<%BANNER%>

Center Characteristics and Kidney Transplant Candidate Outcomes

Permanent Link: http://ufdc.ufl.edu/UFE0021207/00001

Material Information

Title: Center Characteristics and Kidney Transplant Candidate Outcomes
Physical Description: 1 online resource (110 p.)
Language: english
Creator: Schold, Jesse D
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: Health Services Research, Management, and Policy -- Dissertations, Academic -- UF
Genre: Health Services Research thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: There are approximately half of a million patients diagnosed with End-Stage Renal Disease (ESRD) currently in the United States. There are two general classes of treatment for ESRD, maintenance dialysis and kidney transplantation. Kidney transplantation is considered the preferred treatment for ESRD patients who are medically cleared for the surgical procedure as it is associated with an improved quality of life and longer life expectancy. A kidney transplant may derive from a living or a deceased donor. For patients to receive a deceased donation, they must be placed on a waiting list at a transplant center. There are currently approximately 240 kidney transplant centers in the United States. There are numerous characteristics that vary among transplant centers. An important question is whether or not these center factors are associated with prospective transplant candidate survival. Moreover, whether these center factors are applicable to all transplant candidates or have differential effects on certain patients is also unclear. This study examined these questions using a national database containing patient-level information on outcomes for all adult solitary kidney transplant candidates from 1995?2000. Results of this study indicate that transplant center characteristics are significantly associated with patient survival after listing. Of the center characteristics, expected waiting time had the greatest impact on candidate survival. Candidates listed at centers with the longest expected waiting time had a 32% increased hazard for death, translating to approximately two and a half years of reduced life expectancy, as compared to candidates listed at centers with a relatively short expected waiting time. The main findings were applicable to the general candidate population as well as high-risk candidate subgroups. Results of this study may be utilized to inform patients and caregivers about the important impact of characteristics of transplant centers on prospective survival for candidates of kidney transplantation. Investigation into applications of these findings in healthcare policy, for different transplant populations, and in other medical contexts will be examined in follow-up studies deriving from this study.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jesse D Schold.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Harman, Jeffrey S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021207:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021207/00001

Material Information

Title: Center Characteristics and Kidney Transplant Candidate Outcomes
Physical Description: 1 online resource (110 p.)
Language: english
Creator: Schold, Jesse D
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: Health Services Research, Management, and Policy -- Dissertations, Academic -- UF
Genre: Health Services Research thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: There are approximately half of a million patients diagnosed with End-Stage Renal Disease (ESRD) currently in the United States. There are two general classes of treatment for ESRD, maintenance dialysis and kidney transplantation. Kidney transplantation is considered the preferred treatment for ESRD patients who are medically cleared for the surgical procedure as it is associated with an improved quality of life and longer life expectancy. A kidney transplant may derive from a living or a deceased donor. For patients to receive a deceased donation, they must be placed on a waiting list at a transplant center. There are currently approximately 240 kidney transplant centers in the United States. There are numerous characteristics that vary among transplant centers. An important question is whether or not these center factors are associated with prospective transplant candidate survival. Moreover, whether these center factors are applicable to all transplant candidates or have differential effects on certain patients is also unclear. This study examined these questions using a national database containing patient-level information on outcomes for all adult solitary kidney transplant candidates from 1995?2000. Results of this study indicate that transplant center characteristics are significantly associated with patient survival after listing. Of the center characteristics, expected waiting time had the greatest impact on candidate survival. Candidates listed at centers with the longest expected waiting time had a 32% increased hazard for death, translating to approximately two and a half years of reduced life expectancy, as compared to candidates listed at centers with a relatively short expected waiting time. The main findings were applicable to the general candidate population as well as high-risk candidate subgroups. Results of this study may be utilized to inform patients and caregivers about the important impact of characteristics of transplant centers on prospective survival for candidates of kidney transplantation. Investigation into applications of these findings in healthcare policy, for different transplant populations, and in other medical contexts will be examined in follow-up studies deriving from this study.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jesse D Schold.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Harman, Jeffrey S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021207:00001


This item has the following downloads:


