<%BANNER%>

Comparing Prescription Drug Coverage between Medicare Part D and the Federal Employees Health Benefits Program

Permanent Link: http://ufdc.ufl.edu/UFE0042564/00001

Material Information

Title: Comparing Prescription Drug Coverage between Medicare Part D and the Federal Employees Health Benefits Program
Physical Description: 1 online resource (171 p.)
Language: english
Creator: White, Annesha
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: cost, drug, fehbp, formulary, health, managed, medicare
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: This study provides an in-depth empirical comparison of Medicare Part D and the Federal Employees Health Benefits Program (FEHBP) with respect to prescription drug plans. A major difference between these two programs is found in Medicare s required adherence to the United States Pharmacopeia (USP) guidelines. Results revealed that, depending on the method used for analysis, both programs provided broad coverage of top drugs dispensed and sold in the United States. Using the independent samples t-test for the 19 Medicare Part D formularies analyzed, formulary coverage of the top drugs dispensed and sold in the United States ranged from 72-94% (average 84%), while the range was 85-99% (average 94%) for the 5 FEHBP formularies examined (p < .01). Overall, the independent sample t-test findings indicated that the FEHBP plans provided broader drug coverage as compared to Medicare Part D plans. On the other hand, the regression results indicated that, once other factors and interaction effects were taken into account, the programs were shown to be about the same in terms of coverage. In regard to cost sharing, both Medicare Part D plans and the FEHBP plans utilize fixed-dollar copayments more often than coinsurance percentages for generic drugs, but for brand name drugs Medicare Part D plans were more likely to utilize copays while the FEHBP plans were more likely to utilize coinsurance. Furthermore, for the Medicare Part D plans that utilized copayments, the mean copayment for generic drugs was $4.53 (range was $0 - $8) as compared to the FEHBP plans mean copayment of $7.67 (range $5 - $10). Therefore, the Medicare Part D plans provided lower mean copays for generic drugs as compared to the FEHBP plans (p < .05). On the other hand, the finding for brand name drugs was non-significant. For the Medicare Part D plans that utilized coinsurance for generic drugs, mean rates were 17% as compared to the FEHBP plan mean rates of 20% (p < .05). For the Medicare Part D plans that utilized coinsurance for brand name drugs, mean rates were 26% as compared to the FEHBP plan mean rates of 34%. These findings benefit policymakers, health care professionals, and consumers by suggesting lessons that can be learned from both the FEHBP and the Medicare Part D program.
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 Annesha White.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: McKay, Niccie L.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042564:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042564/00001

Material Information

Title: Comparing Prescription Drug Coverage between Medicare Part D and the Federal Employees Health Benefits Program
Physical Description: 1 online resource (171 p.)
Language: english
Creator: White, Annesha
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: cost, drug, fehbp, formulary, health, managed, medicare
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: This study provides an in-depth empirical comparison of Medicare Part D and the Federal Employees Health Benefits Program (FEHBP) with respect to prescription drug plans. A major difference between these two programs is found in Medicare s required adherence to the United States Pharmacopeia (USP) guidelines. Results revealed that, depending on the method used for analysis, both programs provided broad coverage of top drugs dispensed and sold in the United States. Using the independent samples t-test for the 19 Medicare Part D formularies analyzed, formulary coverage of the top drugs dispensed and sold in the United States ranged from 72-94% (average 84%), while the range was 85-99% (average 94%) for the 5 FEHBP formularies examined (p < .01). Overall, the independent sample t-test findings indicated that the FEHBP plans provided broader drug coverage as compared to Medicare Part D plans. On the other hand, the regression results indicated that, once other factors and interaction effects were taken into account, the programs were shown to be about the same in terms of coverage. In regard to cost sharing, both Medicare Part D plans and the FEHBP plans utilize fixed-dollar copayments more often than coinsurance percentages for generic drugs, but for brand name drugs Medicare Part D plans were more likely to utilize copays while the FEHBP plans were more likely to utilize coinsurance. Furthermore, for the Medicare Part D plans that utilized copayments, the mean copayment for generic drugs was $4.53 (range was $0 - $8) as compared to the FEHBP plans mean copayment of $7.67 (range $5 - $10). Therefore, the Medicare Part D plans provided lower mean copays for generic drugs as compared to the FEHBP plans (p < .05). On the other hand, the finding for brand name drugs was non-significant. For the Medicare Part D plans that utilized coinsurance for generic drugs, mean rates were 17% as compared to the FEHBP plan mean rates of 20% (p < .05). For the Medicare Part D plans that utilized coinsurance for brand name drugs, mean rates were 26% as compared to the FEHBP plan mean rates of 34%. These findings benefit policymakers, health care professionals, and consumers by suggesting lessons that can be learned from both the FEHBP and the Medicare Part D program.
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 Annesha White.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: McKay, Niccie L.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042564:00001


