.g., immunization rates to prevent
pneumonia in elderly population

However, there are several limitations of measuring quality of care using the HCUP

QIs. These indicators do not include considerations of severity of risk adjustment; the









denominators of the measurement were based on hospital discharges rather than

populations; and the indicators mainly focus on surgical procedures and did not represent

chronic diseases or pediatric illnesses (83).

The revised Inpatient QIs by the AHRQ include the following areas: volume of

healthcare utilization (esophageal resection volume), mortality indicators for inpatient

procedures, and mortality indicators for inpatient conditions, utilization rates Healthcaree

services overuse, underuse or misuse, e.g., cesarean delivery rate) (84). The

measurements of quality related to drug use mandated to report to the Congress, are the

percentages of persons with outpatient visit and the percentage of persons with

prescription drugs.

It is important to understand what criteria are used to establish those indicators to

be able to justify the selection of the QIs in measuring the quality or appropriateness of

drug utilization in this study.

Reliability and validity criteria for QI selection

Quality indicators to measure quality should be valid and reliable. The AHRQ has

used six criteria (84) to evaluate the validity and reliability of the established quality

indicators as follows:

1. Face validity: the QIs must be established based on sound clinical evidence from the
literature and must be able to indicate quality aspects that apply to providers or the
healthcare system.

2. Precision: the QIs should be able to capture the quality across populations (e.g., types
of providers or health plans)

3. Minimum bias: the QIs should take into account of the disease severity and patient
co-morbidity. The differences of quality must not be biased because of different
patients' disease-related characteristics.

4. Construct validity: the QIs should be consistent with other QIs measures that evaluate
the same aspect of quality. An example for construct validity is that prescribing of









beta-blocker after myocardial infarction should correlate to reduction of
cardiovascular mortality.

5. Fosters real quality improvement: the QIs should facilitate the implementation of
quality improvement programs.

6. Application: the QIs should be able to implement with other indicators and together
provide broader picture of the quality.

Proposed measures of the three targeted quality criteria: efficacy, effectiveness, and

equity will be described, and the operationalization discussed.

Challenges of Measuring Quality of Drug Use

Validity of the measures

In our study, effectiveness of drug therapy will be measured based on only three

health-related-process components: drug choice, dosage, and duration of the therapy.

Other levels of health-related processes are not measured, e.g., severity of the diseases

that affects choices, dosage, and duration of therapy. In addition, the measurement of

quality neither includes health-related structure (e.g., availability of drugs in the hospital

formulary), nor health-related outcomes (e.g., lengths of stay in the hospital, adverse drug

events, and mortality).

Secondly, this unit of measurement of quality of drug use in the unit of proportions

of patients receiving effective drugs based on the US clinical treatment guidelines from

reputable sources may not have some limitations, in the situation that physician

compliance to the guidelines is low. Standard treatment guidelines are not well-

established or well-adopted from the practitioners in primary care hospital level in

Thailand. It seems unfair to use standard dose, duration, and choice of therapy of some

institutions to measure quality of care before introducing them prior to the measurement

of the quality in their settings. However, modern practice of medicine came from









evidence-based that the researcher would assume that the practitioners learn this concept

from medical school, and should update their practice based on current published

evidence. From this reason, variation of effectiveness of drug therapy may be observed

from different hospitals.

Reliability of the measures

The measurement of quality of drug use that is reliable should allow the

comparison of drug use overtime. The measurement unit is not able to capture the

change of effectiveness of drug use over time, if the standard of quality changes (e.g.,

new drugs are approved and recommended by the guideline, or dosage recommendation

is changed). It is important to clarify the standard of effectiveness of drug use when

comparison over time is conducted.

Sensitivity and specificity

Since effectiveness of a drug will be concluded only if all three components

(choice, dosage, and duration of therapy) are correct based on the guideline the measure

have high specificity, but low sensitivity. By this, it means the measure has ability to

capture drug ineffectiveness better than drug effectiveness (i.e., patient received a correct

choice of antibiotic and dosage, but the duration was one day shorter to the

recommendation, it will be categorized as ineffective, however patients might recover as

well as receiving full course of therapy).

Validating Computerized Administrative Databases

Administrative databases have been collected primarily for reimbursement

purposes. Additionally, clinical data in these databases are collected in the electronic

format that facilitates data extraction and statistical analyses. For these reasons, these

databases have become one of the main sources of drug and disease data for









pharmacoepidemiologic research and health policy evaluations(85). The characteristics

of the databases differ depending on the size of the data, timeliness of the data (e.g., how

recent the data is available in the databases), number of variables collected in the

databases, and types of institution that provide healthcare services (e.g., hospital

databases, community pharmacy databases, and claims databases). Most administrative

databases contain patient demographics (e.g., age, gender, marital status, disease

diagnosis, drugs, hospital charges, and health insurance status), which are necessary for

the charge reimbursement and suffice to identify disease condition and drug utilization

for health policy evaluations. Van Eijk et al. suggested in his article about data

requirement for research in 2001 that the data should be accurate and in computerized

and standardized format. In addition, record linkage should be unique and easy to link

the data on patients' characteristics, medical, and prescription drug data (78).

Examples of administrative databases that have been used in

pharmacoepidemiologic and drug utilization research and quality assurance purposes are:

Medicaid Management Information System (38;42;72;76;85-128), the Group Health

Cooperative of Puget Sound(42; 129), the Manitoba Health Service Commission(39; 130-

142), and the Medicare database (94; 143-148).

Data Validation Methods

Data validation methods widely used to validate healthcare data are the external

and internal data validation. External data validation means the data is compared with

other data sources that contain the same data. The comparison data source for external

data validation should be accurate and, thus, is considered the gold standard database.

However, when a gold standard or a validated database is not available, internal

validation methods are used for evaluation. Internal validation method is basically a









cross-check within the same database, or longitudinal comparisons to identify any

inconsistency of the data that are unexpected (e.g., downward peak of prescription drug

claims in a particular month). Methods of data validation, measurements of data quality,

and information from the literature regarding data validation from administrative

databases are discussed below.

External data validation

West et al. describes three major quantitative methods of measurement errors of

the data in a database that is compared with one or more other data sources: 1) reliability,

2) validity, and 3) agreement (149). The reliability measure is used when the same

source of the data is used more than once for the same information on the same person.

An example of reliability is the consistency of blood pressures in repeated measurements

of an individual.

Validity or accuracy of the data can be assessed by comparing specific data from

one database to a superior sources) that is considered the gold standard. There are two

measures for validity testing, sensitivity and specificity. Sensitivity or completeness is

used to represent the extent to which the studied database correctly identifies individuals

who have the characteristics of interest (e.g., benzodiazepine users, or patients with

diabetes) compared with the true occurrence in the gold standard source. Specificity

measures the extent to which the studied database correctly identifies individuals who do

not have the characteristics of interest (e.g., elderly who never used benzodiazepines, or

patients without diabetes) (149). (Figure 2-5)

In general, data that has high sensitivity are more likely to have low specificity

(149). West et al. suggests that absolute values of these measures can be falsely

interpreted. For example, if the true prevalence of insulin use is 5%, and by using claims









data with a specificity of 95% (and 100% sensitivity) the prevalence will double to 10%

(149). For these reasons, the benchmark (or expected values) of the sensitivity and

specificity should be carefully set and interpreted.


Figure 2-4. Formulas for calculating sensitivity, specificity

Adapted from Pharmacoepidemiology, 3rd ed. Strom BL, 2000.(45)

Other terms that have been used in research to identify the "performance" of a

database are positive predictive values (PPV) and negative predictive values (NPV).

Validity cannot be concluded with these measurements because they are calculated using

the denominator from the studied database (not from the gold standard). (Figure 2-4) The

PPV and the NPV rely on sensitivity and specificity (or validity) of the data source and

the prevalence of the exposure or the outcome of interest. For example, even if the

database has the same validity of drug exposure in two populations, the PPV and NPV


Gold standard


Present Absent

Present A B
Database True positive False positive ml
C D
Absent False negative True negative m2


n1 n2 N


Sensitivity = A/nl
Specificity = D/n2
Positive predictive value = A/mi
Negative predictive value = D/m2









will be different if the prevalence of drug exposure in the two populations is not

similar(149).