Full Text
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 E20101113_AAAALG INGEST_TIME 2010-11-13T14:23:08Z PACKAGE UFE0021207_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 9121 DFID F20101113_AACEVP ORIGIN DEPOSITOR PATH schold_j_Page_034thm.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
ae97ef77fafcbe576f5760a9acfd40b5
SHA-1
a41c1a791434cef9e7b971460134d2714cc8e6da
33671 F20101113_AACEWD schold_j_Page_045.QC.jpg
b0e5c395f834b090e8b88c09c5ad651f
ae9fe01baa4a532fac7b5372cf0e63e15593f310
37297 F20101113_AACEVQ schold_j_Page_034.QC.jpg
041544bb8f67e2dcf8ea4689a5adc390
6a69d71b379ad9af38b78ed4dab633b2c1de0ab1
8487 F20101113_AACEWE schold_j_Page_046thm.jpg
58d3ddc7650dec8fd030b376c6eede6b
c7b7d46312e38235aed2e580be3adb1ef065e8d0
13396 F20101113_AACEVR schold_j_Page_036.QC.jpg
b7d379f56010caec8d176e07436fb1e1
89a1ec52e1c3ef55d2e724a623d2d1b41d43a530
34726 F20101113_AACEWF schold_j_Page_046.QC.jpg
6928ee972f2143f3f1691b476b05ec28
8c11c33b1b56d6fd553705856b0d7df4aa2625b3
7940 F20101113_AACEVS schold_j_Page_037thm.jpg
878fdb99dfb2466229c86035ed05a7c7
289a2c04b243dbe1af8cb46501f998471f7cf0c6
20808 F20101113_AACEWG schold_j_Page_047.QC.jpg
9cc6361d6754d6c1618e0359116b4b69
df744b8ecb0a62850e786263a001574464fed64d
33346 F20101113_AACEVT schold_j_Page_037.QC.jpg
6adba9c548bbbbca8cf601b72cc1f77b
1d78d612cf3d4962103874781864fad7aea5648b
2394 F20101113_AACEWH schold_j_Page_048thm.jpg
8e3c24285f08134b07a5cff60b533066
fec6fb0e760216d392820d42e09b75951efa0e5b
9123 F20101113_AACEVU schold_j_Page_038thm.jpg
43c5cba99ab5bf50bca33ef4fdcae18f
8bcc3bb99bb590a9eb5df98a7b3d34c289e5662b
7112 F20101113_AACEWI schold_j_Page_048.QC.jpg
9756af2b1e99ee0f54e8b1cede476500
2c0e20ce5b938e66eca810f9f28f7fbbdeea7acf
38337 F20101113_AACEVV schold_j_Page_038.QC.jpg
a4b03bbea063f5fb3557d845a26babad
87c12b973b01c530d04986a1f662490986f173a1
2306 F20101113_AACEWJ schold_j_Page_049thm.jpg
717420e7fcc0ff95a46331158a50c917
33f4f13500579ca10475df0e8de927f33db36f44
37158 F20101113_AACEVW schold_j_Page_039.QC.jpg
e36ce4e6d9ad47cc2ea90fa4615947c6
cbf2185a2d58c433db998b17e1f1f4d271d2a1f4
7554 F20101113_AACEWK schold_j_Page_049.QC.jpg
333238dfd8cc1c02e588bda6b025730d
8c7663fe9b96a9d25f4ec9306c733f8182f5137f
9063 F20101113_AACEVX schold_j_Page_040thm.jpg
213ac2a1f6347ef1ea4d45d2c3aa3342
845a15b46d58dafd3b7a9735eee1367173d8dac4
2637 F20101113_AACEWL schold_j_Page_050thm.jpg
cfd35776ce87d2bc5ad6fb2481b7dfc9
c4e7046fe616a5dc4ef1b6672fb69bba4efecb34
9059 F20101113_AACEVY schold_j_Page_041thm.jpg
de5f9988155f8d7f2fd01cc7d1cf47bb
c88f4bd59074c4219830b63a5011bc5b10a6df13
8792 F20101113_AACEXA schold_j_Page_062thm.jpg
9f0e6335783b46823c4186e6494344b6
6258d497518f70ea1546c7c6d7397c091927f036
8360 F20101113_AACEWM schold_j_Page_050.QC.jpg
aca7dbe509273850e171859c04306161
2a88a3aeed58a84c3bc31fdf87d9dc6383dfabce
36360 F20101113_AACEXB schold_j_Page_062.QC.jpg
a0f23236658de5c62e6bf0d94e3ed44a
5e91dabf6b6bda1d97297f4110444d7dd2be5b46
2276 F20101113_AACEWN schold_j_Page_051thm.jpg
93615a8e00c556729f67452d847b9ec5
89720d2418e3f078a37dcc105ff21a385ff7c7d2
8830 F20101113_AACEVZ schold_j_Page_042thm.jpg
deccb7b1c321af8203ae2488dbba4413
304efe6a721970ba02167577ae3e75692986d896
3755 F20101113_AACEXC schold_j_Page_063thm.jpg
f4e492a8e17cf6dec3dbcb9ca7ebedb2
3f2118dc76b1086404f4048dbde6c151b034101b
7842 F20101113_AACEWO schold_j_Page_052thm.jpg
c720f0724bdb81e48c08e377c8337b90
0e00fdecfca091b485afa69386fa35320a631cc1
12795 F20101113_AACEXD schold_j_Page_064.QC.jpg
c95705e6fe81a1c1c563ad5ea79818b2
ebb5af608029a5a8b4ee36495a33ef15690b77ff
32317 F20101113_AACEWP schold_j_Page_052.QC.jpg
667df4cfa01fbb7bba0964a71be876bc
4a784592391154f9deba19b97379f3f19438357e
6421 F20101113_AACEXE schold_j_Page_065thm.jpg
5f2dc5a40c74c6b1861d2fc9e90f5b69
9a2cb2d114440ab77fe97a1b1e55e62601f0c519
8481 F20101113_AACEWQ schold_j_Page_053thm.jpg
1cdf28f4c372a5f496679cd534f51b73
799e25444b50bc2d63d62b8bde82956f3e04fcc9
7415 F20101113_AACEXF schold_j_Page_066thm.jpg
aacb0782e2bee21e9c66f86fbcba193b
62d2aa12b14e918bd00c689e6683bb688c4f7456
35822 F20101113_AACEWR schold_j_Page_054.QC.jpg
d8387434c1c510fe2b57ad09029187df
85977fc0adc0d2e8ff28cb2bb1be5205d842df35
29675 F20101113_AACEXG schold_j_Page_066.QC.jpg
9f0646a8c916c06d4f18689aec3e1ab1
140f54992fa7e5873901e05edc0ccb79db79f6f1
8993 F20101113_AACEWS schold_j_Page_056thm.jpg
4a2c3daa7e3766014f7eafc6c8f2b028
7fb0e50dc2e5d46ccb3c8c5881077345869f1b1e
5866 F20101113_AACEXH schold_j_Page_068thm.jpg
deaa391e9608629c8d13607d02ac0d5a
a5a269111a40ef5cd8fac24641ecfef2b38faf2b
38125 F20101113_AACEWT schold_j_Page_056.QC.jpg
b012d9285fed48f9c0aa3cc669b1487d
9731ca74136049a1a723b296439544892efeffd9
1051952 F20101113_AACEAA schold_j_Page_006.jp2
55fd7e51c097fb9d216d468b9ad63244
a08b4bdaf5f909f8c6bafc30cb4b651960afe2cc
22272 F20101113_AACEXI schold_j_Page_068.QC.jpg
2be205f9879efce7831f04ddb8362705
faf119aae88bfdf6d38322718a082460f3df95bc
8919 F20101113_AACEWU schold_j_Page_057thm.jpg
75efc34e48f84136732ab4054c86ed5c
703770ff752e2d00c8e1e6fd7fc81cbf873662b2
8974 F20101113_AACEAB schold_j_Page_020thm.jpg
d2557283d0c68b8dbfcd8d27a47c6cda
98948377dd9717af36cfdc4ddc83f6dabf6440ee
7641 F20101113_AACEXJ schold_j_Page_070thm.jpg
209ed66e6f3f680d528bd2e17c743f41
89177d637969862ff665be6c5971688a9f78becf
36694 F20101113_AACEWV schold_j_Page_057.QC.jpg
09dea5647c1b6fff1db741baae2e1c94
0a04d6a34448efd08b3884cea7fb8bf113cd52cd
46725 F20101113_AACEAC schold_j_Page_069.pro
0ee320a1e29717ac5be60bc96f74d456
0657df66da9c990c5c1d48e842cf511b6f8f6d99
28284 F20101113_AACEXK schold_j_Page_070.QC.jpg
5c03e35e7388c2e5e7f6daed56f6a28b
9e1b2f9a100e6946a43799e8a9d5365af62196d8
8663 F20101113_AACEWW schold_j_Page_058thm.jpg
a8640bdde8905e74b8f06bd5310412f6
5e5221f85f9530e5af58f6491b31e926e817fecc
9075 F20101113_AACEAD schold_j_Page_018thm.jpg
a9a68f5e5a6024629c8969f81b104cbf
9fa440b7a4a67b8357cdc6af2447bad377368d11
27614 F20101113_AACEXL schold_j_Page_071.QC.jpg
049fa63a42561799f547059a5cb27cf3
2a644e4a91e0acc07e4f39d3e8f3fa51c5aebd99
8491 F20101113_AACEWX schold_j_Page_059thm.jpg
1cf6f4d18e2dd7db5b42691bddd9503e
23911a4260c9ff3a3ef754896b34643a2eef79d2
36370 F20101113_AACEYA schold_j_Page_086.QC.jpg
5c440a29c6cac2473426d1391102e912
50b9fe8ebd8439fb4055f571a999a3e19d36c579
6558 F20101113_AACEXM schold_j_Page_072thm.jpg
a0a700f6d066f8e50ce67ef232eaff6e
baa3942f9bb78eb7e3ee02d42fa8ed711a38b595
34414 F20101113_AACEWY schold_j_Page_059.QC.jpg
3f84f6fa0a2658c4fe75bc4110f85c34
7bd0954d7e9311f1048075001c70d62f720f6ba4
14720 F20101113_AACEAE schold_j_Page_067.QC.jpg
5041d12412812a06a38a86423cdbf657
b694f2a0e5c675cd0190893f8e83750ebca15c26
8939 F20101113_AACEYB schold_j_Page_087thm.jpg
40a5e022570bc0723c6416ca540c28a5
d1c52e31ee17d40ddab6361b9afb30dc5964a535
26745 F20101113_AACEXN schold_j_Page_072.QC.jpg
761541eccb4a55a5937a138bdfb31dbd
a5a58ff76497deb4c45af71116345bfc71671c41
34927 F20101113_AACEWZ schold_j_Page_060.QC.jpg
ddb0a91e3bb5cda027608ebed2626d38
e70b18a16408409fe3452b8f424b46f5c384c41f
108637 F20101113_AACEAF schold_j_Page_032.jpg
f165da4c1bb51de7897fbe07382a9723
046e2ad8ace23a0e0e145c2156d3faa19f793d23
9102 F20101113_AACEYC schold_j_Page_088thm.jpg
f89c25312e4bcdd38c3e8d882feac32b
452ad0f583462430a31021eb735f0fa1157323f1
2531 F20101113_AACEXO schold_j_Page_075thm.jpg
447ad7036addd2ee6f69308a9f44ade8
5963fd9c94671490e7f49dd3e05732e279866ef0
1053954 F20101113_AACEAG schold_j_Page_040.tif
65f2d140813b000e855ff54326389b8e
19fff0bdc429d3f92b930e921801cf5de53d98b9
36991 F20101113_AACEYD schold_j_Page_089.QC.jpg
f6ad23ffb1815baa45b0cc6f6ebaa7b8
140feab5552801f4e3793c0fd61ec161093af795
8193 F20101113_AACEXP schold_j_Page_075.QC.jpg
b41c2368d90df770ce83af6ec6df6694
a3e40508073f3e156ddfdd9209ac65d083ed994b
35666 F20101113_AACEAH schold_j_Page_053.QC.jpg
dc402cc2094cc308f15f9ee3d2d6737b
6543730394da2f324441b192ed82fa1212b431f7
36489 F20101113_AACEYE schold_j_Page_091.QC.jpg
bc60d110ae46de7335907bea80317462
319a6ad713757e19d17b9e5316810e18560fc443
2630 F20101113_AACEXQ schold_j_Page_076thm.jpg
4174fd62ad91d4b22c5f03e332ea2df7
a37c096365b6dfb6311a61a9882a612511c93295
57027 F20101113_AACEAI schold_j_Page_057.pro
5cc2e3431496c4c34923eafdf295ae69
ca181f168e5146239f05558a22e80902ab62726b
9031 F20101113_AACEYF schold_j_Page_092thm.jpg
8dac46ad60c6a654d38bfb03ec1c8cf0
754f73932fa4897d8dadad0635b99e511d2588f4
2636 F20101113_AACEXR schold_j_Page_077thm.jpg
8ec1cf0b549dea8b261965d53187c0e3
322258379f0fff8223ec55b1016404836e079df7
119362 F20101113_AACEAJ schold_j_Page_100.jp2
a2d483985d11982e5d52d693cc39630e
244c09ee34b22355728698668a7fbd96f41b7a9b
36286 F20101113_AACEYG schold_j_Page_092.QC.jpg
89f06239a7bcc42e091db555da439b07
1c46bbd1a4ed9ebfe3ed55a6ba0b6910c55c25a7
7686 F20101113_AACEXS schold_j_Page_077.QC.jpg
fb3bf60df6427ba9bc45f24fdb372b60
97245ffcdfca102ce6b7fa1a96b00aad1621e158
334 F20101113_AACEAK schold_j_Page_050.txt
6f402da0bbf908b4107aaa49a348d7c0
c5341ee463b7088073e2a752c5bd373192ed9471
9001 F20101113_AACEYH schold_j_Page_093thm.jpg
dfcf4734da804947ac1a9b8bd267a76c
59b830f92ae4e2ff3bba5f7b02da0724fa478758
8230 F20101113_AACEXT schold_j_Page_078.QC.jpg
9130c845298d7a9acec3dfc841b03832
897bd50b02b494126cb20daf98474bdc8afd723b
121273 F20101113_AACEBA schold_j_Page_085.jp2
5a1d7c166e90b7c1190de00a4be2194d
2f894791a6ebb9a56e3a93aa23e253738dfe7f1e
8866 F20101113_AACEAL schold_j_Page_099thm.jpg
67b963e9b085d2e305dac016301fc7f7
ad866529cf97e9c1f4e1b48f38469dc08ad3f943
35734 F20101113_AACDVF schold_j_Page_008.QC.jpg
e2e33b70db9e07b27853b9b2813376b2
6a36dcf1a812fe455a214990b17629adfbd73079
8958 F20101113_AACEYI schold_j_Page_094thm.jpg
2d220f97fb2c5e9d23f99fdeecceb048
702bda164db274d3fed1c4977f98a88d36b721fb
35480 F20101113_AACEXU schold_j_Page_079.QC.jpg
6e7e893358087b9338992badebbc4a67
bcd3f5855d23b67a7165dfe6983547b30245a40f
8423998 F20101113_AACEBB schold_j_Page_050.tif
da77d666d404479a0c33fa53e95c2958
9df284eebfb7ef733b5e70d87d2b84463f7a9439
4069 F20101113_AACEAM schold_j_Page_011thm.jpg
bf64a499e822f4a4b0a35e37b605661d
ae9764c0103a4d9682a3480267c6e9a4b81beb3f
90 F20101113_AACDVG schold_j_Page_002.txt
5b6298fc762273c4ce79df105ac5700e
2279e3d3dfe06609941f242ced666f181569c989
8836 F20101113_AACEYJ schold_j_Page_095thm.jpg
73e3c7e30739fb67a5366f89c67c9f78
16c3eafd3dfc7ea1c4ce39ca893c0e5b6650a429
36188 F20101113_AACEXV schold_j_Page_080.QC.jpg
7f861c7ca13b099df57a97a567d0b2cd
b876bca294e804e0fd2c230c5e7b554624c8517f
7416 F20101113_AACEBC schold_j_Page_035.QC.jpg
28cdd656f9bd9b4410c44a2ed14d46aa
6018d1d74d077f7e8e079d7b71b57c94076602c9
F20101113_AACEAN schold_j_Page_075.tif
f35a468836d790d01928f377f49d1025
08072904e989aeaa5138ea2b4ef3d7a6caa9beb6
25271604 F20101113_AACDVH schold_j_Page_006.tif
7531b032a2181380c7aedea277cc3b22
5a47106f32b719d66f83e05a16b7a6bfd9803644
35579 F20101113_AACEYK schold_j_Page_095.QC.jpg
823d87d82353eb18985d408a2d1520aa
7b288350e4e410ca8d7990406d7b5d84bb1a95ea
35153 F20101113_AACEXW schold_j_Page_083.QC.jpg
c32d830d65b1f7a99e1daf09357d9362
802726be5c015eb229e05716045f223805910614
114242 F20101113_AACEBD schold_j_Page_055.jpg
0b64951f1d9725b70087f185ec19c9da
371f2d0d85c22b1baba78a90b6157e397fa266a7
116786 F20101113_AACEAO schold_j_Page_054.jp2
7215f33f19ebf9dde2f462f5ad890826
f97fec23553fd7eda4dca7512fb613516ef5dadf
114257 F20101113_AACDVI schold_j_Page_039.jpg
b5b8961e3e725fb927962ddc90d94940
9351154a02f878406a68243c569c016ef5a79fd9
9213 F20101113_AACEYL schold_j_Page_096thm.jpg
cf9f027e92a4b16d48db6fed114d8ad2
1df7e21a7fc0c4e916b0866adbdfaf8ec6222333
9231 F20101113_AACEXX schold_j_Page_084thm.jpg
f728b62d72eebfdc7564cb86c64f4dcd
7c7fa20b3d701e0f6cbed0ecd7f4339896d671a7
36710 F20101113_AACEBE schold_j_Page_081.QC.jpg
309ac912cf60867c3e003d4f66a6e5a5
99c64ff9cbdde95c2f1ee2576ceb0c6560c7fecf
F20101113_AACEAP schold_j_Page_033.tif
b2e5b0739d25a6a68b019ec84c000a10
bc2ee39d055b47b3b67d522189b86a917ce41303
218275 F20101113_AACDVJ schold_j_Page_078.jp2
5f2e0ba05a1e6c05206f435139e1a92e
e15d853bb0f4074b6d7f84c37d1e94898063f416
36989 F20101113_AACEXY schold_j_Page_085.QC.jpg
d340061995af793d4251a517a567f831
bbade869f76c3878b3d078d4127acd025aa38efb
9011 F20101113_AACEAQ schold_j_Page_082thm.jpg
49e6d9a0f17c4393f7379eaea3bc2385
6a6ab986c58bdb97bf9df73b24ca1fccb31168ec
2123 F20101113_AACDVK schold_j_Page_060.txt
7a2837408de21810e8ce43f2ecc1b9bd
0ff4075fbf54aceaff9d0d082026e6f652759e4b
33743 F20101113_AACEYM schold_j_Page_097.QC.jpg
c8be2937d0bec39e73c453eaec1d9d4f
f6533333eb29af6311fe933e68b04b09da6ffc13
9044 F20101113_AACEXZ schold_j_Page_086thm.jpg
093d89e8307776f4e2e7cec3805278d3
7b8b7043583935cff98f6f88d48b40c37135aac6
F20101113_AACEBF schold_j_Page_034.tif
b1168523e53db6ec299314f297d3ed75
7dff758f7066aa3dbbbd6cb09b90516ecc857c83
F20101113_AACEAR schold_j_Page_059.tif
bd3dc9ac35aecdf161c3fca8e901f92c
62ca3182cf31dfcc23eea7619802ebbb202fc8fa
53358 F20101113_AACDVL schold_j_Page_031.pro
8004c4176ab4896c975fe5a2082e8e72
ba41de2e2034628b17681ca797f733b68a434467
35567 F20101113_AACEYN schold_j_Page_099.QC.jpg
b94d22722902b25a2bee9182ee9840dc
f6e165d4dd0440da5b29d2c510a9032fb965c72e
3133 F20101113_AACEBG schold_j_Page_005.pro
50beda04b9a0a7b3d5a91b45a6f3a785
5680c2051a6e01f5505ac11da5cbb959c862930a
2250 F20101113_AACDWA schold_j_Page_042.txt
8425fd26cd5fdef2e979fbb8157c4244
a72b0cd720687ec5984f48e599b88fcc5e59ed91
180146 F20101113_AACEAS schold_j_Page_050.jp2
37e5f8d77076e8d9e6c70c847d193bc1
b41cd3249d1fbe372b867097020f38375749af3b
F20101113_AACDVM schold_j_Page_026.tif
cacd1919ac1828cd23bf2e5c3f435b56
4530c7a344a92d8e70248ea572c208664fb80e16
36003 F20101113_AACEYO schold_j_Page_100.QC.jpg
6b7edc0b29b34ee137af783d56fb0be0
498ea45360081f59edf42384cdda845fe16445bd
F20101113_AACEBH schold_j_Page_036.tif
8821a888cdad2770a547afde79def53a
c6efcbf4ead6702ec650250fe37617e423ef4ec1
116671 F20101113_AACDWB schold_j_Page_022.jp2
8af647e17490496f11ac7aa380fd2353
2d05509e029f60ac13b171f2836e1dec71bb3691
2261 F20101113_AACEAT schold_j_Page_092.txt
17eb3f7511af2083b293ba50b68f8b73
32d5343d1b27edc329ff9a32521189d49ab82b77
119913 F20101113_AACDVN schold_j_Page_026.jp2
b307b60c90d63ae15ea5c49b32b7be85
fd118a15760df53197af3e7495c3725fa88463e5
3301 F20101113_AACEYP schold_j_Page_101thm.jpg
06186b7236ab44cc0f3bd0219da028d7
2923eabd49e6788e73ec775e3dd979e4df9bb16e
17750 F20101113_AACEBI schold_j_Page_009.QC.jpg
6a8996221233fbe5410b5425e4a729e3
b22c87e4455b7523c5b59e0bb8aaddcc942f341b
8505 F20101113_AACDWC schold_j_Page_022thm.jpg
c153fbcc99b8746c53e71553c5d35391
7bc9d67e134793a02773d5ee1be15daa563db57a
107531 F20101113_AACEAU schold_j_Page_014.jp2
6a0153debf6ec3ebc02730ab61f989df
2083e5c278cf52244cfe274855fe04089338ce3c
5684 F20101113_AACDVO schold_j_Page_050.pro
a0a4104930951d1d99df479284a6e0bc
f07977a73c1701dd0dcc47b600ee24fedfabac7a
13374 F20101113_AACEYQ schold_j_Page_101.QC.jpg
f044a0cbe1cf24f6c78b05af8ebf42a3
0aa3584466641fc20113147e8c1257efbe794bfd
8328 F20101113_AACEBJ schold_j_Page_097thm.jpg
247bfb6b7d0787cb0cd6148ba6217ff0
e176c146de8f671f08a967c1f2eb20d9b9345c8e
2231 F20101113_AACDWD schold_j_Page_089.txt
57003113743068e839611014ada87ca8
2901f97418465cdc65dd1b2461c5d8f8c58a1516
123853 F20101113_AACEAV schold_j_Page_020.jp2
2d4a035da8c8434301aeb99138250fb1
e8400206cd28e26a66a6f973d1eefc826898120f
24287 F20101113_AACDVP schold_j_Page_001.jpg
df1b02950a0f542dd4362c105e72a82a
c05accd30d7b84bb53612bdc28525c5989190ab9
35713 F20101113_AACEYR schold_j_Page_102.QC.jpg
f0ecdf72b9f4564319791ecc9781c41a
97c5b88c16396e6c682489a77da0e0509cd4ff0a
47461 F20101113_AACEBK schold_j_Page_067.jpg
c9eab02f36d851403c132777c33cceab
421d61de2884351209e2cdccefb39c58c7aeead9
F20101113_AACDWE schold_j_Page_080.tif
c3ccb71f56a501f4c640474fdf13ed03
e151a06d642a53438ca8630cac91c6d51cbacb61
849 F20101113_AACEAW schold_j_Page_063.txt
a4434d6ee79ef9413c203ea13deed4a4
0642c2d6da8ee6790e9c2492b2d729b03da3bf14
6849 F20101113_AACDVQ schold_j_Page_051.QC.jpg
25abfcbbcfbd5e3b79969eb377311428
ce470bdb11cc35f496f9653fe18a0bb55e0af8ec
34202 F20101113_AACEYS schold_j_Page_103.QC.jpg
4451a0eeb95059f4f067d888121ea0b0
df37519edb8f7a19715d35b7effbdfc13f7af02a
26015 F20101113_AACEBL schold_j_Page_006.QC.jpg
af074590c833e5c691bcc30d5bfce9bd
8a7aaeb02c9f60a4fd8b4c4c663887c60b9020b5
58362 F20101113_AACDWF schold_j_Page_038.pro
c2a17b08309f83f42283e725aece1c6e
cb7d4e33c4e9a0e6fed5c41eaf0c95b0562069b2
21332 F20101113_AACEAX schold_j_Page_048.jpg
7604438f0d0916c65f44643046b99cf8
dbb481ee8142c7e68525f2e36768a41fc4fafb67
3978 F20101113_AACDVR schold_j_Page_067thm.jpg
c170113d70a6ab6a0217b7262ff4545e
7cd109080c1da295720b49388f88a9753abc42f0
35278 F20101113_AACEYT schold_j_Page_105.QC.jpg
aad5e180f3f6003ef08e34b4df291f9b
dd197a6516e1c1341fb8f6a3efb23107739b0215
1241 F20101113_AACECA schold_j_Page_007.txt
7ff1554ba43f406ec709c03d8ab36f23
060726356a1b80981f4fd0c16c54af7eeeeb87e1
2209 F20101113_AACEBM schold_j_Page_103.txt
6851f118e7f6a0fc04c8f8889c1e85ac
81c58583fd938aaad69b5eaba66ff711d5b436e5
2311 F20101113_AACDWG schold_j_Page_061.txt
8ecf88d960987d18566a169c65072129
57a9f23c870a6649f6fd907b3f49ce18c304300e
F20101113_AACEAY schold_j_Page_003.tif
2df99ed92c7b3090b8d01670c2ad0eb5
96d07184617db164e2ff11c23c302a95c64efd0b
113536 F20101113_AACDVS schold_j_Page_025.jpg
d72bd622c7ae9e148cb9436be18eea39
e51f008a9f91e1723697216dc049502a6992a5e9
8817 F20101113_AACEYU schold_j_Page_106thm.jpg
8a7629a1b1fb9ae5fde05bcea4af9b41
b1c73993b875b83c320ee8ca3295dfc07e012f44
123679 F20101113_AACECB schold_j_Page_017.jp2
52b03b914c0c8c6c6e9cceeb3f19ebc8
233540f52fc865ecb6faf95c42dcb8355a2b5982
746 F20101113_AACEBN schold_j_Page_101.txt
b0acfec062d8765746742db0f6cfd38c
deb3f663490a382396ad91e7a3bd6c66e0acae93
6274 F20101113_AACDWH schold_j_Page_110thm.jpg
6fa4e07cd84645a3ac441b2dea37122f
9401d1f293d5c66d9296f88e0b63e1e0b7a6c5c6
8943 F20101113_AACEAZ schold_j_Page_060thm.jpg
fb79b9d2b718c9d4de3daba5d83a50d9
f82ec4b0605ca7ceb75bfa027cff8395467762cd
8795 F20101113_AACDVT schold_j_Page_079thm.jpg
e08ce22cc169de8eb55489b85f2e4373
87ae609d8ca1b4121b2b4c31b1e173b96f033da9
32680 F20101113_AACEYV schold_j_Page_106.QC.jpg
0a4a070dd2f6d385d6d0088ddd5ed63e
87b046794f7144b743c9ea35de1410f319f44c6c
122290 F20101113_AACECC schold_j_Page_039.jp2
a5dff4dc2b12b6cd4c1b80bbe1e6620c
90e33a233229ba8ddcac0d7f7157ed6395710d92
2059 F20101113_AACEBO schold_j_Page_035thm.jpg
af9ca2992c40b12a7d5e867a3c8d6781
3c9178c249ed7a9d6a8faf3f9fe4b513b4bb1568
36701 F20101113_AACDWI schold_j_Page_018.QC.jpg
3efc8508891fec38e876c8286ff98980
c57ec5cb6c6a2d2fff63a2be2566428f0128809c
114368 F20101113_AACDVU schold_j_Page_088.jpg
b197a1044031dd33008c6334cce9a552
0356eb49c7dadae936ca7e92e7675825e75b62fd
36518 F20101113_AACEYW schold_j_Page_107.QC.jpg
a7ee9e3c399109da68c460efd381497d
2d34d6e98b2674f76ff27e1465e733cde0ac3c49
112095 F20101113_AACECD schold_j_Page_096.jpg
40047a83e19973d8d1ce3e6739c10c69
a5a21077c4d9cbfcd32536100a5ed32302ae8cce
381 F20101113_AACEBP schold_j_Page_035.txt
21dcbcefaa6655a43ee9ab5013ab51ed
3b01727711c9f3154c338dc635d7c8e426606aeb
57142 F20101113_AACDWJ schold_j_Page_055.pro
fa6328318b46ed14cd6b217bda520fa5
d9bd6f31e27915a412bc9a4a68d18b9c4abbeb1f
36149 F20101113_AACDVV schold_j_Page_024.QC.jpg
15999c8f6cc8fbf790a80acf3240c3c7
f9e0acaf1adf050eb72ea1e7fbffa9fb0b07dda8
35635 F20101113_AACEYX schold_j_Page_108.QC.jpg
3acfd99506097a8d4232032a1e131d47
305b4b1f09a8431179e60988c35eb638bd65064a
42556 F20101113_AACECE schold_j_Page_101.jp2
e5693ed0a36b2457b9aff302777ecee3
4f169a0fbf9b93625a81bf240345dd47058bfd1f
F20101113_AACEBQ schold_j_Page_031.tif
a3649cb9646b5f781311dd824e8da3cb
0b10c8a2f6a50c62a456d1949971bab8e3b5596c
85833 F20101113_AACDWK schold_j_Page_065.jpg
d1c1010a63ba3018b032b6fdd73f8c79
d1e2fd0533fd9eb42445d55b9d47f24abcdc63ce
11228 F20101113_AACDVW schold_j_Page_098.QC.jpg
3cf6d12390f23b39d8286b122be2b7ba
7c068637e5f1b0f767757e7dde5859e97f668fa9
8321 F20101113_AACEYY schold_j_Page_109thm.jpg
5f509b250deaa7b778a2e7b58c1f3f2f
280c2ae17dbf06d9ac71c9e92c1b9f1649fc8c21
F20101113_AACECF schold_j_Page_088.tif
05846ff08b67ae0749c5332aca839c4b
ea20936e84cf32a6372c37d0b82b17ad4ef8bd1f
F20101113_AACEBR schold_j_Page_076.tif
97c1f426eaf09c2acbc2e002552fd1bc
747c55f35593e73ec2f46ef0feb2f67bfeddc076
57491 F20101113_AACDWL schold_j_Page_096.pro
be76f6dae213317c5314b8cabb1e91f2
bb3a33d42220008bd5a23ba93cb48888a98e3eb7
F20101113_AACDVX schold_j_Page_107.tif
f67727c2bf1f9eb73dce4dbfef33d9c0
c775fe899df260367f036da65df661b72faae3d1
127877 F20101113_AACEYZ UFE0021207_00001.mets FULL
7c62548560c732d165494ead939d6d09
9c30eb72620ae1540ec94e32d85072d2396956c1
1504 F20101113_AACDXA schold_j_Page_065.txt
664b0401ae3db94720b565cf0739f03d
988f6d97d5363b8222882bbcba2895d7a52e6971
F20101113_AACEBS schold_j_Page_102.tif
7c51a424449f2088bdf10e1ae411c831
1326f4e1d9386ecb41236096270a0156314f476b
F20101113_AACDWM schold_j_Page_079.tif
88b10a6554d62c9f659be2584ea7d638
8936f4302a7ff47945384d8f3e794ba41b409bb1
F20101113_AACDVY schold_j_Page_093.tif
bc56d49c4ea74320643091ecc4798bd4
bd6d053ae98f59788442849166d9e688cd6f7a20
5003 F20101113_AACECG schold_j_Page_047thm.jpg
91d01657fa8fb7dc3677330b0be04a49
fc6948a993ab7bf8632464fd82833d59604f7238
1965 F20101113_AACDXB schold_j_Page_033.txt
7a9ec5a0099ba2093e00f438033566f4
94e0d6a01fb9327fc050435f9047add4bc565c6b
2146 F20101113_AACEBT schold_j_Page_044.txt
4a1f72250b330b7ddea0af349e7bf8eb
7c0a02a901d48c403e0cc37de1eb215abc9839d1
102160 F20101113_AACDWN schold_j_Page_014.jpg
bd062b82956604ee161de2413e9eb6fc
bac5ee1976abf50613de34dcba077d3c15b18486
F20101113_AACDVZ schold_j_Page_062.tif
c975d8f5b5a8ddbe94001cc0e4d15d27
0e98dd1968b584866bb56dc54335d40eb73b0180
36432 F20101113_AACECH schold_j_Page_023.QC.jpg
547194485b3393f4b7f8d09cb2e99fae
624a0ca1e7808b071f849aa060e364086987ef1b
69918 F20101113_AACDXC schold_j_Page_068.jpg
2ce200f7299c82a8ebb7ea9452fbed76
cac1805e89f34d7c3037edb040b8b2693dca0cbc
108737 F20101113_AACEBU schold_j_Page_022.jpg
f7d24a54352ec09f381e792bbb23e72a
147b75eaff5a436d8e4f78fc78e9b4f9548d59a3
107701 F20101113_AACDWO schold_j_Page_012.jpg
76e713e21045dc29e0bc89c1da977adc
b38bb3648eee88372585773361363c3ae9d79579
37065 F20101113_AACECI schold_j_Page_088.QC.jpg
f0ee4f8eefe7d26f24af2a2ce4a59c74
90d30906352890448aff0538b8ed43594c12c744
8907 F20101113_AACDXD schold_j_Page_085thm.jpg
7099c18f0edb47a438be38286a819e16
5bb445db6f5a3a01173f13db245dafc65fed4087
97222 F20101113_AACEBV schold_j_Page_074.jp2
8d75e6e66ed36ffce9f7d8f1b0fb239e
5902dbf6968bb74ccfd436a32af147bc0dd85ef2
7210 F20101113_AACDWP schold_j_Page_071thm.jpg
c7a1ea4aaf04c9cb00a74bcd923fc6c4
7e451bdd8fb39e5c2683662d8e7876dfc7a7a66e
2288 F20101113_AACECJ schold_j_Page_088.txt
fd6d439acc3b8c6dc8f4d5606b01f22e
d20d2bbd9ca1d3ddcc7556436c54c6e3eaf872b6
116250 F20101113_AACDXE schold_j_Page_109.jp2
38d64bd4e8ab1f63100ab3d2ca9f7bd3
a5ec3408b5182b7a333cfe0bdfdb75f5c9fa4d63
F20101113_AACEBW schold_j_Page_082.tif
6241231cad6495605d91c67faba7e0ae
c15114d0928d4071aad25178e587a2bd2512bf5d
90918 F20101113_AACDWQ schold_j_Page_065.jp2
9a44e4527050bae29f7026286135cacf
7b31c497aac406dcbdde30c61f2ea28cdbe4d1a0
4455 F20101113_AACECK schold_j_Page_036thm.jpg
6f8c7a1212ba6bcc02f1758adda6d25e
39eb5cd47a1e27c0d8e5e1a35407480560cb79f9
2447 F20101113_AACDXF schold_j_Page_072.txt
b7eda506dfe168b0c6f8eeae286204ea
4f3d2cf524211663b1651d25573a789bc62eed05
25220 F20101113_AACEBX schold_j_Page_110.QC.jpg
14aba2358c1703d69ac8c9b79850fa5a
8cb6a7a3010aa9585f8167717834054d00240af9
14194 F20101113_AACDWR schold_j_Page_098.pro
845c8363d72747c6dd4325d7465e0e78
ee09e3272c7bf21a9eb10727d07c50bf0a2bfac5
110101 F20101113_AACEDA schold_j_Page_079.jpg
cf2df0393ca7dd308fad5684106ca95a
6f9932d2168cb48a6a7d37d46341b44b3742a63c
2241 F20101113_AACECL schold_j_Page_057.txt
3f15ccd4904e12d402c9ed7b0cadc846
063d5aa98869be95b492fcf60ae726a50cc79554
1051984 F20101113_AACDXG schold_j_Page_103.jp2
cca0a5abe2a538408fe3b2d7dfe0673d
c0bab3ef4fe32fb379e7dfa47db6eeadf59fc319
103029 F20101113_AACEBY schold_j_Page_010.jp2
ce5e1b226960fb5bce3ec9878b4f698e
566e635bf4f377b823bf12ebff32836842f60c71
36969 F20101113_AACDWS schold_j_Page_082.QC.jpg
7374cca37f6aeb82c391d578983469c5
6e5592c15d3f1ae9b15f087c097ce03c36846ccc
45623 F20101113_AACEDB schold_j_Page_072.pro
a9fe095a02526e44c7e0de09aff4df2a
dc2334e745de8ed40b4dc9c0c1d62d4b0d185541
8618 F20101113_AACECM schold_j_Page_029thm.jpg
d6f690dab702eb6b2666efabd21c7d71
30918f6c5581460906e3578775668240bdc30387
1051983 F20101113_AACDXH schold_j_Page_008.jp2
5409f90660854a920555a684f1132fe2
217ee3662702b3f0ad9470b9038f8b87416d90a3
111407 F20101113_AACEBZ schold_j_Page_041.jpg
201ab250e0a05cc40fb8c410e1c8e9fc
0484abf3f9deda490ada962c27ec7a0cc8715273
10521 F20101113_AACDWT schold_j_Page_005.jp2
98f70602cf727b6a9557149c5d5ff146
5795c218764a2fe90ffcc9f896fbbcc8c0b8d838
5896 F20101113_AACEDC schold_j_Page_073thm.jpg
c19c1c56e57fbff0b29ef41b4d7f680f
0c28d1b6cbc81ac7bf8a5fc027b18ff53c565a53
2077 F20101113_AACECN schold_j_Page_014.txt
8400eae3e19c31703b4e057f78c1941d
b8bfc89ae663dae615a0c07082d03ed6bddb9812
9325 F20101113_AACDXI schold_j_Page_108thm.jpg
8b31d881b4c5fecaeccfb1399a0afa26
c5fe9368b31bde28fcf548781bad6037bd5b3e43
116173 F20101113_AACDWU schold_j_Page_030.jpg
961955f0c0043ceb9c3cdf6849433414
35cf66314147fa308ee67a9c45556252dd4f45eb
934 F20101113_AACEDD schold_j_Page_002.pro
037ed4af11fa5a4838cd12e5a5d13360
4cb863087c5402fd0d121107b8e7384e8e3fc282
F20101113_AACECO schold_j_Page_038.tif
c643bc37a4c18bbd5cd3382401e94446
85ba3db6d2667aa93fa378ee20e66bbc8df1500b
F20101113_AACDXJ schold_j_Page_032.tif
6fcf3d5ff50df58fff1786e582fcbde0
119dadaad52b1cbbfb1f10807e443865634967b6
3773 F20101113_AACDWV schold_j_Page_064thm.jpg
53ce961aacbda27117b80de4a771b9ff
b8cb597453cc6b73eb4848c67c39aa376caf23d1
9507 F20101113_AACEDE schold_j_Page_035.pro
27d222f8aa61615c076f426d806d8a96
1647179de80b44be630c98792f386f16edfa6923
38166 F20101113_AACECP schold_j_Page_017.QC.jpg
aa279447d5a245a91ff1ff32e3b47e78
5856672b4dedaad5a3bc669d55b8105c160d3577
F20101113_AACDXK schold_j_Page_046.tif
d4fdd87f55afbddf3465f27f5b797039
bbdcdafcd9dbde6c3d8fcc09f1044c14b3c8c3da
117863 F20101113_AACDWW schold_j_Page_099.jp2
030bf34d100c9e0a902655225fecbea7
e213240df514a41c22889554773f80a9bc8281ed
9111 F20101113_AACEDF schold_j_Page_091thm.jpg
089f320eba7430973b232e47ce922ea7
3e7fe12de3c1d6d901b6435bb576477939ec421b
113822 F20101113_AACECQ schold_j_Page_082.jpg
ed08b24c3da8e329cc772687d2857f94
738332f78059f4ad7634c10b12467e0eafdc4186
6063 F20101113_AACDXL schold_j_Page_074thm.jpg
4ba1140a3fca04c666cb2cf965a09e22
122d06cd6d3b931a844c9a6763e6f6a4fffc1723
F20101113_AACDWX schold_j_Page_013.tif
a74a56f74f33a5aacec90f62411819ff
00961ef07b86dde00d9b5c4184081951b948a6ef
2047 F20101113_AACEDG schold_j_Page_028.txt
b3b8ec6cd41f8dd5597c9ee3aa3659d5
e6244c7faa41ff8d966f9076cfa4a0ab69b01311
55172 F20101113_AACECR schold_j_Page_079.pro
621b3a136f5e6a0fc602331581dc3adc
fcaf185d8b142312139102c0c555bef8cd27a84d
8786 F20101113_AACDXM schold_j_Page_104thm.jpg
0b059bcaa5767182b60b6b60aa38d146
b2acc222947bf3f466ae358da904f9ca21f9c908
1051956 F20101113_AACDWY schold_j_Page_108.jp2
a5fa9c21c13fff2cc33e17d6d247c465
8177dfa02cf9384bf57dd624f55bfde5415927c9
25079 F20101113_AACDYA schold_j_Page_065.QC.jpg
553e7f7c53708bbcaae333b1d7ab637f
9ac5c2c354f14670f44f68cc8879376962846931
111421 F20101113_AACECS schold_j_Page_094.jpg
fc1b7fcd326ca1c52a3f0c45acceef6a
a7a2a8bf4d2931845af13fb883940a69ded4ee7a
F20101113_AACDXN schold_j_Page_090.tif
f9f8bcafb7b46d8abe66e59e8e6ae710
a2aab4a7be26826f5b196c19026246378bcf88f0
37452 F20101113_AACDWZ schold_j_Page_055.QC.jpg
fafe6a8185537ff9dcdfd3e167592e54
ed0ed0c9c31b2d85e972d23057b2852e0e95e7ac
120220 F20101113_AACEDH schold_j_Page_096.jp2
8a8c071fc6dbd28b2054e97c2fdedf7a
fa145a6ef0708bc707ed3ddff1f7e99e64386fda
59124 F20101113_AACDYB schold_j_Page_034.pro
f2f157d41eaed20ab9b0ffd5302c2e39
394ffcf83beed11144457c10975a04d290f92384
125214 F20101113_AACECT schold_j_Page_013.jp2
dd15699291aaf4133cab2cf3cc6e1cf7
ff8f11b8adb3ae1204ee85d398a907d5a921314a
110192 F20101113_AACDXO schold_j_Page_029.jpg
ad187d23784843aa2cdcaacfcc4c3097
10a00651c3e3c62306c7398a478c88cef59f28ee
116452 F20101113_AACEDI schold_j_Page_060.jp2
5b72ea21c8b8cbb7f0983df9ad4ab4a8
344982cc09902820ffce9cda952dfb62161a4e7d
8671 F20101113_AACDYC schold_j_Page_044thm.jpg
55fd2dbaf6482f696ad77ad934a86cd6
d5ec950a1d8b1911cb945e08986714eac1aaed2e
F20101113_AACECU schold_j_Page_021.tif
6f165b415661fa10f66c6175def5c84e
2edd1a60adc130903da9dd2539b72d661fca237f
9038 F20101113_AACDXP schold_j_Page_061thm.jpg
d8c7505dcf0d523e4382fb7f2d320171
2f9378c6da368f49cb4af57b7042b993808a0ee6
113715 F20101113_AACEDJ schold_j_Page_057.jpg
9dbc9dad02f0bbfadf33cee28fe8d1c3
c8ae7e17bdcab4718667b50eb31077dd21c57f22
F20101113_AACDYD schold_j_Page_097.tif
90acacf258d8b4b943b8036d5ad78f05
f1b7c59a970dd1c0393ad2bbfa7005942136d962
50927 F20101113_AACECV schold_j_Page_027.pro
825eef53113a5daf660ffab82b3f87a4
9b50e686f7d5e7d3678ecb2781dd7ffae7302a6f
2625 F20101113_AACDXQ schold_j_Page_078thm.jpg
746faa783af9fff1e89f01ebf53729f5
05933f3f4e698577e301860e9e127b3bd847d028
F20101113_AACEDK schold_j_Page_019.tif
26c8820718df1d8c7f3e8b2915a17d8c
74372d4f5069f64c3a50fa7343dffd2711d61858
2352 F20101113_AACDYE schold_j_Page_008.txt
2ab69a330d74b1fed0811c6696e22e62
9036a69b80258d25ea0a0752b8d5cddf8bc14f70
91205 F20101113_AACECW schold_j_Page_073.jp2
0f42b703229b5039089ed8f29f58b03b
5bb256b30170924a008cfb23e7ed1c769225cf4b
121321 F20101113_AACDXR schold_j_Page_055.jp2
490db4f6def61ec31ba21f410193de77
b3dff060c85e6ca0cefca81d887eba6fab397cb4
F20101113_AACEEA schold_j_Page_012.tif
757d76c6683cdb3a99ca1c032514464d
4aed7cb171f59f298745f461bc106a7b339b6dbc
9311 F20101113_AACEDL schold_j_Page_102thm.jpg
6300383cd2e615743084b7cfa6b6bd03
54c7246c2f4fcba73c990176607f8a3015efaa5f
35785 F20101113_AACDYF schold_j_Page_094.QC.jpg
09b035c8d39bb99803e9ad5ec4cab89d
be78c7c2a14a82b443a8fb46347eb42900bb6a14
36570 F20101113_AACECX schold_j_Page_087.QC.jpg
3b63c67a10759517c36936cf71413153
bb15360716c59e3a32209df208757f9910228188
34807 F20101113_AACDXS schold_j_Page_104.QC.jpg
e86975e61b42aea0b07f7b4647bd2c4a
3614a92d3af7e02eae0c447ea40ac02af56ce677
3395 F20101113_AACEEB schold_j_Page_005.QC.jpg
f0120ce4428cc10f4df58f98b584a116
b6402def2c6b329cdcbe11afb32479201febfd71
2239 F20101113_AACEDM schold_j_Page_018.txt
1ce2110ca7545f103d572adcc186bbb0
da2788a0dfc48d06d4260f439adf8ac133c94a66
123921 F20101113_AACDYG schold_j_Page_084.jp2
8903a4b39a26fb432a85a63f00be424d
a8eafbaa8263dc8260e16690b8c8486acfc48547
29129 F20101113_AACECY schold_j_Page_069.QC.jpg
b2e1b6e8be4dec39201b6a4f3952fc66
98710ee21b3dadc7c7e23378e695f6cbe81d3630
122685 F20101113_AACDXT schold_j_Page_025.jp2
31cdaa61e1652cfc5e0f0f08978cca84
fdf7c88af56c5d1e34aca2c026e32e3fb64e0a9f
43046 F20101113_AACEEC schold_j_Page_071.pro
879224ffd6164ee1f36db93251a3c485
34ac5faa001b7e429c7741ec0f1fd3956e6ec758
9126 F20101113_AACEDN schold_j_Page_039thm.jpg
eca21a588d0d5099bd83ebb533bf9b04
3912161dcc565ad64f00a7d3275b80ece9f14414
113164 F20101113_AACDYH schold_j_Page_018.jpg
90e0e1ae776eff963de29b72116802a8
7f3f878263ce6c2ebf9c38ae278e96d4bc008f0c
F20101113_AACECZ schold_j_Page_027.tif
2113db51c6a79c046ccbaa3c4bc6469c
bf22aa225d6e3dc6a8694249de169cb20b2b9df4
117869 F20101113_AACDXU schold_j_Page_019.jpg
61dd99434eee40ad6d4c96c77bbb239e
7cd2482969b902ea3c0eaa60e30500371e558d02
53803 F20101113_AACEED schold_j_Page_060.pro
d121f8efa3d18712990273369ed773a6
40c50a45aba03bdaf75b93690275bd30bd25827e
F20101113_AACEDO schold_j_Page_096.tif
1324da25dc013dc5ce1c7d40a83905b2
5e97dccf5b52320a09e2c5637eed6dd98b70a262
F20101113_AACDYI schold_j_Page_016.tif
ee6affb7758812c7a816f870b0a7419a
0bd9b0340ff7012f4bcd1d7c5f955e920997587c
35243 F20101113_AACDXV schold_j_Page_015.QC.jpg
0130aa4418dca1f58ddc3d6998276a7f
3bdf6be216c3ffbb82d49bf0b50d5cfca4f4781b
122298 F20101113_AACEEE schold_j_Page_081.jp2
16e0756ce4266fdc4f3683382d2ae01d
6fde9a88b748ed14f77c1c68a7dcc82aa60132b3
115275 F20101113_AACEDP schold_j_Page_061.jpg
7d17e920526e97678aac21833515c11b
2f7f957dc849e8695a47319965e47004d70ecdf2
109130 F20101113_AACDYJ schold_j_Page_109.jpg
e3350155015a528ada8f5bd52ac2304e
e69adf51cd73df695ab69e223a44b4b2d8dbbab4
8945 F20101113_AACDXW schold_j_Page_090thm.jpg
aec18a40b62bb99647678c610bdec2e4
1c58ec2b856b10750d0eff921267cabb2237b0ae
8861 F20101113_AACEEF schold_j_Page_004thm.jpg
f1e3e83840f9b600585824cf8a2d1c40
36ffc1b7ce4c8d150d7a0578544cd7358d3a3be3
F20101113_AACEDQ schold_j_Page_100.tif
9ceb355b63b74525b6edabe6e0248fd2
c4ab7a1c52befef31488d838c8b6d17e87d8b274
2504 F20101113_AACDYK schold_j_Page_104.txt
a19c97ae1a5db2ebe83bd1c376fa8d9d
c31f7beae9e9a5ea12861cbdd156d619096fd88e
51985 F20101113_AACDXX schold_j_Page_028.pro
ee850fb4816ea298ccf8e843d564f808
8e25bd7e1fd2e95506873293e14e3650581e5a76
2248 F20101113_AACEEG schold_j_Page_096.txt
6e1ad3daed4f6f4cfcdbcf045c66cc18
1195faa768797673b1c9f9af52e4254ede1bc9dc
18061 F20101113_AACEDR schold_j_Page_051.jpg
b33c10aa3b2d087d705f57a5d40c4da7
611d12f387c08b10c0283fcea91f3abb5c0f2ece
F20101113_AACDYL schold_j_Page_055.tif
b2129c7ecf15adee19fffaebd038ae87
9da0d27600841291f062eaf533cc4e004bdb66af
8308 F20101113_AACDXY schold_j_Page_027thm.jpg
9f1882b3415bfe071300fad39c4df39d
86aaab22ca8629b4e1eeb54608feb4f2899791ce
37638 F20101113_AACEEH schold_j_Page_084.QC.jpg
c1f4664fa9c661976a43765cf41c2739
7a3b4f64bc1d56c647051c42769efd95d11fb189
116140 F20101113_AACDZA schold_j_Page_084.jpg
9e7fe4d06384c02ab83dd84bf17e0766
d7ccbea42b52d86f91b82d0b96a0cd44a2c2376e
57331 F20101113_AACEDS schold_j_Page_087.pro
27aa1c759b002a88697a74cf2c9ab8e4
822df2569fffd1536fcca1021610df9a5455335d
113067 F20101113_AACDYM schold_j_Page_004.jpg
368046d1c85004f3ad40bc59483e980d
74cacbd45fc9ffe782909c93bdeb869658fe4f8e
F20101113_AACDXZ schold_j_Page_084.tif
dc7a45a7b5ba49e1a456f56e9e905afb
2935e4af62d3bff2c627d320c36727a73ff3da6a
8959 F20101113_AACDZB schold_j_Page_055thm.jpg
cc3c9ab6f7de8c56764a507f0f750eb2
66cb89773d175809e1835f876b59b06354d3ffa3
106308 F20101113_AACEDT schold_j_Page_031.jpg
7b1a59610cb336d66a503da79b7ee6ae
6be14a44f632944b8adf567d49f602b3531d2614
56804 F20101113_AACDYN schold_j_Page_094.pro
e9c5890bd1ccbbd70ffd21f4656182e7
1d4c1542a78c407c1eef3d0109794aef239cc72c
54846 F20101113_AACEEI schold_j_Page_029.pro
ab146fe5da7fc5c3b96478ecd5766ac4
031639eb11e3d7842378acd641c490fbf47dedb9
52132 F20101113_AACDZC schold_j_Page_046.pro
df51c389fd37fad9e52718f25e8a6ffe
276fbc1bde6de60c4027ef40e7279cafefbbef1d
F20101113_AACEDU schold_j_Page_091.tif
97621b56129a677ee05019135fc61673
de980754f83a99f20e5f8e5b7f72d84615754f79
55362 F20101113_AACDYO schold_j_Page_032.pro
83ed22f23c74b41f585f6292da3b819a
b513e4370a26d1d79137e29371538280f3a70023
128267 F20101113_AACEEJ schold_j_Page_105.jp2
4b66861776ee42e4f9b533830e25ceaa
ae77a22e94cfbb07a0b64f3277f9a786508f9269
1062 F20101113_AACDZD schold_j_Page_009.txt
74b930b0aca073a479c95b19cf435686
9e33a14896518c592834f000bde845b85b11dc35
F20101113_AACEDV schold_j_Page_052.tif
26efd959bde847ac6392cccc9ca5c0b0
0a2956088940cc854678659bb1c814b841f21c8e
2253 F20101113_AACDYP schold_j_Page_099.txt
e9b4d0def25a31c0533c2cb8b32643f7
3738abf30e8b6bf7167a004eb6650bab533ea420
8807 F20101113_AACEEK schold_j_Page_100thm.jpg
319f25f0e1277c79fb9dec3c06f3dfbe
5551ed780e96a5ec14590b54554614d7ec5505b4
391218 F20101113_AACDZE schold_j.pdf
b2b046dc6d53876889c91390da728d85
c0a52e818424062ff407316f969af59664dbf473
25980 F20101113_AACEDW schold_j_Page_074.QC.jpg
73e8a9fda5fa53a58a7231bae4f3551d
f6cdcf66e62d227ae9c673772d452bfa88dec08b
8395 F20101113_AACDYQ schold_j_Page_054thm.jpg
c806b10aa6c9fa69a0c081b64ba34467
00db2230760c082f4c4ec9f448dd95bdee0306ef
36654 F20101113_AACEEL schold_j_Page_004.QC.jpg
0c35bf48a72346b35521c313045714f5
7da95fa8d0d0a9f750975e66e67064f78e0189f6
2492 F20101113_AACDZF schold_j_Page_107.txt
75e3fd8a3875ac7a8129933b62d930d2
1f5d13f705eb0d95a105ad88845efa2d1343860a
F20101113_AACEDX schold_j_Page_104.tif
8bc5d5bfeec5d69cf14270f4acc12933
f5626e6de7f6ef6c3a4ff4707c96dd02e1b3f5c8
115406 F20101113_AACDYR schold_j_Page_012.jp2
f479d2c3549fe35ff89eea868b5218c3
39da19ffa12527d939a27c18c28482cd25bd8d09
F20101113_AACEFA schold_j_Page_025.tif
72c0b36dfae0f39e9f3cec35ea6779a2
c865783f4b1be32f9b763e2fac17f3f65b468861
121510 F20101113_AACEEM schold_j_Page_089.jp2
fd1bf2872594944c4a058aad4471e9a9
1412da30974d22cfbabe5f3b918930d121423e33
2071 F20101113_AACDZG schold_j_Page_053.txt
4bf5c670d61e1b56fd15cae751322bd6
a756220be715e3f11c4508985818ec38b56e1fe5
115000 F20101113_AACEDY schold_j_Page_016.jpg
7c21889372ab99819230f268121fd429
71ea68c8acab5bf00934945403567b37bff7b5cb
36742 F20101113_AACDYS schold_j_Page_090.QC.jpg
b09edee34d47a195c676e6cf70eb241b
8c10ddb381577d373d7f219ae7187ab756dddb2e
15172 F20101113_AACEFB schold_j_Page_011.QC.jpg
e2d99f215b4477d1b4da0cfd0d819a07
02f6c45c079365ab19d155af290b6abfdc4bc6df
108500 F20101113_AACEEN schold_j_Page_058.jpg
7298e0ffe59cddc644fb42dad7a1cff1
41814109e3c777024c105eb0f97aed8f5a5fa4fc
104707 F20101113_AACDZH schold_j_Page_097.jpg
a5cfbaae9ae5d7ab3859ab5e178b3d9e
6067779df09423492097fcb15986ce612d6ad778
4610 F20101113_AACEDZ schold_j_Page_003.jp2
be4381a99ed3f9a12e40aa30696e0ac4
83f112183f02b190492738ca7f66f05b0e9c3456
363 F20101113_AACDYT schold_j_Page_051.txt
3ccdfafd0c9a262fccfdae143c7933a4
a7c03cc95b957e731ca06a41dab479114b342e81
1494 F20101113_AACEFC schold_j_Page_110.txt
641422883b7d6bca39f3d72c0a2f6ffd
a8116d65cc04dd46ddf07e17da15e7c0eaa9baf8
118181 F20101113_AACEEO schold_j_Page_006.jpg
881922e89650abbc292a82c047b379e0
87431ad60fdd5de5fc91a4ed2607f51c1c459c15
8116 F20101113_AACDZI schold_j_Page_036.pro
2daf168edc468aaff21777e21301de36
4e204203af72dcf2f00ae24dd2e155a2cf37e773
F20101113_AACDYU schold_j_Page_035.tif
90fda5a3516f49d399039c751a656e46
dae0ed9bd596f97827638dd9e34d5fcd44779ff0
96251 F20101113_AACEFD schold_j_Page_072.jp2
8c60ff30047eacd430263398149a389a
f08428166401ddc5371325596ea21bebd5f614ce
F20101113_AACEEP schold_j_Page_109.tif
e371b548faef9dff55fadc9a80d11078
cc88e6a662ce579c0e485cda9dec90ad98661e11
23079 F20101113_AACDZJ schold_j_Page_075.jpg
3d6f3fd54a67d32ddabe8ea29011cdf9
dab85693e0968885853eeca37e99ed464a146a8f
54504 F20101113_AACDYV schold_j_Page_022.pro
11da55ba0c360106890f8a3ef7d0dd83
042ba45e1f59ece3d3980951b714223068f273e3
F20101113_AACEFE schold_j_Page_069.tif
5ba9a003964f088aff3a5ae44a4f1a5d
6036a428c4f27b22a5bfcb76a011408cc3ae6b22
9149 F20101113_AACEEQ schold_j_Page_080thm.jpg
c38b7b98de815baa0a1b649d536b725a
ec79994ad83397bfb985f404416f72e41847be12
116618 F20101113_AACDZK schold_j_Page_056.jpg
94f25b24b5b5daf0620b8a3d7be0433b
fcc90c0d8e682056014fd4b4913f6c0c5f341d55
F20101113_AACDYW schold_j_Page_110.tif
a980006387ac7af46e9be7582c8585c2
aef8c2a4f3420f9a75db80d56775458abbf15ce8
15147 F20101113_AACEFF schold_j_Page_063.QC.jpg
c66779fc029f912d6ae215b181b40690
fe1acc2c763489ec3f39b389c5a2c2e8fbd5e927
F20101113_AACEER schold_j_Page_017.tif
ba3550270e0044357d94778bc2766237
75052c7688adf69b064f2f7805e657ae89fc61c1
F20101113_AACDZL schold_j_Page_057.tif
205c0d600532d88b0a80d192effcf304
c19628ae130c4f8abdd6a4c1dda9570d99ca778a
99902 F20101113_AACDYX schold_j_Page_045.jpg
1d62ac6b14981f99db6370f8b6351b35
7e432c8fdc26025ea05e530c9a35c73d3bf7fc72
3219 F20101113_AACEFG schold_j_Page_007thm.jpg
888699f3f982a8b31b3218490af5ec83
71630e50d213afb1c8002cd7be1babbb860b402d
8744 F20101113_AACEES schold_j_Page_026thm.jpg
c4d49fb2a64fde777c517b5d888d1814
e3f3ad6d07f679c160d70784c6c03f307447e282
F20101113_AACDZM schold_j_Page_067.tif
0c43f0e66a1d6f250fffe8ef44a5bc1d
e30aadb9bd9128c132199374d6aa96d3242cf063
119703 F20101113_AACDYY schold_j_Page_024.jp2
572f9832ca8ab14325fc9eaad806fae8
0a678504d1785cb707232660c56e72272e4a35f5
9039 F20101113_AACEFH schold_j_Page_103thm.jpg
e75e82445d7fa3d53d5b568b37c2c2db
9b399c2bf21cda90bee465d337999bf7c9d8b548
30935 F20101113_AACEET schold_j_Page_047.pro
c0be24f8a9c15df875a9d0fe7cbeb52c
9bfe5efa512e9455fa31493ef427f830c24371cf
34923 F20101113_AACDZN schold_j_Page_058.QC.jpg
9442b247e17b6ad9cfc98e4d52769a58
a4782fb4233a2706f0da031452583bebdef8959e
112099 F20101113_AACDYZ schold_j_Page_086.jpg
985105e0ce17ad6ecce7ca8884007584
10791a41ecacd5953c030c3e945bc84557171acc
F20101113_AACEFI schold_j_Page_039.tif
b0952ac83c22cf390331966d86011cc2
31886afd00e52b4de1381f3dfc2edd94b4fe24d2
F20101113_AACEEU schold_j_Page_097.jp2
4e64e0d1059abaca021cf34e27d4a664
0cd571ad0938440942a2a267594de9b33ff5d372
2129 F20101113_AACDZO schold_j_Page_046.txt
50a40b28607366273eefe86f190db41a
4f06110d5793170af0661fbffc754b465d5fa700
7797 F20101113_AACEEV schold_j_Page_076.QC.jpg
e6048126c3350034d7b1f96c4e4ccc45
3be0a02f0fcfec3f3769b0afe4cba2dd9c59a4ae
33042 F20101113_AACDZP schold_j_Page_027.QC.jpg
8b500e206a32949adb3f7f796d1ad48f
55810042e69db38baaad6f454047905681b98f04
9114 F20101113_AACEFJ schold_j_Page_016thm.jpg
115234c4c28ee67caa1bf91c75d39b9e
1bee25e134200b9deaf219d0b861ef1f318990ef
8661 F20101113_AACEEW schold_j_Page_083thm.jpg
6710ea3cc88ca7ae71e1d8604d94ee59
38adac39612809dcf1e300748bb4e32ffb78f38c
25062 F20101113_AACDZQ schold_j_Page_073.QC.jpg
f22cbf4094b536adc87ca2bbc8da9d12
a3eb78d0ff02247b01e406bf4669ea287fead248
121570 F20101113_AACEFK schold_j_Page_087.jp2
35447344808cd3d0eaa5b7e736fdab83
0b244a1967345f988d57d263c4b664f21d4e2511
52690 F20101113_AACDZR schold_j_Page_015.pro
d55a86a31f1cc4ea150b8291ccacab8e
c309a9aa8574d87bd1e6be516db42797a35b3126
38554 F20101113_AACEGA schold_j_Page_040.QC.jpg
2d5ba6bef91057de6fe9ab0bf3139d70
0e1e0724b1f3528fa34fcb764630cc7465a59b33
101157 F20101113_AACEFL schold_j_Page_037.jpg
de05135eed71a757499b9502e328101d
b374b64a8aa4b6557dd0f731e20feb35bec9ac6d
34641 F20101113_AACEEX schold_j_Page_044.QC.jpg
c546fb4460604ff7f1cc2adad53996c2
0caeb670b9444e6e1853f0bc6bd506b8c26f9f65
F20101113_AACDZS schold_j_Page_011.tif
3d5543e7e59821a52e1d15692b68d047
eda5e899d1fc58811a6f13832ec94b026a8d3c1a
F20101113_AACEGB schold_j_Page_024.tif
67fa1fa06cacbbd6449fc6c209f65fa6
9400bfb3dc74592d9dcf2f2193e524d7fee72f9c
35986 F20101113_AACEFM schold_j_Page_041.QC.jpg
66b1aa4aad31f87704e3e470b7e1f6ff
9563e9441e8aff9c7326e0ec6d20990a836b236c
F20101113_AACEEY schold_j_Page_098.tif
c32fdce7c5cdf04d33034b566d29a62f
d541c2887f852d1f428ae4d9395455f39f8d5c5b
55773 F20101113_AACDZT schold_j_Page_062.pro
74785c66b5dfb63f681b43bad47c430b
be8057768b4d42ff4bbd72f4407716864310bf75
107378 F20101113_AACEGC schold_j_Page_037.jp2
d2a3ae607e7abf11c4a9d6d08f6e0742
742f328d66be1b9614ed930e060e8ee274bb575d
F20101113_AACEFN schold_j_Page_083.tif
f0bc8c9b4e3545b971d0048662cbc699
fbbf0e0c1f988559d8dc69f119a2e52545bd7197
8930 F20101113_AACEEZ schold_j_Page_105thm.jpg
8320d7cb0be3a3636952151d7e82732e
d3205a25291d81032f1f46a5e00c48519d2772f4
108412 F20101113_AACDZU schold_j_Page_095.jpg
214b55cee6b7c9e359ade96f2695ee54
0a18ba698dc4f794b2307b76840b67b50627e6f9
2904 F20101113_AACEGD schold_j_Page_098thm.jpg
ec4aeb3dabb693533cb0c68a6b234db9
eb09f8c840b839c29cfdd03dfb764d069f3e2730
F20101113_AACEFO schold_j_Page_053.tif
ef4d26fe82f536d70bf1a6bc7d113b03
c0418efa29753fea32468a495ef1925e4fd1fcb2
81 F20101113_AACDZV schold_j_Page_003.txt
24c53e1da56989f50dc574cabca73d8b
81d4d9ad5d01ed8bc8440c672284a2bb9cb9c5de
57493 F20101113_AACEGE schold_j_Page_085.pro
ff46b4a919a4890effb640dceef63a65
07e1d6db6a121f259ef066559a1ffdb3777b6aa3
57970 F20101113_AACEFP schold_j_Page_020.pro
0d80d4b79b3ac1e48b850fb9e034c5fa
7b0f3d5b92d099c1e38aeb9a91d2e7e7deeac3db
6389 F20101113_AACDZW schold_j_Page_006thm.jpg
929777b9a1cb9b3e9108a4f026bd2beb
5dc6a40300539c1bc01868e23c36847b1e107f67
113315 F20101113_AACEGF schold_j_Page_089.jpg
74ab4e12acdf09c1c942983f0d080487
a3ab3b1b919c45e87170e5556b9d08a0b5f7cc53
9018 F20101113_AACEFQ schold_j_Page_025thm.jpg
7bcfb3785b2b350808ddd3ca235c5857
49acdf1ad2851d6777cbecd9f63c3ce9b72f59c2
113629 F20101113_AACDZX schold_j_Page_015.jp2
198ccff30310703fce1ee9bcfa99b622
9380e0bea7afc619d6fbd1b083a90195ccb5b218
34754 F20101113_AACEGG schold_j_Page_031.QC.jpg
9067ec9374e275034d07c7127b2fefec
1548d0d8fc69144de5f65312e3709ead4fc2044b
116281 F20101113_AACEFR schold_j_Page_040.jpg
d643a80216f493d45382286bba6be861
2853057da9d910e026012b2a5b73f346fbc3de47
55041 F20101113_AACDZY schold_j_Page_004.pro
eec2e5e9f53f2925102530aabf834507
8bd891f65bebe7755f74daa6e2d2b900a400adaa
9161 F20101113_AACEGH schold_j_Page_107thm.jpg
d8811f232d7ada9bdc462b5f886badef
3da9e767eb3d5af37aa17804e595ab163f962d67
1067 F20101113_AACEFS schold_j_Page_067.txt
c6207a8ecd70365df78cdb6b82e18472
ce9d4e113a42da1e84099eb4d21daf03d38dec8d
128760 F20101113_AACDZZ schold_j_Page_104.jp2
19c0fe1c16d4afc45b3e0eba3c7ae487
03bba64f7fb07f703bd5585621c586c757bf44fc
F20101113_AACEGI schold_j_Page_045.tif
589da98d7421a7e07155472a6b2f2163
f1ac1095db41694927b3683fdebd7543215a9199
56539 F20101113_AACEFT schold_j_Page_024.pro
e6abb3fcf563005d1c92bc55584d94d0
5d7b769f8c1e1e06ab53909e2f28dc584064650e
37281 F20101113_AACEGJ schold_j_Page_016.QC.jpg
73e1741bf089af34761184201edc52f6
6a627ea66e2715dcbc78cf7e4bfb1f8f0e0649d4
F20101113_AACEFU schold_j_Page_023.tif
299d3c2770454795ff6632c88e3beef3
f5ccf3e088fb0ece64896087ca4353748441e44a
62753 F20101113_AACEFV schold_j_Page_047.jpg
19f8323f48100bc64813ebd052b7f1e0
eeef220ae1afadb5e04ad16b41a5a9987434fbeb
2251 F20101113_AACEGK schold_j_Page_087.txt
b2cc72beb5e00016d7e5d1219f1ddc69
d78b43514ed52ab10dc1d97b97343320cea2127e
58351 F20101113_AACEFW schold_j_Page_061.pro
2cc4d01b7bb19d225b0e9ecdbd699062
a17834898d2b652acf4675eead791d916c18a374
36208 F20101113_AACEGL schold_j_Page_096.QC.jpg
5d32b189b64473ba6d3fa4b26ad22f4f
c4f0ce786d8b322620362727a0861e8b925c9013
111164 F20101113_AACEFX schold_j_Page_021.jpg
7b28b0902e914363130f8183df82ac4a
603eb48df92e30ac9632d2310df49b32ba988ece
50569 F20101113_AACEHA schold_j_Page_037.pro
99a8499ba5cb495dfe8cc5cb27f21a37
f3f8fe6eeaea9b359a49e4b8e0e13453c236f167
118011 F20101113_AACEGM schold_j_Page_079.jp2
586a6d8aa22e419081fe045e27551d0b
c75d77e5a1375a421f4a48073566a3089a9c38ac
34407 F20101113_AACEFY schold_j_Page_012.QC.jpg
c1b181a8c91931a219ec067095d41a72
f541004233e9061120d6dd7b2877ccb67c41cdd4
36419 F20101113_AACEHB schold_j_Page_093.QC.jpg
d37052954c4afbfc63a617add69ead23
96a39a7041c6bc0ae5e690d7c88a06852878d26b
18612 F20101113_AACEGN schold_j_Page_101.pro
840511eaae5762d3bd05555db2e8ce0f
55e6d26a20f977d35fa64a9c3b1ce58c22a6f8d6
9032 F20101113_AACEFZ schold_j_Page_081thm.jpg
b59d7411dabc9407af4f01ba271ae3c7
d4dbf014cd412659657055ffa5a2a9c4b33f243a
8362 F20101113_AACEHC schold_j_Page_045thm.jpg
2bf0520c8424d2d623ed05c03965d73d
11d851fe0747d7793f9af07dcbc82b6aceef50dd
53386 F20101113_AACEGO schold_j_Page_054.pro
cb3555e7836155840b5a5d9fadbb731a
65583212b9c48ea775ccfcc33a579c3d86e07b9f
7370 F20101113_AACEHD schold_j_Page_069thm.jpg
0efe7a70d13121762dcd71c4557fc6a6
7356e69f5ebc55edf4f083c66ffe4c920d19b4e7
42116 F20101113_AACEGP schold_j_Page_066.pro
3642554b5f9d4b8a704f9370ecfb6865
2697e435afd243ba7d0a339a507211c9cbd52c79
9117 F20101113_AACEHE schold_j_Page_089thm.jpg
c9ea045e18396d5bcde450aa676765e2
201ef0206dda0c422d4726312ed9ecc739e45937
106545 F20101113_AACEGQ schold_j_Page_053.jpg
200712af729404481219883f8fad92b0
c8a128aa529f136c60c7ec320e2bd1d761e64c05
F20101113_AACEHF schold_j_Page_081.tif
2e85eb0d7157625012fc5d284a294042
70323df4b8f2d209608ef71e868f59ca6a220319
32047 F20101113_AACEGR schold_j_Page_109.QC.jpg
f107dd52b78d037571dad4fbbb81fa4e
100f1f6c57668724e5579fc3d549d0f3b95e1be9
F20101113_AACEHG schold_j_Page_066.tif
9e0af44b5b5664926241c1cad8db26c4
0d9a711fd03f4e8bc92a63c5738e273c958476c8
313 F20101113_AACEGS schold_j_Page_048.txt
034628afa6c0e2727760ed4620ff1b35
96701bd26b5288b23c0852a62bdedaaa3c3deed8
5554 F20101113_AACEHH schold_j_Page_002.jp2
94669f024ef2713f1595e6e61702ccb6
1b2cdfa10e5ed2d6ca7f2860404cc947293e02db
97822 F20101113_AACEGT schold_j_Page_069.jpg
2f677772d53f33961d1abcabf10a7d54
f0f32284a655a93f4f89c920026d288cd466990a
165454 F20101113_AACEHI UFE0021207_00001.xml
bde9c31cccb95a8ea7d6e9394d923dc9
5c8296e9e0e13709d2d181462b92fc644d3f46b3
112290 F20101113_AACEGU schold_j_Page_080.jpg
b881d4737de5115bdb2932b3fb6a1aac
d1ba3fb289a5a77f38dd32866d1ef2469af2b67d
2262 F20101113_AACEGV schold_j_Page_080.txt
3510a8432119f73cbe8c8e1acf9a5415
e271152f8dcbe8c7ec4bb11e3f0eefadafba8ef0
2196 F20101113_AACEGW schold_j_Page_062.txt
0000c0e53fb0d31dd06b869c942d4f43
7402aee1b487770a994158bfbec383e863f91ab3
97093 F20101113_AACEGX schold_j_Page_070.jpg
e2664722373cb7c3233cd91739583824
99d90d177d519d83f16bc720118354c0b109bf09
103487 F20101113_AACEIA schold_j_Page_027.jpg
ed9a35e942ec6a907c05fbf75e11b13f
d0ccd42624f219711ba0a344d55c0b6ff1670388
3939 F20101113_AACEHL schold_j_Page_002.jpg
a0f1caea798171d6647f4840f05f44c9
455633dc063e12c0dd2b5940aad98e2b15d2864b
403 F20101113_AACEGY schold_j_Page_075.txt
02a4f2ad5e35b8a3657db8902bc9ffc5
6a7ba2d46e9f69552793dc5acee330048ed16c46
105218 F20101113_AACEIB schold_j_Page_028.jpg
5821ca8193a4cff118eee49fbe08c17b
8fc3e91a7c9b4169a6cf90a7891f6c343556a20e
3537 F20101113_AACEHM schold_j_Page_003.jpg
f9c1650133782ed3305d9bf8c4f36268
518c888948b7bc8a6a6f2a1830a75d31cf5e81ea
38461 F20101113_AACEGZ schold_j_Page_061.QC.jpg
bbe1b0363e89bff0d5bb6bcfc6f3b333
dd9ff42c206122a656108d1bdc0af7463106fc14
99395 F20101113_AACEIC schold_j_Page_033.jpg
311ab5cd59a89bfc242cdb3c2eb14e33
282b054da0945e57b75ef6c2dc848fcb2b397f11
8916 F20101113_AACEHN schold_j_Page_005.jpg
bf0d859b997e80344e92c24271d8bff9
40d3b484958af967b600bde0076ec8ef4e862836
114811 F20101113_AACEID schold_j_Page_034.jpg
3e791f1c43d87da92aa9507a1428e618
31b2037bbdc6683c36ee51645723b640a18d0aa5
54389 F20101113_AACEHO schold_j_Page_007.jpg
0a988a77138a3bebbadffd63ae9e5f6a
d0c83da4e262b3fcbdb01f826f642358e6b056be
22091 F20101113_AACEIE schold_j_Page_035.jpg
5157d10c1fbbea76892f7485ad1b8a1b
0d4a38f8d1222cf279ec14a82623edc83200e277
122398 F20101113_AACEHP schold_j_Page_008.jpg
f3c93ccb548bb34b78127c0ee4bee9d8
0f70c1280cd2130c7782f983d7b5f4bf2e92c18b
37629 F20101113_AACEIF schold_j_Page_036.jpg
03fceb81dacd7f270687828bcbf17bb6
f7a372893ad6312266fe3457896e99ee781792a9
59979 F20101113_AACEHQ schold_j_Page_009.jpg
b7b74bfb1b3b93058c60c3efb98642a9
89658ff22801a5fba4d6af14bfc94c2764dc2910
114921 F20101113_AACEIG schold_j_Page_038.jpg
e6c3542f06a0797cabaca5901b9a39db
593d0b1d87e75cb42c4633a844e3f71ffbd47097
97610 F20101113_AACEHR schold_j_Page_010.jpg
1b17791321f3e10a7812bb795b55ffe5
242975c26402903496ee3ae5be9cd0477fcb7fbc
113142 F20101113_AACEIH schold_j_Page_042.jpg
409e6c78cb6020370bc3461c45e24827
194e2a36fd62c68298c3380373d09ba100b5d852
46918 F20101113_AACEHS schold_j_Page_011.jpg
62d3998ae90b5ec7ae79ae93e88f001b
7ce22d0b59aec64ecb0a4ec7caa5070cd01392cf
107431 F20101113_AACEII schold_j_Page_043.jpg
38471e1379504260e2f69450800c6af4
b97d8541b419867f5cc7161582b931570b9fd0c7
116947 F20101113_AACEHT schold_j_Page_013.jpg
b0eb84189ea103e4685b27598bf75c44
9a8f65fefb9d0b23450adc5f794655dbf37bf3a5
107105 F20101113_AACEIJ schold_j_Page_044.jpg
65ecfce593fea5c13e549eb8ba12fc7f
c9f71c9733b05f0b0c95cfbdc6001885366cb62d
107359 F20101113_AACEHU schold_j_Page_015.jpg
313e88d622b916d95fa0dfd98b173402
1b8e531ce7c844004e67485a8682a60a75505258
103329 F20101113_AACEIK schold_j_Page_046.jpg
8d04622a21550b669aba8ccfb47070e9
bc1dc27a01911adff9a9159b360b40541dd0c893
115587 F20101113_AACEHV schold_j_Page_017.jpg
a25fb3dead0cc1c27331d3c2f0ce5313
f70a2df908090e213560c903a9f77090fa6ad940
19887 F20101113_AACEIL schold_j_Page_049.jpg
71ba0c221ea92704558d54b041e9af3d
ead39e8da0f413907f8d460f3bd40448f6f19ea5
116827 F20101113_AACEHW schold_j_Page_020.jpg
0441a36fce3e763a44c9829309b78328
89f8640d947bfb2df2b972156bf38223ae849fb3
20597 F20101113_AACEJA schold_j_Page_077.jpg
a952ccd5b1bb00b4d600e1e4226ea962
8d0d6469ead30f91d0d9703bc1f9c50cee890c84
113449 F20101113_AACEHX schold_j_Page_023.jpg
662d401bea89ca055d91646c127dc30c
eca0194a18d3a1336dc5c3bdb7dbc4c4d04e912c
22745 F20101113_AACEJB schold_j_Page_078.jpg
5cb486da7195dfb6270c2c8b32f1d96e
1a6e5a38a9bb2b1cd8418478db2a958545f7da92
24703 F20101113_AACEIM schold_j_Page_050.jpg
117d3aeb348f917ffa94662c82ccde8a
d852e7a77046588b2fac2d13f4d0f4944b243b4d
112803 F20101113_AACEHY schold_j_Page_024.jpg
abfc824fdb1832e2fcaeef9b7bbb9c03
2a59e230cb886a147c007610af7682bfc0e53d22
113882 F20101113_AACEJC schold_j_Page_081.jpg
60fb1e8d0720c86be801c518a852c727
7530a798fc3b17be8586e23d53f3fd77702c4c6a
97192 F20101113_AACEIN schold_j_Page_052.jpg
3a08ac75e81877bd515647a3f4a3cbfe
613869841a1f1ed81d5c37b7a0282f5f5f6cd0b9
111400 F20101113_AACEHZ schold_j_Page_026.jpg
7a4fe38cb70cb7f1aff15b2410abd24b
65b05774b186c3ef621a46e1b2b267cfef5e5699
107213 F20101113_AACEJD schold_j_Page_083.jpg
bd96e4797d55f7eba9703ec6643703c2
804f6eca5d758ca536a6bebebf3151beaf6850b6
108915 F20101113_AACEIO schold_j_Page_054.jpg
350fc5f18572e08e8952fafcb80af1ea
a8a3717fa221f8f534f63451ee34ae91d177b4e8
112675 F20101113_AACEJE schold_j_Page_085.jpg
b98bac3816af7ee4b5a5bf61f4b9d017
c54401ad968468055fb3c6de8b0e0cfe697b74a2
106089 F20101113_AACEIP schold_j_Page_059.jpg
efdf1ee4a1b7ba11a185ea9cdb1bac28
6ee337c764f43c83b602b235acbd426797cf4b2c
112452 F20101113_AACEJF schold_j_Page_087.jpg
4fcfd28a4c12000ef50e27d2084c1b6d
a869ac2d659c2b1455630084ed8fd759eacb93fe
109353 F20101113_AACEIQ schold_j_Page_060.jpg
5c831d6164b3dfb1e1d25ad5965e43a4
aeb8a6f8c97a6cfc1c7121142f746e50a30c7240
112927 F20101113_AACEJG schold_j_Page_090.jpg
cd0323cecc001a8d5eb3826398295692
0a95d15f45e2ebb36f45c9ffff856a6c4a1f4292
111206 F20101113_AACEIR schold_j_Page_062.jpg
a4be26445df1877d5a3af2797def912a
1a0b4acff27fcc11ddcaf8e22ffebca9b26f5c3f
112062 F20101113_AACEJH schold_j_Page_091.jpg
f06e89b5d5c110ed5d1722363ccbea6b
b5682473e49fdc484017f08aa252a9416d6403b0
45830 F20101113_AACEIS schold_j_Page_063.jpg
e0c90460233bfe72c90bcd74d54c654d
07461c5907ddbdc61f7a955a1289a573d8932e1d
112832 F20101113_AACEJI schold_j_Page_092.jpg
9a148e485cf98a36017992f23ec1a188
e37edc058e61dc7463606c16c53dc7b295893405
39050 F20101113_AACEIT schold_j_Page_064.jpg
20c3e772341355a301ec924c20294328
201734e1323d978b18f6efb24e258b5bbe7a73ca
112126 F20101113_AACEJJ schold_j_Page_093.jpg
e2d2d080b6178169932cb3d602c6dfce
32ced02bf32a1598787f5e67c152ba160b6c8278
101092 F20101113_AACEIU schold_j_Page_066.jpg
4189647f7f87fc8628424258b4fad67c
e1863effe68a9495d46e95f96138029ec5f19459
32216 F20101113_AACEJK schold_j_Page_098.jpg
54388b3814f00d4ac969c1ac88aedb89
73895a6b656cd2ae5bffefa104e4e23ff9487710
90923 F20101113_AACEIV schold_j_Page_071.jpg
58a58db6be9bbcb61a27d38ec5a5639f
94c50ce28e3c78b017d42960013d30c395a02b21
110251 F20101113_AACEJL schold_j_Page_099.jpg
051bbf9d25b7d571f1645ac27bef760e
17b261c955a835e892cc221b69e70a7e8ae02ec7
94811 F20101113_AACEIW schold_j_Page_072.jpg
4d0fcdc7d6487c2dd8189af550700697
b4cc036e0bbd3ac64ac5628a35950231f0b796ab
111396 F20101113_AACEJM schold_j_Page_100.jpg
6ae05ca39065c10e485ba4a2dce02fe6
5f36e15b08149b069065803ad098e8a1585b727b
88626 F20101113_AACEIX schold_j_Page_073.jpg
66a8289de5a092c6986efd3879b075e5
193cfe5aca9057f2b729ccc813d54fdf9e8bc48d
49981 F20101113_AACEKA schold_j_Page_011.jp2
d5ea00206347c618a0e00c70f9264004
13d6dac2ac3ab4e5daf8ec223364935104e9ee7d
97417 F20101113_AACEIY schold_j_Page_074.jpg
ce6a79eb126c412038cc5c0146bb30a8
849a81128d8b711b3d76c915ad8795e3cf88f512
123508 F20101113_AACEKB schold_j_Page_016.jp2
ba4a2c2e79582bad66ffea91203d7e90
ddfe222a333366166d203b8d7991e4b9663c1fd3
39430 F20101113_AACEJN schold_j_Page_101.jpg
82cd7a2c78b74339ac5dd302d1ba2c35
3d679898fe925f1ee68d01524e247cdf3756a6bb
22519 F20101113_AACEIZ schold_j_Page_076.jpg
c9dfb8f1730bbbd03db65c8a46338a3a
e4ed2793af1db760a232f5076712147d75f8f25e
122448 F20101113_AACEKC schold_j_Page_018.jp2
054a8426073d6255a9975cc999a7ae6b
fbe7939dd9d477af8135c177779833cf320d41c2
122704 F20101113_AACEJO schold_j_Page_102.jpg
558a56da7de71321d7f5d8f54fbf2eb7
67b66c44ee0c75f5201276dc9a17a7f3046c6085
124935 F20101113_AACEKD schold_j_Page_019.jp2
3404d0956bcf47ba1bdf7256139b0073
3eb395b08605a02aeadd33862935c91d1233ff32
113741 F20101113_AACEJP schold_j_Page_103.jpg
4d6b0c7db31c3709e4683e6cab4119a1
35da12fb81481562e81c4b29b181661248e82b47
120211 F20101113_AACEKE schold_j_Page_021.jp2
b7fe3c155e85394ddd3092b0063ecbf1
b9bd1fa95b912b1d3223287512f93737b01e7519
124977 F20101113_AACEJQ schold_j_Page_104.jpg
90403039af86b215f82dd3eefbff7e71
bb2d86b8bba1f5d6c24caef8f42cc7ef8fb3b409
122695 F20101113_AACEKF schold_j_Page_023.jp2
0e594d73ee8fe25d2aca83c5bc2101af
873381210c3e4071f95e12cb09bdd64ed4b18911
122143 F20101113_AACEJR schold_j_Page_105.jpg
73a03b64634e0dbbe9e95d7404e9b9d0
588c2fd5be9f50e293dc7a4f08e151965460ff42
111249 F20101113_AACEKG schold_j_Page_027.jp2
fe95db7328e9d75cec56f7da9d5b8c74
a56ae6cf3a625fa76b549c433eaefdf9a68e313e
114074 F20101113_AACEJS schold_j_Page_106.jpg
283a30934e31417ec6e4a9511947b2a9
77b29122ec2d722dc2d10952644ee342218d58a9
114412 F20101113_AACEKH schold_j_Page_028.jp2
5ad42e260fddedd92ba2e6b600dbf348
18ac4f8bfe4ea0464e227eca5e26bc69f9999ae6
125297 F20101113_AACEJT schold_j_Page_107.jpg
d482a6ddfe46a4b4b49dbf1c3eeb389d
c0147ca45d94f76f184c94adaca5005ba2970344
116012 F20101113_AACEKI schold_j_Page_029.jp2
864745cdd1250c1cddbdd92fd44dc7ee
32c62c93e0aba1ba171de5101e2748f4d2165e15
124647 F20101113_AACEKJ schold_j_Page_030.jp2
8b8d6ed0c806b7552ad2928efcd81a6f
2678f06c3c1991b73eaca99bfa530c1e1be70e16
125773 F20101113_AACEJU schold_j_Page_108.jpg
4b29becb0510a6957750227ff77bcf2d
81ea283f866fe3350817f6719c6e406aed34393e
111622 F20101113_AACEKK schold_j_Page_031.jp2
ef5819b3f36ec3e23d1844eea8c20263
844488bd12e9ea669c34dd91293d60d2fde201e1
77955 F20101113_AACEJV schold_j_Page_110.jpg
3ddb77ac241f61edf1168ba0f6e173a7
9c9e5b5decbec8654b11372b1f12c918678fcd25
115163 F20101113_AACEKL schold_j_Page_032.jp2
a0ffab492bd0c86217d6685009cbe323
9e2676a4f9e57f6da5c252ea9e283182b1627c74
24087 F20101113_AACEJW schold_j_Page_001.jp2
e3c5a353426c3807bb9f5bb969dcfce1
5a05348cf12b8046b7b65fe737a8db4680e7f33c
144300 F20101113_AACELA schold_j_Page_049.jp2
ebc70fa352a4197a6e97cee72daf950e
d7554a067f1e69ce9f6f9448184a098272142627
106164 F20101113_AACEKM schold_j_Page_033.jp2
bec0817e37c30853250879da64404a4f
a9d6bf6184bac81f0c218f2cb3a90aecc747ad9e
118556 F20101113_AACEJX schold_j_Page_004.jp2
341a0a5ac366cffcf88fba3552c320b7
a578f49728aa3d8026d1fb8c1c51e55c5ef3d304
131615 F20101113_AACELB schold_j_Page_051.jp2
57ef5bb2e12af9724e67452499265a81
7f5283181373fed3ef30918b5feebb3f93afe840
123312 F20101113_AACEKN schold_j_Page_034.jp2
7bda0696ca3fba809c214e878e033280
6ade739e602ee038e5e19a1bcae83a62dfcec720
1001928 F20101113_AACEJY schold_j_Page_007.jp2
ccda396b086e2226a08a218ebf317dda
e9240b9cab7c89397c36b3f0d85b815daf9c4402
104177 F20101113_AACELC schold_j_Page_052.jp2
a90f9fa540465b9aecb4e812d537f4bd
a6b9fbe434b837079e3bf5b3ec11389bf118e3e6
1051971 F20101113_AACEJZ schold_j_Page_009.jp2
fd501f3e27b4bffb4357398915b40bd1
f14209d128610f63c4211646da7fb7ff36ab1aab
113936 F20101113_AACELD schold_j_Page_053.jp2
3112b2cf3477019d8e46c35babdc4bef
35d8a0843ae12865128602c55b18c749fa0ece8c
23814 F20101113_AACEKO schold_j_Page_035.jp2
d1db1ea44a885f218e0bcba3f1313b57
2223bd2e3857963c0ca94cd442c3f9d35d54974a
123804 F20101113_AACELE schold_j_Page_056.jp2
bafb79bf818bc0bafb28010c08c14b23
745fea69c73d346907b7c9907ccb2f80424ffec2
333933 F20101113_AACEKP schold_j_Page_036.jp2
ae6b2506299222c0695ce180bd07f9aa
e8218b3e1a1dd8888bf95437d494295b15817e49
121124 F20101113_AACELF schold_j_Page_057.jp2
11936c15ef8097e9dc89138e04c2f9e4
3ef0fe988043cd22f482af307c184fcf63639ae2
121514 F20101113_AACEKQ schold_j_Page_038.jp2
e3c15e212d67a76b50dd8d6c0f7eb09e
86a06b3bf5818359d585280ed0166f56492f62b8
114888 F20101113_AACELG schold_j_Page_058.jp2
b8611626e01c61449108fc8ea1aa39b3
80ec8ed089f250a03c0a261bdce9b3367faa4322
123404 F20101113_AACEKR schold_j_Page_040.jp2
eaad9baef497bf28526a2a8ca7427f4a
65b931131d16705a9b128e27cda771165f6255b8
113921 F20101113_AACELH schold_j_Page_059.jp2
8f79dcd27fd69568de5c43d6a15a6f31
cdb1eabf76a8ec2698b4bc34e4b0c89f2836c3c5
120130 F20101113_AACEKS schold_j_Page_041.jp2
1cc49f9b7c31916605d4ff2b7f18d00a
3d262b257404e325756be493df4895050e78f8df
124686 F20101113_AACELI schold_j_Page_061.jp2
8c9163a1f2bebae02059b8ad14fa390e
c1cc0a1d1f48ae87c33d9142d10cb43be20e997d
120808 F20101113_AACEKT schold_j_Page_042.jp2
c1e6f350a61ac8e81fe664090ff7eef6
d3b1352874ba109308a0c66792318118a0f6cbf5
120405 F20101113_AACELJ schold_j_Page_062.jp2
365a12d5acc8c23b5b4a8cd88ac6f13c
eb0107df74282c780ae1ad3782768ca4196bc386
113814 F20101113_AACEKU schold_j_Page_043.jp2
0354a4a3496ac9d7e1f27ad8bc9c5605
d80e0ca43be2019e086ac44bfdf009e0d6da17a0
49561 F20101113_AACELK schold_j_Page_063.jp2
cfdd041fd1490c51bcebdbe00252860e
25659b3069be9d29ec2987bc8eab88f1f55fe355
114576 F20101113_AACEKV schold_j_Page_044.jp2
d90a82e60f38a8400f9f4195b1395191
a77900419f81a35803bbd85c31680b54b29e7430
41119 F20101113_AACELL schold_j_Page_064.jp2
f0c68c6e4d924917a10f64cdab160980
fa59abe32419164d85a8a45e0ece6fb5f2780092
105297 F20101113_AACEKW schold_j_Page_045.jp2
7a1d8b5ad3b7958d63dc7892224daaad
647baa4d6b883f938cab281fe6d06d7561f3440f
107411 F20101113_AACELM schold_j_Page_066.jp2
a7a6a5b5cfad8dc74601d02c53ef1d72
98af913e7b70b350f69dc9ab1297183bacb14616
F20101113_AACEKX schold_j_Page_046.jp2
9887d98ddec6582c3a70919862659597
0dbb47883f702529715cd3603dc5052fc45a2be2
121461 F20101113_AACEMA schold_j_Page_090.jp2
08f4a5772034d71c99af63fcefefe26f
5550923f1f0532f83e4043ad1a10ed6f2b7a8efe
48051 F20101113_AACELN schold_j_Page_067.jp2
005e9fa673864da5db95e17727b7cc23
795cd3a5c4b18a4516a7f647eee26b32d2c8f31d
67481 F20101113_AACEKY schold_j_Page_047.jp2
0b9b35ba0f50f6b2493bf37ed934ebe5
5c4904d2880655038f90dd70afe373a8dda5412b
120298 F20101113_AACEMB schold_j_Page_091.jp2
a97806728b4fcbd6419c3027b815a4e3
1a0166ca1d58394483ed9fdd160f07d9b7b8f427
77683 F20101113_AACELO schold_j_Page_068.jp2
5e45ce017101b0438a6d39a33ce7963f
44ddc31e85100379a0a422d0d78b121ef979350f
158034 F20101113_AACEKZ schold_j_Page_048.jp2
0090955dd59c8eb11eff78f231ade721
1c91ff241ef5e7430329d55d9e71c3c942877eb2
120899 F20101113_AACEMC schold_j_Page_092.jp2
68d1cc5e34f0181326105e0d383d5ae1
525a5980466136c2c3d30f12015b44f69ff54907
119810 F20101113_AACEMD schold_j_Page_093.jp2
842520bafa8a30efaf6769ed7df1df9f
96222a193288adbd626667e3b6badc2be8fc4bdb
100115 F20101113_AACELP schold_j_Page_069.jp2
b4c9b908655fdc89ce1ee7ff8045e873
f20700795c5014e23d01e9049652a08e084cba50
119993 F20101113_AACEME schold_j_Page_094.jp2
d6dcd97c4e91fa9986bc6ab46fc33df5
ef0fbabbec15bf9d3e1c10312136672ed6c7a926
99150 F20101113_AACELQ schold_j_Page_070.jp2
c9fc1a73f9f533a7f82ad759731d61e5
878c33b419e00118fd8288f230fd18322cf106d4
118337 F20101113_AACEMF schold_j_Page_095.jp2
a4dba953cd8fb39b55d3286cf120aff5
7e6b8d415681df7c1a4c6b196ba649f930eeed96
94324 F20101113_AACELR schold_j_Page_071.jp2
8b81ab18b8729616e0a7872282a06539
99722345910e87abc36a09afde4986e191bcbb1b
34859 F20101113_AACEMG schold_j_Page_098.jp2
753bff6c475d34bc6ca9c639cab723ae
b5383e31518fb173c07eb5c1ff26682010d8ec35
219386 F20101113_AACELS schold_j_Page_075.jp2
e1aa2691dbc4ad47006dbebd65c16c26
1f60ce74dfa10d3a0c67d2b99f37f24235d5c9dc
1051964 F20101113_AACEMH schold_j_Page_102.jp2
d2c252c107f3e9b45a56c4a62cb85bba
c36f8a13a8e08775af995c7c444348f3ca596f6b
219895 F20101113_AACELT schold_j_Page_076.jp2
9f5073752d41167ef7d8746faa7ec750
5f9764607807485746ca8e85d0a92d5ff10a2729
F20101113_AACEMI schold_j_Page_106.jp2
d2208d20231da3e5f7051377a5280bd4
f71ada5addb17fea00a2d697525d605b74aba05c
187492 F20101113_AACELU schold_j_Page_077.jp2
5b5c211dc8778d27368b218ce259b3ed
656747947b4c7f8428666d9dc513021aa66ba0e6
131214 F20101113_AACEMJ schold_j_Page_107.jp2
882414ed062393fd05c4abf16e9b1318
4889eb39fe4eb526ceaf4c102b0c89c965717110
120769 F20101113_AACELV schold_j_Page_080.jp2
79ef78a45a6740ad10be28fb5bf4c5d9
c5d4c607170e08893ee44b7f43fd6deeebd07b89
84301 F20101113_AACEMK schold_j_Page_110.jp2
a0f75d788cc5399e7fec3118bb2f698e
6ca04cb5f35d31b098237acc6189047e7ffa423a
121990 F20101113_AACELW schold_j_Page_082.jp2
504b806928e8837a19c624f4bfafbe45
1ce0e96473a97d3f4f5929141fff8b9da48e2442
F20101113_AACEML schold_j_Page_001.tif
e85dcca3fdf8651eb61cb16b127f3ce0
a6c02a29acdb606b71927c119d0b46f8de7fe29d
116195 F20101113_AACELX schold_j_Page_083.jp2
07923488416d3fb95d5c8fb8f8b281fe
9fb891eb18030095af19eb0b6ddaca37f4ccc356
F20101113_AACENA schold_j_Page_030.tif
6274b9f9915827875afc9f97cb302321
74d243dc87f41c8bb9466731b6a70d8989a516d2
F20101113_AACEMM schold_j_Page_002.tif
8949696bd14bd0d9597548cc4a5b80b1
cfe0be5cf9fc6eea5d9647d23e86e2e2fe75a99e
119777 F20101113_AACELY schold_j_Page_086.jp2
3d0868377e35d68cb6ba1d9723603678
57d14f84f53975f86c1d05ae42e6ac51b62ab30d
F20101113_AACENB schold_j_Page_037.tif
71e604af9682f0fe85193789d0c99caa
c82c2192619f7544489efb9b088f7833d9e77110
F20101113_AACEMN schold_j_Page_004.tif
5484791f654e299ea5bb8683a4f8a9a0
a351ccbfe6996c9ba370285341db096026887a3a
122872 F20101113_AACELZ schold_j_Page_088.jp2
3c65ecfc77016a2b7fc19d8324f8d3c2
64ae219ab9723ef09aaf43fa9f33881e4e1d7ae5
F20101113_AACENC schold_j_Page_041.tif
a4ed1a290eae41d3d8d1530186873b86
2221821098dd397652229b1c911e1873f96bf61a
F20101113_AACEMO schold_j_Page_005.tif
98eab0e323e98732bc72670c34803ad1
33a4550fca2dc1b74ca5eecca835c5b5a934f227
F20101113_AACEND schold_j_Page_042.tif
4b377535ef73cb04512ba27ee95fcb8c
d68baa0d83f89d6196cf260015392259d5060282
F20101113_AACEMP schold_j_Page_007.tif
4c097ac4321c4c522cf68d0ff484ace0
7799a9339a9e5918e50ee7950c14085416a0e6b1
F20101113_AACENE schold_j_Page_043.tif
9f8cbbf6780afb5798501ede9406eee5
ae1ad744ce302efa8b4e4ccf9a4e7076842326f9
F20101113_AACENF schold_j_Page_044.tif
bc706f4d154414e5dc966ebf2d299221
4ca95a04cb190ac82ef4d3d8e35bd5a869b22142
F20101113_AACEMQ schold_j_Page_008.tif
dd899562e94c84159f50c8aad2fcd62a
cb79f7352e9dc9c996c85d53c59979cc1ae00bad
F20101113_AACENG schold_j_Page_047.tif
e99d984d14935be68deef278228cc17b
49d258a5496b0ce5c454c025ac3a529ba01b7507
F20101113_AACEMR schold_j_Page_009.tif
6b9ddc8798e08e14d73d487d09a32093
cf9105c5e1f803f9a01cc05b8cb2115a5a12d920
F20101113_AACENH schold_j_Page_048.tif
a9d1bf8b3608cc62cb9f71a667759695
0715b8f0946ad1376378e6ecba8a7c18f0ee0a0c
F20101113_AACEMS schold_j_Page_010.tif
82a6e0de14287e3922e0583bf6df29da
efed7ff199d76daa0f9bd5880dc5162242940d2a
F20101113_AACENI schold_j_Page_049.tif
545897fd56835054f65b92d24997cc63
6236ad58c7315abdc9f2c39465cb6e316178f957
F20101113_AACEMT schold_j_Page_014.tif
82ad6d65c4103049a3f79d4ca73129b9
d9b91ca0d214c2e3027418592671629b8bd2c829
F20101113_AACENJ schold_j_Page_051.tif
a0f514e60d3aaa92c9be91c04a511298
0b731530a1ffbb881a2a3cf17c7c841485248b8f
F20101113_AACEMU schold_j_Page_015.tif
e5826c236226b8f1e796cbbe30fe9814
b258990d1657c47589706b213c03f9ce762a60c4
F20101113_AACENK schold_j_Page_054.tif
c8dfcec7f849fa9f7680cd53a70cf0ca
977c926ed84b6e3899d2fbdd5ab380ffa522373d
F20101113_AACEMV schold_j_Page_018.tif
f33aabdb726416e2c8e0a2d99e2eddc0
15c8a658cae403665cabdd986b17b577d041d265
F20101113_AACENL schold_j_Page_056.tif
f47a622c6321b48ca2baddc921032c1f
e1ba4f32a58db4166ab0438d92b68e41596d5455
F20101113_AACEMW schold_j_Page_020.tif
6be79dc54a035a5869c43d73a9da4007
a58ce956d7a290c3b25350457a794786761f2e7e
F20101113_AACEOA schold_j_Page_085.tif
5a9c9bbf2e2694d6de7e06c1816c2a10
df3f16dc73a2b508c473f2d3aa3f07a3d71bfb09
F20101113_AACENM schold_j_Page_058.tif
08a41ad758ac77bde62b61e9a6b0c10b
3c603a57cc1c06e3b282aa85bf01602f0e0a3845
F20101113_AACEMX schold_j_Page_022.tif
17d02af481f90019cd41e2097215d373
464789c1aae3dd5c84435cf01cf6a73b72ff2acd
F20101113_AACEOB schold_j_Page_086.tif
aeaaf4e8f1493486367e92d88d313094
e9749f19287335a73ad0c0aae3ff14115a6e956d
F20101113_AACENN schold_j_Page_060.tif
f8b9328a8ff9cbdec3a8a15650c146d0
3e35df12f020e6bfd66881a3ccb148d3cbac581a
F20101113_AACEMY schold_j_Page_028.tif
b24187a81b6c9fc8e51c2b847585970b
ada4389782065b8675db7fd7712ce7106e79143b
F20101113_AACEOC schold_j_Page_087.tif
d6264fe53fba8352871eab9e757a5c8f
5e9c7f19b5b42453c371ab84aff8c29d7ba56669
F20101113_AACENO schold_j_Page_061.tif
0f60f76e877e1efbe3cdde65912d28a9
84fc6d8065cab4ada1da49c181c70d3ee99f7893
F20101113_AACEMZ schold_j_Page_029.tif
e9c4c2c9040d1156a751b1c675517fcd
c6d1be8d65359dd286ba2029cb6f3c3eace6e723
F20101113_AACEOD schold_j_Page_089.tif
bd0e07bd3e6f2639a21ba7edb15fee9c
181d39fd9b1c673c886274e241ed402cc1392ddf
F20101113_AACENP schold_j_Page_063.tif
be26c8c7e11bf843faedbb8ae076e6cf
82e8c9f1f5f4562465b47a3a4332be628bed113d
F20101113_AACEOE schold_j_Page_092.tif
563f1653295e0718a86e39273a6f39fc
d73d9437210720a01e33e50141af3c2847caf392
F20101113_AACENQ schold_j_Page_064.tif
88a8b49fc9ca90a93dd2865bf8c43035
72ef14129c719ca0b9774dae9aac3e0b8396a43a
F20101113_AACEOF schold_j_Page_094.tif
8ca19c36bd74600bf2ab3ec4bf1cf63d
19f24ac8ed020a1789b5d61ccb0db68054cb80b4
F20101113_AACEOG schold_j_Page_095.tif
049b2bd3e9a3243de1e5b86e9386330d
4a33b755814f5a44ed0eea9d287d846334f65ddc
F20101113_AACENR schold_j_Page_065.tif
82053c7a77108b9fa23d5d142b915e13
ba7ab3bdf5a61e4d96cea0758b95928759950759
F20101113_AACEOH schold_j_Page_099.tif
54fba85586336ba49a2841c4e9475569
d6f87d69145980664d268f0e5a9be569f8bf5ce9
F20101113_AACENS schold_j_Page_068.tif
6a96e10315e329bc8f7de6f124262660
49a3c1478b0e5c0c036e53d56f6a5409762d46d3
F20101113_AACEOI schold_j_Page_101.tif
ef91c9a1bf7ffb28d93315e418f08634
3b18dfdb2010aa5b7c5fe777c9df60eeac1c5837
F20101113_AACENT schold_j_Page_070.tif
5cefcfef8b1bc43b95d690988cfc27e6
1589f186eb68d774a41d75b6642b0078e69d8760
F20101113_AACEOJ schold_j_Page_103.tif
966fc16e9bd26420be90269da1dc9195
96c0ad19a2c1858f083073a0f1059c36be331118
F20101113_AACENU schold_j_Page_071.tif
8b6a9d015d70b4bba4c075b9a336a081
7e3efc2946b31952ecb892935e3bc7c6e5bd5b9e
F20101113_AACEOK schold_j_Page_105.tif
e12657b099608598b15aec19935d0487
0e8bf2ef107ff9964e50d782535133e83f817f7c
F20101113_AACENV schold_j_Page_072.tif
6325056fa6d72cfb5c50c6d2b0087eac
c6d96b028a873f0d63ddfe458061766c6509dd97
F20101113_AACEOL schold_j_Page_106.tif
8bed951a56730017b96ddad018db47e7
fb6a044d0c017c18a5c5877263e485c8d8bbb2e6
F20101113_AACENW schold_j_Page_073.tif
77c84dcf0a245b936b5e47b095cc89cb
6a22a09476fd9a9fe36eecafd3a3277f29805324
F20101113_AACEOM schold_j_Page_108.tif
09a361ea9607c048ce7625a11381f63c
087d77fb3dfb0a14fd57639ac5eb78580b52a764
F20101113_AACENX schold_j_Page_074.tif
1b3bf507f869a3a3afbc5b70a80ad819
78cd249dac30640d75065e0cb318062af8c30c79
57133 F20101113_AACEPA schold_j_Page_018.pro
d89829a60bae71894dda64c73407c133
aea704a44e5b16ab80e0e83604a7c1be43516f22
7488 F20101113_AACEON schold_j_Page_001.pro
059f1de2f4007f77922335b6a9c7ee07
6964b136a9c7540c553842147108cfb9dc78871d
F20101113_AACENY schold_j_Page_077.tif
fb38d652108d58f96f31259f2311da23
b6d6b8a5b4a85cd624194aa0bae1e836a5ee7334
58526 F20101113_AACEPB schold_j_Page_019.pro
9857babfbb7c9a60547ce38b6ec456e7
a2e283e5d26c8afe050ae6a9807e67df9a545429
700 F20101113_AACEOO schold_j_Page_003.pro
ce52e924193ada074a99eee15fe8300b
5b2368610b69cac57add16799c9a176bed78b7c8
F20101113_AACENZ schold_j_Page_078.tif
91994a2d4f29ccaf924841a83bdf5174
0e917ad111bdbfd6c5738fb90834c51cfd0e662a
56199 F20101113_AACEPC schold_j_Page_021.pro
a32056f70194184ed78ebf81a8b27f96
7e5ee92a93496650d9eb8288d952a7b5ccb602ff
78231 F20101113_AACEOP schold_j_Page_006.pro
73ba54c4ff789af7fb09eda1af219719
1f9dda66b80f991250220550f73559cd7b87ed80
56487 F20101113_AACEPD schold_j_Page_023.pro
c907a7bf74d43ccbee3d02d30cd0d529
5849171e0861d291d89944b9ea68ecadad1d0a9b
29453 F20101113_AACEOQ schold_j_Page_007.pro
49e2ea678d84978f9e564edaff8be885
202b3f97a8c7698dac5d49f95cb5591a3e38e817
57095 F20101113_AACEPE schold_j_Page_025.pro
ae7aac6847cd862d03947c62ce804650
a24031109bcdb580de6955e01e70743ed7cb0098
58722 F20101113_AACEOR schold_j_Page_008.pro
fac59a5d0eeaa4d221a9fad5dea9bcdd
8af8f776125a9a178db5082f1588035efe6e57af
54639 F20101113_AACEPF schold_j_Page_026.pro
81344593822f71f95872f78f8d755004
f5de1cd20ccc4ef5700f91faff0e6fce78c650e1
58495 F20101113_AACEPG schold_j_Page_030.pro
566d07f38909fe957d7fca31d8cdb900
05c18e47a8d937529ff1478a9f8489800bf2b1d7
49644 F20101113_AACEPH schold_j_Page_033.pro
e967a932f6874a57bbc1d57f22f81a9d
1385538b2feebb89449ea6a1a1e5b7e7a9f39e57
26317 F20101113_AACEOS schold_j_Page_009.pro
8489d9dd405f549f1c0dc9fb4ce5c0c6
efbe02e5abd4afd2d8e50513f3c0bddeb7c53dd9
58429 F20101113_AACEPI schold_j_Page_039.pro
8eb74affe85276a435ad2b16bb25e3f0
83926305005aa867f385cdb032473ca0fa2e7a7d
47073 F20101113_AACEOT schold_j_Page_010.pro
fc9f1378969a83463ff8284d2b03c5a9
532c358bf5294a66554885d4aac1b7d53306a2d3
58471 F20101113_AACEPJ schold_j_Page_040.pro
845d57d9ffa3a5c26c18a8f691b85850
2edf1234647b23690b27752827d984b08cc511ea
21708 F20101113_AACEOU schold_j_Page_011.pro
a8d1190f0c9f93f0a0cf76ec5c64617f
f67df0c78169f6aa73fbf3ba5eb2521e374da7fa
56082 F20101113_AACEPK schold_j_Page_041.pro
788fae7390e538a5522a875fb3962e20
7ebd99fe23d8faf10fd39e370a1f19f6316ec9d1
53737 F20101113_AACEOV schold_j_Page_012.pro
2e1b16647df48f25bcbb6ea18c37bc2d
18a9b19cc016b424c17afa998a7b7c0065140c47
56835 F20101113_AACEPL schold_j_Page_042.pro
70f0614dce17275df7c2c91e4c6fb83c
86428914ef23fa027d824a1417dd0a34a6940144
59558 F20101113_AACEOW schold_j_Page_013.pro
f58c3df623d500bb8b46f2dd53d5a961
9085de6a6a0e369c672b7acc94a9dff233602e20
F20101113_AACEQA schold_j_Page_067.pro
38822ed4f8d652347679eb5f0d617603
f174769c6b4fc919fa11ca15352aa0610aa3dbc2
53898 F20101113_AACEPM schold_j_Page_043.pro
1ddbcf8d5be42156f697a993180019d8
ae39ef76b5c70931dcc5fd4b50dc666a6f1c6512
51201 F20101113_AACEOX schold_j_Page_014.pro
8419b1145e36d82dd454371beee0b0e7
0ddbfff5ba32d62b62534ff020ec7706ae33604d
38689 F20101113_AACEQB schold_j_Page_068.pro
cf47d633fd760975ce5a1d75194eb561
25ca024d43afde06cc22648a93406f44de212792
54286 F20101113_AACEPN schold_j_Page_044.pro
ae470cea9cdb4583a4b6f7d4618f1714
48fc59fa347c242349ff6caafbd80b6c23190759
57242 F20101113_AACEOY schold_j_Page_016.pro
55e33e67a057389be7853f5c18218c3f
3cceaf765cfa60c13bf57bfb2fc2a2293e23bf2f
45614 F20101113_AACEQC schold_j_Page_070.pro
e5a940954ac6a15f0b2865060054c670
fc0f2958ed920dd90e689197b5b6d13a03aafa36
50558 F20101113_AACEPO schold_j_Page_045.pro
de6437abd6090b76ec970d801747f738
3267527ba91c4c5d0e6d32ee7225e5bdc0751997
57875 F20101113_AACEOZ schold_j_Page_017.pro
7482ce33bb6371aba65be9a5290c6f4d
5d421c2831d7589d97abbd97d1a65ce6d8e3e8f4
43468 F20101113_AACEQD schold_j_Page_073.pro
89429eb45d292b6739f85228988d6baf
0732bffde41f397e702cbfab07babc0942f43fad
5356 F20101113_AACEPP schold_j_Page_048.pro
bd7f6562878523de3d78018b62bcbd70
96c3f40268cff02e4983028fededa332303279f5
47494 F20101113_AACEQE schold_j_Page_074.pro
31723248684b216471c7ab8190911b6d
65a6b5848fd736b83aa6a7b7972b79dfd4d4293f
4676 F20101113_AACEPQ schold_j_Page_049.pro
1d4913260c458bf2228329fe5535a76c
7f4eef6301c1941d177943b8bfbc51b88ad0e41e
6847 F20101113_AACEQF schold_j_Page_075.pro
9811f75f1c02a37d1d397a4e5f3c83dc
9957be1dcfb22fb43326e2d526641bd6e0ed194e
5098 F20101113_AACEPR schold_j_Page_051.pro
f754490d9d3a172500ba75b99c6ca156
634819b007ca894bee2188bae235bcbd7e8ac451
6879 F20101113_AACEQG schold_j_Page_076.pro
b13f0652f0991c7b48fb114b98bb5731
5c67971a376b703ea25f0b340f8c4599309b7361
46001 F20101113_AACEPS schold_j_Page_052.pro
ebe5feca84608a0de1429d621f37ccdd
894c206e81b62fc62fb284048bc8965d8631bfaf
5825 F20101113_AACEQH schold_j_Page_077.pro
8ce8ed62ffe700cfb21be4c8bad15be1
ec24c6b753f9564768c477eabdd47c730e8a2766
6071 F20101113_AACEQI schold_j_Page_078.pro
d6da4feb695eed36fa60f91d445708c4
ca01ae290fd79520c1e4bdc3bfaa73f9f07a4379
50884 F20101113_AACEPT schold_j_Page_053.pro
1b98c40765eaf90afecd8dcd04b26d5d
2997c64c2921b50f74fa3829e29fb56654958871
57661 F20101113_AACEQJ schold_j_Page_080.pro
0fcd1166a44f938f164e0012cd02407f
67afe44b3fa9e97bbeaf1d6f9f15500d9024996d
57909 F20101113_AACEPU schold_j_Page_056.pro
70b0e2e67f4b0daf69c42d0df1ffb560
36e78ddefae190628051eae1aeb32e2e9d626e04
58492 F20101113_AACEQK schold_j_Page_081.pro
9f33c0ccd9cff4f49077f263677e4641
d1b031285414997c875305ee43aaef976438ec50
53962 F20101113_AACEPV schold_j_Page_058.pro
1f0c2a0b62ac03272ca1b810c79153ab
ffb7fc04c6e8b85673a0817bd403c8e584dc7c38
57973 F20101113_AACEQL schold_j_Page_082.pro
11942660fbbddeb9b8e7a68b1af0ad08
4472f89b9d1a08a8781321eb6cb4e87da8c834d8
52112 F20101113_AACEPW schold_j_Page_059.pro
e0c43879bb5b36abd3cb536ec961650b
8b7654e18e589cf65c3cc74acac2ace622d43ff7
54675 F20101113_AACEQM schold_j_Page_083.pro
da610c5f1d3dab329e32c34739d7d442
f5d16fe5a0bf8bb4eda2d554991da29a6066b78f
21224 F20101113_AACEPX schold_j_Page_063.pro
eb6edfbd6fb3107445b636d327f9d595
6c35860062e8fcb0e0e4e30449342e80fcb09162
54886 F20101113_AACERA schold_j_Page_103.pro
033c2b10888ffbd39aec928116ffa14f
43de07a894e011152895b7e7d061e1187079fdeb
58392 F20101113_AACEQN schold_j_Page_084.pro
ebcf3136a6d994c8ccc528701350e296
d6f45bce966156679d8ff25b06dc2be865ca462e
16364 F20101113_AACEPY schold_j_Page_064.pro
020a524650f1f2471496a4e9388fb848
ee28a65a93146dfbaa36dbdb10a02ca650954f0f
61910 F20101113_AACERB schold_j_Page_104.pro
5dbb55a345d21642b176f7ea55e66ca3
db39c5025b56b917ccb638a5e99a529179ef7ed4
56811 F20101113_AACEQO schold_j_Page_086.pro
5d8f98af399227d142c9799ece279dda
5c6e285ba4529fa322e22f9ce76f7790f0f1e787
36848 F20101113_AACEPZ schold_j_Page_065.pro
a3659b3f16446387c94aa99191927b40
a7ffc8c710e5cec6ffbe0fbc2839ece4d369b5c3
60940 F20101113_AACERC schold_j_Page_105.pro
b8a272ed87f1718caf3bc9d474792324
3b1f50ed8cc77a82ee38544f3b9c627469b86b79
58472 F20101113_AACEQP schold_j_Page_088.pro
c9c7de17b364ef9149de110f0209d59e
99242095955b927f9ec54d83c09a2f1476a6f949
53649 F20101113_AACERD schold_j_Page_106.pro
0410827cc5506aef7fc51631da63ef98
da3be1d0b5b3dcb2bc43b912033e27d1970cbfbb
57003 F20101113_AACEQQ schold_j_Page_089.pro
4e2c7e250d8344791dcb784f2d2a1c86
fe9ce2bf3f2779d71ea6e00b1aabd40cde9cda45
62375 F20101113_AACERE schold_j_Page_107.pro
07765adff71ea98d69c9f187a66710d5
68476fe76f56973750f3dccb3334077e06ea0593
56763 F20101113_AACEQR schold_j_Page_090.pro
29ab4758de9a349e6f288cc724a3eb9c
d990438cc39af62f3d1f3b309d9008484c4c4aaf
60321 F20101113_AACERF schold_j_Page_108.pro
e441a4745ef1b9e043dab137aaf85b48
c46e2050df95a133d813c900710a21742d0243df
56551 F20101113_AACEQS schold_j_Page_091.pro
7e84b721e7081242a99d82819823765c
87e55f3c13b822c3472e0aee5d33741e797334fe
54382 F20101113_AACERG schold_j_Page_109.pro
d79a886d13f4e3d3e9584196f4b5c02c
86b5051aae66bccf58d347f6942b74fd0f4450ee
57789 F20101113_AACEQT schold_j_Page_092.pro
a32dc9eb271b5ee6a8d80221c9f86012
355c9e849a84dc59db4be7e8c87f2f052453d4dd
36753 F20101113_AACERH schold_j_Page_110.pro
c6d51e846a01f4773de30965fe7f81b0
b69d70b6af7a9fc62104d7ecabf889480ed2d027
421 F20101113_AACERI schold_j_Page_001.txt
ddc8a993a677529edf21abbe3f9975dc
f841ff07edaca8558f8dde8db3d080bfc8c6b0e7
57182 F20101113_AACEQU schold_j_Page_093.pro
c1d2dad6431c37a91f40f8bbc2b4132c
9b6c13a6e679ac447e577c23389416f840ea55ce
2193 F20101113_AACERJ schold_j_Page_004.txt
8118923333d9e2a9a14d450b047c324c
483fadb82b9fe4a6eec37edfda88cffc5df6dc43
55495 F20101113_AACEQV schold_j_Page_095.pro
7cdefc5f53df34dbb9c5c175435ae51d
cf063ae05521ab4d6dd60ae428b73e33abf6b3d1
131 F20101113_AACERK schold_j_Page_005.txt
cbdf89af433c57b9784da974bf4aa39e
73f174ed6af9959df08224554305b94b92eeaba2
52761 F20101113_AACEQW schold_j_Page_097.pro
7bfb4eba8fafac8af66d496870a5c619
a284d743e52a2be86ba1a503c842112f0c9c1073
3216 F20101113_AACERL schold_j_Page_006.txt
b1ccf427859a4fa466c4bde1d95293a6
fa3e381ff1b12a0bef275ec5c3c15962a8fa14c9
55675 F20101113_AACEQX schold_j_Page_099.pro
afca292f3670d2c81562970d90415500
267e1ab972d0bbae3dc55815332bd8f1286c2ccb
2145 F20101113_AACESA schold_j_Page_026.txt
33e68fe1a171f7a9770f7b032f203809
da68796c992238ecc884a2068e7c7b0ea6565299
2032 F20101113_AACERM schold_j_Page_010.txt
02d9a6deeac2934e0fac4ee96489536b
eeeface5557922c400c5ad060bc9108ebc065438
56236 F20101113_AACEQY schold_j_Page_100.pro
c6dc27826495cd7895e058bfa74c71b0
aa81a8ef45cc495f8915da238c11943b013eabd5
2010 F20101113_AACESB schold_j_Page_027.txt
5168daee3553820d24e122468357f70f
19aa7a92a2e1f02b027b8991cb34219eb8582894
869 F20101113_AACERN schold_j_Page_011.txt
5340ebe3b2283ad45d516ba85c20d3b1
8f711de3d57ad59f6056bf7bf11539aa49fb19fd
60469 F20101113_AACEQZ schold_j_Page_102.pro
6801510e8412ac507ae88fa71cac348c
275a9243467bfc638c4e5d7126052a2de3677177
2152 F20101113_AACESC schold_j_Page_029.txt
0cf59587441b338f3b081f0a91452ec7
0b3009ee98edc88f76cb67e09877c3a0b425f7c7
2188 F20101113_AACERO schold_j_Page_012.txt
be12bd47137a48eb1c9a07a9ce999142
df2a32f9adefa8a51dd6926844c6272ca8896234
2290 F20101113_AACESD schold_j_Page_030.txt
a3442ea2bae8e154f424cab4a70a1082
374ab481b90f237d1ebac843f9ac4f8bdd489cc4
2346 F20101113_AACERP schold_j_Page_013.txt
6078fae4c33dd09d267608a758c871d2
79f5ffbda8064869467d80bc0f1035a3e6a2bf8a
F20101113_AACESE schold_j_Page_031.txt
419c06c95cd5c89b210315aefd1c3a72
e9112b1922a2fe8a041a13de569ad732e80d05f6
2180 F20101113_AACERQ schold_j_Page_015.txt
3d40983a37074ea9515977346044fa52
fdcc0aed38ab10f76031109ecb325a52a0b5f3d2
2187 F20101113_AACESF schold_j_Page_032.txt
de98ceaac0a15e4d4e25cb02c448125a
5d9ec40e871ddc40e54f9dbec816a524e1092220
2243 F20101113_AACERR schold_j_Page_016.txt
cbf5490143b8e19b56955d3c826df8b3
91dbd18e21a8a856e66e1531874b52ceb3b24341
F20101113_AACESG schold_j_Page_034.txt
2375fb5e233f29c45f11a3d799b254ac
c4985c34c71582657d0b58be95359cc71514ffee
2278 F20101113_AACERS schold_j_Page_017.txt
e073b4d7c9f7095160e4ca1d2e53d9b8
241084d2c5f0157dbdd2132658f0fe46247cae2c
709 F20101113_AACESH schold_j_Page_036.txt
3c83f71bbb2fe8a779d4739ea4c8c0eb
a8392926c275c8e3de8d3df94e78678514354bef
2286 F20101113_AACERT schold_j_Page_019.txt
6f3d60d8990f1f84eca75b7d9b3f9e3d
2afe5e8e3984983f0268ad86d9c232c768762bc8
F20101113_AACESI schold_j_Page_037.txt
37aa65f1683f56edd64e5a037caed3aa
3d82ebc90752ffd9fa955709c840553b9a25884a
2265 F20101113_AACERU schold_j_Page_020.txt
e067748444f1faf74a381827e0a6788e
77ed599ad2f17fdb7b858f17cabb3f13a8fe5c87
2329 F20101113_AACESJ schold_j_Page_038.txt
797eaa36ea25d2efd4835207b525fa20
d93ac37b18aeac1030b0d10aef3ef0e1d2f7f411
2285 F20101113_AACESK schold_j_Page_039.txt
b378a6531741ac1dc08f70d9b60d9836
00d7775a439a02221881190d97d773622eba000e
2200 F20101113_AACERV schold_j_Page_021.txt
a0e6a9c94b89534762a339e3059f2eba
48bb2e57219b8052035a163feb3d651f5562566d
2316 F20101113_AACESL schold_j_Page_040.txt
a14327ce9840b28a9f309222760977b0
1ac01de9ae53ee4643f3cd25830f52895199276f
2142 F20101113_AACERW schold_j_Page_022.txt
54a4cdbeee43c946022f1fa8f292adba
375cb1a1c8f6c516a8813b270435619a3dfed92c
2804 F20101113_AACETA schold_j_Page_069.txt
837f03ef98ab9f5da1b7088ee8f8a701
705ac3669473a32c36297be18586712fc8b8019d
2205 F20101113_AACESM schold_j_Page_041.txt
fa74bfbebd07d0de647e991e73bc31af
f612f9780b7d8a6c26d5c6337c2457af11fc5db0
2212 F20101113_AACERX schold_j_Page_023.txt
324e4087fc31ee7e4b76c8442b163b7f
dc74acd61a517f006ec1ede069f3a0f290b766db
3007 F20101113_AACETB schold_j_Page_070.txt
849da06aa36fa03dcde240084818e133
6c90d2731aa35fac9179ba172ae235f57fb97956
2160 F20101113_AACESN schold_j_Page_043.txt
75e10dea14e0bda72e4def0d3b33ee9b
e3d4ec5c9bd1561983e4cb3dff83c350ec0f808b
2221 F20101113_AACERY schold_j_Page_024.txt
9407df5acd840e5ff8e267242f32554f
cebe79455a66e65071e6b3e2a3e67e9c983ae4ab
2879 F20101113_AACETC schold_j_Page_071.txt
86f31f7e398734a13f0e018a864b126f
57f605cd270dfe3c69671b8b82b527c733705fdb
2118 F20101113_AACESO schold_j_Page_045.txt
ef440d9396c3594ec84889c33b266742
719cafdb94b7de335be06daa11e341790bffc111
F20101113_AACERZ schold_j_Page_025.txt
559464846b1d97bbe3b73a667a782d9a
630a405afca90663a2af48861ab72d71718c6fc0
1767 F20101113_AACETD schold_j_Page_073.txt
96bc89f37df77f25e961590569536b3d
a5c001927ead861c118eda85d0d1b6dec664d41d
1234 F20101113_AACESP schold_j_Page_047.txt
5b75322ad6da79b51c9d05bfe3ccc3d0
f2a99d529827101233946706fac07e64ea0db5e2
1849 F20101113_AACETE schold_j_Page_074.txt
a68ebaa082fc81de5e43e444b043d314
8cd0c835f822a1e27e6d4563291ad7ac6b8011b5
301 F20101113_AACESQ schold_j_Page_049.txt
87f87a347658e685dd1f4621638751bb
055938376415bf387fcb80989a244267ce4c47c8
361 F20101113_AACETF schold_j_Page_076.txt
b0f2e60fc7615abad0a6e9a8e26c11c3
e2508196bc143bffec7271fb745705d2b42f5307
1960 F20101113_AACESR schold_j_Page_052.txt
c465b1b99cb62333d947396e7581fe51
9bd8c206f5dd261b62b275fe05ed226b6a5d26d0
294 F20101113_AACETG schold_j_Page_077.txt
11484497007d80f3dcf2948dace0f2d8
e76f9235b0eb2e0f4807037beb2ea34d9af22637
2155 F20101113_AACESS schold_j_Page_054.txt
341115401cab1fef72b21125baf0b178
2ae04072d8154a9ae487ac99349d20da33d39222
325 F20101113_AACETH schold_j_Page_078.txt
82eac30396c36e996b846f4821433477
91524da6cd3a0efd2d86e0094d219cf35320c1f7
2280 F20101113_AACEST schold_j_Page_055.txt
71efde154abdad53a7d1ce879a7c103a
7ad93fdd22128a7dd8f1c9c9e3421c56101b3ac8
2246 F20101113_AACETI schold_j_Page_079.txt
a6d16aeee8a4e58318fa6b0d16142aae
0a749a22eefc31c8743945bde1f3c1869bce3ad3
2281 F20101113_AACESU schold_j_Page_056.txt
b2c2727cd1adddddf6754da584bbd735
bab38b076670988cb6337a9b2ba7cb799ab1191f
F20101113_AACETJ schold_j_Page_081.txt
d20fd94fdf8d61cce3a662e8c74a28ad
6025c65377eb8c43a83ec26588b99d6bd03e8e07
2119 F20101113_AACESV schold_j_Page_058.txt
37781118ac45370716f10041ea05e4b6
2476d5896cea23b45f32615cdf2c87f4478d0335
2268 F20101113_AACETK schold_j_Page_082.txt
935e3668375150a89c0cbe2b77ea3840
ecb8067b5995391ff650eaf3ec63627608d656a6
F20101113_AACETL schold_j_Page_083.txt
0425bcc3051a8be0452bd8cef03f50f9
4388d7e150de8ea69d77004d29e0e6aac570265a
2087 F20101113_AACESW schold_j_Page_059.txt
81b4a1f1f6028c8f27a09d611d4e42ff
65cb048614732a2ab5677014f0e0c2632715a99c
F20101113_AACETM schold_j_Page_084.txt
2be03a427711290f7fd712f4d48aa324
b5474b53132b0b2e1be8bff6ece334a0e9f6ac7b
862 F20101113_AACESX schold_j_Page_064.txt
7be69b4f93aff9d6ad1ccbb6373caf2d
ac3956040471304fa115d7e3b06e7f00f9a40114
2416 F20101113_AACEUA schold_j_Page_108.txt
ef9371c885aa82ee54bfeaa220cd9ed8
2be4c12ff638d9bf54021673f602cb9ea9bcd6a8
F20101113_AACETN schold_j_Page_085.txt
da30f837703ed7ba75ac50d1dedf5a05
943f0314ac5d548c9b62ec49287a71f8234e2174
1890 F20101113_AACESY schold_j_Page_066.txt
7a591a21f0b8c0bf4c911579c8215819
0638921f9066a9df2cf3576a64311b3d0972e600
2220 F20101113_AACEUB schold_j_Page_109.txt
ae832cab0400c8296b58ecc0ac6c324f
ebddde3c9fe2c8bf6cad4da3c5a2dc314bfca3b9
2224 F20101113_AACETO schold_j_Page_086.txt
5ac19c0b5ac1be90f426269e88cbeaca
bbeeda0b3b4c2100373f1be0b6c9a7921736c1c3
F20101113_AACESZ schold_j_Page_068.txt
5eaab6cf25d4e490dd82906afefb318a
9ecc36d88481424c4ec16fbcec40c0694e92459d
2068 F20101113_AACEUC schold_j_Page_001thm.jpg
09b49d3f1e5f39a8199a30de622e36e5
075925773a572e7487430d486b2a244d3b645b2a
7905 F20101113_AACEUD schold_j_Page_001.QC.jpg
d134b510c3ac4608aeb0676d83f09d67
4f3ad9ef6cc581e28eeec0cc891ee3b94ffbcd33
2228 F20101113_AACETP schold_j_Page_090.txt
f6d98c62724fa41b0d28aae74632722b
d52d52779bbf541942ce68b5fab937f210397841
530 F20101113_AACEUE schold_j_Page_002thm.jpg
119b445bbcd5bff996f6f0d82e2fc585
d145cf6fa5cfca2b95411a34064c4e97894d0920
2213 F20101113_AACETQ schold_j_Page_091.txt
bf1ab7a819c96b0fb8d9ffc911dbad60
ec06cf6f5f372d3d7845670b6a13a05d0b6db755
1298 F20101113_AACEUF schold_j_Page_002.QC.jpg
55e3c2aa2c56790fadba0f637a26a0f6
e06f6717636c86a19934b98bed2d17c8fc51e242
2247 F20101113_AACETR schold_j_Page_093.txt
387c780f2ca320b2d231833c6a6925c1
10322cbdf020d994ca69507266d5b5a78fa54862
465 F20101113_AACEUG schold_j_Page_003thm.jpg
4265e51c4ef28ac0e2b95b0e3d71d985
9b9a35ad2e25763ecc93a7309fbb22734a1e6778
F20101113_AACETS schold_j_Page_094.txt
571c841779afa3d2b789eb51929ebbf7
fd56c51dbb629abebdd5b31ec23d79461beb97dc
F20101113_AACEUH schold_j_Page_003.QC.jpg
b575eca51881e9f1c762f9d6da4117c0
eadc18ab8de8616a518bfce9b4a4c2a374e344c7
2172 F20101113_AACETT schold_j_Page_095.txt
2eee545873fe3c5449977293bbb2cafc
b970b6193a707a34e54f3d3d697a72156db7793e
973 F20101113_AACEUI schold_j_Page_005thm.jpg
fd4de94c45e25ac423e151ff351c9edd
1ee4fce4ee37d8b5ea1861d5396c286b0866b0e7
2080 F20101113_AACETU schold_j_Page_097.txt
81930ac16051ec2a2b01587ebe019c31
05c3f81adc1bae4044bc0f3180b1303b24228938
12478 F20101113_AACEUJ schold_j_Page_007.QC.jpg
2295f2a0c4ce444b08a27f8255b3c26b
3fbded5025acbb4c01f4df6d2092d9699c91d267
608 F20101113_AACETV schold_j_Page_098.txt
3a3d8b971c987317e3f7a33d6403dc6b
136e1e681c05dfd551597ca31ab38a3f8092b1fa
8494 F20101113_AACEUK schold_j_Page_008thm.jpg
116c109e849ce6852de335864e59cef8
584c792ab81b03303b16744048cd3fbdb55f9905
2206 F20101113_AACETW schold_j_Page_100.txt
5cd9b57e0b34955c293b003cc9355482
9adbf85b37d808d96acb2bcbb8bb4b9f92544c36
4517 F20101113_AACEUL schold_j_Page_009thm.jpg
26c6b0e63f8bd7f62110a76de613d817
79fb0e7d526bc0e80dbd7dd9d98fa31b936b8f55
35417 F20101113_AACEVA schold_j_Page_022.QC.jpg
84b1e1cbfb174a3fa98fa5dcaab21c78
af3df63bd9e8d3afc872e0200cc3f8c5810fe030
7685 F20101113_AACEUM schold_j_Page_010thm.jpg
4c58a1bb73079c805bb97218026368ed
07bded19ac082fb8d09cc9d7376d4f5676a7617d
2428 F20101113_AACETX schold_j_Page_102.txt
d469e078ebcb0c5ccc75efaf6a6e196e
f2974f4609cdcee91100eda1aadbc61133e39c76
8911 F20101113_AACEVB schold_j_Page_023thm.jpg
779f048260c4e788507726823c968819
ead61dfc414019472d16704c60ded64306f25ac5
29757 F20101113_AACEUN schold_j_Page_010.QC.jpg
7961b7694d064c6ff4730331ab207164
e11c6d7a0649cbe3f22fbada83127f09361ef72c
2435 F20101113_AACETY schold_j_Page_105.txt
319abc298488fa7da860e5b7053f89d6
adc5fa29f65dae4b11d86e517ff368b38759ceb1
8864 F20101113_AACEVC schold_j_Page_024thm.jpg
63bee8c8ae6cc8a757e53e5aaa44c89f
3887946b969d74f5d4848fcf446350d263261823
8685 F20101113_AACEUO schold_j_Page_012thm.jpg
c686e4c69c6127f45160e5b3528d2122
9d858b101f7403b2a4cb5966479d76342329a7b9
2162 F20101113_AACETZ schold_j_Page_106.txt
8c8126a771c570ec725b8e12f9fad152
5527828aa1405bd9e3adb4949be3dae32514d7d5
38350 F20101113_AACEVD schold_j_Page_025.QC.jpg
8eccfad5d39e443004d990e04d8ddb05
1ff6767e832ca1e1ebd83e3876251f20e6de26bf
9037 F20101113_AACEUP schold_j_Page_013thm.jpg
ca32ab35ab7d1d535a996e461b5dea39
75b597aea7fd259e5c77610231ac456e19a235bf
36582 F20101113_AACEVE schold_j_Page_026.QC.jpg
41728b62fcbe8681e80196132b634315
c81c190287a99f6333e0f8a6cb31cb9bdba39819
37438 F20101113_AACEUQ schold_j_Page_013.QC.jpg
c5d707ce651a2f0af7e7aaaa95d200fd
13263cb68d64f605196abb13ed0d0b82b6b36ad2
8473 F20101113_AACEVF schold_j_Page_028thm.jpg
45fe17a682b7504cdf88b5a94672e31e
d93a723bbdb37a5276402e39a0353639b54a11e3
7560 F20101113_AACEUR schold_j_Page_014thm.jpg
858698850b4a7689d205edb8173f52c2
016e2ddacf5fafae1ed214762d93f1136d5787d6
33958 F20101113_AACEVG schold_j_Page_028.QC.jpg
2f415ab19e747e275dc9afc28d42241c
70e3bfa8c21b81b086921bdb34ff20eeccdbe99f
32168 F20101113_AACEUS schold_j_Page_014.QC.jpg
b921a6b946e580c404b3c8369a171048
5d5af29e210c2fdab989b926e7bdb8298cfa7775
35133 F20101113_AACEVH schold_j_Page_029.QC.jpg
0992e79736e4125b80cea7d797c306ba
2aad29291a24cb16c8a74096d0f3e9213c3bd9e2
8467 F20101113_AACEUT schold_j_Page_015thm.jpg
711fbedccc1a289bc5c6a12ce928d25f
92c46df78da2c19fc9820c0818391d5bc8945d72
9201 F20101113_AACEVI schold_j_Page_030thm.jpg
7439aaaea58f717dd34f5a079e8bba66
28c4d65d15c42bc64fc3b5478716722fc4aa68f6
9021 F20101113_AACEUU schold_j_Page_017thm.jpg
eed5e2c8970fc11792999d74b28ab2e2
61292080ac5eb27988d068ba49464fb1e3fd1a03
37852 F20101113_AACEVJ schold_j_Page_030.QC.jpg
408f7c00105ad6d4dc24024d287d2018
716db1691629d7e7e2b7d2f8b26e063b07464466
F20101113_AACEUV schold_j_Page_019thm.jpg
ddb1bcf1cf43c33c501eaa33a874ac6a
60d0b5f134e4a4ee6c24dedf081ecb99fd86e835
8623 F20101113_AACEVK schold_j_Page_031thm.jpg
5fb2ccdd1317c7edb891085ffc87ae0b
4e2c7962e6cb151dd88653683a939723d0006d10
39101 F20101113_AACEUW schold_j_Page_019.QC.jpg
960bc389c0866ca2f4820aeb5fe79ac8
c27eaa4666196bc2a5904515cca41f6d0012cd41
8944 F20101113_AACEVL schold_j_Page_032thm.jpg
bb59e97cd9ba76e161fb5d4f2d6f0512
789c9adf1413a884f9eec9f34b339cabef5a86f9
38302 F20101113_AACEUX schold_j_Page_020.QC.jpg
6160e46378d3b104dc80336875a8461e
6b32aa572521125266505d6c1856a42d7dca5919
35840 F20101113_AACEVM schold_j_Page_032.QC.jpg
f6c4593569d7dfd52d7aea54b34c7b7b
be53650cf3068522ce6d3595925331c7c3d7649e
37729 F20101113_AACEWA schold_j_Page_042.QC.jpg
40ee69d88c359de20bfc8f26a8f9c636
3f5ea8ec7f4dcae2e99346bd2b5e22a55020a291
8100 F20101113_AACEVN schold_j_Page_033thm.jpg
37f596c5ec7fba2e5ee46d2ffc058cda
3ffc9e177875a5b4860f6987c6f58845d88ae1d2
9061 F20101113_AACEUY schold_j_Page_021thm.jpg
2d1e95a6ec1bf74be5a4640b757fe859
eb7e894c5b98f04b8ffe66f7c4767693ed0b86be
8779 F20101113_AACEWB schold_j_Page_043thm.jpg
fd6a0f558b586bd819f547da924096ef
63b09e60fdc2c91b25fb42fb92497ded6f957b98
31983 F20101113_AACEVO schold_j_Page_033.QC.jpg
2e2e96bb28d8b80a810dc2876fc1ed77
94f3e34a19138e2e672097acb69a3efa16732cff
35601 F20101113_AACEUZ schold_j_Page_021.QC.jpg
db9f63cb43b77e845bea04e315a084d7
9ee9ae3cbc2cca0a867e426473372c68b3f6981f
34588 F20101113_AACEWC schold_j_Page_043.QC.jpg
4fa0238d538033249c726d2268f4875a
065aa2c9cb47006532bfe81e1fbf9c373281c5f7