This item has the following downloads:


Full Text

PAGE 1

1 COMPARING PRESCRIPTION DRUG COVERAGE BETWEEN MEDICARE PART D AND THE FEDERAL EMPLOYEES HEALTH BENEFITS PROGRAM By ANNESHA WHITE LOVETT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PAR TIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Annesha White Lovett

PAGE 3

3 T o the memory of my mother Brenda White, whose encouragement went far beyond w hat the eye could see and t o my husband, John, whose love and support provided me the strength and perseverance I needed to press forward and t o all who nurtured my intellectual curiosity, academic interests, and sense of scholarship throughout my lifetime making this milestone possible

PAGE 4

4 ACKNOWLEDGMENTS Thanks go out to the chair and members of my supervisory committee for their expertise and time. A special thanks goes to Dr. McKay for her countless hours of reflecting, reading, and patien ce Thank s g o out to Dr. Ward for the many times that you explain ed econometric equations in detail. Thank s go out to Dr. Harman and Dr. McCarty for agreeing to serve on my committee. Special thanks are extended to the Agency for Health C are Administration. Anne We lls, Ray Aldridge, and Vern Hamilton assisted me in overcoming one of the largest obstacles in the completion of my dissertation. A cknowledge ment s go out to the Ph armaceutical R esearch and M anufacturers of A merica Foundation for believing that this resea r ch was important I appreciate the generous monetary support. A special feeling of gratitude is extended to my husband, John Lovett III, who has shared the many uncertainties, challenges and sacrifices for completing this dissertation. Thanks go out t o my father, Henry White, whose words gave me an extra push to accompli sh my dreams T hank s go out to my sister, Tammy White, whose smile and laughter mo tivated me to complete my study. Additional thanks are extended to my many family and friends who ha ve been my emotional anchors through not only the vagaries of graduate school but my entire life. Thank s also go out to Uncle Mark who emphasized the importance of education and directed me toward this journey back in 1995. Thank s go out t o my mentor, Dr. Odedina, whose life example has spoken to me far louder than words ever could. T hanks are extended t o the faculty at Florida A&M University College of Pharmacy who answered my numerous phone calls for assistance T hanks go out t o my church family for prayer support each day Finally, thanks go out to all t he staff and faculty in the Heal th Services Research Department.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIS T OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIONS ................................ ................................ ........................... 10 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 IN TRODUCTION ................................ ................................ ................................ ........ 14 2 REVIEW OF LITERATURE ................................ ................................ ........................ 18 Overview of Medicare ................................ ................................ ............................. 18 Prescription Drug Coverage under Medicare ................................ .......................... 19 Overview of the Federal Employees Health Benefits Program ............................... 21 Prescription Drug Cove rage under the Federal Employees ................................ .... 25 Health Benefits Program ................................ ................................ ......................... 25 Review of Research Literature ................................ ................................ ................ 26 Medicare Part D ................................ ................................ ................................ ...... 26 Enrollment ................................ ................................ ................................ ........ 26 Dual Eligibility ................................ ................................ ................................ ... 28 Health Professionals ................................ ................................ ........................ 30 Beneficiary Choices ................................ ................................ .......................... 31 Expenditures ................................ ................................ ................................ .... 32 Effect on Outcomes ................................ ................................ .......................... 34 Review of Research Literature ................................ ................................ ................ 35 Federal Employees Health Benefits Program ................................ ......................... 35 Comparison of Medicare Part D and the Federal Employees Health Benefits Program Prescription Drug Coverage ................................ ................................ .. 36 Summary ................................ ................................ ................................ ................ 41 3 CONCEPTUAL FRAMEWORK ................................ ................................ .................. 46 Competition and Regulation in Health Care Markets ................................ .............. 46 Decision Making Process: Medicare Part D vs. FEHBP ................................ ........ 52 Hypotheses ................................ ................................ ................................ ............. 54 4 DATA AND METHODS ................................ ................................ .............................. 58 Data Collection ................................ ................................ ................................ ....... 58 Medicare Part D Prescription Drug Plans ................................ ......................... 58