In this study, percent agreement will be used to identify validity of the diagnosis

and prescription drug data. An example of assessing validity of diagnosis data for

pneumonia from the HI database is to compare the ICD-10 code of J13 and J15 with

prescribed antibiotics (in this database bacterial culture results are not available for

comparison). Using the same method, the reliability of prescribed drug data can be

assessed using the diagnosis codes for any bacterial infections that are sensitive to the

prescribed antibiotics.

Disease diagnosis code

Diagnostic codes that are found in databases are either the International Code of

Diagnosis (ICD) version 9 or 10 (150). The first concern in using the ICD Codes to

identify patients with the disease is whether the ICD Codes are accurately and completely

assigned to the corresponding clinical conditions (151;152). The ICD code system does

not provide any disease description (i.e., standardized diagnostic criteria), which might

produce misclassification of the disease. Diagnosis coding accuracy varies depending on

various factors, e.g., institutional policy of coding for reimbursement purposes, whether

the disease is easy to diagnose or not, incentive for coding, and how well the data coding

personnel are trained (150).

Examples of disease diagnosis coding bias were found in a study by Hsia et al. in

1988, when data accuracy was tested by comparing the claim data with the diagnosis

from original medical records in Medicare patients in 239 hospitals The results show that

20.8% of diagnosis codes was not accurate, and 61.7% of these discrepancies financially









favored the hospitals (153). The study was repeated in 1988 (154), the results showed the

same problem that 14.7% of records contained errors that altered the diagnosis-related

groups; 50.7% of these errors financially benefited hospitals.

The validity of the ICD coding varies across disease states. Payne et al. addresses

that the ICD-9-CM system is not able to capture several clinical problems in outpatient

settings and important functional, socioeconomic, and psychosocial factors. For example,

Alzheimer's disease and related dementias (ADRD) were under-coded in Medicaid,

Medicare and managed care populations (155).

Generally, administrative databases are designed to contain a limited number of

slots for assignment of ICD codes, which might be sufficient for uncomplicated cases,

but inadequate for patient with multiple complications. However, the studies of validity

of disease diagnosis codes in Medicare population (156; 157) suggest an increase in the

number of slots for diagnosis and procedure coding might not assure the improvement in

completeness of the data.

Medicare claims data has been validated for different purposes, e.g., measurement

of drug utilization, disease prevalence, and adverse drug event. Buchmueller validated

the Medicare data for measuring tumor stage by comparing with SEER data as a gold

standard. For the inpatient population, the diagnosis (ICD-9-CM) (158) was obtained

from the Medicare Provider Analysis and Review (MEDPAR) files, while the data for

outpatients were extracted from Medicare claim database. Sensitivity was higher in the

inpatient than the outpatient population. Sensitivity was lower in regional disease than

distant disease. Positive predictive values varied with the affected organs (38.3%-

84.8%). The sensitivity and PPV values were never simultaneously 80% within one









stage of a specific cancer. The findings suggest that Medicare claims data have limited

utility for the study of cancer stage due to the high rate of misclassification (143).

Another data source that has been used in research is the Group Health Cooperative

of Puget Sound (GHC) database. The automated data of GHC was validated to identify

complications and co-morbidities of diabetes patients by comparing the ICD-9-CM and

Current Procedural Terminology codes with medical chart data (gold standard) (129).

The overall sensitivity of diabetes complications was high, but varied by each

complication (79.2% for ostomyelitis to 95.2% for myocardial infarction). The PPV was

low (mean 46.3%, [8.6%-88.5%], except for amputation (82.9%). Even though,

sensitivity was acceptable to detect the complications and co-morbidities of diabetes, the

overall PPV was low.

Prescription drug data

Completeness (or accuracy) of prescription drug data in administrative databases

varies depending upon various factors (e.g., whether it is voluntary data collection or

mandatory) and the complication of data submission.

A study by Kozyrskyj et al. (138), the Drug Programs Information Network (DPIN)

electronic prescription claims database in Manitoba, Canada, confirmed that voluntary

data collection is prone to data incompleteness. The completeness was assessed by

comparing the prescription number from the DPIN with the original pharmacy records.

The results showed that prescriptions reimbursed from the province's drug befit plan,

Pharmacare, were 93% (98%CI 92.4% to 93.6%) with the original records. The

completeness of prescription were lower for the treaty status Indians (Indian Affairs) and

social assistant recipients (Manitoba Family Services/City of Winnipeg Social Services)









compared to the Pharmacare group (79.7%, 98% CI 78.0% to 81.4% and 90.1%, 98% CI

88.8% to 91.4%, respectively). The study found that the completeness of the data was

varied by type of drug benefit plans, where the pharmacists receive drug reimbursement

from Pharmacare based solely on the data entered in the DPIN, while the drug data for

Indian Affairs and the social assistant recipients were voluntarily collected for drug

utilization evaluation purposes (138).

McKenzie et al. validated the accuracy of Medicaid pharmacy claims for estimating

drug use among elderly nursing home residents in Oregon by comparing drug data with

medical charts found that the percent agreement and PPV were above 85% for

antipsychotics, antidepressants, and anxiolytics (kappa = 0.81, 0.63, and .0.52,

respectively) (107). The findings suggest that Medicaid pharmacy claim data are

accurate (107).

Since the computerized healthcare database in this study is used in place of paper

patient records, external data validation (comparing the data from the HI database with

other data sources) is not possible. Thus, internal validation methods will be used.

Internal validation methods

Internal validation of the data has been used widely for disease incidence(159),

diagnosis confirmation(151), and adverse drug event measurement(97; 160-162). The

researchers validated specific data, e.g., disease diagnosis (ICD-9-CM) codes by

comparing with other data that represent the disease such as laboratory results or

prescription drug data. For example, the ICD-10 for community-acquired pneumonia can

be validated with positive results of the blood culture for Streptococcus pneumoniae.

Gerstman et al. used prescription drugs (e.g., anticoagulants) with the disease

diagnosis (deep vein thrombosis or pulmonary embolism) to identify patients with venous









thromboemobolism (163). Movig et al. validated the ICD-9 codes for hyponatremia

(276.1) from the Dutch PHARMO record linkage database by comparing them to

laboratory serum sodium (Na+) measurements. The study found that sensitivity was low

and only 30% of the cases were found using the ICD-9 compared to the Na+

measurements. However, specificity was high (>99%). (151)

Hennessy and his colleagues conducted a macro-level data quality assessment of

Medicaid data, the Computerized On-line Medical Pharmaceutical Analysis and

Surveillance System (COMPASS) (164) from six states. The authors used descriptive

explanation of the data quality based on missing data, unusual presence of the disease in

patients with specific characteristics, and the data inconsistency over time periods. The

study classified potential data errors into 4 types: incomplete claims for certain time

periods; absences of an accurate indicator to identify patient hospitalizations; incomplete

hospitalization data for the beneficiaries; and diagnosis codes in demographic groups that

would be expected to have a low frequency of the diseases (99).

The study first examined missing data by comparing the number of enrollees and

the prescription claims over time periods. Secondly, the incompleteness of the data in a

suspect group (in this study elderly patients because Medicare data may be more

complete) compare with other age groups. Thirdly, the researchers examined the validity

of diagnosis and demographic data. The disease diagnoses were cross-checked with

patient characteristics: 1) the diseases that are specific to a certain age or gender, e.g.,

breast cancer in female, childbirth and pregnancy complications in patients younger than

60, and lung cancer in patients older than 40.









Data quality was identified by using graphs of the number of prescription drug

claims plotted longitudinally to indicate unusual data (upward or downward peaks).

Number of hospitalizations per enrollee was observed by age group. Other analyses were

to assess the accuracy of diagnosis code with patients' demographics. The level of the

graph and its consistency overtime indicate whether the data is valid and reliable, or not.

The macro-level data quality assessment methods used in this study suggests that

despite the comparison data is not available; it is possible to check internal data validity

using various data cross-checking and longitudinal comparisons.














CHAPTER 3
VALIDATING DATA QUALITY

The HI databases contain electronic medical records, pharmacy dispensing data,

and administrative information. The databases were designed by a physician to facilitate

a report of information required by the Ministry of Public Health, to generate disease

statistics, to monitor of healthcare utilization, financing, and lastly for quality

improvement purposes. Accordingly, these databases have comprehensive clinical and

administrative data.