CENTER CHARACTERISTICS AND KIDNEY TRANSPLANT CANDIDATE OUTCOMES


By

JESSE D. SCHOLD













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

2007

































O 2007 by Jesse D. Schold

































To Amy and Caila









ACKNOWLEDGMENTS

I have many individuals to thank for their guidance and support throughout my doctoral

program and professional development. I am most indebted to my doctoral committee,

colleagues, and family for their encouragement and selfless assistance to facilitate my progress

through the program and placing me in a position to fulfill my personal and professional goals. I

would like to particularly thank Dr. Herwig-Ulf Meier-Kriesche, who has served as a supervisor

and friend throughout my experience at the University of Florida. Dr. Meier-Kriesche's most

selfless and devoted mentorship and dedication have been a primary factor in my professional

growth and his instruction has provided me with invaluable knowledge and skills that I will

utilize throughout my career. I am also most grateful to Dr. Jeffrey Harman for his outstanding

leadership within both my academic program and throughout the dissertation process. I am

grateful to Dr. R. Paul Duncan for his supervision in my academic program during which his

personal approach and professional wisdom have provided an optimal format for progress and

growth. I would also like to thank Dr. Neale Chumbler for his encouragement and great

enthusiasm for research.

I have many other professional colleagues whose support, knowledge, and friendship have

had an immeasurable contribution to my personal development and passion to research. I thank

all of these individuals who have enriched my professional growth and experience at the

university. I would also like to thank my academic advisors and instructors for their dedication

and diverse training that have provided me with a great basis for future development.

My deepest gratitude is to my family for their incredible sacrifices and support in allowing

me to pursue my academic and professional goals. I will always be indebted to my wife, Amy,

for her love and support; my daughter, Caila, for teaching me the most important lessons in life;

my father and mother, for always encouraging me to not set limits for achievement; and to many










other family members and friends, whom I have always learned from and who have given me

valuable perspective.












TABLE OF CONTENTS


page

ACKNOWLEDGMENTS .............. ...............4.....


LIST OF TABLES .........__.. ..... .__. ...............8....


LIST OF FIGURES .............. ...............9.....


AB S TRAC T ............._. .......... ..............._ 10...


CHAPTER


1 INTRODUCTION ................. ...............12.......... ......


2 BACKGROUND ................. ...............15.......... .....


Kidney Transplantation for End-Stage Renal Disease Patients ................. ......................15
Variation in Patient Outcomes by Providers of Healthcare ................. .........................15
Factors Associated with Kidney Transplant Center Performance ................. ............... ....17
Kidney Transplant Center Volume and Patient Outcomes................. ......... ..........1
The Impact of Waiting Time on Outcomes for Kidney Transplant Candidates. .............20
Donor Quality and Transplant Recipient Outcomes .............. ...............23....
High Risk Kidney Transplant Patients .............. ...............25....
African-American Transplant Recipients ................. ...............26........... ....
Obese Transplant Recipients .............. ...............28....
Elderly Transplant Recipients .............. ...............29....
Conceptual Framework............... ...............3

3 MATERIALS AND METHODS .............. ...............37....


Overview. ............ ........... ...............37....
Data............... ..... ...............3

Dependent Variables............... .. ..............3
Explanatory Variables of Intere st ................. ...............40....... .....
Additional Explanatory Variables .............. ...............42....
Statistical Analysis............... ...............43
Study Aim I............... ...............45...
Study Aim II .............. ...............45....
Study Aim III ............ ............ ...............46...
Potential Selection Bias .............. ...............46....


4 RE SULT S .............. ...............52....


Study Population............... ... .. ............5
Rate of Transplantation by Center ............ ............ ...............52..
Transplant Center Volume ............ ............ ...............53...












Center Donor Quality .............. ...............53....
Center Performance Ratings ................. ..... ........ .. ...............54.....
Association between Transplant Center Characteristics............... ............5
Reliability of Historical Center Characteristics .................. ...............55........... ...
Kaplan-Meier Candidate Survival by Center Characteristics..........._....._ ........_. ...........55
Multivariate Cox Model for Primary Outcome of Candidate Mortality............... ...............5
Outcomes among High-Ri sk Candi date Group s ................ ...............59..............
Expected Survival by Center Characteristics .............. ...............61....

5 DI SCUS SSION ................. ...............79................

6 CONCLUSIONS AND FUTURE WORK ................. ...............99........... ...


LIST OF REFRENCES .............. ...............102....


BIOGRAPHICAL SKETCH ................. ...............110......... ......










LIST OF TABLES


Table page

4-1 Transplant Candidate Characteristics .............. ...............64....

4-2 Candidate Characteristics by Center Proportion of Transplants within Three Years........65

4-3 Candidate Characteristics by Center Volume Category .............. ....................6

4-4 Candidate Characteristics by Center ECD Proportion Category .............. ....................66

4-5 Candidate Characteristics by Center Performance Category.........._._.... ......_._........66

4-6 Median Levels of Center Characteristics over Time .............. ...............66....

4-7 Correlation Coefficients between Center Characteristics ........._. ...... .___ ..............67

4-8 Center Characteristics at the Time of Transplantation .............. ...............67....

4-9 Adjusted Hazard Ratios for Patient Mortality after Listing for Transplantation ...............68

4-10 Adjusted Hazard Ratios for Receipt of Transplant following Listing .............. ..... ..........69

4-11 Adjusted Hazard Ratios for Post-Transplant Mortality .............. ...............69....

4-12 Adjusted Hazard Ratios for Post-Transplant Overall Graft Loss .............. ...................70

4-13 Adjusted Hazard Ratios for Pre-Transplant Mortality ......___ ........___ ................70

4-14 Adjusted Hazard Ratios for Mortality for African-American Candidates.........................71

4-15 Adjusted Hazard Ratios for Mortality for Elderly Candidates .............. ....................7

4-16 Adjusted Hazard Ratios for Mortality for Obese Candidates .............. ....................7

4-17 Candidate Life Expectancy (in years) after Listing by Levels of Center
Characteri sti cs ........... __..... ._ ...............72....

4-18 Life Expectancy after Listing at Hypothetical Center Characteristic Levels ....................73

4-19 Life Expectancy after Listing for African-American Candidates by Center
Characteristic Levels................ ...............73

4-20 Life Expectancy after Listing for Elderly Candidates by Center Characteristic Levels....74

4-21 Life Expectancy after Listing for Obese Candidates by Center Characteristic Levels......74










LIST OF FIGURES


Figure page

2-1 Conceptual Framework based on Grossman's Health Production Function .....................36

3-1 Distribution of the Annual Number of Deceased Donor Transplants by Center ........._.....48

3-2 Distribution of the Proportion of ECD Transplants by Center .........__.. ... ......._.._......49

3-3 Distribution of the Proportion of Candidates receiving a Deceased Donor Transplant
within Three Years by Center ...........__......___ ...............50...

3-4 Distribution of Standardized Mortality Ratios by Center ....._____ .... ... .. ..............51

4-1 Kaplan-Meier Plot of Candidate Survival by Center Rate of Transplant............._.._.. .......75

4-2 Kaplan-Meier Plot of Candidate Survival by Center Proportion of ECD Transplants......76

4-3 Kaplan-Meier Plot of Candidate Survival by Center Volume ................ .............. .....77

4-4 Kaplan-Meier Plot of Candidate Survival by Center Performance Ratio..........................78









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

CENTER CHARACTERISTICS AND KIDNEY TRANSPLANT CANDIDATE OUTCOMES


By

Jesse D. Schold

August 2007

Chair: Jeffrey Harman
Major: Health Services Research

There are approximately half of a million patients diagnosed with End-Stage Renal Disease

(ESRD) currently in the United States. There are two general classes of treatment for ESRD,

maintenance dialysis and kidney transplantation. Kidney transplantation is considered the

preferred treatment for ESRD patients who are medically cleared for the surgical procedure as it

is associated with an improved quality of life and longer life expectancy. A kidney transplant

may derive from a living or a deceased donor. For patients to receive a deceased donation, they

must be placed on a waiting list at a transplant center. There are currently approximately 240

kidney transplant centers in the United States. There are numerous characteristics that vary

among transplant centers.

An important question is whether or not these center factors are associated with

prospective transplant candidate survival. Moreover, whether these center factors are applicable

to all transplant candidates or have differential effects on certain patients is also unclear. This

study examined these questions using a national database containing patient-level information on

outcomes for all adult solitary kidney transplant candidates from 1995-2000. Results of this

study indicate that transplant center characteristics are significantly associated with patient

survival after listing. Of the center characteristics, expected waiting time had the greatest impact









on candidate survival. Candidates listed at centers with the longest expected waiting time had a

32% increased hazard for death, translating to approximately two and a half years of reduced life

expectancy, as compared to candidates listed at centers with a relatively short expected waiting

time. The main findings were applicable to the general candidate population as well as high-risk

candidate subgroups.

Results of this study may be utilized to inform patients and caregivers about the important

impact of characteristics of transplant centers on prospective survival for candidates of kidney

transplantation. Investigation into applications of these findings in healthcare policy, for

different transplant populations, and in other medical contexts will be examined in follow-up

studies deriving from this study.









CHAPTER 1
INTTRODUCTION

Kidney transplantation is regarded as the most effective treatment modality for End-Stage

Renal Disease (ESRD) patients who are medically cleared for the surgical procedure. As of

October, 2006, there were over 65,000 patients listed to receive a solitary kidney transplant in the

United States. The number of candidate listings has increased by 60% over the past decade and

the number of candidates has increased across all age, ethnic, and gender groups. Over the

corresponding era, the number of available kidneys deriving from deceased donors increased, but

at a much more modest rate. As a result, the difference between the availability of and demand

for kidney transplantation has markedly increased. One of the most important implications of

this temporal shift has been a significant increase in expected waiting times and increased death

rates for kidney transplant candidates awaiting the procedure. This trend is particularly troubling

for transplant candidates and the healthcare sector for two primary reasons. First, kidney

transplantation has been shown to double a patient' s life expectancy as compared to the

alternative treatment modality of maintenance dialysis, and, as such, delayed access to

transplantation places patients at an increased risk of mortality. Second, increased duration of

dialysis prior to transplantation has deleterious effects on recipients following transplantation,

resulting in a cohort of patients with increased risk for morbidity and mortality after the

procedure.

There are approximately 240 transplant centers currently operating in the United States.

Among transplant centers, there is significant variability in expected waiting times for candidates

to receive a transplant. In addition, performance evaluations suggest that outcomes for transplant

recipients vary significantly related to the quality of care at individual transplant centers.

Research also indicates that centers with the highest transplant volume are associated with










improved patient and graft survival following the transplant procedure. Moreover, both short-

term complication rates and long-term survival for transplant recipients are significantly variable

based on the quality of the donor organ. The proportion of high- and low-quality donor organs

also is variable throughout regions of the country and at individual centers. Cumulatively, there

are multiple factors that may potentially influence a patient' s prognosis relative to the selection

of a particular center. The current evidence suggests that patients have incentives to receive a

transplant as early as possible, at the best performing center, and receive the highest quality

donation. However, despite these known effects, there has not been a comprehensive study to

examine the j oint impact of these factors on transplant candidate outcomes. This information is

particularly salient given the expanded candidate waiting list and associated time to

transplantation. In addition, the reliability of historical transplant center factors on prospective

patient outcomes has not been thoroughly evaluated. Transplant candidate characteristics and

prognoses are also widely variable based on age, ethnicity, primary cause of ESRD, and clinical

presentation. Whether certain transplant center profies are more applicable to patients with

different prognoses and care needs has not been previously investigated. In particular, certain

patient groups (including individuals over the age of 65, African-Americans, and the obese) have

differential survival rates on dialysis and following transplantation relative to the general cohort

of patients. In this regard, center characteristics may have a unique impact on these subgroups.

The purpose of this dissertation is to analyze the association of transplant center factors with

patient outcomes, which may ultimately be used to guide candidate and caregiver decision

making to select specific centers for a kidney transplant. The specific aims of this study are to

I. Determine whether characteristics of transplant centers (volume, performance ratings,
waiting time, and donor quality) are significantly associated with candidate mortality
after listing for a solitary kidney transplant.









II. Determine the relative impact of center characteristics and estimate transplant
candidate survival rates at incremental levels that may be utilized as a tool for
selection of a center.

III. Determine the significance and relative impact of transplant center characteristics on
candidate mortality after listing for a solitary kidney transplant in three high-risk
patient groups (elderly, obese, and African-Americans patients).

This study will be carried out utilizing a national database containing the population of

renal transplant candidates listed for transplantation in the United States. The database contains

patient-level information and de-identified indicators of transplant centers. The results of this

study will be applicable for assessing the importance of transplant center factors and potentially

used to help guide patients and caregivers in their selection of a center. Although there are other

factors which may influence the selection of a center, an obj ective measure of the impact of these

transplant center factors may be crucial for informing potentially life-altering decisions.

Secondary outcomes of this study will include descriptive analyses of transplant center

characteristics across the United States and to assess the reliability of historical center factors on

prospective levels.

Chapter 2 provides a brief historical background of kidney transplantation and treatment

of ESRD patients, a summary of the relevant literature in this research area, and a conceptual

framework for this study. Chapter 3 provides descriptive statistics of the candidate population

and transplant centers as well as statistical methodology for conducting this study. Chapter 4

includes the results of the analyses and tables and figures depicting the outcomes of the models.

Chapter 5 includes discussion of the Eindings, and Chapter 6 includes conclusions and potential

follow-up studies deriving from this research.









CHAPTER 2
BACKGROUND

Kidney Transplantation for End-Stage Renal Disease Patients

In 2004 there were nearly half a million ESRD patients in the United States (1). This

count has increased more than three-fold since 1988, partially explained by increased rates of

hypertension, diabetes, improved clinical detection, and an aging population. The most common

treatment modality for ESRD patients is maintenance dialysis. Dialysis is classified as either

hemodialysis or peritoneal, and both modalities require the use of a machine for renal function

replacement either at a designated facility or home-based. The alternative treatment for ESRD

patients is kidney transplantation. The first kidney transplant was performed in 1954 by Dr.

Joseph Murray between identical twins. Significant advances in surgery, immunology,

pharmacy, and clinical care have contributed to widespread application of kidney transplantation

and over 16,000 transplants were performed in the United States fifty years later in 2004. The

procedure is generally classified in two broad forms: deceased donor and living donor

transplantation. Deceased donors are the source of the majority of transplants, although living

transplant rates have increased significantly over the past fifteen years.

Variation in Patient Outcomes by Providers of Healthcare

The notion that patient outcomes vary based on characteristics of the provider of care has

been investigated in a variety of medical contexts. There is extensive evidence in the United

States to suggest that there are differences in patient outcomes as a function of the individual

caregiver, hospital, and region of the country. Schrag et al. investigated the association of

hospital procedure volume and mortality following surgery for colon cancer (2). The study

found significantly improved outcomes associated with hospitals with increased volume,

although it was also noted that the variance in outcomes based on other factors was significantly










higher. Utilizing a national Medicare claims database, a 2005 study concluded that providers

with higher volume were associated with reduced morbidity rates after radical prostatectomy (3).

Examining the treatment of patients with acute myocardial infarction, Krumwell et al. found

significant geographic and physician specialty variation in the use of efficacious drug therapy in

the United States (4). Diminished outcomes have also been demonstrated for individual

surgeons with fewer procedure experiences relative to high and intermediate volume surgeons

(5). A national study also found significantly reduced short- and long-term outcomes for cardiac

transplant recipients at low-volume centers, which comprised more than half of all centers in the

United States (6). Treatment modality and outcomes for patients with myocardial infarction

have also been demonstrated to vary by region of the country (7). Additional research supports

the concept that geography variations may be important based on particular patient groups (8).

There are numerous additional studies that support the notion that variation in treatment

modality and quality of care exists within the United States. These variations suggest that patient

selection of particular providers is important and supports the efforts of agencies charged with

identifying quality of care between healthcare providers. The evidence also implies that there

are potential mechanisms to overcome quality disparities, which include information concerning

processes of care at higher performing providers. As the quality of care and rate of medical

errors has been questioned and highlighted in recent years, the demand for transparency of

information related to outcomes associated with individual institutions and caregivers has

proliferated (9). However, there is also mixed evidence as to whether more transparent

information regarding provider performance is associated with patient selection of providers or

subsequent outcomes. In the context of cardiac bypass surgery, research suggested that report

cards (i.e., performance evaluation of providers) had a positive impact on outcomes and










processes of care (10,11). Furthermore, there was some evidence that positive report grades

influence future utilization for hospitals (12). However, additional reports indicate a lack of

awareness among patients concerning performance ratings and a general failure to incorporate

ratings in provider selection (13-15). In general, there is a perception in the research and

medical communities that variations exist between healthcare providers, but techniques for

measuring performance in an equitable manner, interpreting provider evaluations properly, and

disseminating information regarding provider quality of care in a manner that is useful to both

providers and consumers are both challenging and inexact and will require significant

improvement in the years to come.

Factors Associated with Kidney Transplant Center Performance

In the field of kidney transplantation there has also been significant research dedicated

towards determining "center effects" associated with outcomes for transplant recipients. In the

early days of transplantation these effects were suspected and causes for differences in outcomes

for patients attributed to effects other than known medical risk factors were initially considered.

In 1986, Burdick and Williams found differences in outcomes associated with patient residence

in metropolitan versus those patients that were classified as "out-of-town" (16). This early report

suggested that there may be specific care mechanisms for transplant patients that were associated

with patient outcomes. An additional report determined that significant differences in patient

outcomes were associated with transplant centers that were independent of utilization of specific

immunosuppressive regimens and human leukocyte antigen (HLA)-matched recipients (17).

This report also suggested that center differences were not confined to the immediate post-

transplant period, as graft survival differences related to centers accrued for patients with

function after three months post-transplant. This conclusion inferred that processes of care may

also be implicated as a causal mechanism associated with center variations. Additional research










suggested a "learning curve" associated with immunosuppressive regimens at transplant centers

(18). In this case, centers that were able to incorporate the newest, most efficacious regimens

were associated with superior outcomes. The report also concluded that variation between

centers were minimal for low risk transplants, but derived mostly from treatment of higher risk

patients. In this sense, higher risk patients were more sensitive to care practices, and care for

these patients may be more complex; differences in centers are more observable for these

subgroups. Ogura et al. described characteristics of centers related to different outcomes kidney

transplant recipients (19). One important conclusion of this study was that while complication

rates did not significantly vary between high- and low-performing centers, graft survival

following complications was significantly different. The report found that low-performing

centers did have a significantly higher proportion of at risk patients, but adjusted survival rates

did not obviate differences. Moreover, the maj ority of differences between centers existed in the

initial six months post-transplant rather than long-term divergences. This report found no

significant differences based on the volume of transplants at centers; however, there is extensive

literature suggesting improved outcomes at high volume centers in subsequent reports.

Performance evaluations have been produced and published for kidney transplant centers

since 1991. The Scientific Registry of Transplant Recipients (SRTR) publishes these reports

online and includes classification of centers as either significantly above, below or not

significantly different than expected based on the national experience. The 2005 report indicated

that among the 191 transplant centers with at least 50 kidney transplants between July 2002 and

December 2004, approximately 15% were identified as having statistically higher or lower

survival rates compared to the national experience (20). Along with this information being

freely available to the patients and referring physicians, insurance carriers may utilize this









information for negotiation of contracts with particular centers. A 2006 study, however,

suggested that in aggregate these report cards had no significant influence on selection of centers

for transplant recipients with the possible exception in younger and more highly educated

patients (21). The manuscript found a positive correlation with center performance on a year-to-

year basis but that this association reduced over time. Other questions remain whether

performance ratings are true reflections of center practice and whether adjustment techniques are

sufficient to control for patient variability and associated outcomes (22). As compared to other

contexts in which performance evaluations have been conducted, kidney transplant patients have

less acuity and longer follow-up periods, and the models have significantly less predictive

ability. These factors suggest that performance evaluations in the Hield of kidney transplantation

may be subj ect to influential factors that may not represent center quality care.

Kidney Transplant Center Volume and Patient Outcomes

There is significant variability in the size and scope of kidney transplant centers. Center

volume may be a function of geographic location, presence of competitive centers, academic

affiliation and service to underprivileged patients, inclusion of pediatric programs, use of multi-

organ transplants, and the proportion of living versus deceased donor kidney transplantation.

Early reports indicated that centers with high volume of procedure were associated with

improved outcomes. A comparison of high-, mid-, and low-volume centers utilizing national

data from 1987 to 1991 demonstrated a 19% increase in one-year graft survival in high-volume

centers as compared to low-volume centers (19). This report indicated that there were lower

rates of immediate graft function independent of demographic characteristics of recipients at

lower-volume centers, but no significant differences in acute rejection episodes. A 1997 United

Network of Organ Sharing (UNOS) report concluded that there was significant variation in

outcomes at kidney transplant centers, these effects were most apparent in the first-year post-









transplantation, and, furthermore, that inferior outcomes were most common among small

transplant centers (23). An examination of ten-year survival rates in the United States examined

the association of patient outcomes transplanted at centers with large volume (>1000 transplants

over ten years) versus lower-volume centers (24). Conclusions of this study included that there

was a mildly beneficial effect of large centers, but that these effects were reduced in lower-risk

transplants. In addition, this research suggested that large centers were associated with improved

outcomes, more notably with elderly recipients and spousal living donations, and, in contrast to

previous reports, that differences were most apparent only after two to three years post-

transplantation. In an analysis among living transplant recipients, Gjertson detected improved

outcomes associated with large centers and that among small centers there were significantly

more variable outcomes (25). The most recent evidence of the center volume effect derived from

a 2004 manuscript examining outcomes from 1996 to 2000 for both liver and kidney transplant

recipients. The primary conclusion of the study was that the highest quartile transplant centers

(which transplanted a median number of 167 patients annually) were associated with improved

patient outcomes, but there was no detectable difference among the remaining volume level

centers in terms of the main outcome of overall graft loss (26). The research also suggested that

center effects were most evident within 90 days post-transplantation, but without substantial

clinically significant differences in recipient or donor factors associated with centers at different

volume strata.

The Impact of Waiting Time on Outcomes for Kidney Transplant Candidates

The survival advantage of kidney transplantation as compared to the alternative treatment

of maintenance dialysis was suspected for many years and formalized by Wolfe et al. (27). This

analysis utilized a time-dependent model assessing the impact of a solitary kidney transplant in

reference to a comparable cohort of medically cleared and wait-listed candidates. The results









indicated that transplantation approximately doubled life expectancies for recipients. Wolfe and

colleagues (27) further reported that the relative advantage was observable across age, gender,

and ethnic strata. The analysis also indicated that there was a significant immediate mortality

risk associated with the transplant procedure, but that this risk was obviated within a year of

transplantation and the benefits observable thereafter. The survival advantage of kidney

transplantation has subsequently been reported in other countries with variations in healthcare

structure and patient socio-demographic characteristics (e.g., Australia, New Zealand, and

Canada) (28-30). In addition, this advantage extends to higher risk recipients including

transplantation of lower quality donor kidneys, candidates over seventy years of age, and obese

patients (3 1-33).