PAGE 6

6 Federal Employees Health Benefits Progr am Prescription Drug Plans ............ 60 Construction of Analytic Datasets ................................ ................................ ........... 61 Definition of Study Variables ................................ ................................ ................... 65 Research Questions ................................ ................................ ............................... 71 Data Analysis ................................ ................................ ................................ .......... 74 Descriptive Analysis ................................ ................................ ......................... 74 Independent Samples t test ................................ ................................ .............. 74 Negative Binomial Regression ................................ ................................ ......... 77 5 RESULTS ................................ ................................ ................................ ................... 96 Formulary Coverage ................................ ................................ ............................... 97 Cost Sharing ................................ ................................ ................................ ......... 107 6 SUMMARY, DISCUSSION, AND CONCLUSION ................................ .................... 135 Formulary Coverage ................................ ................................ ............................. 135 Cost Sharing ................................ ................................ ................................ ......... 148 Limitations ................................ ................................ ................................ ............. 151 Conclusion ................................ ................................ ................................ ............ 153 Policy Implications ................................ ................................ ................................ 154 Future Research ................................ ................................ ................................ ... 155 LIST OF REFERENCES ................................ ................................ ............................. 157 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 171

PAGE 7

7 LIST OF TABLES Table page 4 1 Medicare Part D Plans Selected for Analysis ................................ ........................ 80 4 2 FEHBP Prescription Drug Plans Selected for Analysis ................................ .......... 81 4 3 Datasets 1, 2, and 3 Definition of Study Variables ................................ ............... 81 4 4 Datasets Tier Comparison: Medicare Part D and the FEHBP ............................... 84 4 5 Top Therapeutic Classes and List of All Drugs ................................ ..................... 87 5 1 Overview of Data ................................ ................................ ................................ .. 110 5 2 Plans vs. Formularies ................................ ................................ ........................... 110 5 3 Formulary Coverage of 266 Top Drugs in the U.S. ................................ .............. 111 5 4 Independent Samples t Test: Formulary Coverage of 266 Top Drugs in the U.S. ................................ ................................ ................................ .................. 112 5 5 Independent Samples t Test: Formulary Coverage by Therapeutic Class .......... 112 5 6 Independent Samples t Test: Formulary Coverage for each of the 266 Top Drugs in the U.S. ................................ ................................ .............................. 113 5 7 Formulary Coverage of Prescription Drugs: Brand versus Generic by Formulary 122 5 8 Independent Samples t Test: Formulary Coverage Brand versus Generic of Total Drugs Covered ................................ ................................ ........................ 126 5 9 Independent Samples t Test: Formulary Coverage Brand versus Generic as % of All Brand and Generic Drugs ................................ ................................ .... 126 5 10 Negative Binomial Regression Predicting Number of Drugs per Therapeutic Class among Medicare Part D and the FEHBP ................................ ................ 126 5 11 Negative Binomial Regression Predicting Number of Drugs for ADHD Agents .. 127 5 12 Negative Binomial Regression Predicting Number of Drugs for Analgesics ....... 128 5 13 Negative Binomial Regression Predicting Number of Drugs for Anticancer Agents ................................ ................................ ................................ .............. 128 5 14 Negative Binomial Regress ion Predicting Number of Drugs for Antibacterials .. 128