The electronic medical records included patient demographics, type of health

insurance, disease diagnosis (ICD-10), and details of dispensed medications, some

physical examination information and laboratory test values (e.g., blood pressure,

virology and serology test results). However, laboratory data were incomplete and not

usable for this study. Drug names, drug classes, dosage, amount of drug supplies, and the

instruction for use were completely available in the pharmacy dispensing database.

The clinical information appears on the computer screen when healthcare staff

renders hospital services. Physicians are able to view past medical history, previous

diagnosis and treatment, including the medications that have been dispensed. Any new

clinical information obtained during a patient encounter is immediately updated in the

database system. Once the physicians prescribe medications to the patients, the order is

sent electronically to the pharmacy department of the hospital and, as a result, most

prescriptions are dispensed at the hospital pharmacy. Moreover, only prescriptions that

are dispensed at the hospital pharmacy are covered by most health insurance plans,









except the CSMBS group while provide retrospective reimbursement for the prescriptions

dispensed at retail pharmacies. Most CSMBS patients fill prescriptions at the hospital

pharmacy, if the drugs were available in the hospital drug formulary. Because dispensing

activities are directly limited to the electronic record, we assumed that drug dispensing

data from these hospital databases represent hospital drug utilization for both inpatients

and outpatients.

Since the HI database system was implemented in the hospitals, paper medical

records are no longer used. Healthcare providers at the hospitals rely solely on these

electronic medical records for clinical intervention and administrative purposes. For this

reason, data validation could not be performed by comparing the electronic medical

records and pharmacy database with the medical chart in paper format.

The databases were selected for this study because they are able to capture most

aspects of care provided by the community government hospitals, while are the main

provider of primary and secondary care to the 30 HI beneficiaries. By applying the

patient-specific healthcare data from the 19 government hospitals in Ubonratchatani

province, the researcher was able to measure drug utilization covered by the 30-Baht HI

policy at the provincial level.

The researcher used two steps to obtain the datasets from the target hospitals: 1) an

official letter for data request to the Ministry of Public Health, Thailand, who has

authority to suggest the hospitals participate in the research project and to allow the data

to be used for research purposes, 2) interviews wtih hospitals directors and database

managers regarding data quality. The request was approved by the Minister of Public

Health, and the approval communicated to the 19 targeted hospitals in December 2003.









Fourteen of 19 hospitals agreed to provide the data. Two hospitals did not agree to

provide the data, two hospitals did not use the HI database, and one is a tertiary care

hospital that use different database system. One tertiary care hospital in Ubonratchatani

was not included in this study because it was not the main provider of primary care of the

patients covered by 30-Baht HI policy. Officially, services by tertiary care hospitals are

only covered under the 30-Baht HI policy if patients are referred from the community

government hospitals. The HI Database System was used only in the community

government hospitals in Ubonratchatani province.

Healthcare data extracted for data validation analysis in this study were: 1) patient

demographics, 2) disease diagnoses and, 3) dispensed drugs. These data were the core

elements to this study in identifying the effect of the 30-Baht HI policy on hospital drug

utilization in the affected population and in the control group. The data in the HI

database were linked with a unique patient identification number (HN) assigned when

patients first visited the hospital. For demographic data, the researcher included age,

gender, marital status, and job based on the data availability. Disease diagnoses were

extracted separately for inpatient and outpatient services using discharge diagnosis codes.

To be able to understand, interpret and generalize the results of this study, the

researcher interviewed experts that were involved with the HI database regarding data

quality, and the healthcare providers from the hospitals who used the database and were

experienced in patient care. The interview had two major purposes: 1) to explore the

possible errors from data collection, data coding, and data entry, and 2) to identify

extraneous factors that could compromise internal validity of the effect of the policy on

drug use. Details of each expert interview were described as follows:









Interview of Hospital Directors

Interviews with hospital directors of 8 community hospitals were conducted to

gather information of how well the data in the databases represent the actual clinical

practice, clinical interventions, and the treatment. The researcher interviewed one

hospital director of each hospital. If the hospital director was not available, another

knowledgeable practitioner in the hospital was interviewed. The following issues were

discussed:

* Completeness and accuracy of patient-specific data in the database

* Factors affecting prescribing behaviors that might have an effect on drug utilization
rates rather than or in addition to the 30-Baht HI policy (choice of drugs, amount of
medical supplies, hospital admission decision);disease severity, type of patients,
(health insurance benefit), drug price, or physical factors (physical factors, i.e.,
clinic schedule, census)

Interviews of Hospital Database Managers

To explore the validity of the data, interviews of database managers were

conducted to determine the coding system, relationship of the databases (i.e., links

between medical database and pharmacy database), limitations of the software, similarity

of the database between hospitals, and known problems reported from the current users.

Completeness of the data, potential information bias (e.g., from coding errors), frequent

missing data, and other possible factors that lead to lower quality of the data was also

gathered. Finally, eight hospital databases from 19 community government hospitals

were selected for data validation based on data availability and the qualitative

information provided by the hospital directors and the database managers.

Quantitative Assessment of Data Quality

The assessment of the quality of the hospital data was extensively conducted

because the data had not been used for drug utilization research purposes. Quantitative









assessment of the data was performed in regards to four major characteristics: descriptive

analysis of the HI database and data characteristics, face validity, missing data, and

coherence of the data.

If the data quality did not met the criteria specified in each validation process, the

data from that hospital was excluded from the study. Details of how to handle the data

problems and the data exclusion criteria are discussed below.

Database Characteristics

Initially, the researcher explored the data structure and located patient-specific data

on demographics, disease diagnoses, types of health insurance plan, and dispensed drugs.

Next, the linkage was established in the HI database among patient demographics,

disease diagnoses, types of health insurance, and dispensed drugs.

In this study, the database was designed to link patient demographics and diagnosis

data by medical record number and hospital admission number. However, the diagnosis

and prescription drug data from the pharmacy database was linked by prescription

number. The differences of the record linkage posed some difficulties in combining the

data.

Missing Data and Outliers

Descriptive statistics were generated to present types and magnitude of missing

data in the databases. Missing data from the following variables were examined: patient

demographics, types of health insurance, disease diagnosis codes (ICD-10), lists of

prescription drugs, volume of drug dispensed, and directions of drug use. For ICD-10,

the cases that contained missing diagnosis codes were excluded from the data analysis.

Unusual and outlier values resulting from data entry error of patient demographics,

disease diagnosis codes, and prescription drugs and doses were identified based on the









clinical knowledge and experience of the researcher. If unusual values appeared to occur

at random and did not exceed 5% of the total number of values for a given variable, they

were deleted from the dataset (107). If the unusual values exceeded 5%, the whole

database of the hospital would have been excluded from the study, but this was not

necessary.

Face Validity (Plausibility of the Data)

If data were similar among the studied hospitals and the studied years the

researcher assumed that the data were valid and could be combined among the hospitals

for further analysis of the policy. The assessment of the plausibility of the data focused

only on the five most prevalent drugs that were prescribed in the studied period. The five

most prevalent drugs were compared among hospitals, and studied years. Secondly, the

five most prevalent drugs were compared with disease prevalence/incidence provided by

the provincial disease statistics. If the five most dispensed drugs were consistent with the

diseases epidemiology, it was assumed that the hospital drug data was valid.

Data Coherence

The researcher applied the internal data validation method to assess validity of the

data recommended by Hennessy et al., when external validation was not possible.

Hennessy and his colleagues conducted a macro-level data quality assessment of

Medicaid data, the Computerized On-line Medical Pharmaceutical Analysis and

Surveillance System (COMPASS) (164) from six states. The authors used descriptive

explanation of the data quality based on missing data, unusual presence of the disease in

patients with specific characteristics, and the data inconsistency over time periods. The

study classified potential data errors into 4 types: incomplete claims for certain time

periods; absences of an accurate indicator to identify patient hospitalizations; incomplete









hospitalization data for the beneficiaries; and diagnosis codes in demographic groups that

would be expected to have a low frequency of the diseases. (99)

The study first examined missing data by comparing the number of enrollees and

the prescription claims over time periods. Secondly, the incompleteness of the data in a

suspect group (in this study elderly patients because U.S. Medicare data may be more

complete) was compared with other age groups. Thirdly, the researchers examined the

validity of diagnosis and demographic data. The disease diagnoses were cross-checked

with patient characteristics: 1) the diseases that are specific to a certain age or gender,

e.g., breast cancer in female, childbirth and pregnancy complications in patients younger

than 60, and lung cancer in patients older than 40.