The maj ority of patients that are listed as candidates for transplantation have already

initiated dialysis therapy. As such, waiting time associated with the duration of dialysis is a

significant mortality risk for renal transplant candidates prior to the procedure. In addition, there

are a number of retrospective analyses that indicate that the duration of dialysis exposure is

associated with deleterious outcomes following transplantation. In a single center study, Cosio

et al. demonstrated that pre-transplant dialysis was a significant risk factor for post-transplant

patient death and overall graft loss (34). Another single center study utilizing transplant data

from 1980 to 1995 indicated that these effects were particularly noteworthy among living donor

transplant recipients (35). A retrospective analysis deriving from a national database extended

this concept by demonstrating that the time on dialysis had a dose-dependent risk on subsequent

graft loss and patient death after both living and deceased donor transplantation (36). This

analysis reported that even six to twelve months of pre-transplant dialyses was associated with a

37% increase in long-term graft loss as compared to patients without exposure to dialysis. In









addition, this study suggested that there was a relatively similar effect in patients with ESRD

secondary to diabetes, glomerulonephritis, and hypertension. A separate study confirmed this

direct association between the duration of pre-transplant dialysis and the rate of allograft loss in

living transplant recipients utilizing registry data in the United States (37). Still yet another

study investigated the association between duration of pre-transplant dialysis and rate of allograft

loss utilizing a paired-kidney design. This design, though, allowed only outcomes for kidneys

deriving from the same donor transplanted to one individual with short dialysis exposure and

another recipient with extended dialysis exposure. This design allowed for control of donor

characteristics and the potential confounder that patients with extended dialysis had a greater

proclivity to accept lower-quality donations. Results indicated that recipients with less than six

months of pre-transplant dialysis had over two-fold improved ten-year graft survival as

compared with patients with greater than twenty-four months pre-transplant dialysis.

The accumulated evidence highlights the significant benefit of transplantation relative to

dialysis as a treatment modality and additionally emphasizes the importance of rapid acquisition

of a transplant. The growing wait list for transplantation and the extended waiting times to

transplantation from a deceased donor source continue to subj ect patients to dialysis for longer

durations. A transplantation committee report indicated that at the given traj ectories, waiting

times may average ten years in the next decade (3 8). Moreover, the differences in waiting times

have become increasingly variable across regions of the country. As such, a patient' s active

selection of transplant centers may have increased importance in future years. For candidates that

plan on being placed on waiting lists for a deceased donor transplant, the prospective waiting

time may be one of the primary criteria for this decision.









Donor Quality and Transplant Recipient Outcomes

The association of characteristics of donor kidneys with outcomes after transplantation

have been investigated from the early years of transplantation (3 9). Studies have identified the

history of disease among donors, anatomical characteristics of kidneys, biopsy results of the

donor kidney, donor and recipient cytomegalovirus status and HLA matching, and donor

demographic characteristics as significant risk factors for transplant outcomes (40-48). In

aggregate, the combination of risk factors of a donor kidney, most notably the age of the donor,

is acknowledged to convey highly variable life expectancies and graft survival for the

prospective transplant recipient. There have been several attempts to quantify this aggregated

risk of donor characteristics and estimate survival rates based on these levels. The Organ

Procurement and Transplantation Network (OPTN) instituted a formalized definition of marginal

kidneys in 2002 with the advent of the Expanded Criteria Donor (ECD) (44,49). These deceased

donor kidneys were demonstrated to convey a 70% or greater risk for graft loss for transplant

recipients relative to an ideal class of donations and were characterized by a donor age over 60 or

over 50 accompanied with two additional risk factors including a history of hypertension,

elevated terminal donor creatinine, or cerebrovascular cause of death. Nyberg et al. developed a

more granulated scoring system based on four transplant characteristics association with six-

month renal function level and demonstrated significant variability in outcomes by donor risk

groups utilizing a national cohort (43). An alternative risk index has been utilized to create

donor risk groups and demonstrated the association with complication rates as well as long-term

patient and graft survival (50). This alternative risk index derived from a retrospective analysis

indicated that relative to an ideal class of deceased donations (constituting approximately 11% of

donations), the highest risk class of donations had a three-fold risk for graft loss. In addition, the









analysis demonstrated that the risk for graft failure was particularly notable in the first year post-

transplantation, but also persisted for patients with graft retention after one year.

Research indicates that there is significant geographic variability in donor quality with

certain transplant network regions having an over 50% likelihood of low-quality donations (51).

A portion of this effect may be related to the issue that transplant centers in areas with extended

waiting lists may also be more apt to utilize lower-quality kidneys in order to avoid patient

deaths on the waiting list. However, there is clear relationship that transplant candidates who list

for a kidney in certain regions are more likely to receive higher-risk donor kidneys. Beyond

regional effects, there is also likely a transplant center relationship, as centers have variable

acceptance practices of donor kidneys. One of the primary purposes of the ECD policy was to

identify high-risk kidneys and allow patients to consent to receive these kidneys with the

potential benefit of receiving a donation more rapidly. The decision to list for an ECD kidney is

often strongly influenced by transplant physicians and other patient advocates. A 2006 study

demonstrated that patients who listed for an ECD at centers with a high proportion of ECD

candidates had no relative benefit in the amount of waiting time for transplantation (52). In

contrast, candidates who listed for ECD kidneys at centers with a low proportion of ECD

candidates in fact do receive their transplants much more rapidly. In addition, those patients that

listed for a lower-quality organ at centers with a low proportion of ECD candidates were much

more likely to actually receive a lower-quality donation.

There are not any explicit policies directing which renal transplant candidates should

receive lower-quality kidneys. A national transplant committee report suggested that the

candidates most likely to benefit from transplantation are those with decreased life expectancy

on dialysis (38). A retrospective analysis indicated that elderly and diabetic patients often









receive equivalent benefit from an ECD if it can be received with significantly reduced exposure

time to dialysis as compared to a standard criteria donation (SCD) with longer dialysis exposure

time (53). In contrast, younger, and presumably healthier, patients have reduced life

expectancies when accepting lower-quality organs early after transplantation as opposed to

waiting longer for a higher-quality donation. Research also suggests that quality-adjusted life

years (QALYs) may be suboptimal for some patients who accept an ECD kidney early after

ESRD onset (54). This study also reported that these results varied based on transplant candidate

characteristics, inferring that this decision should be patient dependent.

In summary, there is significant evidence indicating that the quality of a donor organ has a

marked impact on patient outcomes. Secondly, there is wide variability in the quality of donor

organs at a regional level. Additionally, candidate listing practices for lower quality organs at a

particular transplant center affect progression along waiting lists and the quality of transplanted

organs. Moreover, the decision to accept lower-quality donations appears to be variable based

on candidate prognosis and health status. In aggregate, the significant influence of donor quality

on recipient outcomes is clear; however, the association of the proportion of high-risk transplants

on a prospective candidates' mortality is unknown.

High-Risk Kidney Transplant Patients

Even though kidney transplantation is a lifesaving procedure across all candidate groups,

outcomes for recipients significantly vary as a function of demographic and clinical

characteristics. Individual differences in outcomes between patient groups have been described

based on known medical prognoses analogous to the general population, but there are also

unique characteristics particular to the transplant population which distinguish patient risk

groups. For this study, three distinct classes of high-risk transplants which have variable

representation at transplant centers in the United States and have known association with inferior









transplant outcomes will be examined: African-Americans, obese, and elderly transplant

candidates.

African-American Transplant Recipients

African-Americans comprised approximately 32% prevalent cases of the ESRD

population in the United States in 2004 (55). Survival rates for African-Americans are improved

on dialysis relative to Caucasians (56). However, among transplant recipients, African-

Americans have significantly higher rates of acute rej section and graft loss (57). There have been

multiple explanations purported regarding these relatively inferior outcomes among African-

Americans including immunological differences, socioeconomic factors, and delayed access to

transplantation (58-60). Alexander et al. distinguished unique steps in the transplant process

including medical suitability, interest in transplant, completion of a transplant workup, and

acceleration on the transplant waiting list. He found racial disparities at several steps in the

process (61). Additional research indicated significant differences in attitudes toward seeking

(or obtaining) a transplant between African-Americans and Caucasians, but that these differences

alone did not explain disparities in access to transplantation (62).

Differences in rates of transplantation to African-Americans are also partially explained

by lower donation rates among African-Americans. Between 1995 and 2004, African-American

donations comprised between 13-15% of all living donations and 10-13% of all deceased

donations (20). In addition, progression along the deceased donor candidate waiting list is

partially predicated on HLA matching (63). HLA-antigens generally are more similar among

race groups, and the reduced donation rates among African-Americans affects longer waiting

times on dialysis and increased deaths on the waiting list for this group. In addition, African-

Americans are more likely than Caucasians to have presensitization or histocompatability

reactivity to particular donors which eliminates potential donor sources (64). A single center










report in 2002 suggested that disparities in outcomes by race could be reduced by encouraging

hepatitis B vaccination and utilizing hepatitis C donor kidneys, subsequently reducing waiting

times for African-American candidates (65).

An alternative explanation for the documented reduced outcomes following

transplantation in African-Americans is more poorly controlled hypertension (66). In addition,

lower economic status may have a strong role in rates of return to dialysis following

transplantation (67). As immunosuppressive medication is a necessary and expensive

maintenance component to graft survival, noncompliance and lack of access to medication

regimens is a significant risk for graft loss. Butkus et al. found higher rates of noncompliance by

African-American recipients at their center, but concluded that racial differences were not fully

attributable to immunosuppressive regimens and more likely due to poor HLA-matching and

socioeconomic factors (68). African-Americans also experience higher rates of delayed graft

function and higher rates of acute rej section (69). Elevated risk of complications often

necessitates more potent levels of immunosuppressive medications and increases the risk of

infection and other medication side effects.

Cumulatively, there appear to be a combination of medical, sociological, and economic

etiologies for reduced outcomes in African-Americans. African-Americans represent a

significant portion of the transplant candidate population and have unique care needs. The

observation that African-American patients have improved survival prior to transplantation on

dialysis yet significantly reduced outcomes following transplantation suggest that there may be

special consideration for selection of a transplant center in this cohort. In this sense, estimated

benefits of listing at centers with the presence of certain factors may be particularly important for









these candidates and specific information regarding estimated survival in this stratum of patients

important.

Obese Transplant Recipients

Rates of obesity in the United States have grown substantially, and obesity has been

implicated as a significant cause of death in the general population (70). Obesity has also been

implicated as an independent cause of ESRD and cardiovascular morbidity (70-74).

Epidemiological evidence suggests that the effect of obesity for ESRD patients is a paradoxical

area of medicine. Among patients that reach ESRD, high body mass index (BMI) is reported to

be protective for dialysis patients (75-79). While the causes) for this association are not clearly

known, there have been multiple hypotheses yielding multiple contributing factors (80). In

contrast, among transplant recipients, studies indicate that recipients with both low and high BMI

have diminished survival following transplantation (81,82). Moreover, obesity has also been

associated with increased complication rates, costs, length of stay, and delayed graft function

following the procedure (81,83-85).

Despite the findings that obese patients have relatively superior survival as compared to

non-obese patients and inferior survival following transplantation, they still receive a significant

survival benefit associated with transplantation (31,86). However, due to the relatively poor

short- and long-term outcomes of obese patients following transplantation, certain transplant

centers exclude patients based on BMI and others advocate aggressive weight loss prior to

transplantation (87,88). Other centers report acceptable results for transplantation in obese and

morbidly obese recipients (89,90). In addition, among ESRD patients, obese patients have lower

rates of listing for transplant then overweight patients despite equivalent survival rates (75).

Collectively, there is an established risk for deleterious outcomes among obese patients









following transplantation and variable acceptance of this population as viable transplant

candidates.

The differential survival rates of obese transplant candidates suggest that the impact of

dialysis time may be different in this cohort. This may be reflected in the incentives to list at

particular transplant centers among this group based on expected waiting times. In addition, the

impact of other center factors including performance ratings, donor quality, and high volume

centers has not been specifically investigated in this population. A limiting factor for these

patients may be the willingness of centers to accept them as viable transplant candidates;

however, among those centers that do list obese patients, listing decisions may be particularly

critical in this population.

Elderly Transplant Recipients

Fifty percent of the incident ESRD cases in the United States in 2004 were patients 65

years of age or older (elderly) (91). Even though the rate of elderly transplant candidates have

doubled in the past decade, only 14% of transplant candidates are over the age of 65 (20).

Therefore, it appears that transplantation is a viable treatment modality in only a small subset of

elderly ESRD patients. However, the entire explanation for the lack of access for older

individuals is unclear. Many of elderly ESRD patients have contraindications to transplant or are

generally not medically viable for the surgical procedure. The greatest proj ected increase in life

span associated with transplantation is among younger recipients; however, even among patients

70 to 74 years old, transplantation is associated with a significant reduction in mortality (27).

Elderly transplant candidates are more likely to receive kidneys from older donors and have

significantly higher death rates on the transplant waiting list (51,53). Elderly recipients are at

higher risk for infectious death following transplantation, rates that are accelerated with

increased exposure to pre-transplant dialysis and varying levels of immunosuppressive regimens










(92). Rates of acute rej section are relatively lower in elderly recipients, but this may also be

indicative of tailored immunosuppressive regimens and immune incompetence (93,94).

Elderly recipients have still been shown to benefit from transplantation; however, the

absolute benefit, in terms of years of life gained, is substantially less than younger recipients. In

fact, evidence suggests a significant loss of scarce resources in terms of donor kidney life-years

and economic loss to payers of ESRD patients due to transplantation of younger donors to older

recipients (95). Even though increased donor age is a risk factor for graft loss among the elderly,

the evidence suggests that only younger transplant recipients survive long enough to receive the

full benefits of a younger donation on average. The Eurotransplant program has been developed

specifically to address this concern by directly allocating older donations to older recipients in

combination with reduced cold ischemia and waiting time on dialysis (96). The distinction in

this program is the older participants are mandated to receive organs from older donors but the

tradeoff of receiving an organ more rapidly and with less ischemic effect may offset the risks

associated with a lower quality donation. Although age is not a factor in OPTN allocation policy

in the US, caregivers have substantial influence in directing the types of donations to particular

transplant candidates. These decision-making processes may vary at the transplant center level

based on experience, the proportion of older donations available in a particular service area, and

length of candidate waiting lists. Also, research has demonstrated that due to the elevated

mortality risk of elderly patients on dialysis, acceptance of lower-quality organs in exchange of

receiving a transplant more rapidly may be a viable option for this cohort (53). In other words,

the relationship of the competing risks of extended waiting time on dialysis and lower quality

donations appears to be unique in the elderly subset of transplant candidates. However, these

decisions depend largely on the amount of waiting time reduced and the availability of quality









donations. These factors are widely variable at transplant centers and the estimated impact of

center characteristics is likely to be unique for this high-risk subset of patients.

Conceptual Framework

The conceptual model for this study is based upon the work of Grossman and the model

of the health production function (97). This model examines the role of individual decision

making in the production of health and the investment of human capital to improve individual

outcomes. Fundamental to this theory is that health is predicated on multiple factors of which

health care represents one of many explanatory factors. Other determining factors include

individual characteristics, environmental conditions, and behavioral components. In this

framework, individuals demonstrate a demand for health care. This demand is not for health

care per se, but for health and the subsequent utility which derives from health. In this sense,

individuals are not solely subj ect to external constraints but are active participants in their own

health status and future prognosis. Additionally, while individuals demand health, they do not

value it over all other goods reflected by choices that have known deleterious impact. In

addition to the utility associated with increased health, there is also an investment component of

health. Health represents an investment as it allows for increased individual production as well

as future utility. Implicit in this model is that individuals value health and always seek to

maximize their level of health and associated utility.

The health production function is represented by a declining slope with potentially

nonlinear deceleration over time. Individuals are endowed with a particular stock of health that

depreciates over time subj ect to effects of aging as well as other potential health shocks such as

disease onset or accidents. Grossman illustrated this depreciation based on individual investment

in health utilizing Equation 2-1.

Hi+i Hi = li GiHi (2-1)









This equation characterizes the loss in health with 6i representing the rate of depreciation in the

ith period, Hi representing the stock of health, and li representing the investment of the individual

in this period. From an economic perspective, the investment component of health is determined

by the marginal benefit of receiving the desired level of health. In this framework, individuals'

health value is also a function of their wage rate and utility as well as the cost of producing the

level of health. In this formulation, the investment component can be more specifically

characterized as a function of the amount of medical care sought in the period as well as the

stock of human capital and time as illustrated in Equation 2-2.

li =li Mi, ~iEi)(2-2)

In this case, Ei represents the stock of human capital, THi represents the time component of

investment in health, and Mi represents medical care for the individual. Utilizing this

framework, this study will examine the effect of center characteristics on outcomes for renal

transplant candidates. In this context, the medical care component of the Grossman model is not

homogenous. The type of medical care, represented by characteristics of the listing center may

have differential impact on patient survival. In this sense this study will investigate the level of

future health that is independently determined by selection of center. Furthermore, under the

Grossman framework in which the investment and utility deriving from investment may vary

based on an individual's prior health status and individual characteristics, this study will

investigate the impact of these decisions in certain high-risk patient groups. Utilizing this

framework, the research question focuses on the effect of the selection of center on outcomes

controlling for a patient' s latent healthcare stock and timing of candidate listing. These

characteristics include the age, race, gender, length of dialysis, insurance, body mass index,

education, and clinical presentation of the individual. Individuals can be viewed as producers of










their health and this production as a function of the type of medical care sought. Therefore,

rather than a single indicator for medical care utilization, the selection of a center can be further

discriminated as a function of center characteristics, in this case the proportion of high-risk

kidneys, center volume, center performance and expected waiting time as depicted in Equation

2-3.

Mi = Mi(Pj, Vj, Qj, Wj) (2-3)

In this extension of the Grossman framework, medical care for individual i is a function of the

proportion of high-risk kidneys (P), transplant volume (V), quality of center (Q), and expected

waiting time(W) at a given center j. Selection of center j will in turn have an impact on future

health status and the associated utility derived from this level of health. Figure 2-1 depicts this

framework for ESRD patients that select a medical care through a specific center. More broadly,

health status in future periods are a function of past health status, depreciation rates, and

investment in health, including the type of medical care attained.

The evidence to date suggests that there is significant association of transplant center

characteristics with outcomes for kidney transplant candidates. However, the joint estimation of

these factors is unclear, but is crucial towards informing potential candidates and caregivers

regarding choices for centers of care. While the individual effects of center performance and

volume, waiting time, and organ quality have been evaluated for transplant recipients, the

interaction of these factors and relative benefit has not been investigated and has not been

examined for prospective transplant candidates. This study will examine the role of these center

characteristics in order to estimate circumstances that offer the best prognoses for transplant

candidates and maximize patient long-term outcomes.









Results of this study will be important at several levels. This study will examine whether

there is an important association of transplant center factors with outcomes for candidates of

transplantation. Failure to Eind any significant associations will suggest that candidates have an

incentive to select centers based on convenience, logistics, or other determinates of their quality

of life. However, significant factors that are found may be important to guide patients and other

caregivers of where to list for a transplant that can affect their mortality risk. In particular,

candidates that have a choice in their selection of centers or have the resources to travel outside a

local area may have an incentive to discriminate between centers based on these characteristics.

Furthermore, this study will highlight whether these factors are uniformly applicable to the

candidate population or have differential impact in high-risk patient groups. A finding of a

significant interaction in this regard would suggest that certain patient groups would be more

advised to list at particular centers based on their mortality risk. Moreover, a significant

interaction would imply that there may be cases in which the entire transplant process could

become more efficient and provide benefit to the full complement of transplant candidates. If

certain center factors are more important to certain patient groups and less important to others,

then it follows that allocating patients to the respective locations may improve outcomes for the

entire population. Identification of the importance of transplant factors that are associated with

candidate mortality may also provide insights for future augmentation of the transplant and

allocation processes. For those factors that are amenable, alteration of these characteristics and

development of infrastructure to support a process that fosters these characteristics would be

important. Finally, results of this study may be generalizable to other areas of medicine in which

there is great variability in patient characteristics and risk level and a choice of potential centers.

In recent years, there has been an increased emphasis on patient decision-making and transparent










information for consumers of healthcare services. This study will elucidate the importance of an

aspect of these guidelines over a broad timeline in a significant patient population. Most

importantly, this study aims to provide important evidence to inform patients and caregivers

about the impact of center characteristics on their long-term prognosis.













DemographiclEnvironmental Factors

Re uairon








Provider selection
ESRD and suitable
and interested in for wait listing with SURVIVAL
a given health
transplant
stock (Hi)









Location
Clinical Factors



Mi = M(Pj, Vj, Qj, Wj)


Figure 2-1. Conceptual Framework based on Grossman's Health Production Function









CHAPTER 3
MATERIALS AND METHODS

Overview

The primary purpose of this study is to examine the association of center characteristics

with mortality for renal transplant candidates. The general approach will be to estimate levels of

four center characteristics utilizing data prior to candidate listing and evaluate the association of

these characteristics with candidate mortality in the years after listing. This study will be a

retrospective analysis of observational data derived from a national registry database. The

outcomes of this study will determine whether each of the tested center characteristics are

significantly associated with candidate mortality and the relative effect of high and low levels of

these factors among those that are significant. Furthermore, this study will estimate life

expectancy for candidates at incremental levels of significant factors that may ultimately be

utilized as a guide for center selection. Finally, this study will reevaluate the significance and

life expectancies for candidates that are of particular high risk and may have unique needs with

respect to center selection.

Data

This study will utilize a national transplant registry which is administered by the SRTR in

order to conduct the analysis. The SRTR database contains information for all transplants

performed in the United States. Information is collected at all centers on a mandatory basis

utilizing data collection forms at several time points for transplant patients corresponding to their

transplant status and has been utilized extensively for government-sponsored reports and peer-

reviewed research publications. Data is collected electronically and files are distributed at cost

for the purpose of research to groups following submission of a research and security plan. For

this study, the data that will be utilized will include information deriving from the form at the









time of transplant candidate listing, the form submitted at the time of transplantation, and follow-

up forms after transplantation. In addition to data collected from transplant centers, files are

enriched with data from Medicare and the Social Security Administration for data fields

pertaining to patient death or re-initiation of dialysis following transplantation.

The database contains patient-level data including donor and recipient demographic

information, primary clinical characteristics, and a numerical (de-identified) code for transplant

centers. The unit of analysis for this study will be adult patients (at least 18 years of age at the

time of listing) that are placed on the waiting list for a solitary kidney transplant as indicated in

the database. Pediatric patients (candidates less than 18 years of age) and candidates listed for

multi-organ transplants will be excluded from this study in order to examine a more homogenous

population. Candidates and recipients at centers with very low volume (less than ten transplants

per year) will also be excluded as estimates of center factors are less reliable. These candidates

represent a significant minority of all national candidates. In addition, recipients of living donor

transplants will be excluded from this study as they are less applicable for candidates placed on

transplant waiting lists for a deceased donor transplant. The study population will consist of all

remaining candidates that were listed for a solitary deceased donor transplant between 1995 and

2000. This period was selected to represent a relatively recent cohort, but with follow-up time

sufficient to evaluate the impact of transplant center factors on long-term patient outcomes.

Current files that will be utilized for the analysis have follow-up information through early 2006.

In addition, data will be utilized from 1992 to 1999 to assess transplant characteristics based on

three years of information accrual prior to candidate listing.

Dependent Variables

The primary outcome variable in the study will be transplant candidate mortality after

listing. Candidates will be followed from the time of listing until the earliest of death or last









follow-up time as indicated in the database. Patient death is indicated in the database with

internal variables derived from center follow-up forms as well as populated from Social Security

master files. The earliest of these dates in cases of discrepancies will be utilized as an endpoint

for the study. Mortality will be compared between study groups utilizing survival models based

on time to death following candidate listing. Models will be censored at the time of living

transplantation to limit results to recipients from deceased donors. A secondary endpoint will be

candidate death prior to the date of transplant. This model will be censored at the time of

transplantation, limiting events to the pre-transplant period. An additional secondary endpoint

will be overall graft loss following transplantation. These models will be initiated at the time of

transplantation for the same study groups and patients will be followed until a last follow-up date

or the date of overall graft loss. Overall graft loss is defined as the composite endpoint of either

patient death or loss of the transplanted kidney indicated by a return to dialysis or re-

transplantation. Additional descriptive information will include the correlation of historical

center characteristics with prospective levels after candidate listing, the association between

center characteristics and the distribution of these characteristics within transplant regions.

The purpose of the primary endpoint is to examine the overall survival rate for patients

based on levels of center characteristics. While other factors are clearly important to patients and

caregivers, measures of death and graft loss are obj ective "hard endpoints" that are clearly

defined. In addition, this data is readily available and reliable, and mortality rates are generally

interpretable for consumers of the research and comparable across study contexts. As the

primary variables of interest have all been shown to have some association with outcomes

following transplantation, the question remains whether the variables will prospectively impact

candidate survival for listed patients and how important the factors are relative to each other.










The purpose of the secondary outcomes is to determine whether differences that exist by center

characteristics are most directly related to pre- or post-transplant mortality.

Explanatory Variables of Interest

The primary explanatory variables of interest will be four transplant level characteristics:

center volume, center performance rating, center proportion of ECDs and the proportion of

patients reaching transplantation. Each of these will be assessed based on data for the three years

prior to the year of listing for the transplant candidate. The variables will be categorized into

quintiles based on their distribution over the study period and implemented as dummy variables

in statistical models. Each of these characteristics will be estimated according to an individual

candidate's year of listing based on retrospective center characteristics over the previous three

years. Transplant volume will be estimated based on the average number of deceased donor

transplants in the three years prior to candidate listing. Performance ratings (also known as

standardized mortality ratios) will be constructed utilizing data for deceased donor transplants

from the prior three years of candidate listings. Ratings will be determined utilizing the

equivalent methodology as the SRTR, which publishes center performance ratings on public

websites (98). These ratings are produced based on the ratio of the number of actual events (in

this case, patient deaths) to the expected number of events based on characteristics of the center

population. The value of this ratio will be utilized for the purpose of this study based on three-

year outcomes. The proportion of high-risk deceased donor transplants will be based on the

dichotomous ECD criteria and defined as the proportion of ECDs (among all deceased donor

transplants) over the prior three years. Transplant center waiting time will be indicated by the

proportion of patients that receive a deceased donor transplant within the three-year period prior

to listing. This proportion will be calculated by Kaplan-Meier models and censored at the time

of patient' s delisting, transfer to other centers, or receipt of a living transplant.









The center characteristics are selected for three primary reasons. First of all, the factors

have been shown to be associated with patient outcomes following transplantation based on

retrospective analysis. As such, it is reasonable to surmise that prospective evaluation of these

factors will have an impact on candidate outcomes following listing. Secondly, the information

regarding these characteristics is all publicly available and as such could be incorporated into

decision-making processes for candidates and their caregivers in order to assist with center

selection. Lastly, these factors have wide regional variability as well as variable representation

among centers within regions. This is important from a methodological perspective as it will

more readily allow for the detection of effects of these factors on patient outcomes. In addition,

the center variability highlights the potential decisions that candidates may make regarding their

choice of center selection.

There are a total of 260 transplant centers with unique identification numbers in the

database within the study period. A portion of these centers discontinued their programs within

the study period and several programs were newly initiated within the time frame. In order to

accurately estimate effects based on transplant center characteristics, centers with less than ten

deceased donor transplants per year of operation will be excluded from the study. Among the

centers, 184 (71%) had at least ten deceased donor transplants per year. These centers accounted

for 97% of all deceased donor transplants over this period. The distribution of the average

annual number of deceased donor transplants is displayed in Figure 3-1. The median (and 25th/

75th percentile) number of deceased donor transplants at these centers was 30 (19 / 48), ranging

from 10 to 177. The median proportion of ECD transplants (of all deceased donor transplants)

was 1 1%, ranging from 0% to 30%. The distribution of the proportion of ECD transplants is

illustrated in Figure 3-2. The proportion of candidates receiving a deceased donor transplant









within three years after listing also varied significantly over the study period. The range of this

proportion was 1 1% to 91% with a median level of 57%. The distribution of this proportion of

candidates receiving a deceased donor transplant is displayed in Figure 3-3. Standardized

mortality ratios ranged from 0 (indicating no events over the time period) to close to 4.0

(indicating four times as many events as would be expected based on characteristics of the

transplant population at the center). The distribution of the mortality ratios by center are

displayed in Figure 3-4.

Additional Explanatory Variables

Multivariate models will incorporate a number of adjustment variables that are

considered to be independently associated with the mortality or waiting time to transplant for

candidates. Candidate age will be utilized as it is strongly associated with death rates as well as

the likelihood to receive a transplant over time. Age will be categorized into groups as 18-44,

45-54, 55-64, and 65+. Candidate race will be used in the models categorized as Caucasian,

African-American, Asian, Hispanic, and Other. Caucasians have significantly higher mortality

rates on the waiting list relative to other race groups as well as greater likelihood to receive a

transplant. Candidate gender will be used in the models; there is some indication that gender is

associated with likelihood to progress more rapidly on the waiting list. Candidate primary cause

of ESRD will be dichotomously represented in the models as diabetes or other. Diabetics with

ESRD have significantly higher death rates and less likelihood to reach transplantation.

Candidate BMI will be categorized into levels representing different health status (Missing, 13-

19, 20-24, 25-29, 30-34, 35+). Obese candidates have been shown to have a slower progression

to transplant and superior survival after dialysis initiation. Additional variables that will be

utilized to control for potential selection bias in center choice will include candidate education

and insurance status. Education will be dichotomously represented as a college degree or higher,









or less than a college degree. Insurance status will be categorized as private, Medicare, or other.

Obese candidates will be defined as those with a calculated BMI greater than 30 kg/m2 and

elderly patients will be classified as those that are 65 years of age or older at the time of listing.

A secondary outcome for the study will include post-transplant graft and patient survival.

These models will include only those candidates that receive a deceased donor transplant and

will be initiated at the time of transplantation with follow-up until death, graft loss, or the last

follow-up period. Cox models will be adjusted for transplant characteristics that are associated

with outcomes after transplantation including HLA mismatching level, donor age, cold ischemia

time, donor race, and pre-transplant dialysis time, in addition to adjustment factors utilized in the

primary outcome models.

Statistical Analysis

The primary analyses used in the dissertation to evaluate the study aims will be

performed with survival models. Survival models (also known as time to event analyses) are

appropriate in contexts with a known time origin, in this case the time of candidate listing, and

with well-defined follow-up period and events. In addition, as in the case for this data, subjects

have variable follow-up periods, and survival models are capable of incorporating these different

follow-up periods in the analyses (as opposed to other types of regression models). Each subj ect

(transplant candidate) has a known candidate listing date as well as a last follow-up date, which

may be death or the last follow-up period. For these analyses death will be treated as the event

of interest and patients that do not die over the study period will be censored at the last follow-up

period. This form of censoring is commonly referred to as right-censoring. Models for

secondary outcomes will be censored at the time of transplantation and additionally an

alternative model will use the time of transplantation as the origin point.









Analyses for examining factors associated with patient outcomes will use a survivor

function S(t), which gives the probability of survival until a given time, t. S(t) is a monotonically

decreasing function which is initiated at a survival probability of one and decreases with

subsequent events over time as indicated graphically on the horizontal axis. The rate of decline

of the survival function is utilized as a measure of the risk at a particular event time and is

generally regarded as the hazard function h(t). The hazard function can generally be written as

shown in Equation 3-1.

h(t) = d/dt (- In [S(t)]) (3-1)

The analyses will also utilize Kaplan-Meier methodology (also known as the product-

limit method) to compare the survival function among sample strata. This methodology is

nonparametric and is appropriate when exact dates are known for events (rather than aggregated

across intervals), which is the case with the current data. Comparisons of the survival function

between study groups will be made with the Log-Rank test which is generally the statistically

most powerful test for making unadjusted comparisons between study groups. Censoring will be

assumed to be noninformative as last follow-up information is generally a product of the most

recent data available and there are few cases of patients that are lost to follow up in this cohort.

To compare outcomes between study groups adjusted for potential confounding factors,

hazard functions will be compared utilizing Cox proportional hazard models. Cox models are

semi-parametric and are not reliant on a specific distributional form of a survival function

supporting the robust nature of the results. An assumption that is applicable for the Cox model is

that the hazard ratio of covariates in the model is proportional over time (i.e., additive changes in

covariates cause multiplicative changes in the hazard functions). For continuous covariates this

will be tested explicitly by entering an interaction term of the covariate in question with time into









the model. In the case of categorical covariates, assessment of proportionality will be made by

visual inspection of the log-log survival function. The Cox model can be written in the form

depicted in Equation 3-2.

hi(t) = ho(t) exp(Plxil + P2Xik + kXik) (3 -2)
where i = individual, t = time, x = covariate of interest, and k = number of covariates

The Wald test will be utilized to test the two-sided hypothesis that a center characteristic

is significantly associated with patient mortality (Ho: P1 = 0 versus Ha: P1 / 0) using the equation

z = P1/(SE)B,. In addition, results will be displayed as adjusted hazard ratios relative to a

reference level of the center factors. Cox models will also be used to estimate adjusted survival

rates following candidate listing for combinations of center-level variables. These results will be

presented in tabular form with applicable standard error estimates.

Study Aim I

Kaplan-Meier and multivariate Cox models will be used to test the hypothesis that the

individual center characteristics are associated with wait-list candidate survival. Kaplan-Meier

models will be constructed for each individual center characteristic based on quintile levels of

the variable utilized as strata in the model. Results will be displayed graphically and tested with

Log-Rank tests. Cox models will be used incorporating all center characteristics simultaneously

along with other potential confounding factors as described previously. Overall significance,

hazard ratios, and confidence intervals for center characteristics will be displayed in tabular

form.

Study Aim II

Cox models will be utilized to estimate survival rates for candidates at multiple

permutations of center characteristic levels. Only center characteristics that were shown to be

significantly associated with candidate mortality will be utilized for this study aim. The baseline










hazard function of the model will be adjusted with incremental levels of the center factors that

span the range of these variables, and all survival estimate combinations will be displayed in

tabular form.

Study Aim III

In order to evaluate the effect of center characteristics in three high-risk candidate groups,

the methodology described in study aims I and II will be utilized strictly in subsets of candidates.

The same explanatory variables will be utilized in Cox models without the variable applicable to

the designation of the high-risk subset. For models restricted to the elderly, age categories will

be reclassified within the elderly cohort. With the subset of African-American patients,

candidate race will be removed from the models. For the obese cohort, new categorization of

BMI will be used for patients with BMI > 30 kg/m2. Survival estimates, as described in study

aim II, will be conducted for each subset of high-risk patient in a similar fashion.

Potential Selection Bias

The primary aim of the study is to evaluate the independent effect of center

characteristics on transplant candidate mortality. However, one of the limitations of this

retrospective analysis will be the possibility that mortality rates at centers are correlated with the

patient' s selection of these centers. In particular, patients that are well informed, have higher

educational background, or have resources that allow them to travel to centers of their choice or

make informed decisions about centers that are of higher quality, and these patients may also

have a lower risk profile. There is data to support the notion that performance measures do not

influence transplant center selection in aggregate; however, this may not fully incorporate more

subtle patient characteristics (21). The effect of this bias may be to overestimate the influence of

center characteristics, assuming that "better" patients select centers that also have better

outcomes.










The primary strategy of the study to account for this potential bias will be to statistically

control for patient characteristics that may influence center selection as well as mortality. These

patient-level variables that will be adjusted for in the outcome models will include age, primary

diagnosis, race, gender, primary insurance coverage, and education level. These factors will be

assumed to provide adequate control for characteristics of patients associated with center

selection related to overall health level, education, affluence level and interaction with the

healthcare system. In addition, examination of the secondary outcome of candidate mortality

prior to transplantation will provide insight on the degree of selection bias that exists. In

particular, mortality rates prior to transplantation are significantly less associated with care

provided by the transplant center, and significant differences in pre-transplant mortality are more

likely attributable to patient characteristics. Evaluation and comparison of pre- and post-

transplant mortality effects by center characteristics will therefore serve as a verification of any

observed associations.
















30


25


P 20


S15


10







0 20 40 60 80 100 120 140 160 180 200

Annual Number of Deceased Donor Transplants by Center

Figure 3-1. Distribution of the Annual Number of Deceased Donor Transplants by Center





















150



125



P 10 0



t 75



50








0 4 8 12 16 20 24 28 32 36 40 44
Proportion of ECD Transplants by Center

Figure 3-2. Distribution of the Proportion of ECD Transplants by Center
















10






c6












0 10 20 30 40 50 60 70 80 90 100
Proportion of Patients Transplanted by Center


Figure 3-3. Distribution of the Proportion of Candidates Receiving a Deceased Donor
Transplant within Three Years by Center
















20 0


17 5



15 0



12 5
P
e

c 10 0
e
n

75



50







200

0.0 0.5 1 .0 1 .5 2.0 2.5 3.0 3.5 4.0


Performance Ratios by Center


Figure 3-4. Distribution of Standardized Mortality Ratios by Center









CHAPTER 4
RESULTS

Study Population

There were 108,928 adult solitary kidney transplant candidates in the study population.

Table 4-1 displays descriptive statistics about the candidate population from 1995-2000. Within

the study period, 8% of candidates were 65 years or older, 59% of candidates were male, 28%

were African-American, 28% had diabetes as a primary cause of ESRD, 15% of candidates had a

college degree or more, 48% had type-O blood, and 55% of candidates had Medicare as their

primary insurance payer. Among candidates with known BMI levels, 21% were obese and

among candidates with known Panel Reactive Antibody (PRA) level at listing, 31% were

sensitized (i.e., PRA > 0).

Rate of Transplantation by Center

Transplant centers for each listing year were categorized based on the proportion of

candidates that received a deceased donor transplant within three years after being placed on the

waiting list. Center categories for this proportion were assigned by quintiles as follows: Q1 =

[7.0-3 8.5], Q2 = [3 8.6-5 1.6], Q3 = [5 1.7-63.1i], Q4 = [63.2-76.7], and Q5 = [76.8-96.0]. Table

4-2 displays candidate characteristics by the center proportion of transplants categorized by

quintile. The most notable differences by center category were among patients with private

primary insurance coverage (Q1=45% vs. Q5=40%), African-American recipients (Q1=32% vs.

Q5=21%), and sensitized patients (Q1=30% vs. Q5=39%). The number of candidates and the

median proportion of patients reaching transplantation at three years for candidates in each group

were as follows: Q1 (n=30478, median = 28.4), Q2 (n=24271, median = 45.4), Q3 (n=22768,

median = 58.1), Q4 (n=18858, median = 67.3), and Q5 (n=12553, median = 83.0).









Transplant Center Volume

Transplant centers for each listing year were categorized based on the average number of

deceased donor transplants three years prior to the year of candidate listing. Center categories

for volume were assigned by quintiles as follows: Q1 = [10.0-18.8], Q2 = [18.9-26.3], Q3 =

[26.4-36.6], Q4 = [36.7-53.7], and Q5 = [53.8-195.0]. Table 4-3 displays candidate

characteristics by center volume quintile. The most notable differences by center category were

among patients with a college degree (Q1=17% vs. Q5=22%), and patients with private primary

insurance coverage (Q1=36% vs. Q5=45%). The number of candidates and median annual

transplant volume for candidates in each group were as follows: Q1 (n=8462, median = 15.7),

Q2 (n=12597, median = 23.0), Q3 (n=16433, median = 31.0), Q4 (n=24370, median = 44.0), Q5

(n=47066, median = 88.0).

Center Donor Quality

Transplant centers for each listing year were categorized based on the proportion of

expanded criteria donor transplants three years prior to the year of candidate listing. Center

categories for the ECD proportion were assigned by quintiles as follows: Q1 = [0-6.4], Q2 =

[6.5-9.9], Q3 = [10.0-13.4], Q4 = [13.5-18.9], Q5 = [19.0-44.0]. Table 4-4 displays candidate

characteristics by center ECD proportion quintile. The most notable differences by center ECD

proportion category were among patients with private primary insurance coverage (Q1=39% vs.

Q5=43%), elderly patients (Q1=7% vs. Q5=10%), African-American patients (Q1=25% vs.

Q5=32%), sensitized patients (Q1=36% vs. Q5=27%), diabetic patients (Q1=30% vs. Q5=26%),

and obese patients (Q1=19% vs. Q5=23%). The number of candidates and the median

proportion of ECD transplants for candidates in each group were as follows: Q1 (n=17980,

median = 4.4), Q2 (n=19229, median = 8.3), Q3 (n=22768, median = 11.8), Q4 (n=24139,

median = 15.7), Q5 (n=24812, median = 24.0).










Center Performance Ratings

Transplant centers for each listing year were categorized based on the ratio of observed to

expected deaths at one year utilizing data for recipients three years prior to the year of candidate

listing. Center categories for volume were assigned by quintiles as follows: Q1 = [0-0.56], Q2

= [0.57-0.83], Q3 = [0.84-1.09], Q4 = [1.10-1.45], Q5 = [1.46-3.80]. Table 4-5 displays

candidate characteristics by center performance quintile. The most notable differences by center

performance category was among patients with private primary insurance coverage (Q1=43% vs.

Q5=35%), African-American patients (Q1=25% vs. Q5=34%) and patients with a college degree

(Q1=21% vs. Q5=18%). The number of candidates and the median ratio of observed to expected

deaths for deceased donor transplants by candidates in each group were as follows: Q1

(n=17939, median = 0.37), Q2 (n=23865, median = 0.72), Q3 (n=27705, median = 0.96), Q4

(n=23972, median = 1.26), Q5 (n=15447, median = 1.70).

Association between Transplant Center Characteristics

The median center volume and performance ratings were relatively stable over the period

(Table 4-6). In contrast, the proportion of patients transplanted at three years significantly

declined throughout the study period, and the proportion of ECD transplants increased over time.

Also indicated in the table, the total number of candidates increased steadily over the study

period.

There was no significant correlation between the volume of deceased donor transplants

with the other three center characteristics investigated in the study (Table 4-7). There was a

significant negative association between the proportion of patients transplanted at three years

with the proportion of ECD transplants; however, the proportion of patients transplanted at three

years was not correlated with performance ratios. The ratio of observed to expected patient

deaths at one year was positively correlated with the ECD proportion between centers. Centers









with a higher proportion of ECD transplants had less patients transplanted at three years and

worse (i.e., higher ratios) performance ratings.

Reliability of Historical Center Characteristics

The volume of transplants was positively associated with several characteristics of the

centers at listing (Table 4-8). On average, listing for a transplant at a center with higher volume

was positively associated with the volume of transplants at the center at the time of

transplantation. Candidates who listed at centers with historically smaller number of patients

transplanted also had longer durations between listing and receiving a transplant. In a similar

fashion, patients who listed at centers with historically higher proportion of ECD transplants also

were more likely to be transplanted at centers with a higher proportion of ECD transplants. In

addition, recipients that listed at centers with better performance ratios also were transplanted at

centers with better ratios during the year of transplantation. In general, characteristics of

transplant centers at the time of listing were still evident at the time of transplantation.

Kaplan-Meier Candidate Survival by Center Characteristics

Figure 4-1 displays candidate survival by quintile level of the proportion of patients

transplanted at three years prior to listing. There was a stepwise and significant association

between a higher proportion of patients transplanted and higher candidate survival over the study

period (p <0.001). The proportion of candidates surviving at ten years following listing was 56%

at centers with the highest proportion of patients reaching transplantation as compared to 50% at

centers with the lowest proportion. Ten-year survival at centers with lowest proportion of ECD

transplants was 54% as compared to 50% at high ECD centers (Figure 4-2). Time to death

following listing by transplant volume was not statistically significant different (Figure 4-3).

The proportion of candidates surviving at ten years based on center performance ratios, ranged

from 51-54% in the lowest-performing and highest-performing centers respectively (Figure 4-4).









Multivariate Cox Model for Primary Outcome of Candidate Mortality

The primary outcome of the study was candidate mortality after listing for

transplantation. Results of the Cox proportional hazard model for patient mortality including the

center study characteristics categorized by quintile level are displayed in Table 4-9. Relative to

Caucasian candidates, all other racial groups had significantly lower mortality after listing.

Older age was significantly associated with increased mortality including a two-fold risk for

candidates over the age of 65 years (adjusted hazard ratio [AHR] = 3.24, 95% confidence

interval [C.I.] 3.13-3.35). Patients with a primary diagnosis of diabetes had an approximate 88%

increase in the hazard ratio for death relative to patients without the diagnosis. Patients with

type-A and type-AB blood had a reduced hazard ratio for death relative to type-O candidates.

Patients that were listed as highly sensitized (PRA > 30) had a 34% increase hazard ratio for

death after listing relative to non-sensitized candidates. Low BMI levels (< 18 kg/m2) and BMI

levels above 35 had an increased hazard ratio for death relative to candidates with BMI between

19-25 kg/m2. Patients with less than a college education had an increased likelihood of mortality

(AHR=1.08, 95% C.I. 1.05-1.12). Candidates with Medicare as a primary insurance payer had

significantly elevated mortality relative to candidates with private insurance. Candidates on

dialysis at the time of listing had an elevated hazard for death relative to candidates that were

listed prior to dialysis initiation.

Among center characteristics, the rate of transplantation had the strongest association

with candidate mortality after listing. Candidates that listed at centers with the lowest percentage

of patients reaching transplantation at three years had a 32% increased relative hazard for death

(AHR = 1.32, 95% C.I. 1.27-1.3 8) relative to candidates listing at centers with the highest

percentage of patients reaching transplantation. In addition, there was a stepwise association

between levels of the center proportion of patients transplanted. Transplant volume was not










significantly associated with mortality between any of the groups. The center proportion of ECD

transplants was associated with candidate mortality between candidates listed at centers with the

highest proportion of ECD transplants (AHR = 1.04, 95% C.I. 1.00-1.08) relative to candidates

listed at centers with the lowest proportion of ECDs. Candidates listed at centers with the lowest

performance measures had a 14% increase in hazard ratio for death (AHR = 1.14, 95% C.I. 1.10-

1.19) relative to candidates listed at centers with the highest performance.

The Cox proportional hazard model results for the outcome of receiving a deceased donor

transplant following listing is displayed in Table 4-10. There was a highly significant

association of receipt of transplant by centers with historically different proportions of candidates

receiving transplants. The hazard for patients receiving a transplant at centers with the lowest

rate was over 5-fold decreased (AHR = 0. 17, 95% C.I. 0. 16-0. 17) relative to centers with highest

rates of transplantation. There was also a stepwise and significant increase in the hazard ratio by

center level. Center volume was also associated with the hazard rate of transplant with the most

notable difference between candidates at the smallest centers less likely to receive a transplant

(AHR=0.94, 95% C.I. 0.90-0.98) relative to candidates at the highest volume centers. There was

a varied association between center levels of ECDs, with candidates at centers with a high

proportion of ECD transplants more likely to receive a transplant (AHR = 1.10, 95% C.I. 1.07-

1.14). Candidates at different performance centers also had mildly different time to

transplantation; however, there was no distinct pattern between levels.

The association of center characteristics with post-transplant mortality for candidates who

acquired a deceased donor transplant within the study period is displayed in Table 4-11. There

was no significant association between hazard ratios for post-transplant mortality with candidates

listed at centers with different rates of transplantation or proportion of ECD transplants.









Candidates listed at centers with small volume (AHR=1.10, 95% C.I. 1.03-1.18) had an elevated

hazard ratio for post-transplant mortality; however this association was not statistically

significant among candidates listed at the smallest centers. Candidates listed at centers with the

lowest performance levels had a 20% increased hazard for death (AHR= 1.20, 95% C.I. 1.11-

1.29). The model for post-transplant graft loss resulted in similar findings. There was no

significant association between graft loss and candidates listed at centers with different rates of

transplant, there was mixed results for center volume, and candidates listed at centers with lower

performance levels had elevated post-transplant graft loss (Table 4-12). One distinction in for

the outcome of graft loss, as compared to patient mortality, was that candidates listed at centers

with highest ECD transplants had a significantly elevated hazard for post-transplant graft loss

(AHR=1.08, 95% C.I. 1.02-1.15) relative to candidates listed at centers with a lowest ECD

proportion. In addition, the model excluding characteristics of the donor organ (e.g., donor age

and ECD) indicated that centers with high ECD levels were more significantly associated with

post-transplant mortality (AHR=1.14, 95% C.I. 1.07-1.23) and graft loss (AHR=1.18, 95% C.I.

1.12-1.25).

Center performance level had a significant and stepwise association with candidate

mortality prior to transplantation (Table 4-13). Specifically, candidates that listed at centers with

lower historical performance levels had higher mortality rates prior to transplantation.

Candidates that listed at centers with the lowest performance levels had a 13% (AHR = 1.13,

95% C.I. 1.08-1.18) elevated hazard for death relative to candidates that listed at centers with the

highest historical performance levels. Alternatively, candidates listed at centers with historically

different rates of transplantation, center volume or ECD proportion had minimal association with

pre-transplant mortality.









Outcomes among High-Risk Candidate Groups

Models were repeated for candidate mortality limited to subsets of high-risk patients.

These models were generated for African-American, elderly, and obese candidates adjusted for

the same covariates with the exception of the variable describing the high-risk characteristic. In

addition, models including interaction terms for the center factors and high-risk characteristics

were generated and hypotheses tested for whether high and low levels of center factors were

more important for high-risk groups relative to their candidate counterparts.

Among African-American candidates, the rate of transplantation had a highly significant

and stepwise association with mortality (Table 4-14). Candidates listed at centers with the

lowest transplant rates had a 28% increased hazard for death (AHR=1.28, 95% C.I. 1.18-1.38)

over the study period. African-American candidates listed at centers with varying volume levels

had no significant difference in mortality. African-American candidates listed at centers with the

highest proportion of ECD transplants had a significantly elevated hazard for death (AHR=1.09,

95% C.I. 1.02-1.16) relative to candidates listed at centers with a low proportion of ECD

transplants. African-American candidates listed at centers with the lowest performance ratio also

had an increased hazard for death (AHR = 1.14, 95% C.I. 1.06-1.22) relative to candidates listed

at centers with the highest performance ratio.

The model including interaction terms between center characteristics and race groups

(limited to African-American and Caucasians) indicated that the association of center effects

with mortality was not different by race. The difference between centers with a high and low

rate of transplantation (p=0.42), high and low volume centers (p=0.83), high and low ECD

centers (p=0.48) and high and low performance ratio (p=0.84) was not different between the two

candidate race groups.









The rate of transplantation had a highly significant and stepwise association with

candidate mortality among elderly candidates (Table 4-15). Elderly candidates listed at centers

with the lowest transplant rates had a 26% increased hazard for death (AHR=1.26, 95% C.I.

1.13-1.40) over the study period. There was not an association between center volume with

mortality for elderly candidates. The proportion of ECD transplants was associated with

mortality with those elderly candidates listed at centers with a mid-level of ECDs having a

reduced hazard for death (AHR=0.90, 95% C.I. 0.82-0.99) relative to candidates listed at centers

with a high proportion of ECDs. Elderly candidates listed at centers with the lowest performance

ratio also had a significantly increased hazard for death (AHR = 1.23, 95% C.I. 1.11-1.36)

relative to candidates listed at centers with the highest performance ratio.

The model including interaction terms between center characteristics and age groups

(limited to 18-44 and 65+ age groups) indicated that the proportion of ECD transplants was

significantly more important in younger age groups than in the elderly (AHR=1.07, 95% C.I.

1.01-1.13). Other center factors did not demonstrate a significantly different impact in younger

versus older patients: rate of transplant (p=0.41), transplant volume (p=0.36), and performance

ratio (p=0.61).

As displayed in Table 4-16, the rate of transplantation had a highly significant and

stepwise association with candidate mortality for obese candidates. Obese candidates listed at

centers with the lowest transplant rates had a 33% increased hazard for death (AHR=1.33, 95%

C.I. 1.22-1.45) relative to candidates listed at centers with the highest transplant rates. There

was no association between center volume and candidate mortality for obese candidates. The

proportion ofECD transplants was also not significantly associated with mortality. Obese

candidates listed at centers with the lowest performance ratio had a significantly increased









hazard for death (AHR = 1.11, 95% C.I. 1.01-1.21) relative to candidates listed at centers with

the highest performance ratio. There was no significant differences in the effect of center listing

for volume (p=0.27), center performance level (p=0.63), transplant rate (p=0.87), or center

proportion of ECD transplants (p=0. 13) between obese and non-obese patients.

Expected Survival by Center Characteristics

Expected survival estimates displayed on Table 4-17 indicate the average expected

survival based on the level of the center characteristic indicated holding all other center and

candidate characteristics at their average level. Results indicate that a candidate's expected

survival is markedly different by levels of center levels of the proportion of patients transplanted

at three years. Average candidates listed at centers with 82% of patients transplanted at three

years have almost 2.5 years increased expected survival relative to patients listed at centers with

10% of patients transplanted within three years. Differences in expected survival between

extreme center volume levels were mildly different ranging from 10.4 to 10.9 years. Candidates

listed at centers with few ECD transplants had an expected survival of 10.8 years as compared to

candidates listed at centers with 32% of ECD transplants, who had an expected survival of 10.5

years on average. Candidates listed at centers with low performance ratings (i.e., high standard

mortality ratios) had a 10.1 year expected survival as compared to candidates listed at centers

with high performance, who had an expected survival of 11.2 years.

Table 4-18 displays expected survival rates for candidates at hypothetical combinations

of center characteristics. The case examples demonstrate differences in expected survival across

ranges of the center proportion transplanted and by extreme levels of the other three center

characteristics. In particular, the estimates demonstrate the differences in life expectancy relative

to the proportion transplanted with either the "worst" combination of other factors as compared

to the "best" combination of the other three factors. Among centers with the lowest proportion









of candidates reaching transplantation, the variation in average life expectancies ranged from 9.3

to 10.3 years. Centers with a mid level of proportion transplanted and the "worst" combination

of additional factors had an average life expectancy of 10.2 years as compared to the best

combination of additional factors at 11.6 years. Average life expectancy was notably higher

within centers with high proportion of transplant candidates, ranging from 11.5 to 13.1 years.

The expected survival for African-American patients following listing ranged from 9.5

years to 11.5 years at centers with a historically low versus high proportion of candidates

reaching transplant at three years respectively (Table 4-19). The variation in life expectancies

varied little by center volume. The average life expectancy ranged from 9.6 to 10.3 years by

centers with a high and low proportion of ECD transplants respectively. Similarly, average life

expectancies for African-Americans ranged from 9.6 to 10.6 years based on the historical center

performance level with higher performance centers associated with longer life expectancies.

The center characteristic with the largest range in life expectancies for elderly candidates

was associated with the center transplant rate (Table 4-20). Specifically, elderly candidates

listed at centers with the highest proportion of candidates transplanted at three years had an

expected 6.5 years of life expectancy as compared to elderly candidates at centers with the

lowest proportion of candidates reaching transplantation, the life expectancy was 5.6 years.

There was mild fluctuation in life expectancy by center volume for elderly candidates, but no

notable pattern among levels. In a similar fashion, life expectancy for elderly candidates was

highest associated with listing at centers with mid-levels of ECD transplants; however, the life

expectancy between low and high centers was similar. Average life expectancies by levels of

center performance ratios ranged from 6.1 years in the highest performance centers to 5.3 years

in the lowest performance centers.









Average expected life years for obese candidates at centers with a low transplant rates

was 8.2 as compared to 10.2 years for candidates listed at centers with the highest proportion of

candidates reaching transplantation (Table 4-21). Life expectancies varied minimally based on

center volume with no detectable pattern for an association of longer life expectancy with higher

of lower volume. Expected life years for obese candidates listed at centers with a low ECD

proportion was 9.0 years as compared to candidates listed at centers with a high ECD proportion

with 8.7 years. Average life expectancy after listing also varied at centers based on performance

level; candidates at the highest performing centers had 9.1 expected life years as compared to 8.4

years at the lowest performing centers.










Table 4-1. Transplant candidate characteristics
Candidate characteristic Level
Age at listing 18-44
45-54
55-64
65+


43
28
21
8
54
28
5
11
2
59
41
28
72)
2
59
16
23
60
15
25
48
52
42
55
3
63
19
10
8
108928


Race


Caucasian
African-American
Asian
Hispanic
Other
Male
Female
Diabetes
Other
13-18
19-30
30+
Missing
Less than college degree
College degree or more
Missing

A, B or AB
Private
Medicare
Other/missing

1-30
30+
Missing


Gender


Primary cause of ESRD

Candidate BMI



Candidate education level


Blood type

Candidate primary insurance


Candidate peak PRA level



Sample Size









Table 4-2. Candidate characteristics by center proportion of transplants within three years
Candidate Center rate of transplantation category *
Level
character stic Q1 Q2 Q3 Q4 Q5
Gender Male (%) 59 59 60 59 59
Blood type Type-O (%) 48 47 48 49 48
Education College degree (%) ^` 21 22 19 20 18
Primary insurance Private (%) 45 42 38 40 40
Age 65+ (%) 8 9 8 8 8
Race African-American (%) 32 30 29 24 21
PRA level > 0(%) ^ 30 29 29 34 39
Primary diagnosis Diabetes 28 28 27 28 27
BMI 30+ (%) ^` 22 22 21 21 21
* Categories represent the quintiles of the proportion of patients receiving a transplant by
three years: Q1 = [7.0-38.5], Q2 = [38.6-51.6], Q3 = [51.7-63.1], Q4 = [63.2-76.7], Q5 =
[76.8-96.0]. ^` The proportions for these categories exclude missing levels.