PAGE 8

8 5 15 Negative Binomial Regression Predicting Number of Drugs for Antic onvulsants ................................ ................................ ................................ 128 5 16 Negative Binomial Regression Predicting Number of Drugs for Antidepressants ................................ ................................ ................................ 129 5 17 Negative Binomial Regression Predicting Number of Drugs for Antipsychotic ... 129 5 18 Negative Binomial Regression Predicting Number of Drugs for Anxiolytics ....... 129 5 19 Negative Binomial Regression Predicting Number of Drugs for Arthritis Agents 129 5 20 Negative Binomial Regression Predicting Number of Drugs for Blood Glucose Regulators ................................ ................................ ................................ ........ 130 5 21 Negative Binomial Regression Predicting Number of Drugs for Blood Products/Modifiers/Volume Expanders ................................ ............................. 130 5 22 Negative Binomial Regression Predicting Number of Drugs for Cardiovascular Agents ................................ ................................ ................................ .............. 130 5 23 Negative Binomial Regression Predicting Number of Drugs for Gastrointestinal Agents ................................ ................................ .................... 130 5 24 Negative Binomial Regression Predicting Number of Drugs for Hormonal Agents ................................ ................................ ................................ .............. 131 5 25 Negative Binomi al Regression Predicting Number of Drugs for Respiratory Tract Agents ................................ ................................ ................................ ..... 131 5 26 Summary of Negative Binomial Regressions Predicting Number of Drugs for Top 15 Therapeutic Classes ................................ ................................ ............. 131 5 27 Co st Sharing Analysis among Medicare Part D (N=19) and FEHBP (N=5) Formularies ................................ ................................ ................................ ....... 132 5 28 Cost Sharing among Top Drugs in the U .S. ................................ ....................... 133 5 29 Independent Samples t Test Comparison of Mean Copay among Plans with Copay by Plan Type ................................ ................................ ......................... 133 5 30 Independent Samples t Test Comparison of Mean Coinsurance among Plans with Coinsurance by Plan Type ................................ .............................. 134

PAGE 9

9 LIST OF FIGURES Figure page 2 1 Estimates of Prescription Drug Coverage a mong Medica re Beneficiaries, 2008 .. 44 2 2 Standard Medicare Drug Benefit, 2010 ................................ ................................ 45 3 1 The Demand Curve ................................ ................................ ............................... 55 3 2 The Supply Curve ................................ ................................ ................................ 55 3 3 Equilibrium ................................ ................................ ................................ ............ 56 3 4 Market Equilibrium ................................ ................................ ................................ 56 3 5 Diagram of the Medicare Part D & FEHBP Decision Making Processes .............. 57

PAGE 10

10 LIST OF ABBREVIATION S FEHBP Federal Emp loyees Health Benefits Program USP United States Pharmacope ia GDP Gross Domestic Product HMO Health Maintenance Organizations PPO Preferred Provider Organizations MMA Medicare Prescription Drug, Improvement, and Modernization Act PDP Prescription Drug Plans MA Medicare Advantage IMS Inter continental Marketing Services HDHP High Deductible Health Plan FFS Fee for Service RUPRI Rural Policy Research Institute GAO Government Accountability Office VA Veterans Administration OPM Office of Personnel Management PBM Pharmac y Benefit Manager P&T Pharmacy and Therapeutics CMS Centers for Medicare and Medicaid Services NDC National Drug Code ADHD Attention Deficit Hyperactivity Disorder FDA Food and Drug Administration SPSS Statistical Package for the Social Sciences Software P ackage STATA Statistics and D ata Software P ackage

PAGE 11

11 TSP Time Series Processor Software P ackage US United States ED Erectile Dysfunction NIDDK National Institute of Diabetes and Digestive and Kidney Diseases

PAGE 12

12 Abstract of Di ssertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COMPARING PRESCRIPTION DRUG COVERAGE BETWEEN MEDICARE PART D AND THE FEDERAL EMPLOYE ES HEALTH BENEFITS PROGRAM By Annesha White Lovett December 2010 Chair: Niccie L. McKay Major: Health Services Research T his study provide s an in depth empirical comparison of Medicare Part D and the Federal Employees Health Benefits Program (FEHBP) with respect to prescription drug adherence to the United States Pharmacopeia ( USP ) guidelines. Results revealed that depending on the method used for analysis, both pr ograms provided broad coverage of top drugs dispensed and sold in the United States. Using the independent samples t test for the 19 Medicare Part D formularies analyzed, formulary coverage of the top drugs dispensed and sold in the United States ranged f rom 72 94% (average 84%), while the range was 85 99% (average 94%) for the 5 FEHBP formularies examined (p<.01). Overall, the independent sample t test findings indicated that the FEHBP plans provided broader drug coverage as compared to Medicare Part D p lans. On the other hand, the regression results indicated that, once other factors and interaction effects were taken into account, the programs were shown to be about the same in terms of coverage.