Data quality was identified by using graphs of the number of prescription drug

claims plotted longitudinally to indicate unusual data (upward or downward peaks).

Number of hospitalizations per enrollee was observed by age group. Other analyses

concerned the accuracy of diagnosis code with patients' demographics.

The macro-level data quality assessment methods used in this study suggests that

even though the comparison data was not available it was possible to check data validity

using various internal data cross-checking and longitudinal comparisons.

The data was considered valid if the percent coherence was equal to or greater than

80% (113). Coherence was measured as a percent agreement between two related

variables, e.g., a particular disease diagnosis and drugs dispensed to treat the disease, or a

gender-specific disease and the corresponding gender.









Validating disease diagnosis codes (ICD-10)

Coherence between ICD-10 and patient demographics

The percent coherence was observed across the studied hospitals and among the

studied years. Graphic plots of the percent coherence of each hospital by study year were

generated and compared with the plots from other hospitals. Coherence between a

gender-specific disease and gender was checked for the following conditions:

* Pregnancy x female
* Prostate cancer x male

Coherence between ICD 10 and dispensed drugs

Three disease diagnosis codes were cross-checked against the drugs usually used to

treat these diseases, including diabetes and antidiabetic drugs, hypertension and

antihypertensive drugs, and bacterial pneumonia and antibiotics. Gastrointestinal tract

infections were not validated because there are varieties of treatment methods depending

on the pathogens, severity and the onset of type symptoms. Coherence level was

cautiously interpreted because it was not only associated with reliability of the data, but

also with the appropriateness of drug prescribing. Even though, the measure was not an

exact measure of validity, it provided valuable information about reliability, in the

absence of a gold standard.

* Diabetes ((ICD10=Elxx) x all antidiabetes drugs in the hospital drug formulary

* Hypertension (ICD10 = Ilxx) x all antihypertensive drugs in the hospital drug
formulary

* Bacterial pneumonia (ICD10= J13x or J15x or J16x or J17x or J18x) x all
prescribed antibiotics that were effective for the treatment of pneumonia






57


Drug data

To identify coherence of drug data, we cross-checked the antidiabetic drugs with

the diagnosis codes of diabetes. The researcher calculated the percent of unmatched data

between drugs and the diabetes diagnosis codes, i.e., the percentage of patients who were

prescribed insulin without a diagnosis code of diabetes. The percent coherence and the

percent mismatched was summarized in Chapter 4.














CHAPTER 4
DATA VALIDATION RESULTS

Hospitals Demographics

The HI database system is used in 16 out of 18 community government hospitals in

Ubonratchatani province (2004).(Table 4-1) The HI databases were installed in the

hospitals at different times depending on the financial and staff readiness to implement

this database system. The completeness and accuracy of the data appeared to be

associated with the length of time the databases have been used.

Of 18 community government community hospitals in Ubonratchatani province,

Thailand, eight hospitals had complete data for disease diagnosis, dispensed drugs, and

types of health insurance available for the entire study period, six hospitals had

incomplete data in 2000, two hospitals did not use the HI database system, and two

hospitals refused to provide the data for our study. Locations of the eight included

hospitals suggest good geographic representation of the government community hospitals

in Ubonratchatani province. (Figure 4-1)

Descriptive information of hospital service utilization, including the number of

medical records in the studied period, hospital visits, hospital admissions, and

prescription drugs is shown in Table 4-1. The capacity of the government community

hospitals varied from 30 to 90 beds. The medical record data or censuses ranged from

36,572 to 179,986 per hospital. This information corresponds with the size of the

hospitals. The numbers of hospital visits, hospital admissions, and prescriptions during

the study period also appeared to correlate with the size of the hospitals.
















































Figure 4-1.


Geographic locations of the included eight community government hospitals
in Ubonratchatani province, Thailand


Table 4-1. Hospital information, numbers of hospital visits, hospital admissions, and
prescriptions from 2000 to 2003
Hospital Number Start HI Number of Number of Number of Number of
of beds database medical hospital visits hospital prescriptions
(year) records admissions
A 30 1999 50,840 264,097 12,832 564,058
B 60 1998 108,131 494,615 31,294 1,147,042
C 30 1999 62,990 254,670 11,766 579,203
D 30 1998 45,075 227,583 11,296 573,720
E 90 1999 179,986 711,957 46,418 1,770,982
F 60 1997 166,116 640,928 34,188 1,373,993
G 60 1998 36,572 568,611 31,839 1,368,213
H 30 1999 63,788 410,585 24,152 902,064










Healthcare providers were able to enter up to 16 diagnosis codes in the HI

database. However in most instances, one to two disease diagnosis codes were used at

each visit/admission. There were higher percentages of patients assigned with two or

more diagnosis codes in inpatient than outpatient settings. For outpatient services, the

percent of patients assigned two or more diagnosis codes were 8.3% in 2000 and

increased to 19.6% in 2003, while the percentage increased in inpatients from 10.03% to

49.3%. (Figure 4-2) The hospital directors commented in the interviews that at the

beginning of the implementation of the HI database system, there was a dual (paper and

electronic) system to keep medical records until the systems were considered stable.

Because of limitations in numbers of data entry personnel, only data that were important

for patient care and billing were entered. In addition, physicians and the data entry

personnel were not familiar with the ICD-10 lists in the databases, thus, it required more

time to enter more diagnosis codes into the database. An increased number of diagnosis

codes does not suggest an increase in co-morbidity, but rather increasing

comprehensiveness in coding as personnel became more familiar with the system.









2000 2001 2002 2003
Year
outpatient impatient


Figure 4-2. Percent of patients with two or more disease diagnosis codes assigned in the
HI database (Hospital F)









Database Characteristics

Healthcare data used for our study were stored in four main databases: 1) TBL

containing table files, e.g., data coding sheets, and patient demographics, 2) IPD

containing data related to inpatient services, 3) OPD containing data related to outpatient

services, and 4) PHM containing dispensed drug data. Patient demographics, types of

health insurance, disease diagnosis, and dispensed drug data were linked using unique

identification numbers (in our study, the identification numbers were deidentified by the

hospitals). (Figure 4-3)

* "TBL" folder containing the following tables:

o PT table containing patient demographic data
o Code book tables for the data in HI database

* "IPD" folder:

o IPTyear containing admission date and time, types of health insurance,
and length of hospital stay
o IPTDXyear: hospital admission date and time, and disease diagnoses

* "OPD" folder:

o OVSTyear: hospital visit date and time, types of health insurance
o OVSTDX: hospital visit date and time, and disease diagnoses

* "PHM" folder:

o PRSCyear: unique patient identifier, prescription number
o PRSCDTyear: prescription number and drug data

The HI database has been continuously updated, e.g., changing disease diagnosis

coding system from ICD-9 CM to ICD-10, adding drug codes for new drugs in the

hospital drug formulary. Additionally, the system was supplementedwith drug

information tools such as drug interaction database to improve quality of service. Based









on its elements, the HI database appears to be a comprehensive source of computerized

patient-specific healthcare data, including drug utilization.


Figure 4-3. Linkage among the selected patient data in the HI database AN: scrambled
hospital admission number; DRUGITEM: drug item code; HN: hospital
number (scrambled medical record number); ICD10: disease diagnosis code;
IPD: inpatient service folder; IPT: inpatient admission folder; IPTDX:
inpatient diagnosis folder; OPD: outpatient service folder; OVST: outpatient
visit folder; OVSTDX: outpatient diagnosis folder; PRSCNO: prescription
number; PHM: pharmacy folder; PRSCD: prescription and patient
information; PRSCDT: dispensed drugs; PTTYPE: type of health insurance;
REGISTERDATE: hospital admission date; REGISTERTIME: hospital
admission time; ; TBL: table folder; VISITDATE: hospital visit date; and
VISITTIME: hospital visit time;


PHM









Face Validity

Comparisons of dispensing frequencies suggested that the most frequent by

dispensed therapeutic categories were similar across hospitals.(Table 4-2) Analgesics,

antidiabetic drugs, antianxiety drugs, antibiotics, vitamin supplements, antihypertensive

drugs, and drugs for gastrointestinal disorders were most prevalent among the hospitals

and across the studied period. Analgesic drugs including acetaminophen in both tablet

and solution preparation, and its combination with muscle relaxants, e.g., orphenadrine.