Table 4-3. Candidate characteristics by center volume category
Candidate Ce
Level
characteristic 01


nter volume category *
Q2 03 04


;9 59
-8 47
7 20
6 36
9 8


Gender
Blood type
Education
Primary insurance
Age
Race
PRA level
Primary diagnosis
BMI


Male (%)
Type-O (%)
College degree (%) ^
Private (%)
65+ (%)
African-American (%)
> 0(%) ^
Diabetes
30+ (%) ^`


* The categories represent the quintile levels of deceased donor transplant volume:
[10.0-18.8], Q2 = [18.9-26.3], Q3 = [26.4-36.6], Q4 = [36.7-53.7], Q5 = [53.8-195.0].
proportions for these categories exclude missing levels.


Q1 =
^` The










Table 4-4. Candidate characteristics by center ECD proportion category
C t ECD ti t *


ener proper on category
Level n, o, n,


Candidate characteristic

Gender
Blood type
Education
Primary insurance
Age
Race
PRA level
Primary diagnosis
BMI


Male (%)
Type-O (%)
College degree (%) ^`
Private (%)
65+ (%)
African-American (%)
> 0(%) ^
Diabetes
30+ (%) ^`


* The categories represent the quintile levels of the proportion of ECD transplants: Q1
6.4], Q2 = [6.5-9.9], Q3 = [10.0-13.4], Q4 = [13.5-18.9], Q5 = [19.0-44.0]. ^ The
proportions for these categories exclude missing levels.

Table 4-5. Candidate characteristics by center performance category
r,,~~,+,,~,,,+,;,+; T ,,,1Center performance category *


[0-






Q5
59
48
18
35
9
34
30
26
23


an aCICitte c aracterst~lS Lc
Gender
Blood type
Education
Primary insurance
Age
Race
PRA level
Primary diagnosis
BMI


etve~


Male (%)
Type-O (%)
College degree (%) ^
Private (%)
65+ (%)
African-American (%)
> 0(%) ^
Diabetes
30+ (%) ^


* The categories assigned by quintile of standard mortality ratios (observed/expected deaths):
Q1 = [0-0.56], Q2 = [0.57-0.83], Q3 = [0.84-1.09], Q4 = [1.10-1.45], Q5 = [1.46-3.80].
^ proportion excludes missing levels


Table 4-6. Median levels of center characteristics over time
1995 1996
Volume 45.7 46.0
Proportion transplanted 61.5 56.6
ECD proportion 8.4 10.4
Performance ratio 0.98 0.98
Candidate listings 16412 16836 1


1997
47.7
54.3
13.1
0.91
7490


1998
47.0
49.1
14.0
0.94
18685


1999
49.3
47.7
14.8
0.96
19270


2000
48.0
45.5
14.7
0.92
20235









Table 4-7. Correlation coefficients between center characteristics
Linear correlation (p- Proportion
value) vllt transplanted p
Volume .01 (0.71)


-0.


ECD
,roportion
0.04 (0.24)
21 (<0.001)

0.08 (0.01)


Performance
ratio
-0.05 (0.09)

0.08 (0.01)


Proportion transplanted
ECD proportion
Performance ratio


0.01 (0.71)
0.04 (0.24)
-0.05 (0.09)


-0.21 (<0.001)
0.04 (0.19)


Table 4-8. Center characteristics at the time of transplantation
Median center characteristics at the year of transplantation
Level at the time of
Time to
listing Performance
Volm transplant ECD prportion rto
oume ~ ~~(month~s) I n 'to
Q1 20 36 9 0.75
Q2 28 27 11 0.79
Q3 34 21 13 0.83
Q4 46 17 16 1.08
Q5 93 12 23 1.13
* based only on patients transplanted during the year candidate was transplanted
















Candidate race (Caucasian)




Candidate age (18-44)


Primary diagnosis (Non-Diabetic)
Blood type (Type O)


PRA level (Zero)



BMI (19-25)






Education level (College)

Primary insurance (Private)



Dialysis status at listing (None)


Proportion transplanted (Highest -



Center volume (Largest Q5)




ECD proportion (Lowest Q1)




Performance Ratio (Best Q1)


African-American
Asian
Hispanic
Other
45-54
55-64
65+
Diabetes

AB

1-30
30+
Missing
13-18
26-30
31-35
36-40
41+
Missing
Less than College
Missing
Medicare
Other
Missing
Hemodialysis
Peritoneal dialysis
Unknown
Q5) Lowest Q1
Low Q2
Mid Q3
High Q4
Smallest Q1
Small Q2
Mid Q3
Large Q4
Low Q2
Mid Q3
High Q4
Highest Q5
Good Q2
Mid Q3
Bad Q4
Worst Q5


Table 4-9. Adjusted hazard ratios for patient mortality after listing for transplantation
Adjusted
95% confidence
Candidate characteristic (reference level) Lev d


\--I----'-- --*--I Y'*-- "UYU'U


interval

0.89
0.56
0.69
0.73
1.69
2.36
3.13
1.84
0.91
0.82
0.99
1.04
1.30
1.38
1.19
0.93
0.97
1.11
1.13
0.96
1.05
1.10
1.31
0.20
0.95
1.31
1.34
1.31
1.27
1.20
1.12
1.12
0.96
1.00
0.96
0.95
0.98
0.96
0.96
1.00
0.99
1.04
1.05
1.10


ratio
0.91
0.59
0.71
0.79
1.73
2.43
3.24
1.88
0.93
0.86
1.02
1.07
1.34
1.43
1.28
0.96
1.00
1.17
1.22
0.99
1.08
1.14
1.34
0.35
1.01
1.36
1.39
1.43
1.32
1.25
1.17
1.17
1.00
1.04
0.99
0.98
1.02
1.00
1.00
1.04
1.02
1.07
1.09
1.14










Table 4-10. Adjusted hazard ratios for receipt of transplant following listing
Ad t d 95% fid


juse con ence


Candidate characteristic (reference level)

Proportion transplanted (Highest Q5)



Center volume (Largest Q5)


Level


hazard ratio
0.17
0.31
0.46
0.59
0.94
0.97
1.02


interval
0.16 0.17
0.29 0.32
0.45 0.48
0.57 0.61
0.90 0.98
0.94 1.00
0.99 1.06


Lowest
Low
Mid
High
Smallest
Small
Mid


Large 0.98 0.95 1.00
ECD proportion (Lowest Q1) Low 1.07 1.04 1.11
Mid 1.09 1.05 1.12
High 1.10 1.07 1.14
Highest 1.02 0.99 1.06
Performance ratio (Best Q1) Good 0.95 0.92 0.98
Mid 0.93 0.90 0.96
Bad 0.95 0.92 0.98
Worst 0.97 0.94 1.01
* Model additionally adjusted for candidate race, age, primary diagnosis, blood type, PRA
level, BMI level, education level, primary insurance payer, and dialysis at the time of listing.

Table 4-11. Adjusted hazard ratios for post-transplant mortality

Candidate characteristic (reference level) Level dut 9 cniee
hazard ratio interval
Proportion transplanted (Highest Q5) Lowest 0.97 0.90 1.04
Low 1.01 0.94 1.08
Mid 1.01 0.95 1.08
High 1.03 0.97 1.10
Center volume (Largest Q5) Smallest 1.05 0.97 1.14
Small 1.10 1.03 1.18
Mid 0.98 0.92 1.05
Large 1.06 1.00 1.11
ECD proportion (Lowest Q1) Low 1.05 0.98 1.12
Mid 1.04 0.98 1.11
High 1.05 0.99 1.13
Highest 1.06 0.99 1.14
Performance ratio (Best Q1) Good 1.01 0.95 1.09
Mid 1.08 1.01 1.15
Bad 1.16 1.08 1.24
Worst 1.20 1.11 1.29
* Model additionally adjusted for recipient race, age, primary diagnosis, PRA level, BMI
level, education level, primary insurance payer, pre-transplant dialysis time, donor age,
expanded criteria donation, HLA-mismatching, cold ischemia time, and recipient and donor
gender.










Table 4-12. Adjusted hazard ratios for post-transplant overall graft loss
Adjusted 95% confidence
Candidate characteristic (reference level) Level
hazard ratio interval
Proportion transplanted (Highest Q5) Lowest 0.94 0.88 1.00
Low 0.97 0.92 1.02
Mid 0.98 0.93 1.03
High 1.02 0.97 1.07
Center volume (Largest Q5) Smallest 0.99 0.93 1.06
Small 1.13 1.07 1.19
Mid 1.05 1.00 1.11
Large 1.10 1.06 1.15
ECD proportion (Lowest Q1) Low 1.05 1.00 1.11
Mid 1.07 1.01 1.12
High 1.08 1.02 1.14
Highest 1.08 1.02 1.15
Performance ratio (Best Q1) Good 1.03 0.97 1.09
Mid 1.09 1.03 1.15
Bad 1.15 1.09 1.22
Worst 1.17 1.10 1.24
* Model additionally adjusted for recipient race, age, primary diagnosis, PRA level, BMI
level, education level, primary insurance payer, pre-transplant dialysis time, donor age,
expanded criteria donation, HLA-mismatching, cold ischemia time, and recipient and donor
gender.

Table 4-13. Adjusted hazard ratios for pre-transplant mortality
Adjusted 95% confidence
Candidate characteristic (reference level) Level
hazard ratio interval
Proportion transplanted (Highest Q5) Lowest 0.99 0.94 1.04
Low 1.01 0.96 1.07
Mid 0.99 0.94 1.05
High 1.05 0.99 1.11
Center volume (Largest Q5) Smallest 1.01 0.96 1.06
Small 1.04 0.99 1.08
Mid 1.01 0.97 1.04
Large 0.96 0.93 0.99
ECD proportion (Lowest Q1) Low 1.01 0.97 1.05
Mid 0.98 0.94 1.02
High 0.97 0.93 1.01
Highest 1.00 0.96 1.04
Performance ratio (Best Q1) Good 1.03 0.99 1.07
Mid 1.07 1.02 1.11
Bad 1.06 1.02 1.10
Worst 1.13 1.08 1.18
*Model additionally adjusted for candidate race, age, primary diagnosis, blood type, PRA
level, BMI level, education level, primary insurance payer, and dialysis at the time of listing.










Table 4-14. Adjusted hazard ratios for mortality for African-American candidates
Adjusted 95% confidence
Candidate characteristic (reference level) Level
hazard ratio interval
Proportion transplanted (Highest Q5) Lowest 1.28 1.18 1.38
Low 1.22 1.12 1.32
Mid 1.14 1.05 1.24
High 1.18 1.09 1.29
Center volume (Largest Q5) Smallest 1.02 0.94 1.11
Small 1.04 0.97 1.11
Mid 1.00 0.95 1.06
Large 0.98 0.94 1.03
ECD proportion (Lowest Q1) Low 1.01 0.95 1.08
Mid 1.01 0.95 1.08
High 1.01 0.95 1.08
Highest 1.09 1.02 1.16
Performance ratio (Best Q1) Good 1.05 0.98 1.13
Mid 1.09 1.02 1.17
Bad 1.10 1.03 1.17
Worst 1.14 1.06 1.22
*Model additionally adjusted for candidate age, primary diagnosis, blood type, PRA level,
BMI level, education level, primary insurance payer, and dialysis at the time of listing.

Table 4-15. Adjusted hazard ratios for mortality for elderly candidates

Candidate characteristic (reference level) Level dut 9 cniee
hazard ratio interval
Proportion transplanted (Highest Q5) Lowest 1.26 1.13 1.40
Low 1.23 1.10 1.36
Mid 1.18 1.06 1.31
High 1.14 1.02 1.26
Center volume (Largest Q5) Smallest 1.01 0.90 1.12
Small 1.05 0.95 1.15
Mid 0.94 0.87 1.03
Large 0.94 0.87 1.01
ECD proportion (Lowest Q1) Low 0.95 0.86 1.05
Mid 0.90 0.82 0.99
High 0.97 0.88 1.07
Highest 0.99 0.91 1.09
Performance ratio (Best Q1) Good 1.02 0.93 1.12
Mid 1.06 0.96 1.16
Bad 1.14 1.04 1.25
Worst 1.23 1.11 1.36
*Model additionally adjusted for candidate race, primary diagnosis, blood type, PRA level,
BMI level, education level, primary insurance payer, and dialysis at the time of listing.










Table 4-16. Adjusted hazard ratios for mortality for obese candidates
Adjusted 95% confidence
Candidate characteristic (reference level) Level
hazard ratio interval
Proportion transplanted (Highest Q5) Lowest 1.33 1.22 1.45
Low 1.23 1.12 1.34
Mid 1.15 1.05 1.26
High 1.14 1.04 1.25
Center volume (Largest Q5) Smallest 0.96 0.88 1.06
Small 0.96 0.89 1.05
Mid 1.02 0.94 1.09
Large 0.95 0.89 1.01
ECD proportion (Lowest Q1) Low 1.02 0.94 1.10
Mid 0.98 0.91 1.06
High 1.01 0.93 1.10
Highest 1.05 0.97 1.14
Performance ratio (Best Q1) Good 0.95 0.88 1.04
Mid 1.02 0.94 1.11
Bad 1.07 0.98 1.16
Worst 1.11 1.01 1.21
* Model additionally adjusted for candidate age, race, primary diagnosis, blood type, PRA
level, education level, primary insurance payer, and dialysis at the time of listing.

Table 4- 17. Candidate life expectancy (in years) after li sting by levels of center character stics
Center characteristics' Q1 Q2 Q3 Q4 Q5
Proportion transplanted at three years 10.0 10.4 11.0 11.0 12.4
Center volume 10.7 10.4 10.8 10.9 10.7
ECD proportion 10.8 10.7 10.8 10.8 10.5
Performance ratio 11.2 11.0 10.6 10.5 10.1
categories assigned by quintile levels of center characteristics: performance ratio
(observed/expected deaths): Q1 = [0-0.56], Q2 = [0.57-0.83], Q3 = [0.84-1.09], Q4 = [1.10-
1.45], Q5 = [1.46-3.80]; the proportion of ECD transplants: Q1 = [0-6.4], Q2 = [6.5-9.9], Q3
= [10.0-13.4], Q4 = [13.5-18.9], Q5 = [19.0-44.0]; the proportion of patients receiving a
transplant by three years: Q1 = [7.0-3 8.5], Q2 = [3 8.6-5 1.6], Q3 = [5 1.7-63.1], Q4= [63.2-
76.7], Q5 = [76.8-96.0]; and deceased donor transplant volume: Q1 = [10.0-18.8], Q2 =
[18.9-26.3], Q3 = [26.4-36.6], Q4 = [36.7-53.7], Q5 = [53.8-195.0].










Table 4-18. Life expectancy after listing at hypothetical center characteristic levels
Proportion ** ECD Standardized Expected
.* Volume ,
transplanted proportion mortality ratio survival
Case #1 Low Low High High 9.3
Case #2 Low Mid Mid Mid 10.1
Case #3 Low High Low Low 10.3
Case #4 Mid Low High High 10.2
Case #5 Mid Mid Mid Mid 11.1
Case #6 Mid High Low Low 11.6
Case #7 High Low High High 11.5
Case #8 High Mid Mid Mid 12.5
Case #9 High High Low Low 13.1
proportion of candidates receiving a deceased donor transplant at three years. ~deceased
donor transplant volume. `proportion of ECD transplants of all deceased donor
transplants. ratio of observed to expected one year patient deaths.


Table 4-19. Life expectancy after listing for African-American candidates by center
characteristic levels


Center characteristics' Q1 Q2 Q3 Q4 Q5
Proportion transplanted at three years 9.5 9.9 10.4 10.1 11.5
Volume 10.0 9.8 10.1 10.2 10.1
ECD proportion 10.3 10.2 10.2 10.2 9.6
Performance ratio 10.6 10.2 10.0 9.9 9.6
categories assigned by quintile levels of center characteristics: performance ratio
(observed/expected deaths): Q1 = [0-0.56], Q2 = [0.57-0.83], Q3 = [0.84-1.09], Q4 =
[1.10-1.45], Q5 = [1.46-3.80]; the proportion of ECD transplants: Q1 = [0-6.4], Q2 =
[6.5-9.9], Q3 = [10.0-13.4], Q4 = [13.5-18.9], Q5 = [19.0-44.0]; the proportion of
patients receiving a transplant by three years: Q1 = [7.0-38.5], Q2 = [38.6-51.6], Q3 =
[51.7-63.1], Q4= [63.2-76.7], Q5 = [76.8-96.0]; and deceased donor transplant volume:
Q1 = [10.0-18.8], Q2 = [18.9-26.3], Q3 = [26.4-36.6], Q4 = [36.7-53.7], Q5 = [53.8-
195.0].










Table 4-20. Life expectancy after listing. for elderly candidates by center characteristic levels
Center characteristics' Q1 Q2 Q3 Q4 Q5
Proportion transplanted at three years 5.6 5.7 5.8 6.0 6.5
Volume 5.7 5.6 6.0 6.0 5.7
ECD proportion 5.7 5.9 6.1 5.8 5.7
Performance ratio 6.1 6.0 5.9 5.6 5.3
Categories assigned by quintile levels of center characteristics: performance ratio
(observed/expected deaths): Q1 = [0-0.56], Q2 = [0.57-0.83], Q3 = [0.84-1.09], Q4 = [1.10-
1.45], Q5 = [1.46-3.80]; the proportion of ECD transplants: Q1 = [0-6.4], Q2 = [6.5-9.9], Q3
= [10.0-13.4], Q4 = [13.5-18.9], Q5 = [19.0-44.0]; the proportion of patients receiving a
transplant by three years: Q1 = [7.0-3 8.5], Q2 = [3 8.6-5 1.6], Q3 = [5 1.7-63.1], Q4= [63.2-
76.7], Q5 = [76.8-96.0]; and deceased donor transplant volume: Q1 = [10.0-18.8], Q2 =
[18.9-26.3], Q3 = [26.4-36.6], Q4 = [36.7-53.7], Q5 = [53.8-195.0].

Table 4-21. Life expectancy after listing for obese candidates by center characteristic levels
Center characteristics' Q1 Q2 Q3 Q4 Q5
Proportion transplanted at three years 8.2 8.7 9.2 9.2 10.2
Volume 9.1 9.1 8.7 9.1 8.8
ECD proportion 9.0 8.9 9.1 8.9 8.7
Performance ratio 9.1 9.4 8.9 8.7 8.4
Categories assigned by quintile levels of center characteristics: performance ratio
(observed/expected deaths): Q1 = [0-0.56], Q2 = [0.57-0.83], Q3 = [0.84-1.09], Q4 = [1.10-
1.45], Q5 = [1.46-3.80]; the proportion of ECD transplants: Q1 = [0-6.4], Q2 = [6.5-9.9], Q3
= [10.0-13.4], Q4 = [13.5-18.9], Q5 = [19.0-44.0]; the proportion of patients receiving a
transplant by three years: Q1 = [7.0-3 8.6], Q2 = [3 8.5-5 1.6], Q3 = [5 1.7-63.1], Q4= [63.2-
76.7], Q5 = [76.8-96.0]; and deceased donor transplant volume: Q1 = [10.0-18.8], Q2 =
[18.9-26.3], Q3 = [26.4-36.6], Q4 = [36.7-53.7], Q5 = [53.8-195.0].















Tranoartoup 10 Year Survival

Q1 (Lowest) 50%

Q2 51%

Q3 54%

Q4 53%

Q5 (Highest) 56%


100





80


S70


S60


50


40


30 Log-Rank p-value < 0.001
0 2 4 6 8 10
Years Post-Listing


Figure 4-1. Kaplan-Meier plot of candidate survival by center rate of transplant



















ECD Proportion 10 Year Survival
Group

Q1 (Lowest) 54%

Q2 53%

Q3 52%

Q4 53%

Q5 (Highest) 50%


100


90


80


70


Log-Rank p-value < 0.001


6

Years Post-Listing


Figure 4-2. Kaplan-Meier plot of candidate survival by center proportion of ECD transplants
















Cent olume 10 Year Survival

Q1 (Lowest) 52%

Q2 52%

Q3 53%

Q4 52%

Q5 (Highest) 53%


30iLog-Rank p-value = 0.73
30


Figure 4-3. Kaplan-Meier plot of candidate survival by center volume


6

Years Post-Listing
















Performance Ratio 10 Year Survival

Q1 (Best) 54%

Q2 54%

Q3 52%

Q4 51%

Q5 (VVorst) 51%


Log-Rank p-value < 0.001

2 4


Figure 4-4. Kaplan-Meier plot of candidate survival by center performance ratio


6
Years Post-Listing









CHAPTER 5
DISCUSSION

ESRD is a pervasive and growing public health concern in the United States. Both the

cumulative number of ESRD patients and proportion of individuals with risk factors for future

ESRD development have increased significantly over the recent era. The past two decades have

also witnessed the growth and acceptance of kidney transplantation as the most effective

treatment modality for patients suffering from ESRD. Correspondingly, the number of patients

selecting transplantation for treatment ofESRD has also significantly increased. The rapid rise

in the number of candidates has increased waiting times to acquire a transplant and increased

mortality among patients on the waiting list for transplantation. In response, there have been

significant efforts on the part of the transplant community to identify efficiencies in the

transplant process, to increase donation rates, and to develop strategies to prolong the lifespan of

donations while simultaneously maintaining equitable access to patients.

The explanatory variables of primary interest in this study, the four center characteristics

(waiting time, performance level, patient volume, and donor risk level), have generally been

shown to be associated with outcomes for transplant recipients. However, the significance and

magnitude of these effects for a prospective transplant candidate has not been specifically

evaluated. Furthermore, the integration of these factors to assess the relative effects provides a

basis by which patients and caregivers can compare potential centers of listing relative to the

impact on candidate prognoses. In fact, the selection of a transplant center is a common

situation, but in many cases, these decisions may be made by default. Patients that are in this

circumstance may often simply follow the advice of primary physicians who, for geographic

purposes or personal preferences, refer patients indiscriminately to a particular transplant center.

Thus, in the predominant number of cases, there is little active decision making on the part of the









transplant candidate or referring physician as to which center is most appropriate for the patient's

needs or if alternative centers may in fact provide a different prognosis for that patient. This

dissertation attempted to provide data to elucidate the impact of center selection that could

potentially inform these decisions. That is, the questions remain, is there a marked difference for

candidate outcomes (e.g., mortality or graft survival) based on the choice of center that they list

for a transplant? Secondly, if there is a difference, what characteristics of the centers are most

important to candidate prognoses? Finally, are these characteristics generalizable to the entire

candidate population or are they relatively more important to select groups that have various

prognoses by treatment modality (i.e., dialysis or transplant)?

The main findings of this study indicate that center characteristics are indeed significantly

associated with patient outcomes. Furthermore, the study demonstrates that these center

characteristics vary in their relative effect and that these effects somewhat differ in high-risk

subsets of the candidate population. The most direct implication of the study is that candidates

and referring physicians have an incentive to incorporate characteristics of transplant centers into

decisions to select centers of care. Further research related to these findings and healthcare

implications of the study results will be discussed in the proceeding section.

For the purpose of simulating the circumstance that new onset transplant candidates must

face in selecting a center, the study was specifically designed to utilize past levels of center

characteristics rather than levels that were associated with the center at the time of

transplantation. Clearly, candidates will not be aware of the characteristics of centers at the time

that a donor organ is offered and must base decisions on levels prior to the time of listing. This

study utilized the aggregated center levels from three years prior to the time of candidate listing

as a representation of these characteristics. In fact, the analysis demonstrated a significant









association of historical levels of center characteristics with levels of these same characteristics

at the time of transplantation. This is important information for prospective candidates to assess

the reliability of center characteristics to assist with their selection of a center. The results

indicate center factors in this study are relatively stable and can be utilized on a prospective basis

for comparing centers. An important caveat to this finding, however, is that, as demonstrated in

the study, waiting times for candidates to reach transplantation are continuing to expand. Thus,

this greater time period is associated with a greater opportunity for center characteristics to alter.

Based on this study period, center volume has been relatively stable over time, waiting times

have increased, and the utilization of ECD donations have increased. It is also worthy of note

that center performance ratios are fairly consistent over time. That is, performance ratios at a

center at the time of listing are similar to the performance ratios at the time of transplant on

average. The performance ratio is presented as a measure of quality of care, and it might be

expected that, due to random events or changes in practice, this value might significantly vary

over time. Results of the analysis suggest that listing at a center with high performance is

generally accompanied by a high performance level at the time of transplantation.

The associations between center characteristics and candidate outcomes found in this

study were relatively consistent with the literature investigating the impact on transplant

recipients. In particular, longer waiting times, better performance ratings, and a lower proportion

of ECD transplants all translated to superior outcomes for transplant candidates from the time of

listing. However, the degree of these associations and the relative importance of these

characteristics are of particular importance in this study. The historical proportion of candidates

that reach transplantation at three years demonstrated a highly significant and dose-response

relationship with outcomes for candidates. Candidates listed at centers with the slowest rate of









transplantation had a 32% elevated hazard for death over the study period relative to candidates

listed at centers with the most rapid rate of transplantation. This finding is perhaps not

surprising, as there are two clear explanations for this association. One is that transplantation has

been demonstrated to convey a significant survival benefit over the alternative treatment

modality of maintenance dialysis which is typically initiated prior to candidate listing.

Therefore, a portion of this effect is likely attributable to the fact that candidates at centers with

longer waiting times accumulate relatively higher cases of mortality prior to reaching

transplantation. In addition, the cumulative effects of dialysis have a significant impact on post-

transplantation mortality. Therefore, those candidates that reach transplantation with longer

waiting times and exposure to dialysis also have poorer outcomes. Cumulatively, findings from

this dissertation demonstrate that for an average candidate, there is a significant survival benefit

that can be attributed to listing at a center with reduced expected time to transplantation.

The relative hazard associated with increased waiting time adjusted for potential

confounding factors is important in demonstrating the independent effect of this characteristic on

candidate mortality. However, the interpretation of this relative risk is not always

straightforward to clinicians or patients. In order to relate the Eindings of this study in terms that

could be useful to a broader audience, the study also estimated the median survival of candidates

from the time of listing. Results of the analysis demonstrate that the average candidate has

approximately two and a half years longer expected survival (12.4 years versus 10.0 years) by

listing at a center with a rapid versus delayed rate of transplantation. Furthermore, as the center

groupings represent quintiles, there are approximately 20% of centers that have the reduced

waiting times, suggesting that they are not uncommon situations. However, there is also likely a

regional component to waiting times, and within certain regions, centers with reduced waiting









times may be scarcer; therefore, future studies may be useful to elucidate the regional availability

of centers with rapid rates of transplant.

Research suggests that centers with the highest transplant volume have superior patient

outcomes. This has been interpreted as related to quality of care associated with greater

experience with the surgical procedure, greater infrastructure and coordination of services to

transplant patients, and availability of ancillary services. In contrast to these research accounts,

center volume was not significantly associated with candidate survival in this study. This

difference may partially reflect the relative unimportance of center volume as compared to other

factors examined in the study as well as the reduced impact of volume for transplant candidates

as compared to recipients. Many candidates for transplantation do not survive to the time at

which a transplant may be offered and others may become unviable for transplant due to health

deterioration. As such, the impact of center volume, which has shown some association between

the largest centers and recipient outcomes, may be diluted. It is also possible that smaller centers

have a greater opportunity to follow patients prior to transplantation and, despite the lack of

facilities, the long-term advantages of larger centers are balanced by increased follow-up care.

Past reports suggest that the effect of volume on recipient outcomes in not linear, but is generally

found to be superior in the top volume centers only. However, this study excluded candidates

listed at very small centers (<10 deceased donor transplants per year) in order to obtain stable

estimates of pre-listing characteristics; in this regard, the impact of volume may also only be

highlighted at extreme levels and less evident among centers with slightly larger volume in this

study .

The quality of deceased donations are highly variable, translating to almost three-fold

hazard ratio for graft loss for a recipient with the lowest quality donors relative to an ideal class









of donations (49,50). As such, there is a clear incentive for transplant candidates to acquire the

highest quality donation available, holding all other factors equal. There are also significant

regional variations in the proportion of high-risk donations. The expanded criteria policy was

initiated in 2002, mandating that candidates prospectively consent as to whether they would be

willing to receive an ECD transplant. Therefore, candidates have some control over whether

they will receive a lower-quality kidney. The potential benefit of listing for an ECD kidney is

that candidates may receive their transplant more rapidly than having to wait for a SCD. In fact,

research indicates that this may be an advisable strategy for some candidates, at least at centers

that selectively list patients (53,99). However, regardless of the listing strategies, candidates that

list at a center with a lower proportion of ECD kidneys should theoretically have an advantage

over candidates that are listed at centers with a high proportion of ECDs. For candidates that are

unwilling to accept an ECD, a lower proportion of ECDs should expedite their acceleration on

the waiting list. In contrast, candidates that are willing to accept ECDs should still have a greater

opportunity to receive a SCD at a center with a lower proportion of ECDs. Moreover, centers

with higher rates of ECD transplants may also have a higher-risk donor pool of organs within

risk classes, though this has not been demonstrated explicitly. This study indicated that

candidates have a significantly elevated mortality risk (AHR = 1.04, p<0.05) by listing at centers

with the highest proportion of ECDs as compared to listing at centers with the lowest proportion

of ECDs. This increased hazard translated to approximately four months of reduced expected

survival among candidates listing at centers with the highest proportion of ECD transplants as

compared to centers with the lowest proportion of ECDs. However, the study also demonstrated

that this effect was significantly higher for post-transplant survival and, in particular, when

eliminating characteristics of donors from the model. Therefore, the study suggests that









candidates that list at centers with lower quality donations are more likely to receive a lower

quality organ for those who survive to the time of the procedure. Furthermore, outcomes for

recipients are significantly reduced at centers with a greater utilization of higher-risk organs.

The study also suggests that there is a correlation between centers that use a greater proportion of

higher-risk organs for centers that have longer waiting times. In other words, centers that have

longer waiting times may be more likely to accept lower-quality organs in order to ameliorate the

candidate volume. From the candidate perspective, waiting time remains the most critical center

factor associated with survival; however, for centers with similar waiting times, the quality of

donor organs may remain a modifier of center selection. In addition, this study did not

specifically examine this effect in patients that actually list for or receive lower-quality organs.

Follow-up studies addressing center selection specifically for candidates willing to accept lower-

quality organs (or for those not willing to accept lower-quality organs) may also elucidate

important center factors that are pertinent to the candidate population.

Performance evaluations are conducted and published by the SRTR and readily

accessible to the public through written reports and the internet. Evaluations are constructed

based on standard mortality ratios which calculate the observed number of events (graft losses or

deaths) at a transplant center relative to what would be expected given the characteristics of the

recipient population over a fixed interval of time. Theoretically, this ratio is indicative of quality

of care and center performance for their recipient population. In fact, the study indicated that

candidates listed at centers with the best historical performance have significantly better

outcomes as compared to candidates listed at centers with the worst performance ratings

(AHR=1.14, p< 0.01). This elevated risk translated to an approximate one year of increased

survival for the average transplant candidate between centers with the best and worst ratings.









However, an important consideration associated with this finding is that a portion of this effect

was observed prior to transplantation. This may imply that centers with higher ratings have

better pre-transplant care, but also may be suggestive of increased patient selection criteria. This

notion will be explored in more detail later in the discussion.

One of the main goals of this study, beyond evaluating the significance of individual

center characteristics, was to ascertain the relative importance of these factors. The general

conclusion that is evident from our study is that the waiting time at a transplant center is the most

important modifier of decisions to select a center among the characteristics examined. The

hazard ratios and estimated survival years from the time of listing for candidates suggest that

while other factors may incrementally impact candidate outcomes, waiting time clearly has the

strongest effect on patient mortality. Even in the presence of a combination of ideal

characteristics of other factors, differences in expected waiting time remain the predominant

determinant of candidate outcomes. In this regard, the average candidate and their caregivers

may have a strong incentive to assess the expected waiting times at centers in the decision-

making process for the selection of a transplant center. The additional knowledge of obj ective

information about the center and characteristics which may influence their outcomes should be

available and disseminated to patients in order to help them make informed decisions. In many

instances, patients simply may rely on the advice of a referring physician and have no significant

participatory role in this decision which may have life-altering implications. Results of this

study and subsequent research deriving from this paradigm may be used to inform patients and

their caregivers about the ramifications of these important decisions and the impact on their

survival. In addition, strategies to disseminate this information to patients and caregivers in an

interpretable fashion are clearly needed in future efforts deriving from this study.









There are certainly additional factors that may affect an individual candidate's selection

of a center that were not incorporated in this analysis, including the geographic location of the

center, the prestige or personal familiarity with the center and personnel, and the comfort level

with the caregivers at the center itself. These factors may be very important to candidates, but

should be considered relative to the impact of other characteristics evaluated in this study.

Moreover, centers clearly cannot be defined only by the characteristics outlined in this study.

The specific factors investigated in this study were based on past research and the degree to

which information was readily available characterizing centers. However, each center is unique

in its physical structure and design, environmental conditions, and medical personnel expertise

and experience. Centers also vary in their academic affiliation, the degree to which they provide

services as part of a safety net program, the coordination and relationship of surgical and medical

departments, physical capacity, the availability of dialysis services, and connection with other

dialysis centers as well as innumerable medical protocols. The degree to which these other

factors may impact candidate outcomes is unknown and this uncertainty must be taken into

consideration. Specifically, this study demonstrates the impact of factors in a typical situation

for an average candidate; however, a broad host of factors may certainly modify these estimates

or be particular pertinent to an individual candidate. Therefore the results of the study must be

interpreted and utilized with these caveats in mind and not be viewed as an omnipotent guide to

center selection that cannot be superseded by other conditions.

Another important aspect of this study was to assess the degree to which center

characteristics were significant and relatively important to all candidates or whether these

characteristics had a particular impact in certain subsets of the candidate population. To this end,

the study examined the association of center characteristics specifically in three "high-risk"









subsets of patients: African-Americans, the elderly, and obese candidates. In general, results

indicated that center characteristics remain important to these high-risk subsets of the population,

particularly the center rate of transplantation. One of the potential implications of this portion of

the analysis was to identify factors that may have differential effect in certain subgroups that

may suggest efficiencies in the transplant process. That is, in the case that specific factors could

be shown to be important to certain groups, but not others, it would be possible that candidates

could simultaneously benefit based on listing at centers with the presence of factors that are

suited to their needs. In contrast, if center factors were relatively equally important to all

subgroups, the analysis may suggest the preferred center characteristics for candidates, but

candidates could obviously not all seek to list at these centers simultaneously. Broadly, the latter

case is suggested by the analysis. That is, while there are factors that appear to have a somewhat

differential impact in the subgroups examined in the study, the main findings are applicable to all

groups. This may imply that results of the study can still be applicable to candidate populations

that have the ability to seek out centers with the preferred characteristics, but cannot be

generalized to the entire candidate population. As the primary characteristic influencing

candidate survival is the rate of transplantation associated with the transplant center, primarily an

increase in donation rates would be needed to significantly reduce waiting times across centers.

The categorization of African-Americans as a "high-risk" subset of the population is

generally a reflection of an increased rate of graft loss for these patients following

transplantation. This elevated risk has been attributed to immunological factors as well as

socioeconomic characteristics of this population (58). However, in contrast, African-Americans

have superior survival on dialysis relative to Caucasians and equivalent patient survival

following transplantation (53,100). Therefore, given these potentially differential effects of









transplant-related factors on African-Americans, the hypothesis that selection of centers may also

be unique to this portion of the candidate population was justified. Results of the analysis

indicate that, as for the general candidate population, listing at a center with the most rapid rate

of transplantation is the most important factor among the variables in the study for African-

American candidates. The 28% elevated increased hazard for candidates that list at centers with

the longest expected waiting times as compared to candidates that list at centers with the shortest

waiting times was the most substantial center characteristic in the analysis. On average, African-

Americans have longer waiting times than their Caucasian counterparts due to the distribution of

HLA antigens in the donor population, which has been a significant component of the organ

allocation algorithm. African-Americans have a smaller likelihood of receiving additional points

that are ascribed to HLA-matching with a potential donor kidney as part of the allocation of

deceased donor organs (101,102).

Even though average waiting times are longer for African-Americans, as this analysis

demonstrates, it is still highly beneficial for African-American candidates to list at centers with

reduced expected waiting times. In fact, the study indicates that expected survival is two years

longer for candidates that list at centers with the shortest time to transplant (11.5 years) as

compared to candidates that list at centers with the longest expected waiting times (9.5 years).

Another interesting result of the analysis is that African-Americans have a significantly increased

survival associated with centers with a low proportion of ECD transplants. The 9% increased

hazard for mortality after listing associated with African-American candidates that list at centers

with the highest ECD proportion was significantly higher than the general population. This is

consistent with the literature that African-Americans may derive a greater benefit by receiving a

higher quality donation even at the expense of additional dialysis exposure as compared to









Caucasian candidates, despite the fact that historically African-Americans are more likely to

receive an ECD transplant (51,53). Candidates that listed at centers with the lowest proportion

of ECD transplants had 10.3 years of average life expectancy after listing as compared to 9.6

years among African-Americans that listed at centers with highest proportion of ECD

transplants. The effect of center volume was not statistically significant in this portion of the

candidate population. Center performance ratios also had a similar estimated impact for African-

American candidates as compared to the general population.

Elderly patients comprise a significant and growing portion of the dialysis population.

Elderly patients comprised 8% of the candidate population over the study period, but have a

three-fold risk for mortality following listing relative to the youngest candidate age group.

Elderly candidates are at significantly higher risk for death while on dialysis as well as after

transplantation. However, elderly candidates still receive a significant benefit from

transplantation with an almost doubling of life expectancy relative to remaining on dialysis

therapy (27). In addition, there is evidence that elderly candidates have heightened incentive to

receive transplants more rapidly, even at the expense of receiving a lower quality donation (53).

The analysis indicates that, as with the general population, the most important center

characteristic for elderly candidates is the rate of transplantation. Elderly candidates listed at

centers with the lowest proportion of candidates reaching transplantation had a 26% increased

hazard for mortality as compared to the candidates listed at centers with the longest expected

waiting times. This translated to a difference of approximately one year in expected survival for

the average elderly candidate. The magnitude of the hazard associated with longer waiting times

was reduced as compared to the general population and may reflect that the impact of transplant

centers is less important to elderly candidates due to higher death rates on dialysis due to other









clinical factors. That is, as a significantly higher proportion of elderly candidates die prior to

reaching transplantation, the accrued benefit from transplantation is observed in a smaller portion

of the population.

As opposed to the general population, for elderly candidates, there was no significant

association between the volume and the ECD proportion at the listing center. As with the

reduced effect of the center rate of transplantation, this may reflect the reduced importance of

center characteristics as compared to the competing risks of other clinically based factors that are

particularly relevant for elderly ESRD patients. However, the non-significant finding associated

with ECD proportions may also demonstrate that elderly candidates that list at these centers may

receive some additional benefit by receiving these organs more rapidly. One notable difference

in the elderly population was the increased association of center performance ratings with

candidate outcomes. The analysis indicates that elderly candidates have a significant difference

in mortality when listing at centers with the highest performance ratios, a 23% increased hazard

for death as compared to listing at centers with the lowest performance ratios. This effect

translated to an estimated one additional year of life expectancy between centers of extreme

performance ratings (6.1 years versus 5.3 years). It is possible that, due to the unique care needs

and relative fragility of elderly candidates, the quality of care between centers is most evident in

these patients. In fact, past reports investigating the nature of "center effects" suggest that

differences between centers are primarily reflective of outcomes in higher-risk populations. As

such, as opposed to younger candidates in whom lower quality of care may result in greater

complication rates or reduced quality of life, the impact on mortality in relatively healthier

subsets may still be marginal. However, another explanation for these findings may relate to

potential selection bias among elderly candidates. The effect of selection bias is a greater risk in









elderly patients for whom many underlying health conditions and co-morbidities that are not

represented in the database can influence outcomes.

As outlined in the methodology portion of the study, there is some potential for selection

bias as a contributing factor to the results. While there are universal contraindications for criteria

for transplant candidacy, some centers may invoke additional screening processes which

contribute to differences in mortality (103,104). In the context of this study, this may result in

bias in results in the case that centers that list a more selective cohort of patients also were more

highly represented by certain center characteristics in the analysis. For instance, in this case if

centers which traditionally had higher center performance ratios also were more selective in

candidate screening, this association may bias certain results of the study. Specifically, the

disproportionate representation of non-codified factors may contribute to differences in outcomes

and inappropriately elevate the effect of performance criteria. As described in the methodology,

there were two fundamental strategies to obviate this potential confounding effect. One strategy

was to adjust for factors that may account for patients that are selectively listed. These include

age, race, patient primary insurance, and education level. In fact, these factors were adjusted for

in the analysis and had statistically significant associations with candidate mortality. The

additional strategy was to partition the follow-up period to pre- and post-transplant survival. The

purpose of this partitioned analysis is that transplant centers typically have significantly less

interaction with candidates prior to transplantation and would be less likely to have an impact on

survival prior to transplantation related to quality of care. In contrast, centers may have a much

stronger association with candidates during the transplant procedure and post-transplant care.

Therefore, differences that occur prior to transplantation are more likely related to factors other

than center quality of care and may reflect a certain degree of patient selection.









The results of the model for pre-transplant mortality indicate a significant association of

standard mortality ratios with death prior to transplantation, while none of the other center

factors examined had a significant association. Candidates listed at centers with the lowest

performance ratios had a 13% elevated mortality prior to transplantation as compared to

candidates listed at centers with the best mortality ratios. In addition, this association was

particularly notable in elderly candidates. This finding must be interpreted carefully and, in

particular, may be indicative of selection bias of "better patients" to centers with higher

performance ratings. In practice, transplant centers are not the primary caregivers of patients

prior to transplantation. In contrast, centers are responsible for listing candidates and

determining whether they meet certain medical criteria, but are not typically responsible for

patient care in the interim between candidacy and the transplant procedure. One of the main

purposes for this sub-analysis was to determine the potential influence of certain selection biases

that may occur and influence the main outcome of patient survival after candidacy. This portion

of the model most likely indicates that centers with better performance ratings either list healthier

candidates, leading to lower pre-transplant mortality, or are located in areas with lower morbidity

among the dialysis population, or there are other environmental factors which are significantly

variable. The fact that this association was stronger among elderly patients, in which selective

listing of patients may be significantly stronger, further suggests the influence of patient

selection. Although center performance was not the most significant factor for candidates in the

primary outcome model, the degree to which an association did exist may in part reflect this

patient selection rather than improved quality of care by the transplant center.

The other important implication of this finding is that models evaluating center

performance may not fully account for underlying health conditions and exogenous factors that










significantly influence mortality but are not associated with quality of care. That is, as candidate

survival rates are positively correlated with transplant outcomes even after adjustment for other

risk factors, this may imply that other non-codified factors are associated with outcomes that are

not accounted for with risk adjustment. Future investigation into the degree to which center

performance is impacted by candidate mortality rates independent of other transplant-related

factors is important to distinguish these effects. The interpretation of center standard mortality

ratios is generally that they are indicative of quality of care. In fact, insurance companies and

government oversight committees may further use these criteria for business purposes and to

investigate centers for inappropriate care practices (22). In contrast, if performance ratios are

merely indicative of selective listing practices, than these additional interpretations relating

quality of care to performance are unwarranted. Furthermore, if this association is not justified

then there are potential dangers that centers will limit access to patients in an effort to improve

performance ratings, but at the same time certain patients may not acquire the benefit of

transplantation as a result. Given the results in this study, performance ratios only had a mild

association with candidate mortality, but even this limited effect should be interpreted with

caution given these potential caveats.

Obese patients represent a unique subset of the transplant candidate pool for several

reasons. The population of obese candidates and transplant recipients is significantly growing

consistent with the growth rates among the general ESRD population (100). Moreover, obese

patients have a unique "paradoxical" survival advantage on maintenance dialysis. However,

obese patients also have significantly elevated risk for post-transplant graft loss and mortality

(105). This study indicates that despite this relatively reduced mortality on dialysis prior to

transplantation and increased mortality following transplantation, the most important center









characteristic of obese candidates remains the rate of transplantation. Obese candidates

significantly benefit from transplantation and as compared to other center characteristics have the

most significantly increased survival expectancy associated with listing at centers with reduced

waiting times as compared to other factors of center volume, performance, or ECD proportion.

Center volume was not associated with differential outcomes for obese patients. Obese

candidates that listed at centers with the highest proportion of ECD transplants had an elevated

risk for death at a similar level to the overall candidate population. Candidates that listed at

centers with the lowest performance ratio also had an elevated risk for mortality. Similar to

elderly patients, the issue of whether this effect is due to transplant center quality of care versus

patient selection is unknown; however, among high-risk patients the opportunity for this

selection bias is likely stronger. On the other hand, centers treating obese candidates and

recipients may require greater quality of care due to the increased potential for surgical

complications and morbidity associated with obese transplant recipients. The most important

results is that expected candidate survival for the average obese candidate ranges from 8.2 years

tol0.2 years for listing at centers with the longest and shortest waiting times respectively.

Consistent with the general candidate population, results of the study indicates that obese

candidates have a significant survival advantage by listing at centers with a rapid rate of

transplantation.

In general, results of the study for higher-risk patients were consistent as for the general

candidate population. As discussed previously, this indicates that broad efficiencies in the

transplantation process are not readily indicated by the study in which certain portions of the

population may be best served by particular centers. In contrast, the study suggests that all

patients have an incentive to list at centers with reduced waiting times above all other center









characteristics. Given this information, the question remains as to which patients will utilize this

information. Intuitively, the answer will likely be those patients who are best informed or

educated or have the means to travel to centers with an ideal set of characteristics. Alternatively,

if the information were well known among the entire candidate population, the net effect would

be that waiting times would eventually be relatively equivalent. That is, the effect of candidates

seeking centers with reduced waiting times would eventually balance these differences. Whether

or not this is realistic is not clear; however, in terms of equity and eliminating geographic

disparities in access to transplantation, these objectives may be ideal. Particularly given the fact

that more vulnerable candidate groups are less likely to benefit from the results of the study,

future efforts to educate these portions of the population and provide access to centers with ideal

characteristics are warranted.

In the framework of Grossman' s health production function, the study suggests that patient

and caregiver behavior, specifically the selection of a center, has a significant impact with the

subsequent deterioration rate of transplant candidate health. As described in Grossman' s model,

patients arrive at a given period (in this case, at the time of candidacy) with a given stock of

health. From that point, decisions to seek health care and the "investment" in future stock of

health are related to their prognosis and subsequent utility. Beyond simply seeking medical care

as outlined as an important factor in Grossman's paradigm, this study suggests that the specific

center at which care is sought is a significant modifier of patient outcomes. In addition,

consistent with the health production function, certain individual characteristics alter the effects

of future health. Whether patients' decision-making processes are related to their own demand

for utility or current stock of health is not fully addressed in this study. However, the study does

suggest that patients that are more informed, motivated, or able to seek care outside of their










immediate region have an improved prognosis through selection of a medical center. An

individual's investment in health through the selection of a center has a profound effect on

transplant candidate survival consistent with the health production function paradigm.

There have been significant efforts from private and public agencies, individual patients,

and patient advocates to provide transparent healthcare information to the population. One of the

significant challenges in these efforts is to disseminate information in a coherent and useful

manner to patients. Certainly, in recent years patients are more likely to interact with healthcare

professionals with a greater arsenal of information obtained through research, personal

communications, or web-based services. The degree to which this information increases the

likelihood to seek appropriate care or facilitates the healthcare interaction is not fully known and

is likely context dependent. However, as the trend towards more informed patients is not likely

to decline in future years, a maj or focus of researchers and caregivers is to disseminate the most

meaningful and interpretable data to consumers of this information. An additional challenge

related to this trend is to dispel increased levels of misinformation that may also be obtained

from the same sources.

Transplantation is somewhat unique to other fields of medicine due to the volume and

granularity of information that exists related to mandatory data collection. In this sense, this

field has a particular opportunity to provide evidence-based recommendations to patients and

their caregivers. However, transplantation is not unique to other medical contexts in that there is

a significantly competitive market, and advertisement of a center may entail numerous

unsubstantiated or at least subjective components. Future work detailing strategies by which to

disseminate information to patients will be a necessary corollary to this study. Furthermore,










given that transplantation is a complex science, relaying the most pertinent information to

concerned parties is a significant obstacle.

Acknowledgment

The data reported here have been supplied by the University Renal Research and Education

Association (URREA) as the contractor for the Scientific Registry of Transplant Recipients

(SRTR). The interpretation and reporting of these data are the responsibility of the author and in

no way should be seen as an official policy of or interpretation by the SRTR or the U.S.

Government.









CHAPTER 6
CONCLUSIONS AND FUTURE WORK

The primary findings of this research suggest that transplant centers are associated with

significantly variable survival for candidates of renal transplantation. This information is

important for prospective candidates and caregivers to facilitate decisions in their selection of a

center of care. There are multiple factors which may ultimately determine the specific center at

which a candidate receives care; however, obj ective information which may have life-altering

ramifications is critical to disseminate to affected individuals. Results of the study indicate that

multiple factors are important, the magnitude of the effects vary significantly, and the most

critical factor for a candidate among the center characteristics examined is the expected duration

of the waiting list. Candidates have a significantly longer expected life span by listing at a center

with a reduced waiting time independent of other center factors. In addition, this finding is valid

across the three high-risk subgroups examined with slight variations in the magnitude of the

effect. For centers with similar expected waiting times, other factors examined in the study also

modify expected survival for prospective candidates.

The results of this study can be summarized to inform transplant candidates that it is

useful to "shop around" for transplant centers, and a key characteristic in these comparisons

should be the expected waiting time to transplant. However, one of the realities in this

population is that not all patients are capable of listing at centers across the country either for

logistical, financial, or insurance constraints. Therefore, those patients that are more affluent,

educated, or generally have fewer barriers to mobilizing are the candidates that can benefit from

this information. Alternatively, patients from lower socioeconomic status or with more

complicated health conditions may not be able to identify centers in a more narrow region with

significantly reduced waiting times. The degree to which centers are available with more









desirable characteristics at a national and regional basis across the country will be an important

investigation for future work. In this regard, one of the potential follow-up projects of this study

is to construct an application that incorporates these results and allows candidates to identify

centers. This could be generated with the capability to incorporate given constraints (e.g., in a

certain region) and allow transparent information regarding the availability of centers and the

associated impact on candidates' prognoses. This could be a critical step in the dissemination of

results to a broad audience rather than through traditional scientific publications by which

information may reach only selected patients and caregivers.

Another interesting and important result of this study is the identification of potential

selection bias associated with center performance ratings. These ratings have important

implications to transplant centers related to government oversight, contracting with insurance

agencies, and advertising to potential candidates and referring physicians as a marker of good

quality of care. However, if these ratings are related to pre-transplant candidate selection (as

suggested by this study) then further investigation of the utility of performance evaluations in the

field of transplantation is warranted. Additional study investigating the association of other

center factors, such as patient proximity to the center, center experience with specific patient

groups, and in relation to the quality of care of dialysis centers associated with a transplant center

may complement the findings of this study. In addition, replication of the analysis in several

other high-risk groups, such as diabetic patients, patients with a history of cardiovascular disease,

or pediatric populations, may also generate important information for subgroups of the growing

ESRD population. Moreover, results may be applicable to other forms of organ transplantation

and more broadly to other forms of health care in which provider selection could have an

important impact on patient prognosis.









The immediate follow-up work for this study will entail disseminating results through

peer-reviewed manuscripts and discussion of these topics in scientific meetings. Beyond

extending the results to other populations and medical contexts, strategic efforts to communicate

findings directly with patients are an important endeavor. This is particularly salient in this

context in which centers may or may not have a vested interest to transmit this information to

patients. Ultimately, the results of the study will be most useful when there is shared information

among patients and caregivers, and it is utilized j ointly to make potentially life-altering decisions

in selection of a transplant center









LIST OF REFERENCES


1. Prevalence of Reported ESRD. Online. Internet. Available:
http://www.usrds. org/2006/reflB_prevalence_06 .pdf. 2006 Annual Data Report (accessed
20 Sep 2006).

2. Schrag D, Cramer LD, Bach PB, Cohen AM, Warren JL, Begg CB. Influence of hospital
procedure volume on outcomes following surgery for colon cancer. JAMA 2000:284:
3028-3035.

3. Bianco FJ, Jr., Riedel ER, Begg CB, Kattan MW, Scardino PT. Variations among high
volume surgeons in the rate of complications after radical prostatectomy: Further
evidence that technique matters. J Urol 2005:173: 2099-2103.

4. Krumholz HM, Radford MJ, Wang Y, Chen J, Heiat A, Marciniak TA. National use and
effectiveness of beta-blockers for the treatment of elderly patients after acute myocardial
infarction: National Cooperative Cardiovascular Project. JAMA 1998:280: 623-629.

5. Harmon JW, Tang DG, Gordon TA et al. Hospital volume can serve as a surrogate for
surgeon volume for achieving excellent outcomes in colorectal resection. Ann Surg
1999:230: 404-411.

6. Hosenpud JD, Breen TJ, Edwards EB, Daily OP, Hunsicker LG. The effect of transplant
center volume on cardiac transplant outcome. A report of the United Network for Organ
Sharing Scientific Registry. JAMA 1994:271: 1844-1849.

7. Stukel TA, Lucas FL, Wennberg DE. Long-term outcomes of regional variations in
intensity of invasive vs medical management of Medicare patients with acute myocardial
infarction. Jama-Journal of the American Medical Association 2005:293: 1329-1337.

8. Baicker K, Chandra A. Medicare spending, the physician workforce, and beneficiaries'
quality of care. Health Affairs 2004:23: W4184-W4197.

9. Institute of Medicine. To Err is Human: Building a Safer Health System. 2000.
Washington, D.C., National Academy Press.

10. Baker DW, Einstadter D, Thomas CL, Husak SS, Gordon NH, Cebul RD. Mortality
trends during a program that publicly reported hospital performance. Medical Care
2002:40: 879-890.

11. Chassin MR. Achieving and sustaining improved quality: Lessons from New York State
and cardiac surgery. Health Affairs 2002:21: 40-51.

12. Mukamel DB, Weimer DL, Zwanziger J, Gorthy SFH, Mushlin AI. Quality report cards,
selection of cardiac surgeons, and racial disparities: A study of the publication of the










New York State Cardiac Surgery Reports. Inquiry--the Journal of Health Care
Organization Provision and Financing 2004:41: 43 5-446.

13. Epstein AM. Rolling down the runway-The challenges ahead for quality report cards.
Jama-Joumal of the American Medical Association 1998:279: 1691-1696.

14. Schauffler HH, Mordaysky JK. Consumer reports in health care: Do they make a
difference? Annual Review of Public Health 2001:22: 69-89.

15. Schneider E, Epstein A. Public performance reports for cardiac surgery--Reply. Jama-
Joumnal of the American Medical Association 1999:281: 135.

16. Burdick JF, Williams GM. What causes center effects in kidney transplantation. Ann
Surg 1986:203: 311-314.

17. Mickey MR. Center effect. Clin Transpl 1986: 165-173.

18. Benlahrache C, Cecka M, Mickey MR, Cicciarelli J. The center effect. Clin Transpl
1987: 325-337.

19. Ogura K, Cecka JM. Center effects in renal transplantation. Clin Transpl 1991: 245-256.

20. 2005 OPTN/SRTR Annual Report .
http://www.u stran spl ant. org/annual~rep orts/current/20O9_don-
race_dc.htm?o=2&g=5&c=1 (accessed 15 Sep 2006)

21. Howard DH, Kaplan B. Do report cards influence hospital choice? The case of kidney
transplantation. Inquiry-the Journal of Health Care Organization Provision and Financing
2006:43: 150-159.

22. Schold JD, Howard RJ. Prediction models assessing transplant center performance: Can a
little knowledge be a dangerous thing? American Joumnal of Transplantation 2006:6:
245-246.

23. Lin HM, Kauffman HM, McBride MA et al. Center-specific graft and patient survival
rates: 1997 United Network for Organ Sharing (UNOS) report. JAMA 1998:280: 1153-
1160.

24. Terasaki PI, Cecka JM. The center effect: Is bigger better? Clin Transpl 1999: 317-324.

25. Gj ertson DW. Center and other factor effects in recipients of living-donor kidney
transplants. Clin Transpl 2001: 209-221.

26. Axelrod DA, Guidinger MK, McCullough KP, Leichtman AB, Punch JD, Merion RM.
Association of center volume with outcome after liver and kidney transplantation. Am J
Transplant 2004:4: 920-927.