NSAIDs (ibuprofen and indomethacin) were commonly found. The most commonly

dispensed antidiabetic drug was glibenclamide, while metformin was prevalent in only

some hospitals (6 out of 8 hospitals). Antianxiety drugs included diazepam in various

strengths (2-10 mg), furazepam and amitriptyline. In the antibiotic class, only

amoxicillin was found within the top five prescribed drugs. Vitamin B complex,

multivitamin, and ferrous sulfate were the most prevalent vitamin supplements. For

antihypertensive drugs, hydrochlorothiazide and enalapril were prescribed frequently.

Lastly, antacid and H2-receptor antagonist (ranitidine) were found within the five most

prevalent dispensed drugs across hospitals and over time.














Table 4-2. Five most prevalent dispensed drugs for inpatient and outpatient use of the eight included hospitals (A-H) during from


2000 to 2003
Hospital 2000
A Para* 500 Mg Tab
Para Syr. 120 Mg/5ml
Diazepam 2 Mg
Ibuprofen 400 Mg Tab.
Glyceryl Guiacolate Syr
B 2000
Paracetamol Tab. 500 Mg
Paracetamol Syr. 120 Mg/5ml
Para 450 +Orphenadrine 25 Mg
Bromhexine Tab. 8 Mg
Ascorbic Acid Tab. 100 Mg
C 2000
Paracetamol Tab. 500 Mg
Paracetamol Syr. 120 Mg/5ml
Diazepam Tab. 2mg
Antacid Suspension 240 Ml
Indomethacin Cap. 25 Mg
D 2000
Para 500 Mg Tab.
Diazepam 2 Mg
Multivitamin Tab
Par Syr. 120 Mg/5ml
Diazepam 5 Mg


2001
Para 500 Mg Tab
Para Syr. 120 Mg/5ml
Ibuprofen 400 Mg Tab.
Diazepam 2 Mg
Ammonium Carbonate Mixture
2001
Paracetamol Tab. 500 Mg
Ibuprofen Tab. 400 Mg
Para 450 +Orphenadrine 25 Mg
Paracetamol Syr. 120 Mg/5ml
Amoxycillin 125mg/5ml
2001
Paracetamol Tab. 500 Mg
Paracetamol Syr. 120 Mg/5ml
Vitamin B Complex Tab.
Diazepam Tab. 2mg
Antacid Suspension
2001
Para 500 Mg Tab.
Para Syr. 120 Mg/5ml
Diazepam 2 Mg
Multivitamin Tab
Diazepam 5 Mg


2002
Para 500 Mg Tab
Para Syr. 120 Mg/5ml
Diazepam 2 Mg
Mtv Coated Tab.
Para 325 Mg Tab.
2002
Paracetamol Tab. 500 Mg
Ibuprofen Tab. 400 Mg
Para 450 +Orphenadrine 25 Mg
Paracetamol Syr.120 Mg/5ml
Amoxycillin Cap. 500 Mg
2002
Paracetamol Tab. 500 Mg
Paracetamol Syr. 120 Mg/5ml
Vitamin B Complex Tab.
Diazepam Tab. 2mg
Antacid Suspension 240 Ml
2002
Para 500 Mg Tab.
Par Syr. 120 Mg/5ml
Ibuprofen 200 Mg Tab.
Balm
Ammonium Carbonate


2003
Para 500 Mg Tab
Para Syr. 120 Mg/5ml
Diazepam 2 Mg
Mtv Coated Tab.
Para 325 Mg Tab.
2003
Paracetamol Tab. 500 Mg
Ibuprofen Tab. 400 Mg
Para 450 +Orphenadrine 25 Mg
Paracetamol Syr. 120 Mg/5ml
Ranitidine 150 Mg Tab
2003
Paracetamol Tab.500 Mg
Paracetamol Syr. 120 Mg/5ml
Vitamin B Complex Tab.
Glibenclamide Tab. 5 Mg
Diazepam Tab. 2mg
2003
Para 500 Mg Tab.
Par Syr. 120 Mg/5ml
Norgesic
Balm
Ibuprofen 400 Mg














Table 4-2. Continued
Hospital 2000
E Paracetamol Tab. 500 Mg
Alumina And Magnesia Susp.
Diazepam Tab. 2mg
Diazepam Tab. 5mg
Ibuprofen Tab. 400 Mg
F 2000
Para 500 Mg Tab.
Glibenclamide 5 Mg
Hctz Tab. 50 Mg
Par Syr.120 Mg/5ml 60 Ml
Ammonium Carbonate
G 2000
Para 500 Mg Tab.
Diazepam 2 Mg
Par Syr.120 Mg/5ml 60 Ml
Ranitidine 150 Mg. Tab.
Furapam Tab. 500 Mg
H 2000
Paracetamol Tab. 500 Mg
Paracetamol Syr. 120 Mg/5ml
Paracetamol+Orphenadine
Antacid Suspension 240 Ml
Multivitamin Tab.


*PARA: paracetamol or acetaminophen


2001
Paracetamol Tab. 500 Mg
Diazepam Tab. 5mg
Diazepam Tab. 2mg
Alumina And Magnesia
Ibuprofen Tab. 400 Mg
2001
Para 500 Mg Tab.
Para Syr. 120 Mg/5ml 60 Ml
Glibenclamide 5 Mg
Ammonium Carbonate Mixture
Hctz Tab. 50 Mg
2001
Para 500 Mg Tab.
Diazepam 2 Mg
Ranitidine 150 Mg. Tab.
Para Syr. 120 Mg/5ml
Furapam Tab. 500 Mg
2001
Paracetamol Tab. 500 Mg
Paracetamol Syr. 120 Mg/5ml
Diazepam Tab. 2mg
Multivitamin Tab.
Antacid Suspension


2002
Paracetamol Tab. 500 Mg
Diazepam Tab. 5mg
Diazepam Tab. 2mg
Alumina And Magnesia
Ibuprofen Tab. 400 Mg
2002
Para 500 Mg Tab.
Para Syr. 120 Mg/5ml 60 Ml
Amoxy 250 Mg
Balm
Ammonium Carbonate
2002
Para 500 Mg Tab.
Diazepam 2 Mg
Para Syr. 120 Mg/5ml 60 Ml
Furapam Tab. 500 Mg
Ranitidine 150 Mg. Tab.
2002
Paracetamol Tab. 500 Mg
Paracetamol Syr. 120 Mg/5ml
Antacid Suspension 240 Ml
Multivitamin Tab.
Ibuprofen Tab. 400 Mg


2003
Paracetamol Tab. 500 Mg
Alumina And Magnesia Susp.
Diazepam Tab. 5mg
Diazepam Tab. 2mg
Paracet.500 +Orphenadrine
2003
Para 500 Mg Tab.
Para Syr. 120 Mg/5ml 60 Ml
Balm
Ranitidine 150mg
Amoxy 250 Mg
2003
Para 500 Mg Tab.
Diazepam 2 Mg
Para Syr. 120 Mg/5ml
Ranitidine 150 Mg. Tab
Furapam Tab. 500 Mg
2003
Paracetamol Tab. 500 Mg
Paracetamol Syr. 120 Mg/5ml
Ranitidine 150 Mg Tab.
Ibuprofen Tab. 400 Mg
Antacid Suspension










Drug utilization pattern were consistent with the data from Provincial Disease

Statistics Report in 2002 (165). (Table 4-3) The 10 most common causes of morbidity for

outpatient services include respiratory tract and gastrointestinal tract disorders.

musculoskeletal disorders, and infectious diseases.