27. Wolfe RA, Ashby VB, Milford EL et al. Comparison of mortality in all patients on
dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric
transplant. New England Joumnal of Medicine 1999:341: 1725-1730.

28. McDonald SP, Russ GR. Survival of recipients of cadaveric kidney transplants compared
with those receiving dialysis treatment in Australia and New Zealand, 1991-2001.
Nephrology Dialysis Transplantation 2002: 17: 2212-2219.

29. Rabbat CG, Thorpe KE, Russell JD, Churchill DN. Comparison of mortality risk for
dialysis patients and cadaveric first renal transplant recipients in Ontario, Canada. Journal
of the American Society of Nephrology 2000: 11: 917-922.

30. Schnuelle P, Lorenz D, Trede M, van der Woude FJ. Impact of renal cadaveric
transplantation on survival in end-stage renal failure: Evidence for reduced mortality risk
compared with hemodialysis during long-term follow-up. Journal of the American
Society of Nephrology 1998:9: 2135-2141.

31. Glanton CW, Kao TC, Cruess D, Agodoa LY, Abbott KC. Impact of renal transplantation
on survival in end-stage renal disease patients with elevated body mass index. Kidney Int
2003:63: 647-653.

32. Ojo AO, Hanson JA, Meier-Kriesche HU et al. Survival in recipients of marginal
cadaveric donor kidneys compared with other recipients and wait-listed transplant
candidates. Journal of the American Society of Nephrology 2001:12: 589-597.

33. Oniscu GC, Brown H, Forsythe JL. How old is old for transplantation? Am J Transplant
2004:4: 2067-2074.

34. Cosio FG, Alamir A, Yim S et al. Patient survival after renal transplantation: I. The
impact of dialysis pre-transplant. Kidney Int 1998:53: 767-772.

35. Asderakis A, Augustine T, Dyer P et al. Pre-emptive kidney transplantation: The
attractive alternative. Nephrol Dial Transplant 1998:13: 1799-1803.

36. Meier-Kriesche HU, Port FK, Oj o AO et al. Effect of waiting time on renal transplant
outcome. Kidney Intemnational 2000:58: 1311-1317.

37. Mange KC, Joffe MM, Feldman HI. Effect of the use or nonuse of long-term dialysis on
the subsequent survival of renal transplants from living donors. N Engl J Med 2001:344:
726-731.

38. Gaston RS, Danovitch GM, Adams PL et al. The report of a national conference on the
wait list for kidney transplantation. Am J Transplant 2003:3: 775-785.










39. Lenhard V, Dreikorn K, Rohl L. Results of kidney transplantation in relation to HLA-A,
B, DR matching and quality of donor organ. Proc Eur Dial Transplant Assoc 1980: 17:
450-456.

40. Alexander JW, Zola JC. Expanding the donor pool: Use of marginal donors for solid
organ transplantation. Clin Transplant 1996:10: 1-19.

41. Faenza A, Sestigliani E, Zambianchi L, Ridolfl L. Utilization of suboptimal kidney
donors. Transplant Proc 2004:36: 485-487.

42. Meier-Kriesche HU, Port FK, Oj o AO et al. Effect of waiting time on renal transplant
outcome. Kidney International 2000:58: 1311-1317.

43. Nyberg SL, Matas AJ, Kremers WK et al. Improved scoring system to assess adult
donors for cadaver renal transplantation. Am J Transplant 2003:3: 715-721.

44. Port FK, Bragg-Gresham JL, Metzger RA et al. Donor characteristics associated with
reduced graft survival: An approach to expanding the pool of kidney donors.
Transplantation 2002:74: 1281-1286.

45. Su XM, Zenios SA, Chertow GM. Incorporating recipient choice in kidney
transplantation. Journal of the American Society of Nephrology 2004:1 5: 1656-1663.

46. Swanson SJ, Hypolite IO, Agodoa LYC et al. Effect of donor factors on early graft
survival in adult cadaveric renal transplantation. American Journal of Transplantation
2002:2: 68-75.

47. Schnitzler MA, Whiting JF, Brennan DC et al. The expanded criteria donor dilemma in
cadaveric renal transplantation. Transplantation 2003:75: 1940-1945.

48. Ratner LE, Kraus E, Magnuson T, Bender JS. Transplantation of kidneys from expanded
criteria donors. Surgery 1996: 119: 372-377.

49. Metzger RA, Delmonico FL, Feng S, Port FK, Wynn JJ, Merion RM. Expanded criteria
donors for kidney transplantation. Am J Transplant 2003:3 Suppl 4: 114-125.

50. Schold JD, Kaplan B, Baliga RS, Meier-Kriesche HU. The broad spectrum of quality in
deceased donor kidneys. Am J Transplant 2005:5: 757-765.

51. Schold JD, Kaplan B, Chumbler NR et al. Access to quality: Evaluation of the allocation
of deceased donor kidneys for transplantation. J Am Soc Nephrol 2005:16: 3121-3127.

52. Schold JD, Howard RJ, Scicchitano MJ, Meier-Kriesche HU. The expanded criteria
donor policy: An evaluation of program obj ectives and indirect ramifications. American
Journal of Transplantation 2006:6: 1689-1695.










53. Schold J.D., Meier-Kriesche HU. Which Renal Transplant Candidates Should Accept
Marginal Kidneys in Exchange for Shorter Waiting Time on Dialysis? Clinical Joumnal of
the American Society of Nephrology 1[3]. 5-1-2006.

54. Schnitzler MA, Whiting JF, Brennan DC et al. The expanded criteria donor dilemma in
cadaveric renal transplantation. Transplantation 2003:75: 1940-1945.

55. Prevalance of Reported ESRD. Online. Intemet. Available:
http://www.usrds. org/2006/reflB_prevalence_06 .pdf. 2006 Annual Data Report .
(accessed 20 Sep 2006).

56. Patient Survival. Online. Intemet. Available:
http://www.usrds. org/2006/ref/I_survival_06 .pdf. 2006 Annual Data Report (accessed
20 Sep 2006).

57. Young CJ, Gaston RS. Renal transplantation in black Americans. N Engl J Med
2000:343: 1545-1552.

58. Gaston RS. Factors affecting renal allograft survival in African Americans. Blood Purif
1996:14: 327-333.

59. Gaston RS, Benfield M. The relationship between ethnicity and outcomes in solid organ
transplantation. J Pediatr 2005:147: 721-723.

60. Young CJ, Gaston RS. Renal transplantation in black Americans. N Engl J Med
2000:343: 1545-1552.

61. Alexander GC, Sehgal AR. Barriers to cadaveric renal transplantation among blacks,
women, and the poor. Jama-Journal of the American Medical Association 1998:280:
1148-1152.

62. Ayanian JZ, Cleary PD, Weissman JS, Epstein AM. The effect of patients' preferences
on racial differences in access to renal transplantation. New England Joumnal of Medicine
1999:341: 1661-1669.

63. Organ Distribution: Allocation of Deceased Kidneys. Online. Internet. Available:
http://www.optn. org/PoliciesandBylaws/policies/pdfs/polic_.d. (accessed 11-20-
2004)

64. Kerman RH, Kimball PM, Van Buren CT et al. Influence of race on crossmatch outcome
and recipient eligibility for transplantation. Transplantation 1992:53: 64-67.

65. Foster CE, III, Philosophe B, Schweitzer EJ et al. A decade of experience with renal
transplantation in African-Americans. Ann Surg 2002:236: 794-804.









66. Cosio FG, Alamir A, Yim S et al. Patient survival after renal transplantation: I. The
impact of dialysis pre-transplant. Kidney Int 1998:53: 767-772.

67. Kalil RS, Heim-Duthoy KL, Kasiske BL. Patients with a low income have reduced renal
allograft survival. Am J Kidney Dis 1992:20: 63-69.

68. Butkus DE, Meydrech EF, Raju SS. Racial differences in the survival of cadaveric renal
allografts. Overriding effects of HLA matching and socioeconomic factors. N Engl J Med
1992:327: 840-845.

69. Neylan JF, Sayegh MH, Coffman TM et al. The allocation of cadaver kidneys for
transplantation in the United States: Consensus and controversy. Journal of the American
Society of Nephrology 1999:10: 2237-2243.

70. Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW, Jr. Body-mass index and
mortality in a prospective cohort of U. S. adults. N Engl J Med 1999:341: 1097-1 105.

71. Ejerblad E, Fored CM, Lindblad P, Fryzek J, McLaughlin JK, Nyren O. Obesity and risk
for chronic renal failure. J Am Soc Nephrol 2006:17: 1695-1702.

72. Allison DB, Fontaine KR, Manson JE, Stevens J, VanItallie TB. Annual deaths
attributable to obesity in the United States. JAMA 1999:282: 1530-1538.

73. Manson JE, Willett WC, Stampfer MJ et al. Body weight and mortality among women. N
Engl J Med 1995:333: 677-685.

74. Stevens J, Plankey MW, Williamson DF et al. The body mass index-mortality
relationship in white and African American women. Obes Res 1998:6: 268-277.

75. Abbott KC, Glanton CW, Agodoa LY. Body mass index and enrollment on the renal
transplant waiting list in the United States. J Nephrol 2003:16: 40-48.

76. Fung F, Sherrard DJ, Gillen DL et al. Increased risk for cardiovascular mortality among
malnourished end-stage renal disease patients. Am J Kidney Dis 2002:40: 307-314.

77. Kakiya R, Shoji T, Tsujimoto Y et al. Body fat mass and lean mass as predictors of
survival in hemodialysis patients. Kidney Int 2006:70: 549-556.

78. Kalantar-Zadeh K, Kopple JD, Kilpatrick RD et al. Association of morbid obesity and
weight change over time with cardiovascular survival in hemodialysis population. Am J
Kidney Dis 2005:46: 489-500.

79. Wiesholzer M, Harm F, Schuster K et al. Initial body mass indexes have contrary effects
on change in body weight and mortality of patients on maintenance hemodialysis
treatment. J Ren Nutr 2003:13: 174-185.










80. Kalantar-Zadeh K. Causes and consequences of the reverse epidemiology of body mass
index in dialysis patients. J Ren Nutr 2005:15: 142-147.

81. Gore JL, Pham PT, Danovitch GM et al. Obesity and outcome following renal
transplantation. Am J Transplant 2006:6: 357-363.

82. Meier-Kriesche HU, Vaghela M, Thambuganipalle R, Friedman G, Jacobs M, Kaplan B.
The effect of body mass index on long-term renal allograft survival. Transplantation
1999:68: 1294-1297.

83. Johnson DW, Isbel NM, Brown AM et al. The effect of obesity on renal transplant
outcomes. Transplantation 2002:74: 675-681.

84. Drafts HH, Anjum MR, Wynn JJ, Mulloy LL, Bowley JN, Humphries AL. The impact of
pre-transplant obesity on renal transplant outcomes. Clin Transplant 1997: 11: 493-496.

85. Armstrong KA, Campbell SB, Hawley CM, Nicol DL, Johnson DW, Isbel NM. Obesity
is associated with worsening cardiovascular risk factor profiles and proteinuria
progression in renal transplant recipients. Am J Transplant 2005:5: 2710-2718.

86. Pelletier SJ, Maraschio MA, Schaubel DE et al. Survival benefit of kidney and liver
transplantation for obese patients on the waiting list. Clin Transpl 2003: 77-88.

87. Holley JL, Monaghan J, Byer B, Bronsther O. An examination of the renal transplant
evaluation process focusing on cost and the reasons for patient exclusion. Am J Kidney
Dis 1998:32: 567-574.

88. Modlin CS, Flechner SM, Goormastic M et al. Should obese patients lose weight before
receiving a kidney transplant? Transplantation 1997:64: 599-604.

89. Bennett WM, McEvoy KM, Henell KR, Valente JF, Douzdjian V. Morbid obesity does
not preclude successful renal transplantation. Clin Transplant 2004: 18: 89-93.

90. Howard RJ, Thai VB, Patton PR et al. Obesity does not portend a bad outcome for kidney
transplant recipients. Transplantation 2002:73: 53-55.

91. 2006 USRDS Annual Data Report. http://www.usrds. org/2006/ref/A~incidence_06.pdf .
(accessed 20 Sep 2006).

92. Meier-Kriesche HU, Ojo AO, Hanson JA, Kaplan B. Exponentially increased risk of
infectious death in older renal transplant recipients. Kidney Int 2001:59: 1539-1543.

93. Doyle SE, Matas AJ, Gillingham K, Rosenberg ME. Predicting clinical outcome in the
elderly renal transplant recipient. Kidney Int 2000:57: 2144-2150.










94. Ismail N, Hakim RM, Helderman JH. Renal replacement therapies in the elderly: Part II.
Renal transplantation. Am J Kidney Dis 1994:23: 1-15.

95. Meier-Kriesche HU, Schold JD, Gaston RS, Wadstrom J, Kaplan B. Kidneys from
deceased donors: Maximizing the value of a scarce resource. Am J Transplant 2005:5:
1725-1730.

96. Smits JM, Persijn GG, van Houwelingen HC, Claas FH, Frei U. Evaluation of the
Eurotransplant Senior Program. The results of the first year. Am J Transplant 2002:2:
664-670.

97. Grossman M. Concept of Health Capital and Demand for Health. Journal of Political
Economy 1972:80: 223-225.

98. Dickinson DM, Shearon TH, O'Keefe J et al. SRTR center-specific reporting tools:
Posttransplant outcomes. Am J Transplant 2006:6: 1198-1211.

99. Schold JD, Howard RJ, Scicchitano MJ, Meier-Kriesche HU. The expanded criteria
donor policy: An evaluation of program obj ectives and indirect ramifications. Am J
Transplant 2006:6: 1689-1695.

100. Schold JD, Srinivas TR, Guerra G et al. A 'weight-listing' paradox for candidates of
renal transplantation? Am J Transplant 2007:7: 550-559.

101. Rebellato LM, Arnold AN, Bozik KM, Haisch CE. HLA matching and the United
Network for Organ Sharing Allocation System: Impact of HLA matching on African-
American recipients of cadaveric kidney transplants. Transplantation 2002:74: 1634-
1636.

102. Zachary AA, Braun WE, Hayes JM et al. Effect of HLA matching on organ distribution
among whites and African-Americans. Transplantation 1994:57: 1115-1119.

103. Holley JL, Monaghan J, Byer B, Bronsther O. An examination of the renal transplant
evaluation process focusing on cost and the reasons for patient exclusion. Am J Kidney
Dis 1998:32: 567-574.

104. Kendrick E. Evaluation of the Transplant Recipient. In: Danovitch GM, ed. Handbook of
Kidney Transplantation. 2001: 130-145.

105. Meier-Kriesche HU, Arndorfer JA, Kaplan B. The impact of body mass index on renal
transplant outcomes: A significant independent risk factor for graft failure and patient
death. Transplantation 2002:73: 70-74.









BIOGRAPHICAL SKETCH

Jesse D. Schold, Ph.D., M. Stat., M.Ed., completed his doctorate program in 2007 in the

department of Health Services Research, Management and Policy at the University of Florida in

the College of Public Health and Health Professions. Dr. Schold received his undergraduate

training at Emory University. After receiving a B.A., he worked in the healthcare setting for

several years and then enrolled in graduate studies at North Carolina State University. Dr.

Schold received both a Master of Statistics and a Master of Education in the year 2000. After

two years working in industry as a statistician, Dr. Schold began work as a research coordinator

in the Department of Medicine at the University of Florida. During his experience, he also

enrolled as a doctoral student in the Health Services Research program. Dr. Schold is currently

an Associate Instructor in the Department of Medicine and has had peer-reviewed scientific

articles published in j ournals including Transplantation, the Journal of the American Society of

Nephrology, Diabetes Care, the Clinical Journal of the American Society ofNephrology,

Seminars in Dialysis, Blood, Clinical T ansplant, Biology of Blood and Marrow

T ansplantation, and the American Journal of T ansplantation. Dr. S chol d pl ans to continue

work in an academic setting in the fields of transplantation and health services research.





PAGE 1

1 CENTER CHARACTERISTICS AND KIDNEY TRANSPLANT CANDIDATE OUTCOMES By JESSE D. SCHOLD 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 2007

PAGE 2

2 2007 by Jesse D. Schold

PAGE 3

3 To Amy and Caila

PAGE 4

4 ACKNOWLEDGMENTS I have many individuals to thank for their guidance and support th roughout my doctoral program and professional development. I am most indebted to my doctoral committee, colleagues, and family for their encouragement a nd selfless assistance to facilitate my progress through the program and placing me in a position to fulfill my personal and professional goals. I would like to particularly thank Dr. Herwig-Ulf Meier-Kriesche, w ho has served as a supervisor and friend throughout my experience at the Universi ty of Florida. Dr. Meier-Kriesches most selfless and devoted mentorship and dedication have been a prim ary factor in my professional growth and his instruction has provided me w ith invaluable knowledge and skills that I will utilize throughout my career. I am also most gr ateful to Dr. Jeffrey Harman for his outstanding leadership within both my academic program and throughout the dissertation process. I am grateful to Dr. R. Paul Duncan for his supe rvision in my academic program during which his personal approach and professional wisdom have provided an optimal format for progress and growth. I would also like to thank Dr. Neal e Chumbler for his encouragement and great enthusiasm for research. I have many other professional colleagues w hose support, knowledge, and friendship have had an immeasurable contribution to my personal development and passion to research. I thank all of these individuals who ha ve enriched my professional growth and experience at the university. I would also like to thank my academic advisors and instructors for their dedication and diverse training that have provided me with a great basis for future development. My deepest gratitude is to my family for thei r incredible sacrifices and support in allowing me to pursue my academic and professional goals. I will always be indebted to my wife, Amy, for her love and support; my daught er, Caila, for teaching me the most important lessons in life; my father and mother, for always encouraging me to not set limits for achievement; and to many

PAGE 5

5 other family members and friends, whom I have always learned from and who have given me valuable perspective.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES................................................................................................................ .........9 ABSTRACT....................................................................................................................... ............10 CHAPTER 1 INTRODUCTION..................................................................................................................12 2 BACKGROUND....................................................................................................................15 Kidney Transplantation for EndStage Renal Disease Patients..............................................15 Variation in Patient Outcomes by Providers of Healthcare....................................................15 Factors Associated with Kidney Tr ansplant Center Performance..........................................17 Kidney Transplant Center Vo lume and Patient Outcomes..............................................19 The Impact of Waiting Time on Outcomes for Kidney Transplant Candidates..............20 Donor Quality and Transplant Recipient Outcomes.......................................................23 High Risk Kidney Transplant Patients...................................................................................25 African-American Transplant Recipients........................................................................26 Obese Transplant Recipients...........................................................................................28 Elderly Transplant Recipients.........................................................................................29 Conceptual Framework...........................................................................................................31 3 MATERIALS AND METHODS...........................................................................................37 Overview....................................................................................................................... ..........37 Data........................................................................................................................... ..............37 Dependent Variables............................................................................................................ ...38 Explanatory Variables of Interest...........................................................................................40 Additional Explanatory Variables..........................................................................................42 Statistical Analysis........................................................................................................... .......43 Study Aim I.................................................................................................................... .........45 Study Aim II................................................................................................................... ........45 Study Aim III.................................................................................................................. ........46 Potential Selection Bias....................................................................................................... ...46 4 RESULTS........................................................................................................................ .......52 Study Population............................................................................................................... ......52 Rate of Transplantation by Center..........................................................................................52 Transplant Center Volume......................................................................................................53

PAGE 7

7 Center Donor Quality........................................................................................................... ..53 Center Performance Ratings...................................................................................................54 Association between Transplant Center Characteristics.........................................................54 Reliability of Historical Center Characteristics......................................................................55 Kaplan-Meier Candidate Survival by Center Characteristics.................................................55 Multivariate Cox Model for Primary Outcome of Candidate Mortality.................................56 Outcomes among High-Risk Candidate Groups.....................................................................59 Expected Survival by Center Characteristics.........................................................................61 5 DISCUSSION..................................................................................................................... ....79 6 CONCLUSIONS AND FUTURE WORK.............................................................................99 LIST OF REFRENCES.............................................................................................................. .102 BIOGRAPHICAL SKETCH.......................................................................................................110

PAGE 8

8 LIST OF TABLES Table page 4-1 Transplant Candidate Characteristics................................................................................64 4-2 Candidate Characteristics by Center Proportion of Transplants within Three Years........65 4-3 Candidate Characteristics by Center Volume Category....................................................65 4-4 Candidate Characteristics by Center ECD Proportion Category.......................................66 4-5 Candidate Characteristics by Center Performance Category.............................................66 4-6 Median Levels of Center Characteristics over Time.........................................................66 4-7 Correlation Coefficients be tween Center Characteristics..................................................67 4-8 Center Characteristics at th e Time of Transplantation......................................................67 4-9 Adjusted Hazard Ratios for Patient Mort ality after Listing for Transplantation...............68 4-10 Adjusted Hazard Ratios for Receipt of Transplant following Listing...............................69 4-11 Adjusted Hazard Ratios for Post-Transplant Mortality.....................................................69 4-12 Adjusted Hazard Ratios for Post-Transplant Overall Graft Loss......................................70 4-13 Adjusted Hazard Ratios for Pre-Transplant Mortality.......................................................70 4-14 Adjusted Hazard Ratios for Mortal ity for African-American Candidates.........................71 4-15 Adjusted Hazard Ratios for Mo rtality for Elderly Candidates..........................................71 4-16 Adjusted Hazard Ratios for Mortality for Obese Candidates............................................72 4-17 Candidate Life Expectancy (in year s) after Listing by Levels of Center Characteristics................................................................................................................ ....72 4-18 Life Expectancy after Listing at H ypothetical Center Characteristic Levels....................73 4-19 Life Expectancy after Listing fo r African-American Candidates by Center Characteristic Levels..........................................................................................................73 4-20 Life Expectancy after Listing for Elderl y Candidates by Center Ch aracteristic Levels....74 4-21 Life Expectancy after Listing for Obese Candidates by Center Characteristic Levels......74

PAGE 9

9 LIST OF FIGURES Figure page 2-1 Conceptual Framework based on Gr ossmans Health Production Function.....................36 3-1 Distribution of the Annual Number of Deceased Donor Transplants by Center...............48 3-2 Distribution of the Proportion of ECD Transplants by Center..........................................49 3-3 Distribution of the Proportion of Candidat es receiving a Deceased Donor Transplant within Three Years by Center............................................................................................50 3-4 Distribution of Standardi zed Mortality Ratios by Center..................................................51 4-1 Kaplan-Meier Plot of Candidate Su rvival by Center Rate of Transplant..........................75 4-2 Kaplan-Meier Plot of Candidate Surviv al by Center Proportion of ECD Transplants......76 4-3 Kaplan-Meier Plot of Candida te Survival by Center Volume...........................................77 4-4 Kaplan-Meier Plot of Candidate Survival by Center Performance Ratio..........................78

PAGE 10

10 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CENTER CHARACTERISTICS AND KIDNEY TRANSPLANT CANDIDATE OUTCOMES By Jesse D. Schold August 2007 Chair: Jeffrey Harman Major: Health Services Research There are approximately half of a million patients diagnosed with End-Stage Renal Disease (ESRD) currently in the United States. There are two general classes of treatment for ESRD, maintenance dialysis and kidney transplantati on. Kidney transplantati on is considered the preferred treatment for ESRD patients who are medi cally cleared for the surgical procedure as it is associated with an improved quality of life and longer life expectanc y. A kidney transplant may derive from a living or a deceased donor. Fo r patients to receive a deceased donation, they must be placed on a waiting list at a transplant center. There are cu rrently approximately 240 kidney transplant centers in the United States. There are numer ous characteristics that vary among transplant centers. An important question is whether or not these center factors ar e associated with prospective transplant candidate survival. Moreover, whether thes e center factors are applicable to all transplant candidates or have differential e ffects on certain patients is also unclear. This study examined these questions using a national da tabase containing patient -level information on outcomes for all adult solitary kidney transplant candidates from 1995. Results of this study indicate that transplant ce nter characteristics are significantly associated with patient survival after listing. Of the cen ter characteristics, expected wa iting time had the greatest impact

PAGE 11

11 on candidate survival. Candidates listed at centers with the longe st expected waiting time had a 32% increased hazard for death, translating to appr oximately two and a half years of reduced life expectancy, as compared to candidates listed at centers with a relativel y short expected waiting time. The main findings were applicable to th e general candidate population as well as high-risk candidate subgroups. Results of this study may be utilized to info rm patients and caregivers about the important impact of characteristics of tr ansplant centers on prospective survival for candidates of kidney transplantation. Investigation into applications of these findings in healthcare policy, for different transplant populations, and in other medical contexts will be examined in follow-up studies deriving from this study.

PAGE 12

12 CHAPTER 1 INTRODUCTION Kidney transplantation is regarded as the most effective treatment modality for End-Stage Renal Disease (ESRD) patients who are medically cleared for the surgical procedure. As of October, 2006, there were over 65,000 patients listed to receive a solitary ki dney transplant in the United States. The number of candidate listings has increased by 60% over the past decade and the number of candidates has increased across a ll age, ethnic, and gender groups. Over the corresponding era, the number of available kidneys deriving from deceased donors increased, but at a much more modest rate. As a result, the difference between the availability of and demand for kidney transplantation has markedly increased. One of the most important implications of this temporal shift has been a significant incr ease in expected waiting times and increased death rates for kidney transplant candida tes awaiting the procedure. This trend is particularly troubling for transplant candidates and the healthcare s ector for two primary re asons. First, kidney transplantation has been shown to double a patients life expectancy as compared to the alternative treatment modality of maintenance dialysis, and, as such, delayed access to transplantation places patients at an increased ri sk of mortality. Sec ond, increased duration of dialysis prior to transp lantation has deleterious effects on re cipients following transplantation, resulting in a cohort of patient s with increased risk for morb idity and mortality after the procedure. There are approximately 240 tran splant centers currently opera ting in the United States. Among transplant centers, there is significant variab ility in expected waiting times for candidates to receive a transplant. In a ddition, performance evaluations sugge st that outcomes for transplant recipients vary significantly rela ted to the quality of care at in dividual transplant centers. Research also indicates that centers with the highest transplant volume are associated with

PAGE 13

13 improved patient and graft survival following the transplant procedure. Moreover, both shortterm complication rates and long-term survival for transplant recipients ar e significantly variable based on the quality of the donor organ. The prop ortion of highand lo w-quality donor organs also is variable throughout regions of the count ry and at individual cente rs. Cumulatively, there are multiple factors that may poten tially influence a patients prognos is relative to the selection of a particular center. The current evidence sugg ests that patients have incentives to receive a transplant as early as possible, at the best performing center, and receive the highest quality donation. However, despite these known effects, there has not been a comprehensive study to examine the joint impact of thes e factors on transplant candidate outcomes. This information is particularly salient given the expanded ca ndidate waiting list and associated time to transplantation. In addition, the re liability of historical transpla nt center factors on prospective patient outcomes has not been thoroughly evaluate d. Transplant candidate characteristics and prognoses are also widely variable based on age, ethnicity, primary cause of ESRD, and clinical presentation. Whether certain tr ansplant center profiles are more applicable to patients with different prognoses and care needs has not been pr eviously investigated. In particular, certain patient groups (including individuals over the age of 65, African-Americans, and the obese) have differential survival rates on dial ysis and following transplantation relative to the general cohort of patients. In this regard, center characteristics may have a unique impact on these subgroups. The purpose of this dissertation is to analyze th e association of transpla nt center factors with patient outcomes, which may ultimately be used to guide candidate and caregiver decision making to select specific centers for a kidney transplant. The speci fic aims of this study are to I. Determine whether characteristics of transp lant centers (volume, performance ratings, waiting time, and donor quality) are significantly associated with candidate mortality after listing for a soli tary kidney transplant.

PAGE 14

14 II. Determine the relative impact of center ch aracteristics and estimate transplant candidate survival rates at incremental leve ls that may be utilized as a tool for selection of a center. III. Determine the significance and relative impact of transplant center characteristics on candidate mortality after listing for a solitary kidney transplant in three high-risk patient groups (elderly, obese, a nd African-Americans patients). This study will be carried out utilizing a na tional database containing the population of renal transplant candidates listed for transplantati on in the United States. The database contains patient-level information and de-identified indicators of transplant centers. The results of this study will be applicable for assessing the importan ce of transplant center factors and potentially used to help guide patients and caregivers in their se lection of a center. Although there are other factors which may influence the sele ction of a center, an objective measure of the impact of these transplant center factors may be crucial for informing potentially life-altering decisions. Secondary outcomes of this study will include descriptive analyses of transplant center characteristics across the United States and to asse ss the reliability of hist orical center factors on prospective levels. Chapter 2 provides a brief historical backgr ound of kidney transplant ation and treatment of ESRD patients, a summary of the relevant lite rature in this research area, and a conceptual framework for this study. Chapter 3 provides de scriptive statistics of the candidate population and transplant centers as well as statistical methodology for conducting this study. Chapter 4 includes the results of the analyses and tables and figures depicting the outcomes of the models. Chapter 5 includes discussion of the findings, an d Chapter 6 includes conclusions and potential follow-up studies deriving from this research.

PAGE 15

15 CHAPTER 2 BACKGROUND Kidney Transplantation for EndStage Renal Disease Patients In 2004 there were nearly half a million ESRD patients in the United States (1). This count has increased more than three-fold since 1988, partially explained by increased rates of hypertension, diabetes, improved clinical detection, and an ag ing population. The most common treatment modality for ESRD patient s is maintenance dialysis. Dial ysis is classified as either hemodialysis or peritoneal, and both modalities require the use of a machine for renal function replacement either at a designate d facility or home-based. The alternative treatment for ESRD patients is kidney transplantation. The first kidney transplant was pe rformed in 1954 by Dr. Joseph Murray between identical twins. Si gnificant advances in surgery, immunology, pharmacy, and clinical care have contributed to widespread applic ation of kidney transplantation and over 16,000 transplants were performed in the United States fifty years later in 2004. The procedure is generally classi fied in two broad forms: deceased donor and living donor transplantation. Deceased donors are the source of the majority of transplants, although living transplant rates have increased signifi cantly over the past fifteen years. Variation in Patient Outcomes by Providers of Healthcare The notion that patient outcomes vary based on characteristics of the provider of care has been investigated in a variety of medical contexts. There is extensive evidence in the United States to suggest that there are differences in patient outcomes as a function of the individual caregiver, hospital, and region of the country. Sc hrag et al. investigat ed the association of hospital procedure volume and mortality following surgery for colon cancer (2). The study found significantly improved outcomes associated with hospitals with increased volume, although it was also noted that th e variance in outcomes based on other factors was significantly

PAGE 16

16 higher. Utilizing a national Medicare claims da tabase, a 2005 study concluded that providers with higher volume were associated with reduced morbidity rates af ter radical prostatectomy (3). Examining the treatment of patients with acute myocardial infarcti on, Krumwell et al. found significant geographic and physician specialty varia tion in the use of efficacious drug therapy in the United States (4). Diminished outcomes ha ve also been demonstrated for individual surgeons with fewer procedure experiences rela tive to high and intermediate volume surgeons (5). A national study also found significantly reduced shortand long-term outcomes for cardiac transplant recipients at low-volume centers, which comprised more than half of all centers in the United States (6). Treatment modality and outco mes for patients with myocardial infarction have also been demonstrated to vary by region of the country (7). Additional research supports the concept that geography variations may be impor tant based on particular patient groups (8). There are numerous additional st udies that support the notion that variation in treatment modality and quality of care exists within the Unite d States. These variations suggest that patient selection of particular providers is important and supports the e fforts of agencies charged with identifying quality of care between healthcare providers. The evidence also implies that there are potential mechanisms to overcome quality disp arities, which include information concerning processes of care at higher performing providers As the quality of care and rate of medical errors has been questioned and highlighted in recent years, the dema nd for transparency of information related to outcomes associated with individual institutions and caregivers has proliferated (9). However, there is also mixed evidence as to whether more transparent information regarding provider perf ormance is associated with patie nt selection of providers or subsequent outcomes. In the context of cardiac bypass surgery, research suggested that report cards (i.e., performance evaluation of provi ders) had a positive impact on outcomes and

PAGE 17

17 processes of care (10,11). Furthermore, there was some evidence that positive report grades influence future utilization for hospitals (12). However, additional reports indicate a lack of awareness among patients concerning performance ratings and a general fa ilure to incorporate ratings in provider selection (13). In genera l, there is a perception in the research and medical communities that variations exist between healthcare providers, but techniques for measuring performance in an equitable manner, interpreting provider eval uations properly, and disseminating information regarding provider quality of care in a manner that is useful to both providers and consumers are both challenging and inexact and will require significant improvement in the years to come. Factors Associated with Kidney Transplant Center Performance In the field of kidney transplantation there ha s also been significant research dedicated towards determining center effects associated w ith outcomes for transplant recipients. In the early days of transplantation these effects were suspected and causes for differences in outcomes for patients attributed to effects other than known medical risk fact ors were initially considered. In 1986, Burdick and Williams found differences in outcomes associated with patient residence in metropolitan versus those patients that were cl assified as out-of-town (16). This early report suggested that there may be specific care mechanisms for transplant patients that were associated with patient outcomes. An additional report de termined that significant differences in patient outcomes were associated with transplant centers that were independent of utilization of specific immunosuppressive regimens and human leukocyte antigen (HLA)matched recipients (17). This report also suggested that center differe nces were not confined to the immediate posttransplant period, as graft survival differences related to centers accrued for patients with function after three months post-transplant. This conclusion inferred that processes of care may also be implicated as a causal mechanism associat ed with center variations. Additional research

PAGE 18

18 suggested a learning curve asso ciated with immunosuppressive re gimens at transplant centers (18). In this case, centers that were able to incorporate the newest, most efficacious regimens were associated with superior outcomes. The report also concluded th at variation between centers were minimal for low risk transplants, bu t derived mostly from treatment of higher risk patients. In this sense, higher risk patients we re more sensitive to care practices, and care for these patients may be more complex; differen ces in centers are more observable for these subgroups. Ogura et al. described characteristics of cen ters related to different outcomes kidney transplant recipients (19). On e important conclusion of this study was that while complication rates did not significantly vary between highand low-performing cen ters, graft survival following complications was signi ficantly different. The report found that low-performing centers did have a significantly higher proportion of at risk patien ts, but adjusted survival rates did not obviate differences. Moreover, the majority of differences between centers existed in the initial six months post-transplant rather th an long-term divergences. This report found no significant differences based on the volume of tran splants at centers; however, there is extensive literature suggesting improved out comes at high volume centers in subsequent reports. Performance evaluations have been pr oduced and published for ki dney transplant centers since 1991. The Scientific Registry of Transpla nt Recipients (SRTR) publishes these reports online and includes classification of centers as either signifi cantly above, below or not significantly different than exp ected based on the national experi ence. The 2005 report indicated that among the 191 transplant centers with at l east 50 kidney transplants between July 2002 and December 2004, approximately 15% were identifie d as having statistically higher or lower survival rates compared to the national experi ence (20). Along with this information being freely available to the patient s and referring physicians, insura nce carriers may utilize this

PAGE 19

19 information for negotiation of contracts w ith particular centers. A 2006 study, however, suggested that in aggregate thes e report cards had no significant influence on selection of centers for transplant recipients with the possibl e exception in younger and more highly educated patients (21). The manuscript found a positive correlation with center performance on a year-toyear basis but that this association reduced over time. Other questions remain whether performance ratings are true reflections of cente r practice and whether adjustment techniques are sufficient to control for patient variability and a ssociated outcomes (22). As compared to other contexts in which performance evaluations have been conducted, kidney tr ansplant patients have less acuity and longer follow-up periods, and th e models have signifi cantly less predictive ability. These factors suggest th at performance evaluations in th e field of kidney transplantation may be subject to influential factors that may not represen t center quality care. Kidney Transplant Center Volume and Patient Outcomes There is significant variability in the size and scope of kidney transpla nt centers. Center volume may be a function of geographic location, presence of competitive centers, academic affiliation and service to underprivileged patients inclusion of pediatric programs, use of multiorgan transplants, and the propor tion of living versus deceased donor kidney transplantation. Early reports indicated that cen ters with high volume of proc edure were associated with improved outcomes. A comparison of high-, mid, and low-volume centers utilizing national data from 1987 to 1991 demonstrated a 19% increa se in one-year graft su rvival in high-volume centers as compared to low-volum e centers (19). This report i ndicated that there were lower rates of immediate graft function independent of demographic char acteristics of recipients at lower-volume centers, but no significant differen ces in acute rejection episodes. A 1997 United Network of Organ Sharing (UNOS) report conclude d that there was sign ificant variation in outcomes at kidney transplant centers, these eff ects were most apparent in the first-year post-

PAGE 20

20 transplantation, and, furthermore, that infe rior outcomes were most common among small transplant centers (23). An exam ination of ten-year survival rates in the United States examined the association of patient outcomes transplanted at centers with large volume (>1000 transplants over ten years) versus lower-volume centers (24). Conclusions of this st udy included that there was a mildly beneficial effect of large centers, but that these effects were reduced in lower-risk transplants. In addition, this re search suggested that large centers were associated with improved outcomes, more notably with elde rly recipients and s pousal living donations, and, in contrast to previous reports, that differe nces were most apparent only after two to three years posttransplantation. In an analysis among living transplant recipien ts, Gjertson detected improved outcomes associated with large centers and th at among small centers there were significantly more variable outcomes (25). The most recent evidence of the center volume effect derived from a 2004 manuscript examining outcomes from 1996 to 2000 for both liver and kidney transplant recipients. The primary conclusi on of the study was that the highe st quartile tran splant centers (which transplanted a median number of 167 pa tients annually) were asso ciated with improved patient outcomes, but there was no detectable difference among the remaining volume level centers in terms of the main outco me of overall graft loss (26). Th e research also suggested that center effects were most evident within 90 days post-transplantation, but without substantial clinically significant differences in recipient or donor f actors associated with centers at different volume strata. The Impact of Waiting Time on Outc omes for Kidney Transplant Candidates The survival advantage of kidney transplantati on as compared to the alternative treatment of maintenance dialysis was suspected for many year s and formalized by Wolfe et al. (27). This analysis utilized a time-dependent model assessing the impact of a solitary kidney transplant in reference to a comparable cohort of medically cl eared and wait-listed candidates. The results

PAGE 21

21 indicated that transplantation approximately doubled life expectancies fo r recipients. Wolfe and colleagues (27) further reported that the relative advantage was observable across age, gender, and ethnic strata. The analysis also indicated that there wa s a significant immediate mortality risk associated with the transplant procedure, bu t that this risk was obviated within a year of transplantation and the benefits observable thereafter. The su rvival advantage of kidney transplantation has subsequently been reported in other countries with vari ations in healthcare structure and patient socio-de mographic characteristics (e.g., Australia, New Zealand, and Canada) (28). In addition, this advantage extends to higher risk recipients including transplantation of lower quality donor kidneys, candidates over seve nty years of age, and obese patients (31). The majority of patients that are listed as candidates for transplantation have already initiated dialysis therapy. As such, waiting tim e associated with the duration of dialysis is a significant mortality risk for renal transplant candidates prior to th e procedure. In addition, there are a number of retrospective analyses that indica te that the duration of dialysis exposure is associated with deleterious outcomes following tr ansplantation. In a si ngle center study, Cosio et al. demonstrated that pre-tran splant dialysis was a significant risk factor for post-transplant patient death and overall graft loss (34). Anothe r single center study utiliz ing transplant data from 1980 to 1995 indicated that these effects were particularly noteworthy among living donor transplant recipients (35). A retrospective analysis deriving fr om a national database extended this concept by demonstrating that the time on di alysis had a dose-dependent risk on subsequent graft loss and patient death after both living and deceased donor transplantation (36). This analysis reported that even six to twelve months of pre-transplant dialyses was associated with a 37% increase in long-term graft loss as compared to patients without exposure to dialysis. In

PAGE 22

22 addition, this study suggested that there was a relatively similar effect in patients with ESRD secondary to diabetes, glomerulonephritis, and h ypertension. A separate study confirmed this direct association between the dura tion of pre-transplant dialysis and the rate of allograft loss in living transplant recipients utiliz ing registry data in the United States (37). Still yet another study investigated the association between duration of pre-transplant dialysis and rate of allograft loss utilizing a paired-kidney design. This design, though, al lowed only outcomes for kidneys deriving from the same donor transplanted to on e individual with short dialysis exposure and another recipient with extended dialysis exposur e. This design allowed for control of donor characteristics and the potential confounder that patients with extended dialysis had a greater proclivity to accept lower-quality donations. Results indicated that recipients with less than six months of pre-transplant dial ysis had over two-fold improved ten-year graft survival as compared with patients with greater than tw enty-four months pretransplant dialysis. The accumulated evidence highlights the signific ant benefit of transp lantation relative to dialysis as a treatment modali ty and additionally emphasizes th e importance of rapid acquisition of a transplant. The growing wait list for tr ansplantation and the extended waiting times to transplantation from a deceased donor source conti nue to subject patients to dialysis for longer durations. A transplantation co mmittee report indicated that at the given trajectories, waiting times may average ten years in the next decade ( 38). Moreover, the differences in waiting times have become increasingly variable across regions of the country. As such, a patients active selection of transplant centers may have increased importance in future years. For candidates that plan on being placed on waiting lists for a deceas ed donor transplant, the prospective waiting time may be one of the primary crit eria for this decision.

PAGE 23

23 Donor Quality and Transplant Recipient Outcomes The association of characteristics of donor kidneys with outcomes after transplantation have been investigated from the early years of tr ansplantation (39). Studi es have identified the history of disease among donors, anatomical char acteristics of kidneys, biopsy results of the donor kidney, donor and recipient cytomegalovi rus status and HLA matching, and donor demographic characteristics as significant risk factors for tr ansplant outcomes (40). In aggregate, the combination of risk factors of a donor kidney, most notab ly the age of the donor, is acknowledged to convey highly variable life expectancies and graft survival for the prospective transplant recipient. There have been several attemp ts to quantify this aggregated risk of donor characteristics and estimate surviv al rates based on thes e levels. The Organ Procurement and Transplantation Network (OPTN) in stituted a formalized definition of marginal kidneys in 2002 with the advent of the Expa nded Criteria Donor (ECD) (44,49). These deceased donor kidneys were demonstrated to convey a 70% or greater risk for graft loss for transplant recipients relative to an ideal class of donations and were char acterized by a donor age over 60 or over 50 accompanied with two additional risk factors including a history of hypertension, elevated terminal donor creatinine, or cerebrovascul ar cause of death. Nybe rg et al. developed a more granulated scoring system based on four transplant characteristics association with sixmonth renal function level and demonstrated sign ificant variability in outcomes by donor risk groups utilizing a national cohort (43). An altern ative risk index has been utilized to create donor risk groups and demonstrated the associati on with complication rate s as well as long-term patient and graft survival (50). This alternative risk index derive d from a retrospective analysis indicated that relative to an ideal class of deceased donations (c onstituting approximately 11% of donations), the highest risk class of donations had a three-fold risk for graft loss. In addition, the

PAGE 24

24 analysis demonstrated that the ri sk for graft failure was particular ly notable in the first year posttransplantation, but also persis ted for patients with graft re tention after one year. Research indicates that there is significant geographic vari ability in donor quality with certain transplant network regions having an ove r 50% likelihood of low-quality donations (51). A portion of this effect may be re lated to the issue that transplant centers in areas with extended waiting lists may also be more apt to utilize lo wer-quality kidneys in order to avoid patient deaths on the waiting list. Howeve r, there is clear relationship th at transplant candidates who list for a kidney in certain regions are more like ly to receive higher-r isk donor kidneys. Beyond regional effects, there is also li kely a transplant center relatio nship, as centers have variable acceptance practices of donor kidne ys. One of the primary purposes of the ECD policy was to identify high-risk kidneys and allow patients to consent to receive th ese kidneys with the potential benefit of receiving a donation more rapidl y. The decision to list for an ECD kidney is often strongly influenced by tr ansplant physicians and other patient advocates. A 2006 study demonstrated that patients who listed for an ECD at centers with a high proportion of ECD candidates had no relative benefit in the amount of waiting time for transplantation (52). In contrast, candidates who listed for ECD kidneys at centers with a low proportion of ECD candidates in fact do receive their transplants much more rapidly. In addition, those patients that listed for a lower-quality organ at centers with a low proportion of ECD candidates were much more likely to actually receive a lowerquality donation. There are not any explicit policies directi ng which renal transplant candidates should receive lower-quality kidneys. A national tr ansplant committee report suggested that the candidates most likely to benefit from transplant ation are those with decreased life expectancy on dialysis (38). A retrospectiv e analysis indicated that elde rly and diabetic patients often

PAGE 25

25 receive equivalent benefit from an ECD if it can be received with signifi cantly reduced exposure time to dialysis as compared to a standard cr iteria donation (SCD) with longer dialysis exposure time (53). In contrast, younger, and presumab ly healthier, patients have reduced life expectancies when accepting lower-quality organs early after transpla ntation as opposed to waiting longer for a highe r-quality donation. Research also suggests that quality-adjusted life years (QALYs) may be suboptimal for some pa tients who accept an ECD kidney early after ESRD onset (54). This study also reported that these results vari ed based on transplant candidate characteristics, inferring that this decision should be patient dependent. In summary, there is signifi cant evidence indicating that the quality of a donor organ has a marked impact on patient outcomes. Secondly, ther e is wide variability in the quality of donor organs at a regional level. Additionally, candidate listing practices for lower quality organs at a particular transplant center affect progression al ong waiting lists and the quality of transplanted organs. Moreover, the decision to accept lower-quality donations appears to be variable based on candidate prognosis and health status. In aggr egate, the significant infl uence of donor quality on recipient outcomes is clear; how ever, the association of the proportion of high-risk transplants on a prospective candidates mortality is unknown. High-Risk Kidney Transplant Patients Even though kidney transplanta tion is a lifesaving procedure across all candidate groups, outcomes for recipients signifi cantly vary as a function of demographic and clinical characteristics. Individual di fferences in outcomes between patie nt groups have been described based on known medical prognoses analogous to the general popul ation, but there are also unique characteristics particular to the transp lant population which dist inguish patient risk groups. For this study, three dis tinct classes of high-risk tran splants which have variable representation at transplant centers in the United States and have known asso ciation with inferior

PAGE 26

26 transplant outcomes will be examined: African-Americans, obese, and elderly transplant candidates. African-American Transplant Recipients African-Americans comprised approximat ely 32% prevalent cases of the ESRD population in the United States in 2004 (55). Su rvival rates for African-Americans are improved on dialysis relative to Caucasians (56). However, among transplant recipients, AfricanAmericans have significantly higher rates of acute rejection and graft loss (57). There have been multiple explanations purported regarding thes e relatively inferior outcomes among AfricanAmericans including immunological differences, socioeconomic fact ors, and delayed access to transplantation (58). Alexande r et al. distinguished unique st eps in the transplant process including medical suitability, in terest in transplant, completio n of a transplant workup, and acceleration on the transplant waiting list. He fo und racial disparities at several steps in the process (61). Additional research indicated sign ificant differences in attitudes toward seeking (or obtaining) a transplant between African-America ns and Caucasians, but that these differences alone did not explain disparities in access to transplantation (62). Differences in rates of transplantation to Af rican-Americans are also partially explained by lower donation rates among African-Americans. Between 1995 and 2004, African-American donations comprised between 13% of all li ving donations and 10% of all deceased donations (20). In addition, progression along th e deceased donor candidate waiting list is partially predicated on HLA matching (63). HLA-antigens generally are more similar among race groups, and the reduced donation rates am ong African-Americans affects longer waiting times on dialysis and increased deaths on the waiting list for this group. In addition, AfricanAmericans are more likely than Caucasians to have presensitization or histocompatability reactivity to particular donors which eliminates potential donor sources (64). A single center

PAGE 27

27 report in 2002 suggested that disparities in ou tcomes by race could be reduced by encouraging hepatitis B vaccination and util izing hepatitis C donor kidneys subsequently reducing waiting times for African-American candidates (65). An alternative explanation for the documented reduced outcomes following transplantation in Afri can-Americans is more poorly contro lled hypertension (66) In addition, lower economic status may have a strong role in rates of return to dialysis following transplantation (67). As immunosuppressive medication is a necessary and expensive maintenance component to graft survival, nonc ompliance and lack of access to medication regimens is a significant risk for graft loss. Butkus et al. found higher rates of noncompliance by African-American recipients at their center, but c oncluded that racial diffe rences were not fully attributable to immunosuppressi ve regimens and more likely due to poor HLA-matching and socioeconomic factors (68). African-Americans also experience higher ra tes of delayed graft function and higher rates of acute rejection (69). Elevated risk of complications often necessitates more potent levels of immunosuppressive medications and increases the risk of infection and other medica tion side effects. Cumulatively, there appear to be a combina tion of medical, sociological, and economic etiologies for reduced outcom es in African-Americans. African-Americans represent a significant portion of the transp lant candidate popul ation and have unique care needs. The observation that African-American patients have improved survival prior to transplantation on dialysis yet significantly reduced outcomes following transpla ntation suggest that there may be special consideration for selection of a transplant center in this c ohort. In this sense, estimated benefits of listing at centers with the presence of certain factors may be particularly important for

PAGE 28

28 these candidates and specific info rmation regarding estimat ed survival in this stratum of patients important. Obese Transplant Recipients Rates of obesity in the United States have grown substantially, and obesity has been implicated as a significant cause of death in the general populati on (70). Obesity has also been implicated as an independent cause of ESRD and cardiovascula r morbidity (70). Epidemiological evidence suggests that the effect of obesity fo r ESRD patients is a paradoxical area of medicine. Among patients that reach ESRD, high body mass index (BMI) is reported to be protective for dialysis patien ts (75). While the cause(s) for this association are not clearly known, there have been multiple hypotheses yiel ding multiple contributing factors (80). In contrast, among transplant recipien ts, studies indicate that recipi ents with both low and high BMI have diminished survival following transplantat ion (81,82). Moreover, obesity has also been associated with increased complication rates, co sts, length of stay, a nd delayed graft function following the procedure (81,83). Despite the findings that obese pa tients have relatively superi or survival as compared to non-obese patients and inferior surv ival following transplantation, th ey still receive a significant survival benefit associated with transplanta tion (31,86). However, due to the relatively poor shortand long-term outcomes of obese patients following transplantation, certain transplant centers exclude patients based on BMI and othe rs advocate aggressive weight loss prior to transplantation (87,88). Other cen ters report acceptable results fo r transplantation in obese and morbidly obese recipients (89,90). In addition, among ESRD patients, obese patients have lower rates of listing for transplant th en overweight patients despite equivalent survival rates (75). Collectively, there is an established risk for deleterious outcomes among obese patients

PAGE 29

29 following transplantation and variable acceptanc e of this population as viable transplant candidates. The differential survival rates of obese transp lant candidates suggest that the impact of dialysis time may be different in this cohort. This may be refl ected in the incentives to list at particular transplant centers among this group ba sed on expected waiting times. In addition, the impact of other center factor s including performance ratings donor quality, and high volume centers has not been specifically investigated in this population. A limiting factor for these patients may be the willingness of centers to ac cept them as viable transplant candidates; however, among those centers that do list obese patie nts, listing decisions may be particularly critical in this population. Elderly Transplant Recipients Fifty percent of the incident ESRD cases in the United States in 2004 were patients 65 years of age or older (e lderly) (91). Even though the rate of elderly transplant candidates have doubled in the past decade, only 14% of transpla nt candidates are over the age of 65 (20). Therefore, it appears that transplantation is a vi able treatment modality in only a small subset of elderly ESRD patients. However, the entire explanation for the lack of access for older individuals is unclear. Many of elderly ESRD patients have cont raindications to tr ansplant or are generally not medically viable for the surgical pr ocedure. The greatest projected increase in life span associated with transplantation is am ong younger recipients; however, even among patients 70 to 74 years old, transplantation is associated with a significant reduction in mortality (27). Elderly transplant candidates are more likely to receive kidneys from older donors and have significantly higher death rates on the transplant waiting list (51,53). Elderly recipients are at higher risk for infectious death following tran splantation, rates that are accelerated with increased exposure to pre-transpla nt dialysis and varying levels of immunosuppressive regimens

PAGE 30

30 (92). Rates of acute rejection are relatively lowe r in elderly recipients, but this may also be indicative of tailored immunos uppressive regimens and imm une incompetence (93,94). Elderly recipients have stil l been shown to benefit from transplantation; however, the absolute benefit, in terms of years of life gaine d, is substantially less th an younger recipients. In fact, evidence suggests a significan t loss of scarce resources in terms of donor kidney life-years and economic loss to payers of ESRD patients due to transplantation of younger donors to older recipients (95). Even though increased donor age is a risk factor for graft loss among the elderly, the evidence suggests that only younger transplant recipients su rvive long enough to receive the full benefits of a younger donation on average. Th e Eurotransplant program has been developed specifically to address this conc ern by directly allocating older donations to older recipients in combination with reduced cold ischemia and waiti ng time on dialysis (96) The distinction in this program is the older partic ipants are mandated to receive organs from older donors but the tradeoff of receiving an organ more rapidly and w ith less ischemic effect may offset the risks associated with a lower quality donation. Although age is not a factor in OPTN allocation policy in the US, caregivers have substantial influence in directing the types of donations to particular transplant candidates. These de cision-making processes may vary at the transplant center level based on experience, the proportion of older donations available in a particular service area, and length of candidate waiting lists. Also, research has demonstrat ed that due to the elevated mortality risk of elderly patient s on dialysis, acceptance of lowerquality organs in exchange of receiving a transplant more rapidl y may be a viable option for this cohort (53). In other words, the relationship of the competi ng risks of extended wa iting time on dialysis and lower quality donations appears to be unique in the elderly subset of transpla nt candidates. However, these decisions depend largely on the amount of waiting time reduced and the availability of quality

PAGE 31

31 donations. These factors are widely variable at transplant cente rs and the estimated impact of center characteristics is likely to be unique for this high-risk subset of patients. Conceptual Framework The conceptual model for this study is base d upon the work of Grossman and the model of the health production function (97). This mo del examines the role of individual decision making in the production of health and the inve stment of human capital to improve individual outcomes. Fundamental to this theory is that health is predicated on multiple factors of which health care represents one of many explanatory factors. Other determining factors include individual characteristics, e nvironmental conditions, and behavi oral components. In this framework, individuals demonstrate a demand for h ealth care. This demand is not for health care per se, but for health and the subsequent uti lity which derives from health. In this sense, individuals are not solely subject to external constraints but are active participants in their own health status and future prognosis. Additionall y, while individuals demand health, they do not value it over all other goods reflected by choices that have known deleterious impact. In addition to the utility associated with increased h ealth, there is also an investment component of health. Health represents an investment as it allows for increased individual production as well as future utility. Implicit in this model is th at individuals value health and always seek to maximize their level of health and associated utility. The health production function is represente d by a declining slope with potentially nonlinear deceleration over time. Individuals are e ndowed with a particular stock of health that depreciates over time subject to e ffects of aging as well as other potential health shocks such as disease onset or accidents. Grossman illustrated this depreciation based on individual investment in health utilizing Equation 2-1. Hi+i Hi = Ii iHi (2-1)

PAGE 32

32 This equation characterizes the loss in health with i representing the rate of depreciation in the ith period, Hi representing the stock of health, and Ii representing the invest ment of the individual in this period. From an economic perspective, th e investment component of health is determined by the marginal benefit of receiving the desired le vel of health. In this framework, individuals health value is also a function of their wage rate and utility as well as the cost of producing the level of health. In this formulation, the i nvestment component can be more specifically characterized as a function of th e amount of medical ca re sought in the period as well as the stock of human capital and time as illustrated in Equation 2-2. Ii = Ii (Mi, THi, Ei) (2-2) In this case, Ei represents the stock of human capital, THi represents the time component of investment in health, and Mi represents medical care for the individual. Utilizing this framework, this study will examine the effect of center characteristic s on outcomes for renal transplant candidates. In this context, the medical care comp onent of the Grossman model is not homogenous. The type of medical care, represen ted by characteristics of the listing center may have differential impact on patient survival. In th is sense this study will investigate the level of future health that is independently determined by selection of center. Furthermore, under the Grossman framework in which the investment a nd utility deriving from investment may vary based on an individuals prior health status and in dividual characteristics, this study will investigate the impact of these decisions in cer tain high-risk patient groups. Utilizing this framework, the research question focuses on the e ffect of the selection of center on outcomes controlling for a patients latent healthcare st ock and timing of candi date listing. These characteristics include the age, race, gender, length of dialysis, in surance, body mass index, education, and clinical presentati on of the individual. Individuals can be viewed as producers of

PAGE 33

33 their health and this production as a function of the type of medical care sought. Therefore, rather than a single indicator fo r medical care utilization, the selec tion of a center can be further discriminated as a function of center characteris tics, in this case the proportion of high-risk kidneys, center volume, center pe rformance and expected waiting time as depicted in Equation 2-3. Mi = Mi(Pj, Vj, Qj, Wj) (2-3) In this extension of the Grossman framework, me dical care for individual i is a function of the proportion of high-risk kidneys (P), transplant volume (V), qualit y of center (Q), and expected waiting time(W) at a given center j. Selection of center j will in turn have an impact on future health status and the associated utility derived from this level of health. Figure 2-1 depicts this framework for ESRD patients that select a medica l care through a specific center. More broadly, health status in future periods are a function of past health status, depreciation rates, and investment in health, including the type of medical care attained. The evidence to date suggests that th ere is significant associat ion of transplant center characteristics with outcomes for kidney transplant candidates. However, the joint estimation of these factors is unclear, but is crucial toward s informing potential candidates and caregivers regarding choices for centers of care. While the individual eff ects of center performance and volume, waiting time, and organ quality have be en evaluated for transplant recipients, the interaction of these factors and relative benefit has not been investigated and has not been examined for prospective transplant candidates. This study will examine the role of these center characteristics in order to estimate circumstances that offer the best prognoses for transplant candidates and maximize pati ent long-term outcomes.

PAGE 34

34 Results of this study will be important at se veral levels. This study will examine whether there is an important associati on of transplant center factors with outcomes for candidates of transplantation. Failure to find any significant as sociations will suggest that candidates have an incentive to select centers based on convenience, l ogistics, or other determ inates of their quality of life. However, significant f actors that are found may be importa nt to guide patients and other caregivers of where to list for a transplant that can affect their mortalit y risk. In particular, candidates that have a choice in th eir selection of centers or have the resources to travel outside a local area may have an incentive to discriminate between centers based on these characteristics. Furthermore, this study will highlight whether th ese factors are uniformly applicable to the candidate population or have differential impact in high-risk patient groups. A finding of a significant interaction in this re gard would suggest that certain patient groups would be more advised to list at particular centers based on their mortality risk. Moreover, a significant interaction would imply that ther e may be cases in which the en tire transplant process could become more efficient and provide benefit to the full complement of transplant candidates. If certain center factors are more important to cert ain patient groups and less important to others, then it follows that allocating patients to the respective locations may improve outcomes for the entire population. Identification of the importance of transplant factors that are associated with candidate mortality may also provide insights for future augmentation of the transplant and allocation processes. For those factors that are amenable, alteration of these characteristics and development of infrastructure to support a proc ess that fosters these characteristics would be important. Finally, results of this study may be generalizable to ot her areas of medicine in which there is great variability in patient characteristic s and risk level and a choice of potential centers. In recent years, there has been an increased em phasis on patient decision -making and transparent

PAGE 35

35 information for consumers of healthcare services This study will elucidate the importance of an aspect of these guidelines over a broad timelin e in a significant patient population. Most importantly, this study aims to provide important evidence to inform pa tients and caregivers about the impact of cen ter characteristics on th eir long-term prognosis.

PAGE 36

36 SURVIVAL ESRD and suitable and interested in transplant Provider selection for wait listing with a given health stock (Hi) Vj Pj Wj Qj Location Center Reputation Demographic/Environmental FactorsClinical Factors Accessing medical care through certain providers may be associated with a differential health deterioration rate: Mi= M(Pj, Vj, Qj, Wj) Figure 2-1. Conceptual Fr amework based on Grossmans Health Production Function

PAGE 37

37 CHAPTER 3 MATERIALS AND METHODS Overview The primary purpose of this study is to examin e the association of center characteristics with mortality for renal transplant candidates. Th e general approach will be to estimate levels of four center characteristics utilizing data prior to candidate listing and eval uate the association of these characteristics with candida te mortality in the years after listing. This study will be a retrospective analysis of observa tional data derived from a na tional registry database. The outcomes of this study will determine whether e ach of the tested cente r characteristics are significantly associated with candidate mortality a nd the relative effect of high and low levels of these factors among those that are significant. Furthermore, this study will estimate life expectancy for candidates at incremental levels of significant factors th at may ultimately be utilized as a guide for center selection. Finally, this study will reevaluate the significance and life expectancies for candidates that are of part icular high risk and may have unique needs with respect to center selection. Data This study will utilize a national transplant registry which is administered by the SRTR in order to conduct the analysis. The SRTR database contains information for all transplants performed in the United States. Information is collected at all centers on a mandatory basis utilizing data collection forms at several time poi nts for transplant patient s corresponding to their transplant status and has been utilized exte nsively for government-spons ored reports and peerreviewed research publications. Da ta is collected electronically a nd files are distributed at cost for the purpose of research to groups following su bmission of a research and security plan. For this study, the data that will be utilized will include information deriving from the form at the

PAGE 38

38 time of transplant candidate listing, the form s ubmitted at the time of transplantation, and followup forms after transplantation. In addition to da ta collected from transplant centers, files are enriched with data from Medi care and the Social Security Administration for data fields pertaining to patient death or re-initiation of dialysis following transplantation. The database contains patie nt-level data including donor and recipient demographic information, primary clinical characteristics, and a numerical (de-identified ) code for transplant centers. The unit of analysis for this study will be adult patients (at least 18 years of age at the time of listing) that are placed on the waiting list fo r a solitary kidney transplant as indicated in the database. Pediatric patients (candidates less than 18 years of age) and candidates listed for multi-organ transplants will be excluded from this study in order to exam ine a more homogenous population. Candidates and recipients at centers with very low vol ume (less than ten transplants per year) will also be excluded as estimates of center factors are less reli able. These candidates represent a significant minority of all national candidates. In a ddition, recipients of living donor transplants will be excluded from this study as they are less applicable for candidates placed on transplant waiting lists for a deceased donor tr ansplant. The study popula tion will consist of all remaining candidates that were listed for a so litary deceased donor tran splant between 1995 and 2000. This period was selected to represent a re latively recent cohort, but with follow-up time sufficient to evaluate the impact of transplant center factors on long-te rm patient outcomes. Current files that will be utilized for the anal ysis have follow-up information through early 2006. In addition, data will be utili zed from 1992 to 1999 to assess tran splant characteristics based on three years of information accrual prior to candidate listing. Dependent Variables The primary outcome variable in the study will be transplant candidate mortality after listing. Candidates will be followed from the time of listing until the earliest of death or last

PAGE 39

39 follow-up time as indicated in the database. Pa tient death is indicated in the database with internal variables derive d from center follow-up forms as well as populated from Social Security master files. The earliest of th ese dates in cases of discrepancie s will be utilized as an endpoint for the study. Mortality will be compared between study groups utilizing survival models based on time to death following candidate listing. Mode ls will be censored at the time of living transplantation to limit results to recipients from deceased donors. A secondary endpoint will be candidate death prior to the date of transplant. This model will be censored at the time of transplantation, limiting events to the pre-transp lant period. An additi onal secondary endpoint will be overall graft loss following transplantation. These models will be initiated at the time of transplantation for the same study groups and patien ts will be followed until a last follow-up date or the date of overall graft loss. Overall graft lo ss is defined as the com posite endpoint of either patient death or loss of the transplanted kidne y indicated by a return to dialysis or retransplantation. Additional descriptive informa tion will include the correlation of historical center characteristics with pros pective levels after candidate listing, the association between center characteristics and the distribution of thes e characteristics within transplant regions. The purpose of the primary endpoint is to exam ine the overall survival rate for patients based on levels of center characteristics. While other factors are clearly important to patients and caregivers, measures of death and graft loss are objective har d endpoints that are clearly defined. In addition, this data is readily availa ble and reliable, and mortality rates are generally interpretable for consumers of the research and comparable across study contexts. As the primary variables of interest have all been shown to have some a ssociation with outcomes following transplantation, the question remains wh ether the variables will prospectively impact candidate survival for listed patien ts and how important the factors are relative to each other.

PAGE 40

40 The purpose of the secondary outcomes is to dete rmine whether differences that exist by center characteristics are most directly relate d to preor post-transplant mortality. Explanatory Variables of Interest The primary explanatory variables of interest will be four transplant le vel characteristics: center volume, center performance rating, cente r proportion of ECDs and the proportion of patients reaching transplantation. Ea ch of these will be assessed ba sed on data for the three years prior to the year of listing for the transplant can didate. The variables will be categorized into quintiles based on their distribut ion over the study period and im plemented as dummy variables in statistical models. Each of these characteri stics will be estimated according to an individual candidates year of listin g based on retrospective center charac teristics over the previous three years. Transplant volume will be estimated based on the average number of deceased donor transplants in the three years pr ior to candidate listing. Perf ormance ratings (also known as standardized mortality ratios) will be construc ted utilizing data for deceased donor transplants from the prior three years of candidate listings Ratings will be determined utilizing the equivalent methodology as the SRTR, which pub lishes center performan ce ratings on public websites (98). These ratings are produced based on the ratio of the numbe r of actual events (in this case, patient deaths) to the expected number of events based on charac teristics of the center population. The value of this rati o will be utilized for the purpose of this study based on threeyear outcomes. The proportion of high-risk deceased donor transplants will be based on the dichotomous ECD criteria and defined as the proportion of ECDs (among all deceased donor transplants) over the prior three years. Transp lant center waiting time will be indicated by the proportion of patients that receive a deceased donor transplant within the three-year period prior to listing. This proportion will be calculated by Kaplan-Meier models and censored at the time of patients delisting, transfer to other centers, or rece ipt of a living transplant.

PAGE 41

41 The center characteristics are selected for three primary reas ons. First of all, the factors have been shown to be associated with patie nt outcomes following transplantation based on retrospective analysis. As suc h, it is reasonable to surmise that prospective evaluation of these factors will have an impact on candidate outco mes following listing. Secondly, the information regarding these characteristics is all publicly ava ilable and as such could be incorporated into decision-making processes for candidates and their caregivers in order to assist with center selection. Lastly, these factors ha ve wide regional vari ability as well as variable representation among centers within regions. This is important from a methodological perspective as it will more readily allow for the detection of effects of these factors on patient outcomes. In addition, the center variability highlights the potential deci sions that candidates ma y make regarding their choice of center selection. There are a total of 260 transplant centers with unique identification numbers in the database within the study period. A portion of these centers discontinue d their programs within the study period and several programs were newly initiated within th e time frame. In order to accurately estimate effects based on transplant cent er characteristics, centers with less than ten deceased donor transplants per year of opera tion will be excluded from the study. Among the centers, 184 (71%) had at least ten deceased donor transplants per year. These centers accounted for 97% of all deceased donor tr ansplants over this period. Th e distribution of the average annual number of deceased donor transplants is displayed in Figure 3-1. The median (and 25th / 75th percentile) number of deceased donor transplants at these centers was 30 (19 / 48), ranging from 10 to 177. The median proportion of ECD tr ansplants (of all deceased donor transplants) was 11%, ranging from 0% to 30%. The distribu tion of the proportion of ECD transplants is illustrated in Figure 3-2. The proportion of ca ndidates receiving a deceased donor transplant

PAGE 42

42 within three years after listing al so varied significantly over the study period. The range of this proportion was 11% to 91% with a median level of 57%. The dist ribution of this proportion of candidates receiving a deceased donor transplant is displayed in Figure 3-3. Standardized mortality ratios ranged from 0 (indicating no ev ents over the time period) to close to 4.0 (indicating four times as many events as would be expected based on ch aracteristics of the transplant population at the cen ter). The distribution of the mortality ratios by center are displayed in Figure 3-4. Additional Explanatory Variables Multivariate models will incorporate a num ber of adjustment variables that are considered to be independently associated with the mortality or waiting time to transplant for candidates. Candidate age will be utilized as it is strongly associated with death rates as well as the likelihood to receive a transplant over time Age will be categorized into groups as 18, 45, 55, and 65+. Candidate race will be used in the models categorized as Caucasian, African-American, Asian, Hispanic, and Other. Ca ucasians have significantly higher mortality rates on the waiting list relative to other race groups as well as greater likelihood to receive a transplant. Candidate gender will be used in the models; there is some indication that gender is associated with likelihood to progress more rapi dly on the waiting list. Candidate primary cause of ESRD will be dichotomously repr esented in the models as diabet es or other. Diabetics with ESRD have significantly higher death rates and less likelihood to reach transplantation. Candidate BMI will be categorized into levels representing different health status (Missing, 13 19, 20, 25, 30, 35+). Obese candidates have b een shown to have a slower progression to transplant and superior survival after dialys is initiation. Additional variables that will be utilized to control for potential selection bias in center choice will include candidate education and insurance status. Education will be dichotomous ly represented as a college degree or higher,

PAGE 43

43 or less than a college degree. Insu rance status will be categorized as private, Medicare, or other. Obese candidates will be defined as those with a calculated BMI greater than 30 kg/m2 and elderly patients will be classified as those that are 65 years of age or older at the time of listing. A secondary outcome for the study will include po st-transplant graft and patient survival. These models will include only those candidates that receive a deceased donor transplant and will be initiated at the time of transplantation with follow-up until death, graft loss, or the last follow-up period. Cox models will be adjusted for transplant characteristics that are associated with outcomes after transplanta tion including HLA mismatching leve l, donor age, cold ischemia time, donor race, and pre-transplant dialysis time, in addition to adjustment factors utilized in the primary outcome models. Statistical Analysis The primary analyses used in the disser tation to evaluate th e study aims will be performed with survival models. Survival mode ls (also known as time to event analyses) are appropriate in contexts with a known time origin, in this case the time of candidate listing, and with well-defined follow-up period and events. In addition, as in the case for this data, subjects have variable follow-up periods, and survival mode ls are capable of incorporating these different follow-up periods in the analyses (as opposed to other types of regr ession models). Each subject (transplant candidate) has a known candidate listi ng date as well as a la st follow-up date, which may be death or the last followup period. For these analyses deat h will be treated as the event of interest and patients that do not die over the st udy period will be censored at the last follow-up period. This form of censoring is commonly re ferred to as right-cen soring. Models for secondary outcomes will be censored at the time of transplantation and additionally an alternative model will use th e time of transplantation as the origin point.

PAGE 44

44 Analyses for examining factors associated with patient outcomes will use a survivor function S(t), which gives the probab ility of survival until a given time, t. S(t) is a monotonically decreasing function which is init iated at a survival probability of one and decreases with subsequent events over time as i ndicated graphically on the horizont al axis. The rate of decline of the survival function is utilized as a measure of the risk at a particular event time and is generally regarded as the hazard function (t). The hazard function can generally be written as shown in Equation 3-1. (t) = d/dt ( ln [S(t)]) (3-1) The analyses will also utilize Kaplan-Meier methodology (also known as the productlimit method) to compare the survival functi on among sample strata. This methodology is nonparametric and is appropriate when exact date s are known for events (rather than aggregated across intervals), which is the case with the curre nt data. Comparisons of the survival function between study groups will be made with the LogRank test which is generally the statistically most powerful test for making unadjusted compar isons between study groups. Censoring will be assumed to be noninformative as last follow-up in formation is generally a product of the most recent data available and there are few cases of patients that are lost to follow up in this cohort. To compare outcomes between study groups ad justed for potential confounding factors, hazard functions will be compared utilizing C ox proportional hazard models. Cox models are semi-parametric and are not reliant on a specifi c distributional form of a survival function supporting the robust nature of the results. An assumption that is applicable for the Cox model is that the hazard ratio of covari ates in the model is proportional over time (i.e., additive changes in covariates cause multiplicative changes in the hazar d functions). For continuous covariates this will be tested explicitly by enteri ng an interaction term of the covariate in question with time into

PAGE 45

45 the model. In the case of categorical covari ates, assessment of propor tionality will be made by visual inspection of the log-log survival function. The Cox model can be written in the form depicted in Equation 3-2. hi(t) = ho(t) exp( 1xi1 + 2xik + + kxik) (3-2) where i = individual, t = time, x = covariate of interest, and k = number of covariates The Wald test will be utilized to test the two-sided hypothesi s that a center characteristic is significantly associated with patient mortality (Ho: 1 = 0 versus Ha: 1 0) using the equation z = 1/(SE)1. In addition, results will be displayed as adjusted hazard ratios relative to a reference level of the center factors. Cox models will also be used to estimate adjusted survival rates following candidate listing for combinations of center-level variables. These results will be presented in tabular form with applic able standard error estimates. Study Aim I Kaplan-Meier and multivariate Cox models will be used to test the hypothesis that the individual center characteristics are associated with wait-list ca ndidate survival. Kaplan-Meier models will be constructed for each individual cent er characteristic based on quintile levels of the variable utilized as strata in the model. Re sults will be displayed graphically and tested with Log-Rank tests. Cox models will be used incorp orating all center charact eristics simultaneously along with other potential confoundi ng factors as described previ ously. Overall significance, hazard ratios, and confidence intervals for cente r characteristics will be displayed in tabular form. Study Aim II Cox models will be utilized to estimate survival rates for candidates at multiple permutations of center characteristic levels. On ly center characteristics that were shown to be significantly associated with candidate mortality w ill be utilized for this study aim. The baseline

PAGE 46

46 hazard function of the model will be adjusted with incremental levels of the center factors that span the range of these variables, and all surv ival estimate combinations will be displayed in tabular form. Study Aim III In order to evaluate the effect of center char acteristics in three highrisk candidate groups, the methodology described in study aims I and II will be utilized strictly in subsets of candidates. The same explanatory variables will be utilized in Cox models wit hout the variable applicable to the designation of the high-risk subset. For models restricted to the elderly, age categories will be reclassified within the el derly cohort. With the subset of African-American patients, candidate race will be removed fr om the models. For the obese cohort, new categorization of BMI will be used for patients with BMI > 30 kg/m2. Survival estimates, as described in study aim II, will be conducted for each subset of high-risk patient in a similar fashion. Potential Selection Bias The primary aim of the study is to eval uate the independent effect of center characteristics on transplant candidate mortalit y. However, one of th e limitations of this retrospective analysis will be the possibility that mortality rates at centers are correlated with the patients selection of these cente rs. In particular, patients that are well informed, have higher educational background, or have res ources that allow them to travel to centers of their choice or make informed decisions about centers that are of higher quality, and these patients may also have a lower risk profile. There is data to support the notion that performance measures do not influence transplant center selection in aggregat e; however, this may not fully incorporate more subtle patient characteristics (21). The effect of this bias may be to overestimate the influence of center characteristics, assuming that better pa tients select centers that also have better outcomes.

PAGE 47

47 The primary strategy of the study to account for this potential bias will be to statistically control for patient characteristics that may influe nce center selection as we ll as mortality. These patient-level variables that will be adjusted for in the outcome m odels will include age, primary diagnosis, race, gender, primary insurance coverage and education level. These factors will be assumed to provide adequate control for charac teristics of patients a ssociated with center selection related to overall health level, educ ation, affluence level and interaction with the healthcare system. In addition, examination of the secondary outcome of candidate mortality prior to transplantation will provide insight on the degree of selection bias that exists. In particular, mortality rates prior to transplanta tion are significantly less associated with care provided by the transplant center, and significant differences in pretransplant mortality are more likely attributable to patient characteristics. Evaluation and compar ison of preand posttransplant mortality effects by center characteristic s will therefore serve as a verification of any observed associations.

PAGE 48

48 P e r c e n t 0 5 10 15 20 25 30 35 Annual Number of Deceased Donor Transplants by Center 020406080100120140160180200 Figure 3-1. Distribution of the Annual Numb er of Deceased Donor Transplants by Center

PAGE 49

49 P e r c e n t 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 Proportion of ECD Transplants by Center 048121620242832364044 Figure 3-2. Distribution of the Pr oportion of ECD Transplants by Center

PAGE 50

50 P e r c e n t 0 2 4 6 8 10 12 Proportion of Patients Transplanted by Center 0102030405060708090100 Figure 3-3. Distribution of the Proportion of Candidates Receiving a Deceased Donor Transplant within Three Years by Center

PAGE 51

51 P e r c e n t 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 Performance Ratios by Center 0.00.51.01.52.02.53.03.54.0 Figure 3-4. Distribution of Standardized Mortality Ratios by Center

PAGE 52

52 CHAPTER 4 RESULTS Study Population There were 108,928 adult solitary kidney transp lant candidates in the study population. Table 4-1 displays descriptive statistics abou t the candidate population from 1995. Within the study period, 8% of candidates were 65 years or older, 59% of candidates were male, 28% were African-American, 28% had diabetes as a primary cause of ESRD, 15% of candidates had a college degree or more, 48% had type-O blood, a nd 55% of candidates had Medicare as their primary insurance payer. Among candidates with known BMI levels, 21% were obese and among candidates with known Panel Reactive Antibody (PRA) level at listing, 31% were sensitized (i.e., PRA > 0). Rate of Transplantation by Center Transplant centers for each listing year we re categorized based on the proportion of candidates that received a deceased donor transplant within three years after being placed on the waiting list. Center categories for this proportion were assigned by quintiles as follows: Q1 = [7.0.5], Q2 = [38.6.6], Q3 = [51.7.1], Q4 = [63.2.7], a nd Q5 = [76.8.0]. Table 4-2 displays candidate charac teristics by the cente r proportion of transplants categorized by quintile. The most notable differences by cen ter category were among patients with private primary insurance coverage (Q1=45% vs. Q5=40 %), African-American recipients (Q1=32% vs. Q5=21%), and sensitized patients (Q1=30% vs. Q5=39%). The number of candidates and the median proportion of patients reaching transplant ation at three years for candidates in each group were as follows: Q1 (n=30478, median = 28.4), Q2 (n=24271, median = 45.4), Q3 (n=22768, median = 58.1), Q4 (n=18858, median = 67.3), and Q5 (n=12553, median = 83.0).

PAGE 53

53 Transplant Center Volume Transplant centers for each listing year were categorized based on the average number of deceased donor transplants three years prior to th e year of candidate lis ting. Center categories for volume were assigned by quintiles as follows: Q1 = [10.0.8], Q2 = [18.9.3], Q3 = [26.4.6], Q4 = [36.7.7], and Q5 = [53.8.0] Table 4-3 displays candidate characteristics by center volume quintile. The mo st notable differences by center category were among patients with a college degr ee (Q1=17% vs. Q5=22%), and pa tients with private primary insurance coverage (Q1=36% vs. Q5=45%). The number of candidates and median annual transplant volume for candidates in each group were as follows: Q1 (n=8462, median = 15.7), Q2 (n=12597, median = 23.0), Q3 (n=16433, medi an = 31.0), Q4 (n=24370, median = 44.0), Q5 (n=47066, median = 88.0). Center Donor Quality Transplant centers for each listing year we re categorized based on the proportion of expanded criteria donor tran splants three years prior to the y ear of candidate listing. Center categories for the ECD proportion were assigned by quintiles as follows: Q1 = [0.4], Q2 = [6.5.9], Q3 = [10.0.4], Q4 = [ 13.5.9], Q5 = [19.0.0]. Table 4-4 displays candidate characteristics by center ECD propor tion quintile. The most nota ble differences by center ECD proportion category were among patients with priv ate primary insurance coverage (Q1=39% vs. Q5=43%), elderly patients (Q1=7% vs. Q5= 10%), African-American patients (Q1=25% vs. Q5=32%), sensitized patients (Q1=36% vs. Q5= 27%), diabetic patients (Q1=30% vs. Q5=26%), and obese patients (Q1=19% vs. Q5=23%). The number of candida tes and the median proportion of ECD transplants for candidates in each group were as follows: Q1 (n=17980, median = 4.4), Q2 (n=19229, median = 8.3), Q3 (n=22768, median = 11.8), Q4 (n=24139, median = 15.7), Q5 (n=24812, median = 24.0).

PAGE 54

54 Center Performance Ratings Transplant centers for each listing year were categorized based on the ratio of observed to expected deaths at one year utiliz ing data for recipients three year s prior to the year of candidate listing. Center categories for volume were assi gned by quintiles as follows: Q1 = [0.56], Q2 = [0.57.83], Q3 = [0.84.09], Q4 = [1.10.45], Q5 = [1.46.80]. Table 4-5 displays candidate characteristics by center performance quintile. The most notable differences by center performance category was among patients with priv ate primary insurance coverage (Q1=43% vs. Q5=35%), African-American patients (Q1=25% vs. Q5=34%) and patients with a college degree (Q1=21% vs. Q5=18%). The number of candidates and the median ratio of observed to expected deaths for deceased donor transplants by candidates in each group were as follows: Q1 (n=17939, median = 0.37), Q2 (n=23865, median = 0.72), Q3 (n=27705, median = 0.96), Q4 (n=23972, median = 1.26), Q5 (n=15447, median = 1.70). Association between Transplant Center Characteristics The median center volume and performance rati ngs were relatively stable over the period (Table 4-6). In contrast, the proportion of patie nts transplanted at th ree years significantly declined throughout the study period and the proportion of ECD tran splants increased over time. Also indicated in the table, the total number of candidates increased steadily over the study period. There was no significant correlation between the volume of deceased donor transplants with the other three center characteristics invest igated in the study (Tab le 4-7). There was a significant negative association between the propor tion of patients transp lanted at three years with the proportion of ECD transplants; however, th e proportion of patients transplanted at three years was not correlated with performance ratios. The ratio of observe d to expected patient deaths at one year was positively correlated with the ECD proportion between centers. Centers

PAGE 55

55 with a higher proportion of ECD transplants had le ss patients transplanted at three years and worse (i.e., higher ratios) performance ratings. Reliability of Historical Center Characteristics The volume of transplants was positively associ ated with several characteristics of the centers at listing (Table 48). On average, listing for a transpla nt at a center with higher volume was positively associated with the volume of transplants at the center at the time of transplantation. Candidates who lis ted at centers with historica lly smaller number of patients transplanted also had longer dur ations between listing and receivi ng a transplant. In a similar fashion, patients who listed at cen ters with historically higher pr oportion of ECD transplants also were more likely to be transplanted at centers with a higher proportion of ECD transplants. In addition, recipients that li sted at centers with bett er performance ratios also were transplanted at centers with better ratios during the year of tran splantation. In genera l, characteristics of transplant centers at the time of listing were still evident at th e time of transplantation. Kaplan-Meier Candidate Survival by Center Characteristics Figure 4-1 displays candidate survival by quintile level of the proportion of patients transplanted at three years prio r to listing. There was a step wise and significant association between a higher proportion of patients transplant ed and higher candidate survival over the study period (p <0.001). The proportion of candidates surviving at ten years following listing was 56% at centers with the highest proportion of patients reaching transplantation as compared to 50% at centers with the lowest proportion. Ten-year surv ival at centers with lowest proportion of ECD transplants was 54% as compared to 50% at hi gh ECD centers (Figure 4-2). Time to death following listing by transplant volume was not sta tistically significant different (Figure 4-3). The proportion of candidates surviv ing at ten years based on cent er performance ratios, ranged from 51% in the lowest-performing and highest-pe rforming centers respectivel y (Figure 4-4).

PAGE 56

56 Multivariate Cox Model for Primary Outcome of Candidate Mortality The primary outcome of the study was ca ndidate mortality after listing for transplantation. Results of the Cox proportional hazard model for patient mortality including the center study characteristics categor ized by quintile level are displa yed in Table 4-9. Relative to Caucasian candidates, all other racial groups ha d significantly lower mortality after listing. Older age was significantly associated with incr eased mortality including a two-fold risk for candidates over the age of 65 years (adjuste d hazard ratio [AHR] = 3.24, 95% confidence interval [C.I.] 3.13.35). Patients with a primary diagnosis of di abetes had an approximate 88% increase in the hazard ratio for death relative to patients without the diagnosis. Patients with type-A and type-AB blood had a reduced hazard rati o for death relative to type-O candidates. Patients that were listed as highly sensitized (P RA > 30) had a 34% increase hazard ratio for death after listing relative to non-sens itized candidates. Low BMI levels ( 18 kg/m2) and BMI levels above 35 had an increased hazard ratio for death relative to candidates with BMI between 19 kg/m2. Patients with less than a college educati on had an increased lik elihood of mortality (AHR=1.08, 95% C.I. 1.05.12). Candidates with Medicare as a primary insurance payer had significantly elevated mortality relative to candidates with private insurance. Candidates on dialysis at the time of listing had an elevated h azard for death relative to candidates that were listed prior to di alysis initiation. Among center characteristics, the rate of tran splantation had the st rongest association with candidate mortality after list ing. Candidates that listed at cen ters with the lowest percentage of patients reaching transplanta tion at three years had a 32% in creased relative hazard for death (AHR = 1.32, 95% C.I. 1.27.38) rela tive to candidates listing at centers with the highest percentage of patients reaching transplantation. In addition, th ere was a stepwise association between levels of the center proportion of pa tients transplanted. Transplant volume was not

PAGE 57

57 significantly associated with mo rtality between any of the groups. The center proportion of ECD transplants was associated with candidate mortalit y between candidates listed at centers with the highest proportion of ECD transplants (AHR = 1.04, 95% C.I. 1.00.08) relative to candidates listed at centers with the lowest proportion of ECDs. Candidates lis ted at centers with the lowest performance measures had a 14% increase in hazard ratio for death (AHR = 1.14, 95% C.I. 1.10 1.19) relative to candidates listed at cent ers with the highest performance. The Cox proportional hazard model results for the outcome of receiving a deceased donor transplant following listing is displayed in Table 4-10. There wa s a highly significant association of receipt of transplant by centers wi th historically different proportions of candidates receiving transplants. The hazard for patients re ceiving a transplant at ce nters with the lowest rate was over 5-fold decreased (AHR = 0.17, 95% C.I. 0.16.17) relative to centers with highest rates of transplantation. There was also a stepwise and significant increase in the hazard ratio by center level. Center volume was also associated w ith the hazard rate of transplant with the most notable difference between candidate s at the smallest centers less likely to receive a transplant (AHR=0.94, 95% C.I. 0.90.98) relative to candidates at the highest volume centers. There was a varied association between cen ter levels of ECDs, with candi dates at centers with a high proportion of ECD transplants more likely to receive a transplant (AHR = 1.10, 95% C.I. 1.07 1.14). Candidates at different performance cen ters also had mildly different time to transplantation; however, there was no distinct pattern between levels. The association of center charac teristics with post-transplant mortality for candidates who acquired a deceased donor transplant within the study period is displayed in Table 4-11. There was no significant association between hazard ratios for post-transplant mortality with candidates listed at centers with different rates of transp lantation or proportion of ECD transplants.

PAGE 58

58 Candidates listed at centers with small vol ume (AHR=1.10, 95% C.I. 1.03.18) had an elevated hazard ratio for post-transplant mortality; howev er this association was not statistically significant among candidates listed at the smallest centers. Candidate s listed at centers with the lowest performance levels had a 20% increased hazard for death (AHR= 1.20, 95% C.I. 1.11 1.29). The model for post-transp lant graft loss resulted in similar findings. There was no significant association between graf t loss and candidates listed at cen ters with different rates of transplant, there was mixed results for center volum e, and candidates listed at centers with lower performance levels had elevated post-transplant graft loss (Table 4-12). One distinction in for the outcome of graft loss, as compared to patien t mortality, was that candidates listed at centers with highest ECD transplants had a significantly elevated hazard for post-transplant graft loss (AHR=1.08, 95% C.I. 1.02.15) relative to candidate s listed at centers with a lowest ECD proportion. In addition, the model excluding char acteristics of the donor organ (e.g., donor age and ECD) indicated that centers with high ECD levels were more significantly associated with post-transplant mortality (AHR=1.14, 95% C.I. 1.07.23) and graft loss (AHR=1.18, 95% C.I. 1.12.25). Center performance level had a significant and stepwise associat ion with candidate mortality prior to transplantation (Table 4-13). Sp ecifically, candidates that listed at centers with lower historical performance levels had highe r mortality rates prior to transplantation. Candidates that listed at cente rs with the lowest performan ce levels had a 13% (AHR = 1.13, 95% C.I. 1.08.18) elevated hazard for death relative to candidates that listed at centers with the highest historical performance levels. Alternativ ely, candidates listed at centers with historically different rates of transplantat ion, center volume or ECD proportion had minimal association with pre-transplant mortality.

PAGE 59

59 Outcomes among High-Risk Candidate Groups Models were repeated for candidate mortalit y limited to subsets of high-risk patients. These models were generated for African-Ameri can, elderly, and obese candidates adjusted for the same covariates with the exception of the vari able describing the high-ri sk characteristic. In addition, models including interact ion terms for the center factors and high-risk characteristics were generated and hypotheses te sted for whether high and low le vels of center factors were more important for high-risk groups rela tive to their candi date counterparts. Among African-American candidates, the rate of transplantat ion had a highly significant and stepwise association with mortality (Table 4-14). Candidates listed at centers with the lowest transplant rates had a 28% increased hazard for death (AHR=1.28, 95% C.I. 1.18.38) over the study period. African-American candidates listed at centers with varying volume levels had no significant difference in mortality. AfricanAmerican candidates listed at centers with the highest proportion of ECD transplants had a sign ificantly elevated hazard for death (AHR=1.09, 95% C.I. 1.02.16) relative to candidates listed at centers with a low proportion of ECD transplants. African-American ca ndidates listed at centers with th e lowest performance ratio also had an increased hazard for death (AHR = 1.14, 95% C.I. 1.06.22) relative to candidates listed at centers with the highest performance ratio. The model including interaction terms betw een center characteristics and race groups (limited to African-American and Caucasians) indi cated that the associa tion of center effects with mortality was not different by race. The difference between centers with a high and low rate of transplantation (p= 0.42), high and low volume centers (p=0.83), high and low ECD centers (p=0.48) and high and low performance ra tio (p=0.84) was not different between the two candidate race groups.

PAGE 60

60 The rate of transplantation had a highly significant and stepwise association with candidate mortality among elderly candidates (Table 4-15). Elderly candidates listed at centers with the lowest transplant rates had a 26% increased hazard for death (AHR=1.26, 95% C.I. 1.13.40) over the study period. There was not an association between center volume with mortality for elderly candidates. The proporti on of ECD transplants was associated with mortality with those elderly can didates listed at centers with a mid-level of ECDs having a reduced hazard for death (AHR=0.90, 95% C.I. 0.82.99) relative to candidates listed at centers with a high proportion of ECDs. El derly candidates listed at centers with the lowest performance ratio also had a significantly increased hazard for death (AHR = 1.23, 95% C.I. 1.11.36) relative to candidates listed at centers with the highest performance ratio. The model including interaction terms betw een center characteri stics and age groups (limited to 18 and 65+ age groups) indicated that the proportion of ECD transplants was significantly more important in younger age groups than in the elderly (AHR=1.07, 95% C.I. 1.01.13). Other center factors did not demonstrat e a significantly different impact in younger versus older patients: rate of transplant (p=0.41), transplant volume (p=0.36), and performance ratio (p=0.61). As displayed in Table 4-16, the rate of tr ansplantation had a highly significant and stepwise association with candida te mortality for obese candidate s. Obese candidates listed at centers with the lowest transplant rates had a 33% increased hazard fo r death (AHR=1.33, 95% C.I. 1.22.45) relative to candidates listed at centers with the highe st transplant rates. There was no association between center volume and candidate mortality for obese candidates. The proportion of ECD transplants wa s also not significantly associ ated with mortality. Obese candidates listed at centers with the lowest pe rformance ratio had a significantly increased

PAGE 61

61 hazard for death (AHR = 1.11, 95% C.I. 1.01.21) rela tive to candidates listed at centers with the highest performance ratio. There was no signif icant differences in the effect of center listing for volume (p=0.27), center performance level (p =0.63), transplant rate (p=0.87), or center proportion of ECD transplant s (p=0.13) between obese and non-obese patients. Expected Survival by Center Characteristics Expected survival estimates displayed on Table 4-17 indicate the average expected survival based on the level of the center charac teristic indicated holdi ng all other center and candidate characteristics at thei r average level. Results indica te that a candidates expected survival is markedly different by levels of center levels of the proportion of patients transplanted at three years. Average candidates listed at cente rs with 82% of patients transplanted at three years have almost 2.5 years increased expected surviv al relative to patients listed at centers with 10% of patients transplanted with in three years. Differences in expected survival between extreme center volume levels were mildly diffe rent ranging from 10.4 to 10.9 years. Candidates listed at centers with few ECD transplants had an expected survival of 10.8 years as compared to candidates listed at centers with 32% of ECD transplants, who ha d an expected survival of 10.5 years on average. Candidates liste d at centers with low performan ce ratings (i.e., high standard mortality ratios) had a 10.1 year expected survival as compared to candidates listed at centers with high performance, who had an exp ected survival of 11.2 years. Table 4-18 displays expected survival rates for candidates at hypothetical combinations of center characteristics. The case examples dem onstrate differences in expected survival across ranges of the center pro portion transplanted and by extreme le vels of the other three center characteristics. In particular, the estimates demonstrate the differences in life expectancy relative to the proportion transplanted with either the w orst combination of other factors as compared to the best combination of th e other three factors. Among cen ters with the lowest proportion

PAGE 62

62 of candidates reaching transplantation, the varia tion in average life expe ctancies ranged from 9.3 to 10.3 years. Centers with a mid level of prop ortion transplanted and the worst combination of additional factors had an av erage life expectancy of 10.2 years as compared to the best combination of additional factors at 11.6 years. Average life expectancy was notably higher within centers with high proporti on of transplant candidates, ra nging from 11.5 to 13.1 years. The expected survival for African-American patients following listing ranged from 9.5 years to 11.5 years at centers with a historically low versus high proportion of candidates reaching transplant at three years respectively (Tab le 4-19). The variation in life expectancies varied little by center vo lume. The average life expectancy ranged from 9.6 to 10.3 years by centers with a high and low proportion of ECD tr ansplants respectively. Similarly, average life expectancies for African-Americans ranged from 9.6 to 10.6 years based on the historical center performance level with higher performance centers associated with longe r life expectancies. The center characteristic with the largest rang e in life expectancies for elderly candidates was associated with the center transplant rate (Table 4-20). Specifi cally, elderly candidates listed at centers with the highest proportion of candidates transp lanted at three years had an expected 6.5 years of life expectancy as compar ed to elderly candidates at centers with the lowest proportion of candidates reaching transplantation, the life expectancy was 5.6 years. There was mild fluctuation in life expectancy by center volume for elderly candidates, but no notable pattern among levels. In a similar fash ion, life expectancy for elderly candidates was highest associated with listing at centers with mid-levels of ECD transplants; however, the life expectancy between low and high centers was sim ilar. Average life expectancies by levels of center performance ratios ranged from 6.1 years in the highest performance centers to 5.3 years in the lowest performance centers.

PAGE 63

63 Average expected life years fo r obese candidates at centers with a low transplant rates was 8.2 as compared to 10.2 years for candidates lis ted at centers with the highest proportion of candidates reaching transplantation (Table 4-21). Life expectancies varied minimally based on center volume with no detectable pattern for an a ssociation of longer life expectancy with higher of lower volume. Expected life years for obese candidates listed at centers with a low ECD proportion was 9.0 years as compared to candidate s listed at centers w ith a high ECD proportion with 8.7 years. Average life expectancy after list ing also varied at centers based on performance level; candidates at the highest performing centers had 9.1 expected life year s as compared to 8.4 years at the lowest performing centers.

PAGE 64

64 Table 4-1. Transplant ca ndidate characteristics Candidate characteristic Level % Age at listing 18 43 45 28 55 21 65+ 8 Race Caucasian 54 African-American 28 Asian 5 Hispanic 11 Other 2 Gender Male 59 Female 41 Primary cause of ESRD Diabetes 28 Other 72) Candidate BMI 13 2 19 59 30+ 16 Missing 23 Candidate education level Less than college degree 60 College degree or more 15 Missing 25 Blood type O 48 A, B or AB 52 Candidate primary insurance Private 42 Medicare 55 Other/missing 3 Candidate peak PRA level 0 63 1 19 30+ 10 Missing 8 Sample Size 108928

PAGE 65

65 Table 4-2. Candidate characteristics by center proportion of transplant s within three years Center rate of transplantation category Candidate characteristic Level Q1Q2Q3 Q4 Q5 Gender Male (%) 595960 59 59 Blood type Type-O (%) 484748 49 48 Education College degree (%) ^ 212219 20 18 Primary insurance Private (%) 454238 40 40 Age 65+ (%) 898 8 8 Race African-American (%) 323029 24 21 PRA level > 0 (%) ^ 302929 34 39 Primary diagnosis Diabetes 282827 28 27 BMI 30+ (%) ^ 222221 21 21 Categories represent the quintiles of the pr oportion of patients rece iving a transplant by three years: Q1 = [7.0 38.5], Q2 = [38.6.6], Q3 = [51.7 .1], Q4 = [63.2.7], Q5 = [76.8.0]. ^ The proportions for these categories exclude missing levels. Table 4-3. Candidate characteri stics by center volume category Center volume category Candidate characteristic Level Q1Q2Q3 Q4 Q5 Gender Male (%) 595959 59 59 Blood type Type-O (%) 484748 48 48 Education College degree (%) ^ 172020 19 22 Primary insurance Private (%) 363639 41 45 Age 65+ (%) 988 8 8 Race African-American (%) 232929 33 26 PRA level > 0 (%) ^ 293032 32 32 Primary diagnosis Diabetes 272727 28 28 BMI 30+ (%) ^ 222222 22 21 The categories represent the quintile levels of deceased donor transplant volume: Q1 = [10.0.8], Q2 = [18.9.3], Q3 = [26.4.6], Q4 = [36.7.7], Q5 = [53.8.0]. ^ The proportions for these categor ies exclude missing levels.

PAGE 66

66 Table 4-4. Candidate characterist ics by center ECD proportion category Center ECD proportion category Candidate characteristic Level Q1Q2Q3 Q4 Q5 Gender Male (%) 585959 59 60 Blood type Type-O (%) 484848 48 47 Education College degree (%) ^ 202219 21 20 Primary insurance Private (%) 394140 43 43 Age 65+ (%) 778 8 10 Race African-American (%) 252727 29 32 PRA level > 0 (%) ^ 363233 30 27 Primary diagnosis Diabetes 302828 27 26 BMI 30+ (%) ^ 192122 21 23 The categories represent the quin tile levels of the proportion of ECD transplants: Q1 = [0 6.4], Q2 = [6.5.9], Q3 = [10.0.4], Q4 = [13.5.9], Q5 = [19.0.0]. ^ The proportions for these categor ies exclude missing levels. Table 4-5. Candidate characterist ics by center performance category Center performance category Candidate characteristic Level Q1Q2Q3 Q4 Q5 Gender Male (%) 595859 60 59 Blood type Type-O (%) 484848 48 48 Education College degree (%) ^ 212222 19 18 Primary insurance Private (%) 434344 40 35 Age 65+ (%) 888 9 9 Race African-American (%) 252728 29 34 PRA level > 0 (%) ^ 303034 32 30 Primary diagnosis Diabetes 272728 28 26 BMI 30+ (%) ^ 212021 22 23 The categories assigned by quintil e of standard mortality ratios (observed/expected deaths): Q1 = [0.56], Q2 = [0.57.83], Q3 = [0.84.09], Q4 = [1.10.45], Q5 = [1.46.80]. ^ proportion excludes missing levels Table 4-6. Median levels of center characteristics over time 1995199619971998 1999 2000 Volume 45.746.047.747.0 49.3 48.0 Proportion transplanted 61.556.654.349.1 47.7 45.5 ECD proportion 8.410.413.114.0 14.8 14.7 Performance ratio 0.980.980.910.94 0.96 0.92 Candidate listings 16412168361749018685 19270 20235

PAGE 67

67 Table 4-7. Correlation coefficients between center characteristics Linear correlation (pvalue) Volume Proportion transplanted ECD proportion Performance ratio Volume 0.01 (0.71)0.04 (0.24) -0.05 (0.09) Proportion transplanted 0.01 (0.71) -0.21 (<0.001) ECD proportion 0.04 (0.24)-0.21 (<0.001) 0.08 (0.01) Performance ratio -0.05 (0.09)0.04 (0.19)0.08 (0.01) Table 4-8. Center characteristic s at the time of transplantation Median center characteristics at the year of transplantation Level at the time of listing Volume Time to transplant (months) ECD proportion Performance ratio* Q1 20 36 9 0.75 Q2 28 27 11 0.79 Q3 34 21 13 0.83 Q4 46 17 16 1.08 Q5 93 12 23 1.13 based only on patients transplanted duri ng the year candidate was transplanted

PAGE 68

68 Table 4-9. Adjusted hazard ratios for patient mortality after listing for transplantation Candidate characteristic (reference level) Level Adjusted hazard ratio 95% confidence interval Candidate race (Caucasian) African-American 0.91 0.89 0.93 Asian 0.59 0.56 0.63 Hispanic 0.71 0.69 0.74 Other 0.79 0.73 0.85 Candidate age (184) 4554 1.73 1.69 1.78 5564 2.43 2.36 2.50 65+ 3.24 3.13 3.35 Primary diagnosis (Non-Diabetic) Diabetes 1.88 1.84 1.92 Blood type (Type O) A 0.93 0.91 0.95 AB 0.86 0.82 0.91 B 1.02 0.99 1.05 PRA level (Zero) 130 1.07 1.04 1.10 30+ 1.34 1.30 1.39 Missing 1.43 1.38 1.48 BMI (1925) 1318 1.28 1.19 1.38 2630 0.96 0.93 0.98 3135 1.00 0.97 1.04 3640 1.17 1.11 1.23 41+ 1.22 1.13 1.33 Missing 0.99 0.96 1.02 Education level (College) Less than College 1.08 1.05 1.12 Missing 1.14 1.10 1.18 Primary insurance (Private) Medicare 1.34 1.31 1.37 Other 0.35 0.20 0.63 Missing 1.01 0.95 1.08 Dialysis status at listing (None) Hemodialysis 1.36 1.31 1.40 Peritoneal dialysis 1.39 1.34 1.45 Unknown 1.43 1.31 1.56 Proportion transplanted (Highest Q5) Lowest Q1 1.32 1.27 1.38 Low Q2 1.25 1.20 1.30 Mid Q3 1.17 1.12 1.21 High Q4 1.17 1.12 1.21 Center volume (Largest Q5) Smallest Q1 1.00 0.96 1.04 Small Q2 1.04 1.00 1.07 Mid Q3 0.99 0.96 1.02 Large Q4 0.98 0.95 1.00 ECD proportion (Lowest Q1) Low Q2 1.02 0.98 1.05 Mid Q3 1.00 0.96 1.03 High Q4 1.00 0.96 1.03 Highest Q5 1.04 1.00 1.08 Performance Ratio (Best Q1) Good Q2 1.02 0.99 1.06 Mid Q3 1.07 1.04 1.11 Bad Q4 1.09 1.05 1.13 Worst Q5 1.14 1.10 1.19

PAGE 69

69 Table 4-10. Adjusted hazard ratios for re ceipt of transplant following listing Candidate characteristic (reference level) Level Adjusted hazard ratio 95% confidence interval Proportion transplanted (Hi ghest Q5) Lowest 0.17 0.16 0.17 Low 0.31 0.29 0.32 Mid 0.46 0.45 0.48 High 0.59 0.57 0.61 Center volume (Largest Q5) Smallest 0.94 0.90 0.98 Small 0.97 0.94 1.00 Mid 1.02 0.99 1.06 Large 0.98 0.95 1.00 ECD proportion (Lowest Q1) Low 1.07 1.04 1.11 Mid 1.09 1.05 1.12 High 1.10 1.07 1.14 Highest 1.02 0.99 1.06 Performance ratio (Best Q1) Good 0.95 0.92 0.98 Mid 0.93 0.90 0.96 Bad 0.95 0.92 0.98 Worst 0.97 0.94 1.01 Model additionally adjusted for candidate race, age, primary diagnosis, blood type, PRA level, BMI level, education level, primary insu rance payer, and dialysis at the time of listing. Table 4-11. Adjusted hazard ratio s for post-transplant mortality Candidate characteristic (reference level) Level Adjusted hazard ratio 95% confidence interval Proportion transplanted (Hi ghest Q5) Lowest 0.970.90 1.04 Low 1.010.94 1.08 Mid 1.010.95 1.08 High 1.030.97 1.10 Center volume (Largest Q5) Smallest 1.050.97 1.14 Small 1.101.03 1.18 Mid 0.980.92 1.05 Large 1.061.00 1.11 ECD proportion (Lowest Q1) Low 1.050.98 1.12 Mid 1.040.98 1.11 High 1.050.99 1.13 Highest 1.060.99 1.14 Performance ratio (Best Q1) Good 1.010.95 1.09 Mid 1.081.01 1.15 Bad 1.161.08 1.24 Worst 1.201.11 1.29 Model additionally adjusted for recipient ra ce, age, primary diagnosis, PRA level, BMI level, education level, primary insurance paye r, pre-transplant dialysis time, donor age, expanded criteria donation, HLA-mismatching, cold ischemia time, and recipient and donor gender.

PAGE 70

70 Table 4-12. Adjusted hazard ratios fo r post-transplant overall graft loss Candidate characteristic (reference level) Level Adjusted hazard ratio 95% confidence interval Proportion transplanted (Hi ghest Q5) Lowest 0.940.88 1.00 Low 0.970.92 1.02 Mid 0.980.93 1.03 High 1.020.97 1.07 Center volume (Largest Q5) Smallest 0.990.93 1.06 Small 1.131.07 1.19 Mid 1.051.00 1.11 Large 1.101.06 1.15 ECD proportion (Lowest Q1) Low 1.051.00 1.11 Mid 1.071.01 1.12 High 1.081.02 1.14 Highest 1.081.02 1.15 Performance ratio (Best Q1) Good 1.030.97 1.09 Mid 1.091.03 1.15 Bad 1.151.09 1.22 Worst 1.171.10 1.24 Model additionally adjusted for recipient ra ce, age, primary diagnosis, PRA level, BMI level, education level, primary insurance paye r, pre-transplant dialysis time, donor age, expanded criteria donation, HLA-mismatching, cold ischemia time, and recipient and donor gender. Table 4-13. Adjusted hazard ratio s for pre-transplant mortality Candidate characteristic (reference level) Level Adjusted hazard ratio 95% confidence interval Proportion transplanted (Hi ghest Q5) Lowest 0.990.94 1.04 Low 1.010.96 1.07 Mid 0.990.94 1.05 High 1.050.99 1.11 Center volume (Largest Q5) Smallest 1.010.96 1.06 Small 1.040.99 1.08 Mid 1.010.97 1.04 Large 0.960.93 0.99 ECD proportion (Lowest Q1) Low 1.010.97 1.05 Mid 0.980.94 1.02 High 0.970.93 1.01 Highest 1.000.96 1.04 Performance ratio (Best Q1) Good 1.030.99 1.07 Mid 1.071.02 1.11 Bad 1.061.02 1.10 Worst 1.131.08 1.18 *Model additionally adjusted for candidate r ace, age, primary diagnosis, blood type, PRA level, BMI level, education level, primary insu rance payer, and dialysis at the time of listing.

PAGE 71

71 Table 4-14. Adjusted hazard ratios for mo rtality for African-American candidates Candidate characteristic (reference level) Level Adjusted hazard ratio 95% confidence interval Proportion transplanted (Hi ghest Q5) Lowest 1.281.18 1.38 Low 1.221.12 1.32 Mid 1.141.05 1.24 High 1.181.09 1.29 Center volume (Largest Q5) Smallest 1.020.94 1.11 Small 1.040.97 1.11 Mid 1.000.95 1.06 Large 0.980.94 1.03 ECD proportion (Lowest Q1) Low 1.010.95 1.08 Mid 1.010.95 1.08 High 1.010.95 1.08 Highest 1.091.02 1.16 Performance ratio (Best Q1) Good 1.050.98 1.13 Mid 1.091.02 1.17 Bad 1.101.03 1.17 Worst 1.141.06 1.22 *Model additionally adjusted fo r candidate age, primary diag nosis, blood type, PRA level, BMI level, education level, primary insurance payer, and dialysis at the time of listing. Table 4-15. Adjusted hazard ratios fo r mortality for elderly candidates Candidate characteristic (reference level) Level Adjusted hazard ratio 95% confidence interval Proportion transplanted (Hi ghest Q5) Lowest 1.261.13 1.40 Low 1.231.10 1.36 Mid 1.181.06 1.31 High 1.141.02 1.26 Center volume (Largest Q5) Smallest 1.010.90 1.12 Small 1.050.95 1.15 Mid 0.940.87 1.03 Large 0.940.87 1.01 ECD proportion (Lowest Q1) Low 0.950.86 1.05 Mid 0.900.82 0.99 High 0.970.88 1.07 Highest 0.990.91 1.09 Performance ratio (Best Q1) Good 1.020.93 1.12 Mid 1.060.96 1.16 Bad 1.141.04 1.25 Worst 1.231.11 1.36 *Model additionally adjusted for candidate race, primary diagnosis, blood type, PRA level, BMI level, education level, primary insurance payer, and dialysis at the time of listing.

PAGE 72

72 Table 4-16. Adjusted hazard ratios fo r mortality for obese candidates Candidate characteristic (reference level) Level Adjusted hazard ratio 95% confidence interval Proportion transplanted (Hi ghest Q5) Lowest 1.331.22 1.45 Low 1.231.12 1.34 Mid 1.151.05 1.26 High 1.141.04 1.25 Center volume (Largest Q5) Smallest 0.960.88 1.06 Small 0.960.89 1.05 Mid 1.020.94 1.09 Large 0.950.89 1.01 ECD proportion (Lowest Q1) Low 1.020.94 1.10 Mid 0.980.91 1.06 High 1.010.93 1.10 Highest 1.050.97 1.14 Performance ratio (Best Q1) Good 0.950.88 1.04 Mid 1.020.94 1.11 Bad 1.070.98 1.16 Worst 1.111.01 1.21 Model additionally adjusted for candidate age, race, primary diagnosis, blood type, PRA level, education level, primary insurance payer, and dialysis at the time of listing. Table 4-17. Candidate life expect ancy (in years) after listing by levels of center characteristics Center characteristics* Q1 Q2 Q3 Q4 Q5 Proportion transplanted at three years 10.010.411.0 11.0 12.4 Center volume 10.710.410.8 10.9 10.7 ECD proportion 10.810.710.8 10.8 10.5 Performance ratio 11.211.010.6 10.5 10.1* categories assigned by quintile levels of center characteri stics: performance ratio (observed/expected deaths): Q1 = [ 0.56], Q2 = [0.57.83], Q3 = [0.84.09], Q4 = [1.10 1.45], Q5 = [1.46.80]; the proportion of ECD transplants: Q1 = [0.4], Q2 = [6.5.9], Q3 = [10.0.4], Q4 = [13.5.9], Q5 = [19.0.0]; th e proportion of patients receiving a transplant by three years: Q1 = [ 7.0.5], Q2 = [38.6.6], Q3 = [51.7.1], Q4= [63.2 76.7], Q5 = [76.8.0]; and deceased donor tran splant volume: Q1 = [10.0.8], Q2 = [18.9.3], Q3 = [26.4.6], Q4 = [36.7.7], Q5 = [53.8.0].

PAGE 73

73 Table 4-18. Life expectancy after listing at hypothetical center ch aracteristic levels Proportion transplanted* Volume** ECD proportion^Standardized mortality ratio^^ Expected survival Case #1 Low Low High High 9.3 Case #2 Low Mid Mid Mid 10.1 Case #3 Low High Low Low 10.3 Case #4 Mid Low High High 10.2 Case #5 Mid Mid Mid Mid 11.1 Case #6 Mid High Low Low 11.6 Case #7 High Low High High 11.5 Case #8 High Mid Mid Mid 12.5 Case #9 High High Low Low 13.1 *proportion of candidates receiving a deceased donor transplant at three years. **deceased donor transplant volume. ^proportion of ECD transplant s of all deceased donor transplants. ^^ratio of observed to expected one year patient deaths. Table 4-19. Life expectancy after listi ng for African-American candidates by center characteristic levels Center characteristics* Q1 Q2 Q3 Q4 Q5 Proportion transplanted at three years 9.59.910.410.1 11.5 Volume 10.09.810.110.2 10.1 ECD proportion 10.310.210.210.2 9.6 Performance ratio 10.610.210.09.9 9.6* categories assigned by quintile levels of center characteri stics: performance ratio (observed/expected deaths): Q1 = [0.56], Q2 = [0.57.83], Q3 = [0.84.09], Q4 = [1.10.45], Q5 = [1.46.80]; the proportion of EC D transplants: Q1 = [0.4], Q2 = [6.5.9], Q3 = [10.0.4], Q4 = [13.5.9], Q5 = [19.0.0]; the proportion of patients receiving a transpla nt by three years: Q1 = [7.0.5], Q2 = [38.6.6], Q3 = [51.7.1], Q4= [63.2.7], Q5 = [76.8.0]; and deceased donor transplant volume: Q1 = [10.0.8], Q2 = [18.9.3], Q3 = [26.4.6], Q4 = [36.7.7], Q5 = [53.8 195.0].