Table 4-3. Provincial Disease Statistics (2002) on the 10 most common causes of
morbidity for outpatient services, Ubonratchatani province, Thailand
Cause of morbidity Number of cases Incidence/1,000 population
Respiratory tract disorders 857,007 483.22
Gastrointestinal tract disorders 572,429 322.76
Musculoskeletal system disorders 301,827 170.18
Infectious diseases 249,033 140.42
Endocrine disorders 230,662 130.06
Unindentified causes 213,488 120.37
Skin and connective tissue disorders 195,440 110.20
Urogenital disorders 168,024 94.74
Metal and behavioral disorders 161,917 91.30
Cardiovascular diseases 147,528 83.18

Source: Annual Epidemiological Surveillance Report 2002. Ministry of Public Health, Thailand (165)

The results suggest that the patterns of drug prescribing were not different among

the studied hospitals. Similar patterns of dispensed drugs across the studied years

indicate that validity of drug data was maintained over time. The congruence between

the most frequently dispensed drugs and disease incidences suggests that the data were

accurate.

Missing Data

Small numbers of missing data were observed in patient demographics, types of

health insurance, disease diagnosis codes for outpatient and inpatient service (defined as

presence of at least 1 disease diagnosis code), and prescription drugs among hospitals and

across the studied period. The data in Table 4-4 shows that age (date of birth) contains









0.02-0.22% missing data, while gender, marital status, and occupation have 0% missing,

with an exception of hospital A, C and D, which have a small percentage of missing data

in one or two characteristics. This suggests the demographic data were sufficient for

subgroup analyses.

There was no missing data on disease diagnosis of inpatient data in most

hospitals, except hospital F (0.02%) (Table 4-5). However, it is unknown whether all

disease diagnoses related to a patient's conditions were recorded in the database. There

were higher percentages of missing data in the outpatient dataset with a median of 1.83%

[0.05, 3.13] compared with the inpatient data. This result might suggest that missing data

was not specific to a particular hospital, but rather the process of data entry in outpatient

and inpatient care.

Pharmacy databases of the HI database systems contained small amounts of

missing data with a median of 0.04% [0.01, 0.08] for drug codes assigned to dispensed

drugs. The percent missing data on drug quantity were higher (0.25% vs. 0.04%).

Hospital A and G had complete data on drug quantity. (Table 4-6) Again, we could not

verify whether all dispensed drugs were entered into the database.

Overall, patient demographics, disease diagnosis, and drug data contained small

percentage of missing data, which was acceptable by common standards (<5%). The

results from the analysis of missing data suggest that the HI database had sufficient data

quality and could be used for the analyses in our study.













Table 4-4. Missing data on patient demographics Hospital Missing data
Data/Hospital A B C D E F G H
Numbers of patients N=50,840 N=108,117 N=62,974 N=4,900 N= 179,986 N= 166,116 N=36,572 N=63,788
Age 10 (0.02) 240 (0.22) 11 (0.02) 0 40 (0.02) 63 (0.04) 24 (0.06) 80 (0.12)
Gender 0 0 0 0 1 (<0.01) 0 0 0
Marital status 0 0 154 (0.24) 0 0 0 0 0
Occupation 2 (<0.01) 0 32 (0.05) 0 0 0 0 0

Table 4-5. Missing data on disease diagnosis codes for inpatient and outpatient data of eight studied hospitals (2000-2003)
Missing data (%)
Hospital A B C D E F G H
Inpatient 0/16,253 (0) 0/44,779 (0) 0/16,647 (0) 1/16043 (0) 2/63923 (0) 7/34461 (0.02) 0/41756 (0) 0/34323 (0)
ICD-10
Outpatient 140/264,086 11,940/443,989 7585/242,558 5874/305,427 13447/792,365 5143/470,923 9497/556,875 8130/407,121
ICD-10 (0.05) (2.69) (3.13) (1.92) (1.70) (1.09) (1.70) (2.00)

Table 4-6. Missing data on prescribed drugs of eight studied hospitals over four years (2000-2003)
Missing data (%)
Hospital A B C D E F G H
N= 564,058 N= 960,965 N= 429,641 N= 573,720 N= 1,770,976 N= 1,030,151 N= 1,368,213 N= 902,064
Drug name 279 (0.05) 255 (0.03) 152 (0.04) 164 (0.03) 652 (0.01) 601 (0.06) 1081 (0.08) 411 (0.05)
Prescribed 0 2573 (0.27) 2066 (0.48) 1433 (0.25) 3919 (0.22) 12248 (1.19) 2 (0) 2479 (0.03)
quantity









Data Coherence

Disease Diagnosis and Gender

The percent coherence between gender-specific diagnosis and gender was high for

pregnancy and female among eight hospitals over the studied period with a median of

98.89% [95.42, 99.86]. Inpatient data tended to have a higher percent coherence than

outpatient data (median 99.18% [96.63, 99.86] vs. 98.68% [94.74, 99.83]). The accuracy

remained high from the beginning of the study period to the end. There was no significant

change of the percent coherence over time.

Disease diagnosis with benign prostate hyperplasia and prostate conditions

validated with male gender yielded 100% accuracy in every hospital, except hospital E in

the year of 2000 (87.5%). There were two hospitals (Hospital A and D) that did not have

any cases of the above conditions in 2001 and 2002, thus the percent coherence could not

be calculated.

Disease Diagnosis and Drugs

The researcher validated the accuracy of the disease diagnosis codes by calculating

the percent of persons who were diagnosed with the disease and received medications.

The diagnosis codes for three disease states, diabetes, hypertension, and pneumonia, were

included in the validation of the accuracy of the diagnosis code. The disease diagnosis

codes for gastrointestinal tract infections were not validated because there were a variety

of drug therapy regimens for these conditions, and which in tern are not specific enough

to identify the percent coherence between the disease diagnosis codes and the drugs.

Diabetes and antidiabetic drugs

The type of antidiabetic drugs available in the hospital formulary varied among

the hospitals. However, every hospital included oral antidiabetics and insulin. Most of









the hospitals contained only glibenclamide in the sulfonylurea class. Chlorpropamide

were found in only 6 hospitals (except A and H). Glipizide and glyburide were found in

only 2 hospitals (Hospital B and G). In the biguanide group, only metformin was

included in the hospital drug formulary in all hospitals. Regular and intermediate-acting

insulin (NPH) were included in the drug formulary of every hospital.

Disease diagnosis codes and antidiabetic drugs showed a median percent

coherence of 89.81% [74.56, 97.71]. (Figure 4-4) The percent coherence of the inpatient

population was lower than the findings for the outpatient populations, 80.44% [65.20,

89.47] (Figure 4-5). The findings for both inpatients and outpatients appear reasonable as

a small proportion of diabetes patients is typically managed with dietary restrictions only.

In our study, the researcher used only ICD-10 codes to select patients with the

disease for the measurement of drug utilization and hospital visits/admission. Thus,

accuracy of the diseases diagnosis is crucial for patient identification of the disease.

Using disease diagnosis with low accuracy might underestimate drug use in a population

and impose misclassification bias.











100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00


2000


2001


2002


---A
-.- A
--- B
C
D
-*- E
F-- F
-G
- H


2003


Year



Figure 4-4. Percent data coherence between disease diagnosis of diabetes and
antidiabetic drugs of outpatient data among eight hospitals, 2000-2003


100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00


2000


2001


2002


-- -A
--- B
C
D
--- E
----F
-i-G
-H


2003


Year



Figure 4-5. Percent data coherence between disease diagnosis of diabetes and
antidiabetic drugs of inpatient data among eight hospitals, 2000-2003.


III";









Hypertension and antihypertensive drugs

Antihypertensive drugs included in every hospital formulary were diuretics, beta-

blockers, alpha-blockers, and ACEIs. Most of the drugs were diuretics, including

hydrochlorothiazide (HCTZ), furosemide, amiloride and the combination of the two

drugs (amiloride and HCTZ). Calcium channel blockers verapamill and felodipine) were

included in only three hospitals (B, E and H). There were limited choices of beta-

blockers propranololl and atenolol) and ACEIs (enalapril and lisinopril). However, they

were available in every hospital.

The median percent coherence between the diagnosis of hypertension (Ilxx) and

the above antihypertensive medication was 89.53% [80.81, 95.56] for the outpatient

population. Similar coherence levels were found across hospitals and studied years.

(Figure 4-6) For the inpatient population, the percent coherence varied with a median of

80.24% [58.46, 100] (Figure 4-7). Four hospitals showed lower coherence in 2000 and

2001 than in the following years.

A median of 10-20% of the patients who had a diagnosis of hypertension did not

receive any antihypertensive drugs, which appears reasonable as some patients may try

behavioral changes prior to drug treatment. Alternatively, patients may not be able to

afford drugs, or patients may have obtained antihypertensive drugs from other sources.

Therefore, it is not appropriate to conclude misclassification for diagnosis for

diabetes or hypertension. While the coherence levels appear realistic no final decision

about diagnosis accuracy can be made. Coherence levels varied more for outpatient data

when individual hospitals were compared, suggesting that either diagnosis coding

accuracy or the completeness of drug records may vary.













100.00

90.00

80.00

70.00

60.00

50.00

40.00

30.00

20.00

10.00

0.00


2000


2001


2002


-A--A

--- B

C

D

-- E

-F-- F

G

--H


2003


Year




Figure 4-6. Percent data coherence between disease diagnosis of hypertension and
antihypertensive drugs of outpatient data among eight hospitals, 2000-2003


100.00


90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00


2000


2001


2002


-.- A

--- B

C

D

--E

-.-- F

G

-H


2003


Year




Figure 4-7. Percent data coherence between disease diagnosis of hypertension and
antihypertensive drugs of inpatient data among eight hospitals, 2000-2003


_

-

-

-

-

-

-

-


~i--









Bacterial pneumonia and antibiotics

Since there was no official treatment guideline for the treatment of bacterial

pneumonia implemented in the community hospitals, we selected all antibiotics that had

an indication for the treatment of bacterial pneumonia based on Clinical Pharmacology

Online 2005 drug information database (166). These include beta-lactams (e.g.,

amoxicillin, penicillin, and cloxacillin), cepalosporins (e.g., cephazolin, ceftriazone),

macrolides (e.g., erythromycin, tetracycline, doxycycline, and roxithromycin),

fluoroquinolone (e.g., norfloxacin, ciprofloxacin), aminoglycosides (e.g., gentamicin),

sulfamethoxazole and trimethoprim, and lincomycin. The selected antibiotics had

indications for multiple infectious diseases and thus, were non-specific for the treatment

of bacterial pneumonia. Bacterial cultures were not commonly ordered in the community

hospitals, i.e., empirical treatment was common.

Most of the hospitals similar coherence levels between disease diagnosis of

bacterial pneumonia and antibiotics with a median of 92.52% [64.25, 100] for outpatient

service and 89.20%, [42.67, 100] for inpatient services. Hospital E had distinctively

lower coherence compared with other hospitals 72.08%, [64.25, 72.47] and 53.95%,

[42.67, 59.33] for outpatient and inpatient service, respectively.

All patients who were diagnosed with bacterial pneumonia should be given an

antibiotic. From this analysis, 7% of the outpatient population did not receive any

antibiotics. The number was higher in the inpatient population. These discrepancies

might be the results of coding errors. Thus, using the ICD-10 codes in identifying

patients with bacterial pneumonia approximately 10% of the patients may be falsify

classified as having pneumonia, or may have missing medication records.













100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00


2000


2001


2002


-.- A
--- B

C
D
E
-*- F

G
H-


2003


Year


Figure 4-8.






1'




e-

I-
e
S ,
c.
o
&


Percent data coherence between disease diagnosis of bacterial pneumonia
and antibiotics of outpatient data among eight hospitals, 2000-2003


00.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00


2000


2001


2002


----A
--- B

C
D
E
--F

-- G
-H


2003


Year


Figure 4-9. Percent data coherence between disease diagnosis of bacterial pneumonia
and antibiotics of inpatient data among eight hospitals, 2000-2003


p


-
-
-
_
-
-










Drugs and Disease Diagnosis

Antidiabetic drugs and diabetes

The median percent data coherence of antidiabetic drugs with the diagnosis of

diabetes had a median of 99.16% [95.09, 100], indicating good data accuracy. There

were small variations of the percent coherence of antidiabetic drugs and disease diagnosis

codes, however they were not significant. The percent coherence of Hospital E decreased

slightly over the studied period (98.76%-95.46%). (Figure 4-10) Since antidiabetic drugs

have few non-diabetes indications, only small discrepancies between codes for

antidiabetic drugs and diagnosis codes of diabetes should be expected.



100.00

99.00 D
--- A
98.00 B
97.00- C
C D
96.00

8 95.00- F

94.00 G
-H
93.00

92.00
2000 2001 2002 2003
Year



Figure 4-10. Percent data coherence between antidiabetic drugs and disease diagnosis of
diabetes among eight hospitals, 2000-2003

No comparisons between drugs and diagnosis were conducted for antihypertensive

drugs and hypertension, and bacterial pneumonia and antibiotics, because both drug

groups have multiple indications and thus low coherence levels would be expected.









The results from data validation analysis using data coherence method suggest that

disease diagnosis code had sufficient accuracy to be used to identify patients with

diseases. The disease diagnosis codes captured approximately 90% of the patient who

received corresponding drug therapy. Dispensed drugs in this database were a good

indicator only for patients with diabetes. At least 10% of diabetes patients were not

captured by antidiabetic drugs.

In conclusion, the results of the validation of data quality, including the observation

of database characteristics, the assessment of missing data, face validity of drug data, and

the assessment of data coherence, suggest that the data contains sufficient quality for the

analysis of the effect of the 30-Baht HI policy on hospital drug utilization. As stated

earlier, the assessment lacked comparison with a gold standard, thus, it is recommended

to revalidate data quality before using the database for other purpose.

Results from the Expert Interviews

A summary of the interviews with representatives of 8 study hospitals is

summarized below. Details of the interviews are presented in Appendix E. Regarding

major health problems in the community, diabetes, hypertension, acute illness related to

seasons, e.g., cold or muscle soreness in rainy season related to working in the rice field,

and infectious diseases, e.g., tuberculosis, Infectious diarrhea, urinary tract infections,

were common. HIV/AIDS was one of the prevalent conditions in this area.

Most hospital directors commented that the hospitals did not have any specific

action plan to accommodate increases in census due to the 30-Baht HI policy, except in

one hospital where the P&T committee had set up plans for patient referral and drug

dispensing (allowed local public health office to dispense some antibiotics (e.g.,

amoxicillin, tetracycline). Regarding the impact of the 30-Baht HI policy, the hospital









directors found there was a natural increase in numbers of patients visiting the hospitals,

but no significant change in the numbers of hospital visits by diabetes patients. There

was no difference in drug choice amongst different health insurance plans. Prescribing

patterns were reported as being"patient centered" that focused on the patients' needs. In

addition, prescribing choice was mostly restricted by the hospital drug formulary. One

hospital reported that CSMBS beneficiaries received more drug supply per visit. Another

hospital reduced the duration of supplied drugs (e.g., dispensing of antidiabetic drugs was

changed from 3-month to 2-month supply after the 30-Baht HI policy was implemented.

Regarding other factors that might affect drug utilization rates, every hospital

director referred to the Hospital Accreditation program that helps foster hospitals to set

standards and assure equity of care regardless of patients' health insurance status. One

hospital initiated patient education programs for patients with chronic diseases (diabetes

and hypertension). In addition, this hospital adapted nine clinical practice guidelines

from the Royal Medical School, Thailand for the treatment of common illnesses (e.g.,

infectious diseases, and asthma).

Experts from three from eight interviewed hospitals reported no data available in

2000. Of the five hospitals that had data during the study period, the experts commented

that patient data in the HI database were accurate and useful for providing information on

hospital services and administrative tasks. However, several hospital directors

commented that physicians might not know all the available disease diagnosis codes in

the database. One hospital instituted a data validation team to randomly check

completeness and accuracy of the data (only for inpatient service by comparing the

computerized patient data with the medication administration record (MAR).






79


In summary, the interviewees felt that an increasing numbers of patients eligible for

30-Baht policy were encountered after implementation of the 30-Baht policy. There was

no different in the patterns of drug prescribing between 30-Baht beneficiaries and those

with others health insurance plans. Most hospitals did not have any plan to support the

implementation of the 30-Baht HI policy. Accuracy and completeness of the data in the

electronic database was rated high, however, formal quantitative validation was broadly

missing.