PAGE 74

74 Table 4-20. Life expectancy after listing for el derly candidates by center characteristic levels Center characteristics* Q1 Q2 Q3 Q4 Q5 Proportion transplanted at three years 5.65.75.8 6.0 6.5 Volume 5.75.66.0 6.0 5.7 ECD proportion 5.75.96.1 5.8 5.7 Performance ratio 6.16.05.9 5.6 5.3*Categories assigned by quintile levels of cen ter characteristics: performance ratio (observed/expected deaths): Q1 = [ 0.56], Q2 = [0.57.83], Q3 = [0.84.09], Q4 = [1.10 1.45], Q5 = [1.46.80]; the proportion of ECD transplants: Q1 = [0.4], Q2 = [6.5.9], Q3 = [10.0.4], Q4 = [13.5.9], Q5 = [19.0.0]; th e proportion of patients receiving a transplant by three years: Q1 = [ 7.0.5], Q2 = [38.6.6], Q3 = [51.7.1], Q4= [63.2 76.7], Q5 = [76.8.0]; and deceased donor tran splant volume: Q1 = [10.0.8], Q2 = [18.9.3], Q3 = [26.4.6], Q4 = [36.7.7], Q5 = [53.8.0]. Table 4-21. Life expectancy after listing for obese candidates by center characteristic levels Center characteristics* Q1 Q2 Q3 Q4 Q5 Proportion transplanted at three years 8.28.79.2 9.2 10.2 Volume 9.19.18.7 9.1 8.8 ECD proportion 9.08.99.1 8.9 8.7 Performance ratio 9.19.48.9 8.7 8.4*Categories assigned by quintile levels of cen ter characteristics: performance ratio (observed/expected deaths): Q1 = [ 0.56], Q2 = [0.57.83], Q3 = [0.84.09], Q4 = [1.10 1.45], Q5 = [1.46.80]; the proportion of ECD transplants: Q1 = [0.4], Q2 = [6.5.9], Q3 = [10.0.4], Q4 = [13.5.9], Q5 = [19.0.0]; th e proportion of patients receiving a transplant by three years: Q1 = [ 7.0.6], Q2 = [38.5.6], Q3 = [51.7.1], Q4= [63.2 76.7], Q5 = [76.8.0]; and deceased donor tran splant volume: Q1 = [10.0.8], Q2 = [18.9.3], Q3 = [26.4.6], Q4 = [36.7.7], Q5 = [53.8.0].

PAGE 75

75 Years Post-ListingSurvival (%)56% Q5 (Highest) 53% Q4 54% Q3 51% Q2 50% Q1 (Lowest) 10 Year Survival Proportion Transplanted Group Log-Rank p-value < 0.0010 2 468 1030 40 50 60 70 80 90 100 Figure 4-1. Kaplan-Meier plot of candidate survival by ce nter rate of transplant

PAGE 76

76 50% Q5 (Highest) 53% Q4 52% Q3 53% Q2 54% Q1 (Lowest) 10 Year Survival ECD Proportion Group Log-Rank p-value < 0.0010 2 468 10 Years Post-ListingSurvival (%)30 40 50 60 70 80 90 100 Figure 4-2. Kaplan-Meier plot of candidate survival by center proportion of ECD transplants

PAGE 77

77 53% Q5 (Highest) 52% Q4 53% Q3 52% Q2 52% Q1 (Lowest) 10 Year Survival Center Volume Group 02 468 10 Years Post-ListingSurvival (%)30 40 50 60 70 80 90 100 Log-Rank p-value = 0.73 Figure 4-3. Kaplan-Meier plot of ca ndidate survival by center volume

PAGE 78

78 02 468 10 Years Post-ListingSurvival (%)30 40 50 60 70 80 90 100 51% Q5 (Worst) 51% Q4 52% Q3 54% Q2 54% Q1 (Best) 10 Year Survival Performance Ratio Log-Rank p-value < 0.001 Figure 4-4. Kaplan-Meier pl ot of candidate survival by center performance ratio

PAGE 79

79 CHAPTER 5 DISCUSSION ESRD is a pervasive and growing public health concern in the United States. Both the cumulative number of ESRD patients and proportion of individuals with risk factors for future ESRD development have increased significantly over the recent era. The past two decades have also witnessed the growth and acceptance of ki dney transplantation as the most effective treatment modality for patients suffering from ESRD. Correspondingly, the number of patients selecting transplantation for treatment of ESRD ha s also significantly incr eased. The rapid rise in the number of candidates has increased waitin g times to acquire a tr ansplant and increased mortality among patients on the wai ting list for transplantation. In response, there have been significant efforts on the part of the transplant community to identify efficiencies in the transplant process, to increase don ation rates, and to develop strate gies to prolong the lifespan of donations while simultaneously maintaini ng equitable access to patients. The explanatory variables of primary interest in this study, the four center characteristics (waiting time, performance level, patient volume, and donor risk level), have generally been shown to be associated with outcomes for transpla nt recipients. However, the significance and magnitude of these effects for a prospective tr ansplant candidate has not been specifically evaluated. Furthermore, the inte gration of these factors to asse ss the relative effects provides a basis by which patients and caregivers can comp are potential centers of listing relative to the impact on candidate prognoses. In fact, the se lection of a transplant center is a common situation, but in many cases, these decisions may be made by default. Patients that are in this circumstance may often simply follow the advi ce of primary physicians who, for geographic purposes or personal preferences, refe r patients indiscriminately to a particular transp lant center. Thus, in the predominant number of cases, there is little active decision ma king on the part of the

PAGE 80

80 transplant candidate or referring physician as to which center is mo st appropriate for the patients needs or if alternative centers may in fact provide a different prognosis for that patient. This dissertation attempted to provide data to elucidate the impact of center selection that could potentially inform these decisions. That is, the questions remain, is ther e a marked difference for candidate outcomes (e.g., mortality or graft survival) based on the c hoice of center that they list for a transplant? Secondly, if th ere is a difference, what characte ristics of the centers are most important to candidate prognoses? Finally, are these characteristics generalizable to the entire candidate population or are they relatively more important to select gr oups that have various prognoses by treatment modality (i.e., dialysis or transplant)? The main findings of this study indicate that center characteristics are indeed significantly associated with patient outcomes. Furthermor e, the study demonstrates that these center characteristics vary in their re lative effect and that these eff ects somewhat differ in high-risk subsets of the candidate population. The most direct implication of the study is that candidates and referring physicians have an incentive to incor porate characteristics of transplant centers into decisions to select centers of care. Further research related to these findings and healthcare implications of the study results will be discussed in the proceeding section. For the purpose of simulating the circumstan ce that new onset transplant candidates must face in selecting a center, the study was specifically designed to utilize past levels of center characteristics rather than levels that were associated with the center at the time of transplantation. Clearly, candidate s will not be aware of the charac teristics of centers at the time that a donor organ is offered and must base decisi ons on levels prior to th e time of listing. This study utilized the aggregated center levels from th ree years prior to the time of candidate listing as a representation of these characteristics. In fact, the analysis demonstrated a significant

PAGE 81

81 association of historical levels of center characteristics with leve ls of these same characteristics at the time of transplantation. This is important information fo r prospective candidates to assess the reliability of center characteristics to assist with their selection of a center. The results indicate center factors in this st udy are relatively stable and can be utilized on a prospective basis for comparing centers. An important caveat to th is finding, however, is that as demonstrated in the study, waiting times for candidates to reach transplantation are continuing to expand. Thus, this greater time period is associated with a grea ter opportunity for center characteristics to alter. Based on this study period, center volume has be en relatively stable over time, waiting times have increased, and the utilizati on of ECD donations have increase d. It is also worthy of note that center performance ratios ar e fairly consistent over time. Th at is, performance ratios at a center at the time of listing are similar to the performance ratios at the time of transplant on average. The performance ratio is presented as a measure of quality of care, and it might be expected that, due to random events or changes in practice, this value might significantly vary over time. Results of the analys is suggest that listing at a cen ter with high performance is generally accompanied by a high performance level at the time of transplantation. The associations between center characteris tics and candidate outcomes found in this study were relatively consistent with the literature investigating the impact on transplant recipients. In particular, l onger waiting times, better performance ratings, and a lower proportion of ECD transplants all translated to superior outcomes for transp lant candidates from the time of listing. However, the degree of these associat ions and the relative importance of these characteristics are of particular importance in this study. The historical proportion of candidates that reach transplantation at three years demonstrated a hi ghly significant and dose-response relationship with outcomes for candida tes. Candidates listed at centers with the slowest rate of

PAGE 82

82 transplantation had a 32% elevated hazard for death over the study period relative to candidates listed at centers with the most rapid rate of tran splantation. This finding is perhaps not surprising, as there are two clear ex planations for this association. One is that transplantation has been demonstrated to convey a significant survival benefit over the alternative treatment modality of maintenance dialysis which is typi cally initiated prior to candidate listing. Therefore, a portion of this effect is likely attributable to the fact that candidates at centers with longer waiting times accumulate relatively hi gher cases of mortality prior to reaching transplantation. In addition, the cumulative effects of dialysis have a significant impact on posttransplantation mortality. Ther efore, those candidates that re ach transplantation with longer waiting times and exposure to dialysis also have poorer outcomes. Cumulatively, findings from this dissertation demonstrate that for an average candidate, there is a sign ificant survival benefit that can be attributed to listi ng at a center with reduced expe cted time to transplantation. The relative hazard associated with incr eased waiting time adjusted for potential confounding factors is important in demonstrating the independent effect of this characteristic on candidate mortality. However, the interpretation of this relative risk is not always straightforward to clinicians or patients. In orde r to relate the findings of this study in terms that could be useful to a broader audience, the study al so estimated the median survival of candidates from the time of listing. Results of the analysis demonstrate that the averag e candidate has approximately two and a half years longer expe cted survival (12.4 years versus 10.0 years) by listing at a center with a rapid ve rsus delayed rate of transplanta tion. Furthermore, as the center groupings represent quintiles, ther e are approximately 20% of cen ters that have the reduced waiting times, suggesting that they are not uncommon situations. Howe ver, there is also likely a regional component to waiting times, and within certain regions, centers with reduced waiting

PAGE 83

83 times may be scarcer; therefore, future studies may be useful to elucidate the regional availability of centers with rapid rates of transplant. Research suggests that centers with the highe st transplant volume have superior patient outcomes. This has been interpreted as relate d to quality of care a ssociated with greater experience with the surgical procedure, greater infrastructure and coordi nation of services to transplant patients, and availabil ity of ancillary servic es. In contrast to these research accounts, center volume was not significantly associated with candidate survival in this study. This difference may partially reflect the relative unimpo rtance of center volume as compared to other factors examined in the study as well as the redu ced impact of volume fo r transplant candidates as compared to recipients. Many candidates fo r transplantation do not survive to the time at which a transplant may be offered and others may become unviable for transplant due to health deterioration. As such, the impact of center volume, which has shown some association between the largest centers and recipient out comes, may be diluted. It is al so possible that smaller centers have a greater opportunity to follow patients prio r to transplantation a nd, despite the lack of facilities, the long-term advantag es of larger centers are balan ced by increased follow-up care. Past reports suggest that the eff ect of volume on recipient outcomes in not linear, but is generally found to be superior in the top volume centers only. However, this study excluded candidates listed at very small centers (<10 deceased donor tr ansplants per year) in or der to obtain stable estimates of pre-listing characteristics; in this regard, the impact of volume may also only be highlighted at extreme levels and less evident am ong centers with slightly larger volume in this study. The quality of deceased donati ons are highly variable, transl ating to almost three-fold hazard ratio for graft loss for a recipient with the lowest quality donors relative to an ideal class

PAGE 84

84 of donations (49,50). As such, there is a clear in centive for transplant can didates to acquire the highest quality donation available, holding all other fact ors equal. There ar e also significant regional variations in the propor tion of high-risk donations. The expanded criteria policy was initiated in 2002, mandating that candidates prospectivel y consent as to whether they would be willing to receive an ECD transplant. Therefor e, candidates have some control over whether they will receive a lower-quality kidney. The potential benefit of listing for an ECD kidney is that candidates may receive their transplant more rapidly than having to wait for a SCD. In fact, research indicates that this may be an advisable strategy for some candidates, at least at centers that selectively list patients (53,99) However, regardless of the listing strategies, candidates that list at a center with a lower pr oportion of ECD kidneys should theo retically have an advantage over candidates that are listed at centers with a high proportion of ECDs. For candidates that are unwilling to accept an ECD, a lower proportion of ECDs should expedite their acceleration on the waiting list. In contrast, can didates that are willing to accept ECDs should still have a greater opportunity to receive a SCD at a center with a lower proportion of ECDs. Moreover, centers with higher rates of ECD transplants may also have a higher-risk donor pool of organs within risk classes, though this has not been demonstrated explicitly This study indicated that candidates have a significantly elevated mortality risk (AHR = 1.04, p<0.05) by listing at centers with the highest proportion of ECDs as compared to listing at centers w ith the lowest proportion of ECDs. This increased hazard translated to approximately four months of reduced expected survival among candidates listing at centers with the highest pr oportion of ECD transplants as compared to centers with the lowest proportion of ECDs. However, the study also demonstrated that this effect was significan tly higher for post-transplant su rvival and, in particular, when eliminating characteristics of donors from the model. Therefore, the study suggests that

PAGE 85

85 candidates that list at centers with lower quali ty donations are more likely to receive a lower quality organ for those who surv ive to the time of the procedure. Furthermore, outcomes for recipients are significantly reduced at centers w ith a greater utilization of higher-risk organs. The study also suggests that there is a correlation between centers that use a greater proportion of higher-risk organs for centers th at have longer waiting times. In other words, centers that have longer waiting times may be more likely to accept lower-quality organs in order to ameliorate the candidate volume. From the candi date perspective, waiting time re mains the most critical center factor associated with surviva l; however, for centers with similar waiting times, the quality of donor organs may remain a modifier of cente r selection. In add ition, this study did not specifically examine this effect in patients that actually list for or receive lower-quality organs. Follow-up studies addressing center selection spec ifically for candidates willing to accept lowerquality organs (or for those not willing to accept lower-quality organs) may also elucidate important center factors that are pert inent to the candidate population. Performance evaluations are conducted and published by the SRTR and readily accessible to the public through written reports a nd the internet. Evaluations are constructed based on standard mortality ratios which calculate the observed number of events (graft losses or deaths) at a transplant center re lative to what would be expected given the characteristics of the recipient population over a fixed interval of time. Theoretically, this ratio is indicative of quality of care and center performance fo r their recipient population. In fact, the study indicated that candidates listed at centers with the best hist orical performance have significantly better outcomes as compared to candidates listed at centers with the worst performance ratings (AHR=1.14, p< 0.01). This elevated risk translated to an appr oximate one year of increased survival for the average transplant candidate be tween centers with the best and worst ratings.

PAGE 86

86 However, an important consideration associated with this finding is that a portion of this effect was observed prior to transplantation. This ma y imply that centers wi th higher ratings have better pre-transplant care, but also may be suggestiv e of increased patient selection criteria. This notion will be explored in more deta il later in the discussion. One of the main goals of this study, be yond evaluating the significance of individual center characteristics, was to ascertain the rela tive importance of these factors. The general conclusion that is evident from our study is that th e waiting time at a transplant center is the most important modifier of decisions to select a center among the ch aracteristics examined. The hazard ratios and estimated survival years from the time of listing for candidates suggest that while other factors may incrementally impact candidate outcomes, waiting time clearly has the strongest effect on patient mortality. Even in the presence of a combination of ideal characteristics of other factors, differences in expected waiting time remain the predominant determinant of candidate outcomes. In this re gard, the average candidate and their caregivers may have a strong incentive to assess the expe cted waiting times at centers in the decisionmaking process for the selection of a transplant center. The additional knowledge of objective information about the center and characteristics which may influence their outcomes should be available and disseminated to patients in order to help them make informed decisions. In many instances, patients simply may rely on the advice of a referring physician and have no significant participatory role in this deci sion which may have life-altering implications. Results of this study and subsequent research deriving from this paradigm may be used to inform patients and their caregivers about the ramifi cations of these important deci sions and the impact on their survival. In addition, strategies to disseminate th is information to patients and caregivers in an interpretable fashion are clea rly needed in future efforts deriving from this study.

PAGE 87

87 There are certainly additional factors that ma y affect an individual candidates selection of a center that were not incorporated in this analysis, including the ge ographic location of the center, the prestige or personal familiarity with the center and personnel, and the comfort level with the caregivers at the center itself. These factors may be very important to candidates, but should be considered relative to the impact of other characteris tics evaluated in this study. Moreover, centers clearly cannot be defined only by the characteristics outlined in this study. The specific factors investigated in this study were based on past research and the degree to which information was readily available characteri zing centers. However, each center is unique in its physical structure and design, environmen tal conditions, and medical personnel expertise and experience. Centers also vary in their acad emic affiliation, the degree to which they provide services as part of a safety net program, the c oordination and relationship of surgical and medical departments, physical capacity, the availability of dialysis services, a nd connection with other dialysis centers as well as innumerable medical protocols. The degr ee to which these other factors may impact candidate outcomes is unknow n and this uncertainty must be taken into consideration. Specificall y, this study demonstrates the impact of factors in a typical situation for an average candidate; however, a broad host of factors may certainly modify these estimates or be particular pertinent to an individual candi date. Therefore the results of the study must be interpreted and utilized with th ese caveats in mind and not be vi ewed as an omnipotent guide to center selection that cannot be superseded by other conditions. Another important aspect of this study was to assess the degree to which center characteristics were significant and relatively important to all candi dates or whether these characteristics had a particular impact in certain subsets of the candidate population. To this end, the study examined the associati on of center characteristics speci fically in three high-risk

PAGE 88

88 subsets of patients: African-Americans, the elderl y, and obese candidates. In general, results indicated that center char acteristics remain important to thes e high-risk subsets of the population, particularly the center rate of transplantation. One of the poten tial implications of this portion of the analysis was to identify fact ors that may have differential e ffect in certain subgroups that may suggest efficiencies in the tr ansplant process. That is, in the case that specific factors could be shown to be important to certain groups, but no t others, it would be pos sible that candidates could simultaneously benefit based on listing at centers with the presence of factors that are suited to their needs. In contra st, if center factors were rela tively equally important to all subgroups, the analysis may suggest the preferre d center characteristics for candidates, but candidates could obviously not all se ek to list at these centers si multaneously. Broadly, the latter case is suggested by the analysis. That is, while th ere are factors that appear to have a somewhat differential impact in the subgroups examined in th e study, the main findings are applicable to all groups. This may imply that results of the study can still be applicable to candidate populations that have the ability to seek out centers with the preferred characteristics, but cannot be generalized to the entire candi date population. As the primar y characteristic influencing candidate survival is the rate of transplantation associated with the transplant center, primarily an increase in donation rates would be needed to si gnificantly reduce waiting times across centers. The categorization of African-Americans as a high-risk subset of the population is generally a reflection of an increased rate of graft loss for these patients following transplantation. This elevated risk has been attributed to immunologi cal factors as well as socioeconomic characteristics of this population (58). However, in contrast, African-Americans have superior survival on dial ysis relative to Caucasians a nd equivalent patient survival following transplantation (53,100). Therefore, given these potentially differential effects of

PAGE 89

89 transplant-related factors on African-Americans, th e hypothesis that selection of centers may also be unique to this portion of th e candidate population was justifie d. Results of the analysis indicate that, as for the genera l candidate population, listing at a cen ter with the most rapid rate of transplantation is the most important fact or among the variables in the study for AfricanAmerican candidates. The 28% elevated increase d hazard for candidates that list at centers with the longest expected waiting times as compared to ca ndidates that list at ce nters with the shortest waiting times was the most substantial center charact eristic in the analysis. On average, AfricanAmericans have longer waiting times than their Ca ucasian counterparts due to the distribution of HLA antigens in the donor population, which has been a significant component of the organ allocation algorithm. African-Ame ricans have a smaller likelihood of receiving additional points that are ascribed to HLA-matching with a poten tial donor kidney as part of the allocation of deceased donor organs (101,102). Even though average waiting times are longer for African-Americans, as this analysis demonstrates, it is still highly beneficial for Afri can-American candidates to list at centers with reduced expected waiting times. In fact, the stud y indicates that expected survival is two years longer for candidates that list at centers with the shortest time to transplant (11.5 years) as compared to candidates that list at centers with the longest expected waiting times (9.5 years). Another interesting result of the analysis is that African-Ameri cans have a significantly increased survival associated with centers with a low pr oportion of ECD transplants. The 9% increased hazard for mortality after listing associated with African-American candidates that list at centers with the highest ECD proportion was significantly higher than the general population. This is consistent with the literature that African-Ame ricans may derive a grea ter benefit by receiving a higher quality donation even at the expense of additional dialysis exposure as compared to

PAGE 90

90 Caucasian candidates, despite the fact that hist orically African-Americans are more likely to receive an ECD transplant (51,53). Candidates that listed at cen ters with the lowest proportion of ECD transplants had 10.3 years of average life expectancy after listing as compared to 9.6 years among African-Americans that listed at centers with highest proportion of ECD transplants. The effect of center volume was not statistically significant in this portion of the candidate population. Center performance ratios al so had a similar estimated impact for AfricanAmerican candidates as compar ed to the general population. Elderly patients comprise a significant a nd growing portion of th e dialysis population. Elderly patients comprised 8% of the candida te population over the study period, but have a three-fold risk for mortality following listi ng relative to the youngest candidate age group. Elderly candidates are at significan tly higher risk for death while on dialysis as well as after transplantation. However, elderly candidate s still receive a signi ficant benefit from transplantation with an almost doubling of life expectancy rela tive to remaining on dialysis therapy (27). In addition, there is evidence that elderly candidate s have heightened incentive to receive transplants more rapidly, even at the e xpense of receiving a lower quality donation (53). The analysis indicate s that, as with the general population, the most important center characteristic for elderly candidates is the rate of transplantation. Elde rly candidates listed at centers with the lowest proportion of candidate s reaching transplantation had a 26% increased hazard for mortality as compared to the candidate s listed at centers with the longest expected waiting times. This translated to a difference of approximately one year in expected survival for the average elderly candidate. The magnitude of the hazard associated with longer waiting times was reduced as compared to the general population and may reflect that the impact of transplant centers is less important to elderly candidates due to higher death rates on dialysis due to other

PAGE 91

91 clinical factors. That is, as a significantly higher proportion of elderly candidates die prior to reaching transplantation, the accrued benefit from transplantation is observed in a smaller portion of the population. As opposed to the general population, for elde rly candidates, there was no significant association between the volume and the ECD pr oportion at the listing ce nter. As with the reduced effect of the center rate of transplant ation, this may reflect th e reduced importance of center characteristics as compared to the competing risks of other clinically based factors that are particularly relevant for elderl y ESRD patients. However, the non-significant finding associated with ECD proportions may also demonstrate that el derly candidates that list at these centers may receive some additional benefit by receiving these organs more ra pidly. One notable difference in the elderly population was the increased asso ciation of center perf ormance ratings with candidate outcomes. The analysis indicates that elderly candidate s have a significant difference in mortality when listing at centers with the highest performance ratios, a 23% increased hazard for death as compared to listing at centers with the lowest performance ratios. This effect translated to an estimated one additional year of life expectancy between centers of extreme performance ratings (6.1 years vers us 5.3 years). It is possible th at, due to the unique care needs and relative fragility of elderly ca ndidates, the quality of care betw een centers is most evident in these patients. In fact, past reports investiga ting the nature of center effects suggest that differences between centers are primarily reflective of outcomes in higher-risk populations. As such, as opposed to younger candidates in whom lower quality of care may result in greater complication rates or reduced quality of life, the impact on mortality in relatively healthier subsets may still be marginal. However, anothe r explanation for these findings may relate to potential selection bias among elderl y candidates. The effect of sel ection bias is a greater risk in

PAGE 92

92 elderly patients for whom many underlying health conditions and co-morbidities that are not represented in the database can influence outcomes. As outlined in the methodology portion of the st udy, there is some potential for selection bias as a contributing factor to the results. While there are univer sal contraindications for criteria for transplant candidacy, some centers may invoke additional screen ing processes which contribute to differences in mort ality (103,104). In the context of this study, this may result in bias in results in the case that ce nters that list a more selective c ohort of patients also were more highly represented by certain center characteristics in the analysis. For instance, in this case if centers which traditionally had higher center performance ratios also were more selective in candidate screening, this associ ation may bias certain results of the study. Specifically, the disproportionate representation of non-codified fact ors may contribute to differences in outcomes and inappropriately elevate the effect of perfor mance criteria. As described in the methodology, there were two fundamental strategies to obviat e this potential confounding effect. One strategy was to adjust for factors that may account for pati ents that are selectively listed. These include age, race, patient primary insuranc e, and education level. In fact these factors were adjusted for in the analysis and had statistically significan t associations with ca ndidate mortality. The additional strategy was to partition the follow-up period to preand post-transplant survival. The purpose of this partitioned analysis is that tran splant centers typically have significantly less interaction with candidates prior to transplantation and would be less likely to have an impact on survival prior to transplantation related to quality of care. In contrast, centers may have a much stronger association with candida tes during the transplant proce dure and post-transplant care. Therefore, differences that occur prior to transp lantation are more likely related to factors other than center quality of care and may reflect a certain degree of patient selection.

PAGE 93

93 The results of the model for pre-transplant mo rtality indicate a significant association of standard mortality ratios with death prior to transplantation, while none of the other center factors examined had a significan t association. Candidates listed at centers with the lowest performance ratios had a 13% elevated mortalit y prior to transplantation as compared to candidates listed at centers with the best mortal ity ratios. In additi on, this association was particularly notable in elderly candidates. This finding must be interpreted carefully and, in particular, may be indicative of selection bias of better patients to centers with higher performance ratings. In practice, transplant centers are not th e primary caregivers of patients prior to transplantation. In contrast, cente rs are responsible for listing candidates and determining whether they meet certain medical criteria, but ar e not typically responsible for patient care in the interim betw een candidacy and the transplant procedure. One of the main purposes for this sub-analysis was to determine th e potential influence of certain selection biases that may occur and influence the main outcome of patient survival after candidacy. This portion of the model most likely indicates that centers with better performan ce ratings either list healthier candidates, leading to lo wer pre-transplant mortality, or are located in areas with lower morbidity among the dialysis population, or there are other environmental factors wh ich are significantly variable. The fact that this association was stronger among elde rly patients, in which selective listing of patients may be significantly stronge r, further suggests the influence of patient selection. Although center performance was not the most significant factor for candidates in the primary outcome model, the degree to which an a ssociation did exist may in part reflect this patient selection rather than improved quali ty of care by the transplant center. The other important implication of this finding is that models evaluating center performance may not fully account for underlying health conditions and exogenous factors that

PAGE 94

94 significantly influence mortality but are not associated with quality of care. That is, as candidate survival rates are positively correl ated with transplant outcomes even after adjustment for other risk factors, this may imply that other non-codified factors are a ssociated with outcomes that are not accounted for with risk adjustment. Future investigation into the degree to which center performance is impacted by candidate mortality rates independent of ot her transplant-related factors is important to distinguish these effects. The interpreta tion of center standard mortality ratios is generally that they ar e indicative of quality of care. In fact, insurance companies and government oversight committees may further use these criteria for business purposes and to investigate centers for inappropriate care practices (22). In contrast, if performance ratios are merely indicative of selective listing practices, than these additional interpretations relating quality of care to performance are unwarranted. Furthermore, if th is association is not justified then there are potential dangers that centers will limit access to patients in an effort to improve performance ratings, but at the same time certa in patients may not acquire the benefit of transplantation as a result. Given the results in this study, performance ratios only had a mild association with candidate mortality, but even this limited effect should be interpreted with caution given these po tential caveats. Obese patients represent a unique subset of the transplant candidate pool for several reasons. The population of obese candidates and tr ansplant recipients is significantly growing consistent with the growth rates among the ge neral ESRD population ( 100). Moreover, obese patients have a unique paradoxical survival ad vantage on maintenance dialysis. However, obese patients also have signifi cantly elevated risk for post-tran splant graft loss and mortality (105). This study indicates that despite this relative ly reduced mortality on dialysis prior to transplantation and increased mortality following transplantation, the most important center

PAGE 95

95 characteristic of obese candidates remains the rate of transplantati on. Obese candidates significantly benefit from transplant ation and as compared to other center characteristics have the most significantly increased survival expectancy associated with listing at centers with reduced waiting times as compared to other factors of center volume, performan ce, or ECD proportion. Center volume was not associat ed with differential outcomes for obese patients. Obese candidates that listed at centers with the highest proportion of EC D transplants had an elevated risk for death at a similar level to the overall candidate population. Candidates that listed at centers with the lowest performance ratio also ha d an elevated risk for mortality. Similar to elderly patients, the issue of whethe r this effect is due to transpla nt center quality of care versus patient selection is unknown; however, among high-risk patients the opportunity for this selection bias is likely stronger. On the other hand, centers trea ting obese candidates and recipients may require greater quality of car e due to the increased potential for surgical complications and morbidity associated with obese transplant recipients. The most important results is that expected candida te survival for the average obese candidate ranges from 8.2 years to10.2 years for listing at centers with the longe st and shortest waiting times respectively. Consistent with the general candidate population, results of the study indicates that obese candidates have a significant su rvival advantage by listing at ce nters with a rapid rate of transplantation. In general, results of the study for higher-risk patients were consistent as for the general candidate population. As discusse d previously, this indicates th at broad efficiencies in the transplantation process are not readily indicated by the study in which certain portions of the population may be best served by particular cente rs. In contrast, the study suggests that all patients have an incentive to list at centers wi th reduced waiting times above all other center

PAGE 96

96 characteristics. Given this information, the quest ion remains as to which patients will utilize this information. Intuitively, the answer will likely be those patients who are best informed or educated or have the means to travel to centers w ith an ideal set of charac teristics. Alternatively, if the information were well known among the en tire candidate population, the net effect would be that waiting times would eventual ly be relatively equivalent. Th at is, the effect of candidates seeking centers with reduced waiting times would eventually balance these differences. Whether or not this is realistic is not clear; however, in terms of equity and eliminating geographic disparities in access to transplantat ion, these objectives may be ideal. Particularly given the fact that more vulnerable candidate groups are less li kely to benefit from the results of the study, future efforts to educate these portions of the population and provide access to centers with ideal characteristics are warranted. In the framework of Grossmans health produ ction function, the study suggests that patient and caregiver behavior, specifically the selection of a center, has a significant impact with the subsequent deterioration rate of transplant candidate he alth. As described in Grossmans model, patients arrive at a given period (in this case, at the time of candidacy) with a given stock of health. From that point, decisions to seek hea lth care and the investment in future stock of health are related to their prognosis and subse quent utility. Beyond simply seeking medical care as outlined as an important fact or in Grossmans paradigm, this study suggests that the specific center at which care is sought is a significant modifier of patient outcomes. In addition, consistent with the health production function, cert ain individual characteris tics alter the effects of future health. Whether patients decisionmaking processes are rela ted to their own demand for utility or current stock of health is not fu lly addressed in this study. However, the study does suggest that patients that are mo re informed, motivated, or able to seek care outside of their

PAGE 97

97 immediate region have an improved prognosis th rough selection of a me dical center. An individuals investment in h ealth through the selec tion of a center ha s a profound effect on transplant candidate survival consistent with the health production function paradigm. There have been significant efforts from pr ivate and public agencies, individual patients, and patient advocates to provide transparent healthcare information to the population. One of the significant challenges in these efforts is to di sseminate information in a coherent and useful manner to patients. Certainly, in re cent years patients are more likely to interact with healthcare professionals with a greater arsenal of in formation obtained through research, personal communications, or web-based services. The de gree to which this information increases the likelihood to seek appropriate care or facilitates the h ealthcare interaction is not fully known and is likely context dependent. However, as the tr end towards more informed patients is not likely to decline in future years, a major focus of res earchers and caregivers is to disseminate the most meaningful and interpreta ble data to consumers of this in formation. An additional challenge related to this trend is to dispel increased leve ls of misinformation that may also be obtained from the same sources. Transplantation is somewhat unique to other fields of medicine due to the volume and granularity of information that ex ists related to mandatory data collection. In this sense, this field has a particular opportunity to provide evidence-based recommendations to patients and their caregivers. However, transplantation is not unique to other medical cont exts in that there is a significantly competitive market, and advertisement of a center may entail numerous unsubstantiated or at least subjec tive components. Future work de tailing strategies by which to disseminate information to patients will be a n ecessary corollary to this study. Furthermore,

PAGE 98

98 given that transplantation is a complex scien ce, relaying the most pertinent information to concerned parties is a significant obstacle. Acknowledgment The data reported here have been supplied by the University Renal Research and Education Association (URREA) as the contractor for the Sc ientific Registry of Transplant Recipients (SRTR). The interpretation and reporting of these da ta are the responsibility of the author and in no way should be seen as an official policy of or interpretation by the SRTR or the U.S. Government.

PAGE 99

99 CHAPTER 6 CONCLUSIONS AND FUTURE WORK The primary findings of this research suggest that transplant center s are associated with significantly variable su rvival for candidates of renal tran splantation. This information is important for prospective candidate s and caregivers to facilitate de cisions in their selection of a center of care. There are multiple factors which may ultimately determine the specific center at which a candidate receives care; however, objective informati on which may have life-altering ramifications is critical to disseminate to affected individuals. Results of the study indicate that multiple factors are important, the magnitude of the effects vary significantly, and the most critical factor for a candidate among the center ch aracteristics examined is the expected duration of the waiting list. Candidates have a significan tly longer expected life span by listing at a center with a reduced waiting time independent of other cen ter factors. In addition, this finding is valid across the three high-risk subgroups examined with slight varia tions in the magnitude of the effect. For centers with similar expected waiti ng times, other factors examined in the study also modify expected survival for prospective candidates. The results of this study can be summarized to inform transplant candidates that it is useful to shop around for tran splant centers, and a key characteristic in these comparisons should be the expected waiting time to transplant However, one of the realities in this population is that not all patients are capable of listing at centers across the country either for logistical, financial, or insuran ce constraints. Therefore, those patients that are more affluent, educated, or generally have fewer barriers to m obilizing are the candidates that can benefit from this information. Alternatively, patients from lower socioeconomic status or with more complicated health conditions may not be able to identify centers in a more narrow region with significantly reduced waiting times. The degr ee to which centers are available with more

PAGE 100

100 desirable characteristics at a national and region al basis across the country will be an important investigation for future work. In this regard, on e of the potential follow-up projects of this study is to construct an application that incorporates these results and allows candidates to identify centers. This could be generated with the capab ility to incorporate give n constraints (e.g., in a certain region) and allow transparent information regarding the availabil ity of centers and the associated impact on candidates prognoses. This c ould be a critical step in the dissemination of results to a broad audience rather than thr ough traditional scientific publications by which information may reach only selected patients and caregivers. Another interesting and impor tant result of this study is the identificati on of potential selection bias associated with center performance ratings. These ratings have important implications to transplant centers related to government oversight, cont racting with insurance agencies, and advertising to pot ential candidates and referring physicians as a marker of good quality of care. However, if these ratings are related to pre-transplant candidate selection (as suggested by this study) then further investigation of the utility of performance evaluations in the field of transplantation is wa rranted. Additional study investig ating the association of other center factors, such as patient proximity to the center, center experience with specific patient groups, and in relation to the quality of care of dialysis centers asso ciated with a transplant center may complement the findings of this study. In a ddition, replication of th e analysis in several other high-risk groups, such as diabetic patients, patients with a history of cardiovascular disease, or pediatric populations, may also generate important inform ation for subgroups of the growing ESRD population. Moreover, results may be applicable to other forms of organ transplantation and more broadly to other forms of health car e in which provider selection could have an important impact on patient prognosis.

PAGE 101

101 The immediate follow-up work for this st udy will entail disseminating results through peer-reviewed manuscripts and discussion of th ese topics in scientific meetings. Beyond extending the results to other popu lations and medical contexts, st rategic efforts to communicate findings directly with patients ar e an important endeavor. This is particularly salient in this context in which centers may or may not have a vested interest to transmit this information to patients. Ultimately, the results of the study will be most useful wh en there is shared information among patients and caregivers, and it is utilized jointly to make pot entially life-altering decisions in selection of a transplant center

PAGE 102

102 LIST OF REFRENCES 1. Prevalence of Reported ESRD. Online. Internet. Available: http://www.usrds.org/2006/ref/B_prevalence_06.pdf 2006 Annual Data Report (accessed 20 Sep 2006). 2. Schrag D, Cramer LD, Bach PB, Cohen AM, Warren JL, Begg CB. Influence of hospital procedure volume on outcomes following surgery for colon cancer. JAMA 2000:284: 3028. 3. Bianco FJ, Jr., Riedel ER, Begg CB, Ka ttan MW, Scardino PT. Variations among high volume surgeons in the rate of complicati ons after radical prostatectomy: Further evidence that technique ma tters. J Urol 2005:173: 2099. 4. Krumholz HM, Radford MJ, Wang Y, Chen J, Heiat A, Marciniak TA. National use and effectiveness of beta-blockers for the treatmen t of elderly patients after acute myocardial infarction: National Cooperative Card iovascular Project. JAMA 1998:280: 623. 5. Harmon JW, Tang DG, Gordon TA et al. Ho spital volume can serv e as a surrogate for surgeon volume for achieving excellent outco mes in colorectal resection. Ann Surg 1999:230: 404. 6. Hosenpud JD, Breen TJ, Edwards EB, Daily OP Hunsicker LG. The ef fect of transplant center volume on cardiac transplant outcome. A report of the United Network for Organ Sharing Scientific Registry. JAMA 1994:271: 1844. 7. Stukel TA, Lucas FL, Wennberg DE. Longterm outcomes of regional variations in intensity of invasive vs medical management of Medicare patients w ith acute myocardial infarction. Jama-Journal of the Amer ican Medical Association 2005:293: 1329. 8. Baicker K, Chandra A. Medicare spendi ng, the physician workforce, and beneficiaries quality of care. Health Affairs 2004:23: W4184W4197. 9. Institute of Medicine. To Err is Hu man: Building a Safer Health System. 2000. Washington, D.C., National Academy Press. 10. Baker DW, Einstadter D, Thomas CL Husak SS, Gordon NH, Cebul RD. Mortality trends during a program th at publicly reported hospital performan ce. Medical Care 2002:40: 879. 11. Chassin MR. Achieving and sustaining im proved quality: Lessons from New York State and cardiac surgery. Health Affairs 2002:21: 40. 12. Mukamel DB, Weimer DL, Zwanziger J, Go rthy SFH, Mushlin AI. Quality report cards, selection of cardiac surgeons, and racial disparities: A st udy of the publication of the

PAGE 103

103 New York State Cardiac Surgery Reports. Inquirythe Journal of Health Care Organization Provision a nd Financing 2004:41: 435. 13. Epstein AM. Rolling down the runwayThe challenges ahead for quality report cards. Jama-Journal of the American Medical Association 1998:279: 1691. 14. Schauffler HH, Mordavsky JK. Consumer reports in health care: Do they make a difference? Annual Review of Public Health 2001:22: 69. 15. Schneider E, Epstein A. Public perfor mance reports for cardiac surgeryReply. JamaJournal of the American Me dical Association 1999:281: 135. 16. Burdick JF, Williams GM. What causes cen ter effects in kidney transplantation. Ann Surg 1986:203: 311. 17. Mickey MR. Center effect. Clin Transpl 1986: 165. 18. Benlahrache C, Cecka M, Mickey MR, Ci cciarelli J. The center effect. Clin Transpl 1987: 325. 19. Ogura K, Cecka JM. Center effects in renal transplantation. Clin Transpl 1991: 245. 20. 2005 OPTN/SRTR Annual Report http://www.ustransplant.org /annual_reports/current/209_donrace_dc.htm?o=2&g=5&c=1 (accessed 15 Sep 2006) 21. Howard DH, Kaplan B. Do report cards influence hospital choice? The case of kidney transplantation. Inquiry-the J ournal of Health Care Organi zation Provision and Financing 2006:43: 150. 22. Schold JD, Howard RJ. Prediction models a ssessing transplant center performance: Can a little knowledge be a dangerous thing? Amer ican Journal of Transplantation 2006:6: 245. 23. Lin HM, Kauffman HM, McBride MA et al Center-specific graft and patient survival rates: 1997 United Network for Organ Sh aring (UNOS) report. JAMA 1998:280: 1153 1160. 24. Terasaki PI, Cecka JM. The center effec t: Is bigger better? Clin Transpl 1999: 317. 25. Gjertson DW. Center and other factor effects in recipients of living-donor kidney transplants. Clin Transpl 2001: 209. 26. Axelrod DA, Guidinger MK, McCullough KP, Leichtman AB, Punch JD, Merion RM. Association of center volume w ith outcome after liver and ki dney transplantation. Am J Transplant 2004:4: 920.

PAGE 104

104 27. Wolfe RA, Ashby VB, Milford EL et al. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and reci pients of a first cadaveric transplant. New England Journa l of Medicine 1999:341: 1725. 28. McDonald SP, Russ GR. Survival of recipi ents of cadaveric kidney transplants compared with those receiving dialysis treatmen t in Australia and New Zealand, 1991. Nephrology Dialysis Tran splantation 2002:17: 2212. 29. Rabbat CG, Thorpe KE, Russell JD, Church ill DN. Comparison of mortality risk for dialysis patients and cadaveric first renal tran splant recipients in Ontario, Canada. Journal of the American Societ y of Nephrology 2000:11: 917. 30. Schnuelle P, Lorenz D, Trede M, van der Woude FJ. Impact of renal cadaveric transplantation on survival in end-stage renal failure: Eviden ce for reduced mortality risk compared with hemodialysis during longterm follow-up. Journal of the American Society of Nephrology 1998:9: 2135. 31. Glanton CW, Kao TC, Cruess D, Agodoa LY, Abbott KC. Impact of renal transplantation on survival in end-stage renal disease patien ts with elevated body mass index. Kidney Int 2003:63: 647. 32. Ojo AO, Hanson JA, Meier-Kriesche HU et al. Survival in recipients of marginal cadaveric donor kidneys compared with othe r recipients and wait-listed transplant candidates. Journal of the American Society of Nephrology 2001:12: 589. 33. Oniscu GC, Brown H, Forsythe JL. How old is old for transplantation? Am J Transplant 2004:4: 2067. 34. Cosio FG, Alamir A, Yim S et al. Patient survival after renal tr ansplantation: I. The impact of dialysis pre-tran splant. Kidney Int 1998:53: 767. 35. Asderakis A, Augustine T, Dyer P et al. Pre-emptive kidney transplantation: The attractive alternative. Nephrol Dial Transplant 1998:13: 1799. 36. Meier-Kriesche HU, Port FK, Ojo AO et al. Effect of waiting time on renal transplant outcome. Kidney International 2000:58: 1311. 37. Mange KC, Joffe MM, Feldman HI. Effect of the use or nonuse of long-term dialysis on the subsequent survival of renal transpla nts from living donors. N Engl J Med 2001:344: 726. 38. Gaston RS, Danovitch GM, Adams PL et al The report of a national conference on the wait list for kidney transplanta tion. Am J Transplant 2003:3: 775.

PAGE 105

105 39. Lenhard V, Dreikorn K, Rohl L. Results of kidney transplantation in relation to HLA-A, B, DR matching and quality of donor organ. Proc Eur Dial Transplant Assoc 1980:17: 450. 40. Alexander JW, Zola JC. Expanding the donor pool: Use of marginal donors for solid organ transplantation. Clin Transplant 1996:10: 1. 41. Faenza A, Sestigliani E, Zambianchi L, Ridolfi L. Utilization of suboptimal kidney donors. Transplant Proc 2004:36: 485. 42. Meier-Kriesche HU, Port FK, Ojo AO et al. Effect of waiting time on renal transplant outcome. Kidney International 2000:58: 1311. 43. Nyberg SL, Matas AJ, Kremers WK et al. Improved scoring system to assess adult donors for cadaver renal transplantat ion. Am J Transplant 2003:3: 715. 44. Port FK, Bragg-Gresham JL, Metzger RA et al. Donor characteristics associated with reduced graft survival: An approach to expanding the pool of kidney donors. Transplantation 2002:74: 1281. 45. Su XM, Zenios SA, Chertow GM. In corporating recipien t choice in kidney transplantation. Journal of the American Societ y of Nephrology 2004:15: 1656. 46. Swanson SJ, Hypolite IO, Agodoa LYC et al. Effect of donor f actors on early graft survival in adult cadaveric renal transplantation. American Journal of Transplantation 2002:2: 68. 47. Schnitzler MA, Whiting JF, Brennan DC et al. The expanded criteria donor dilemma in cadaveric renal transplantati on. Transplantation 2003:75: 1940. 48. Ratner LE, Kraus E, Magnuson T, Bender JS Transplantation of kidneys from expanded criteria donors. Surgery 1996:119: 372. 49. Metzger RA, Delmonico FL, Feng S, Po rt FK, Wynn JJ, Merion RM. Expanded criteria donors for kidney transplantation. Am J Transplant 2003:3 Suppl 4: 114. 50. Schold JD, Kaplan B, Baliga RS, Meier-Kriesche HU. The broad spectrum of quality in deceased donor kidneys. Am J Transplant 2005:5: 757. 51. Schold JD, Kaplan B, Chumbler NR et al. Access to quality: Evaluation of the allocation of deceased donor kidneys for transpla ntation. J Am Soc Nephrol 2005:16: 3121. 52. Schold JD, Howard RJ, Scicchitano MJ, Meier-Kriesche HU. The expanded criteria donor policy: An evaluation of program objectiv es and indirect ramifications. American Journal of Transplantation 2006:6: 1689.

PAGE 106

106 53. Schold J.D., Meier-Kriesche HU. Whic h Renal Transplant Ca ndidates Should Accept Marginal Kidneys in Exchange for Shorter Wa iting Time on Dialysis? Clinical Journal of the American Society of Nephrology 1[3]. 5-1-2006. 54. Schnitzler MA, Whiting JF, Brennan DC et al. The expanded criteria donor dilemma in cadaveric renal transplantati on. Transplantation 2003:75: 1940. 55. Prevalance of Reported ESRD. Online. Internet. Available: http://www.usrds.org/2006/ref/B_prevalence_06.pdf 2006 Annual Data Report (accessed 20 Sep 2006). 56. Patient Survival. Onli ne. Internet. Available: http://www.usrds.org/2006/ref/I_survival_06.pdf 2006 Annual Data Report (accessed 20 Sep 2006). 57. Young CJ, Gaston RS. Renal transplanta tion in black Americans. N Engl J Med 2000:343: 1545. 58. Gaston RS. Factors affecti ng renal allograft surv ival in African Americans. Blood Purif 1996:14: 327. 59. Gaston RS, Benfield M. The relationship between ethnicity and outcomes in solid organ transplantation. J Pediatr 2005:147: 721. 60. Young CJ, Gaston RS. Renal transplanta tion in black Americans. N Engl J Med 2000:343: 1545. 61. Alexander GC, Sehgal AR. Barriers to cad averic renal transpla ntation among blacks, women, and the poor. Jama-Journal of the American Medical Association 1998:280: 1148. 62. Ayanian JZ, Cleary PD, Weissman JS, Epst ein AM. The effect of patients preferences on racial differences in access to renal transplantation. New England Journal of Medicine 1999:341: 1661. 63. Organ Distribution: Allocation of Deceas ed Kidneys. Online. Internet. Available: http://www.optn.org/PoliciesandB ylaws/policies/pdfs/policy_7.pdf (accessed 11-202004) 64. Kerman RH, Kimball PM, Van Buren CT et al. Influence of race on crossmatch outcome and recipient eligibility for transpla ntation. Transplantation 1992:53: 64. 65. Foster CE, III, Philosophe B, Schweitzer EJ et al. A decade of experience with renal transplantation in African-Am ericans. Ann Surg 2002:236: 794.

PAGE 107

107 66. Cosio FG, Alamir A, Yim S et al. Patient survival after renal tr ansplantation: I. The impact of dialysis pre-tran splant. Kidney Int 1998:53: 767. 67. Kalil RS, Heim-Duthoy KL, Kasiske BL. Patie nts with a low income have reduced renal allograft survival. Am J Kidney Dis 1992:20: 63. 68. Butkus DE, Meydrech EF, Raju SS. Racial di fferences in the survival of cadaveric renal allografts. Overriding effects of HLA matchi ng and socioeconomic factors. N Engl J Med 1992:327: 840. 69. Neylan JF, Sayegh MH, Coffman TM et al. The allocation of cadaver kidneys for transplantation in the United States: Consensu s and controversy. Journal of the American Society of Nephrology 1999:10: 2237. 70. Calle EE, Thun MJ, Petrelli JM, Rodri guez C, Heath CW, Jr. Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 1999:341: 1097. 71. Ejerblad E, Fored CM, Lindblad P, Fryzek J, McLaughlin JK, Nyren O. Obesity and risk for chronic renal failure. J Am Soc Nephrol 2006:17: 1695. 72. Allison DB, Fontaine KR, Manson JE, Stevens J, VanItallie TB. Annual deaths attributable to obesity in th e United States. JAMA 1999:282: 1530. 73. Manson JE, Willett WC, Stampfer MJ et al. Body weight and mo rtality among women. N Engl J Med 1995:333: 677. 74. Stevens J, Plankey MW, Williamson DF et al. The body mass index-mortality relationship in white and African Am erican women. Obes Res 1998:6: 268. 75. Abbott KC, Glanton CW, Agodoa LY. B ody mass index and enrollment on the renal transplant waiting list in the United States. J Nephrol 2003:16: 40. 76. Fung F, Sherrard DJ, Gillen DL et al. In creased risk for cardiovascular mortality among malnourished end-stage renal disease patients. Am J Kidney Dis 2002:40: 307. 77. Kakiya R, Shoji T, Tsujimoto Y et al. Body fat mass and lean mass as predictors of survival in hemodialysis patie nts. Kidney Int 2006:70: 549. 78. Kalantar-Zadeh K, Kopple JD, Kilpatrick RD et al. Association of morbid obesity and weight change over time with cardiovascular survival in hemodial ysis population. Am J Kidney Dis 2005:46: 489. 79. Wiesholzer M, Harm F, Schuster K et al. Initial body mass indexes have contrary effects on change in body weight and mortality of patients on maintenance hemodialysis treatment. J Ren Nutr 2003:13: 174.

PAGE 108

108 80. Kalantar-Zadeh K. Causes and conseque nces of the reverse epidemiology of body mass index in dialysis patients. J Ren Nutr 2005:15: 142. 81. Gore JL, Pham PT, Danovitch GM et al. Obesity and outcome following renal transplantation. Am J Tr ansplant 2006:6: 357. 82. Meier-Kriesche HU, Vaghela M, Thambuganipa lle R, Friedman G, Jacobs M, Kaplan B. The effect of body mass index on long-term re nal allograft survival. Transplantation 1999:68: 1294. 83. Johnson DW, Isbel NM, Brown AM et al. The effect of obesity on renal transplant outcomes. Transplantation 2002:74: 675. 84. Drafts HH, Anjum MR, Wynn JJ, Mulloy LL, Bowley JN, Humphries AL. The impact of pre-transplant obesity on re nal transplant outcomes. C lin Transplant 1997:11: 493. 85. Armstrong KA, Campbell SB, Hawley CM, Nicol DL, Johnson DW, Isbel NM. Obesity is associated with worsening cardiovascula r risk factor profiles and proteinuria progression in renal transplant recipi ents. Am J Transplant 2005:5: 2710. 86. Pelletier SJ, Maraschio MA, Schaubel DE et al. Survival benefit of kidney and liver transplantation for obese patients on th e waiting list. Clin Transpl 2003: 77. 87. Holley JL, Monaghan J, Byer B, Bronsther O. An examination of the renal transplant evaluation process focusing on cost and the reasons for patient exclusion. Am J Kidney Dis 1998:32: 567. 88. Modlin CS, Flechner SM, Goormastic M et al. Should obese patients lose weight before receiving a kidney transplant? Transplantation 1997:64: 599. 89. Bennett WM, McEvoy KM, Henell KR, Valente JF, Douzdjian V. Morbid obesity does not preclude successful renal transpla ntation. Clin Transplant 2004:18: 89. 90. Howard RJ, Thai VB, Patton PR et al. Ob esity does not portend a bad outcome for kidney transplant recipients. Tr ansplantation 2002:73: 53. 91. 2006 USRDS Annual Data Report. http://www.usrds.org/2006/ref/A_incidence_06.pdf (accessed 20 Sep 2006). 92. Meier-Kriesche HU, Ojo AO, Hanson JA, Kaplan B. Exponentially increased risk of infectious death in older renal tran splant recipients. Kidney Int 2001:59: 1539. 93. Doyle SE, Matas AJ, Gillingham K, Rosenbe rg ME. Predicting clinical outcome in the elderly renal transplant reci pient. Kidney Int 2000:57: 2144.

PAGE 109

109 94. Ismail N, Hakim RM, Helderman JH. Renal re placement therapies in the elderly: Part II. Renal transplantation. Am J Kidney Dis 1994:23: 1. 95. Meier-Kriesche HU, Schold JD, Gaston RS Wadstrom J, Kaplan B. Kidneys from deceased donors: Maximizing the value of a scarce resource. Am J Transplant 2005:5: 1725. 96. Smits JM, Persijn GG, van Houwelingen HC, Claas FH, Frei U. Evaluation of the Eurotransplant Senior Program. The results of the first year. Am J Transplant 2002:2: 664. 97. Grossman M. Concept of Health Capita l and Demand for Health. Journal of Political Economy 1972:80: 223. 98. Dickinson DM, Shearon TH, OKeefe J et al. SRTR center-specific reporting tools: Posttransplant outcomes. Am J Transplant 2006:6: 1198. 99. Schold JD, Howard RJ, Scicchitano MJ, Meier-Kriesche HU. The expanded criteria donor policy: An evaluation of program objec tives and indirect ramifications. Am J Transplant 2006:6: 1689. 100. Schold JD, Srinivas TR, Guerra G et al A weight-listing paradox for candidates of renal transplantation? Am J Transplant 2007:7: 550. 101. Rebellato LM, Arnold AN, Bozik KM, Haisch CE. HLA matching and the United Network for Organ Sharing Allocation Syst em: Impact of HLA matching on AfricanAmerican recipients of cadaveric kidney transplants. Transp lantation 2002:74: 1634 1636. 102. Zachary AA, Braun WE, Hayes JM et al. Effect of HLA matchi ng on organ distribution among whites and African-Americans. Transplantation 1994:57: 1115. 103. Holley JL, Monaghan J, Byer B, Bronsther O. An examination of the renal transplant evaluation process focusing on cost and the reasons for patient exclusion. Am J Kidney Dis 1998:32: 567. 104. Kendrick E. Evaluation of the Transplant Recipient. In: Danovitch GM, ed. Handbook of Kidney Transplantation 2001: 130. 105. Meier-Kriesche HU, Arndorfer JA, Kaplan B. The impact of body mass index on renal transplant outcomes: A significant independent risk factor for graft failure and patient death. Transplantation 2002:73: 70.

PAGE 110

110 BIOGRAPHICAL SKETCH Jesse D. Schold, Ph.D., M.Stat., M.Ed., comp leted his doctorate program in 2007 in the department of Health Services Research, Manageme nt and Policy at the Univ ersity of Florida in the College of Public Health and Health Prof essions. Dr. Schold received his undergraduate training at Emory University. Af ter receiving a B.A., he worked in the healthcare setting for several years and then enrolled in graduate st udies at North Carolina State University. Dr. Schold received both a Master of Statistics and a Master of Edu cation in the year 2000. After two years working in industry as a statistician, Dr. Schold began wo rk as a research coordinator in the Department of Medicine at the University of Florida. During hi s experience, he also enrolled as a doctoral student in the Health Serv ices Research program. Dr. Schold is currently an Associate Instructor in the Department of Medicine and has had pe er-reviewed scientific articles published in journals including Transplantation, the Journal of the American Society of Nephrology, Diabetes Care, the Clinical Journal of the American Socie ty of Nephrology, Seminars in Dialysis, Blood, Clinical Transplant, Biology of Blood and Marrow Transplantation, and the American Journal of Transplantation Dr. Schold plans to continue work in an academic setting in the fields of transplantation and health services research.