CHAPTER 5
METHODS

Our study applies a cross-sectional observational design using segmented

regression analysis of time series data with a concurrent control group to evaluate the

impact the 30-Baht HI policy on hospital drug utilization in Ubonratchatani province,

Thailand. In other words, only observations of patients who visited or were admitted to

the hospital at a given month were measured; no cohort of the patients was followed

during the study period. To identify the effect of this policy, we selected drug utilization

rates and drug utilization quality as primary measures, and hospital visit rates or hospital

admission rates as secondary measures. These measures were calculated based upon data

extracted for the 30-Baht HI beneficiaries (considered the intervention group) and the

CSMBS group (control group) from the patient-specific healthcare data from the HI

database from the eight community government hospitals.

The impact of the 30-Baht HI policy was evaluated in four of the most prevalent

diseases that affect the Thai population. The four diseases states included community-

acquired pneumonia, gastrointestinal tract infections, hypertension, and diabetes.

Pneumonia and gastrointestinal tract infections were selected to identify whether the

policy had an effect on acute conditions, while hypertension and diabetes were selected to

identify the effect of the policy on chronic conditions.

The measurement of drug utilization included all drugs that were available in the

hospital drug formulary and had an indication to treat the selected diseases based on

official labeling by the US Food and Drug Administration. Drug utilization rates of each









selected disease were measured using the DDD unit per population. Drug utilization

quality was assessed based on the percent appropriateness of prescribing. Hospital

visit/admission rates were estimated based on the eligible population registered for each

of the study hospitals. Linear regression analysis of time series data was used to analyze

the effect of the policy on drug utilization rates, drug utilization quality, hospital visit

rates or hospital admission rates. By using a time series technique with a concurrent

control group, any immediate change (level) and a gradual change (trend) of the measures

associated with the 30 Bath HI policy could be identified. The concurrent control group

was included in the regression model by differencing the study measures between the

intervention and the control group. Including a concurrent control group, we were able to

control for external factors that might influence the measures at the same time when the

30-Baht HI policy was implemented.

In addition to the above measures, drug use disparity (i.e., differences of drug

utilization between the two populations) was calculated based on the assumption that the

previously uninsured population had lower rates of drug utilization. Changes of drug use

disparity related to the 30-Baht HI policy was identified by comparing the differences of

drug utilization rate and quality between the two groups one year before and one year

after the implementation of the policy using T-test and Chi-Square test at an alpha level

of .05.

Hypotheses

There were six hypotheses for statistical tests for research questions 1 to 6. We

chose an Seasonal Autoregressive Integrated Moving average (SARIMA) linear

regression model because it incorporated associations (a trend) of the measures during the

pre-policy period and intervention (input) variables, including a pulse function variable to









identify an abrupt effect, and a step function variable to identify trend change. The

model was set up to integrate seasonality or the association of the measures in a repeated

interval (e.g., annually). If trend is taken into account in the analysis, history and

maturation biases on the measures can be reduced. A constant in the regression model

was not included by centering the data with the mean of the measures of the whole series

to simplify the model. We allowed a delay in the effect of the 30-Baht Hi policy, if

visual observations of the plots suggested changes of the study measures after the

implementation of the policy. However to simplify the model, we did not include a

decay parameter (the rate that the policy took effect) into the model. Equation 5-1, 5-2,

and 5-3 were chosen for testing the hypothesis 1 to 4. Hypothesis 5 and 6 were tested

with t-test and Chi-Square, respectively. All hypotheses were tested at the significance

level of 0.05. Equation 5-1 is the full regression model that includes historical trends,

immediate impact variable, and trend impact variables. Equation 5-2 illustrates the

regression model for autoregressive process, while Equation 5-3 illustrates the moving

average process. We used logit in the regression analysis for the measure of appropriate

prescribing.

Yt = 11Tt + 2 Pt + 03 St +et (5-1)

Autoregressive process

Yt = (piyt-1 + (p2yt-2 + (p3yt-3+... (pnyt-n + et ...... (5-2)

Moving Average

Yt = et 91e t-1 62e t-2 83e t-3 ... One t-n ... (5-3)


Yt: A difference of the studied measure between the 30-Baht HI beneficiaries and the CSMBS group after
the policy in month t









Tt: A continuous variable indicating time in months from the beginning of the observation period to the
month before the implementation of the policy (prepolicy period from January 2000 to May 2001)

Pt: A nominal variable for time t indicating the presence of the policy only at month 18. t < 18, Pt = 0; t =
18, Pt =1; t >18, Pt = 0

St: A nominal variable for time t indicating the presence of the policy at month 18 and after. t < 18, St = 0;
t >=18, St =1

31: A regression parameter of the slope of the measure in prepolicy period (prepolicy trend)

P2: A regression parameter estimating the level change in the measure immediately after the policy (abrupt
change)

P3: A regression parameter estimating the trend change in the measure after the policy

9t: Autoregressive coefficient

Ot: Moving Average coefficient

n: numbers of observation



For infectious diarrhea, bacterial pneumonia, diabetes and hypertension:

Research question 1: Did drug utilization rates of the 30-Baht beneficiaries (DRi)

change after the 30-Baht HI policy was implemented, controlling for the DRi before the

policy and the drug utilization rates of the control group (DRc)?

Hypothesis: DRi changed after the 30-Baht HI policy was implemented, controlling for

the DRi before the policy and the DRc

HO: 2 =0; 3 =0

HI: 032 0; 03 3 0

Research question 2: Did the percent prescribing appropriateness of the 30-Baht

beneficiaries (DQi) change after the 30-Baht HI policy was implemented, controlling for

the DQi before the policy and those of the control group (DQc)?

Hypothesis2: DQi changed after the 30-Baht HI policy was implemented, controlling for

the DQi before the policy and the DQc









HO: 2 =0; 3 =0

HI: 0132 0; 03 0

Research question 3: Did hospital admission rates for inpatient services for bacterial

pneumonia and infectious diarrhea change in the 30-Baht beneficiaries (ARi) change

after the 30-Baht HI policy was implemented, controlling for admission rates of the

control group (ARc)?

Hypothesis 3: ARi change after the 30-Baht HI policy was implemented, controlling for

the ARi before the policy and ARc

HO: 2 =0; 3 =0

HI: 021 0 0; 03 0

Research question 4: Did hospital visit rates for outpatient services for diabetes and

hypertension of the 30-Baht beneficiaries (HRi) change after the 30-Baht HI policy was

implemented, controlling for the HRi before the policy and the rates of the control group

(HRc)?

Hypothesis 4: HRi change after the 30-Baht HI policy was implemented, controlling for

the HRi before the policy and HRc

HO: 02 =0; 3 =0

HI: 021 0 0; 03 0

Research question 5: Did the difference of drug utilization rates (dDA) between the 30

Bath HI and the CSMBS groups change after the 30-Baht HI policy?

Hypothesis 5: dDA change after the 30-Baht HI policy

HO: before Pafter

HI: before Pafter









before: dDA before the 30-Baht HI policy

before: dDA after the 30-Baht HI policy

Research question 6: Did the difference of the percentages of prescribing

appropriateness (dDQ) between the 30 Bath HI and the CSMBS groups change after the

30-Baht HI policy?

Hypothesis 6: dDQ change after the 30-Baht HI policy

HO: before Pafter

HI: before Pafter

before: dDQ before the 30-Baht HI policy

before: dDQ after the 30-Baht HI policy

Patient Selection

The primary and secondary measures were extracted for patients 18 and older, who

visited the hospitals in Ubonratchatani for outpatient services and/or were admitted to the

hospitals for inpatient services from January 1, 2000 to December 31, 2003. Patients

who were admitted to the hospital for inpatient services were excluded if they were

referred to other healthcare facilities with less than 1 day from the admission time.

Patients who visited outpatient clinics, but did not have any assigned disease diagnosis

code were not included, because these patients might have "self discharged" prior to

seeing the healthcare providers for examination, diagnosis, or treatment. We included

patients who were diagnosed with at least one of the four most prevalent

diseases/conditions in Thailand, according to The Thai National Disease Statistics 2002

(110; 165).

The selected diseases defined by the International Statistical Classification of

Diseases and Related Health Problems, Version 10 (ICD10) (167) were: 1) bacterial