PAGE 13

13 In regard to cost sharing both Medicare Part D plans and the FEHBP plans utilize fixed dollar copayments more often than coinsurance percentages for generic drugs but f or brand name drugs Medicare Part D plans were more likely to utilize copays while the FEHBP plans were more likely to utilize coinsurance. Furthermore, for the Medicare Part D plans that utilized copayments the mean copayment for generic drugs was $4.53 (range was $0 $8) as compared to the FEHBP plans mean copa yment of $7.67 (range $5 $10) Therefore, the Medicare Part D plans provi ded lower mean copays for generic drugs as compared to the FEHBP plans (p<.05). On the oth er hand, the finding for brand name drugs was non signifi cant. For the Medicare Part D plans that utilized coinsurance for generic drugs, mean rates were 17% as compar ed to the FEHBP plan mean rates of 20% (p<.05). For the Medicare Part D plans that utilized coinsurance for brand name drugs, mean rates were 26% as compared to the FEHB P plan mean rates of 34%. These findings benefit policymakers, health care profession als, and consumers by suggesting lessons that can be learned from both the FEHBP and the Medicare Part D program.

PAGE 14

14 CHAPTER 1 INTRODUCTION The Medicare program was created in 1965 to provide health care for people age 65 and older. Over time, costs associated with the Medicare program have increased steadily and have contributed to the growth in national health expenditures For example, health care expenditures in the United States were $253 billion in 1980, $714 billion in 1990, and rose to $2 tr illion in 2006 (Centers for Medicare and Medicaid Services 2008). In 2006, health care spending was about $7,026 per resident for the (Catlin, Cowan, Hartman, et al 2008) F or Medicare alone, total expenditures in 2006 were approximately 3.1% of GDP or $408 billion (Annual Report of the Boards of Trustees 2007). Although the original Medicare program provided access to numerous health services, prescription drugs were not co vered. To address rising drug costs and to provide more comprehensive health care coverage, the Medicare Prescription Drug, Improvement, and Modernization Act, which created Medicare Part D, was implemented in 2006. Medicare Part D is a voluntary drug be nefit offered by private health plans that covers a wide range of prescription medications listed on the Medicare Part D formulary. Anyone who is eligible for Medicare is eligible for Medicare Part D Upon its inception, over 20 million seniors enroll ed to receive prescription drug coverage (Henry J. Kaiser Family Foundation 2007). The purpose of Medicare Part D was to provide seniors with access to prescription drugs. Yet many questions surrounding coverage for prescription drugs have emerged The re are questions among state policy makers as to the projected cost savings in

PAGE 15

15 switching enrollees from other plans to Medicare Part D plans. Enrollees question whether Medicare Part D plans provide an expansion in prescription drug coverage compared to o ther plans. In addition, concerns about Medicare funding persist as costs continue to rise at alarming rates. Will funding for Medicare eventually run out? What are plans to address rising prescription costs? It has been proposed to reform Medicare to make it similar to the Federal Employees Health Benefits Program ( National Bipartisan Commission 1999) The Federal Employees Health Benefits Program (FEHBP) was created by Congress in 1959 (Butler and Moffit 1995). The program provides health care cove rage to active and retired federal and postal workers and their families, in addition to active and retired members of Congress and congressional staff. Over 400 private plans compete to provide coverage to m ore than 9 million people. Some have cited the saving techniques and provision of quality services as markers of a truly successful program (Heritage Foundation 2003; Francis 2003). Specific advantages include optional enrollment and broad eligibility requirements. In addition, the F EHBP uses community ratings as a disincentive for plans to determine coverage on the basis of beneficiary risk. Furthermore, the program offers more p rovider choice and access, greater rural access, greater achievements in cost control, and better health benefits by covering preventive services, dental services and health care costs incurred abroad (Hoff 1998; Medicare Payment Advisory Commission 2002 ; McBride 2003; Peck 2003; and Francis 1993). For these reasons, the FEHB program has been cited as super ior compared to the Medicare program (Butler and Moffit 1995).

PAGE 16

16 FEHBP type model. Several studies dispute the claim that adopting a FEHBP model would offer an improvement t o the Medicare program (Moon 2002; Oberlander 2000). Prior to the implementation of Medicare Part D, the lack of coverage for prescription drugs made the FEHBP particularly attractive in comparison to Medicare. Now that Medicare Part D has been implemente d, however, there is renewed debate on the desirability of switching enrollees from Medicare to an FEHB type plan. An unanswered question central to the debate is how Medicare Part D and the FEHB prescription drug plans compare. To answer this question, this dissertation will compare Medicare Part D and the Federal Employees Health Benefits program with respect to prescription drug plans. The analysis will focus on the consumer perspective by examining differences in drug coverage and cost sharing. The following research questions will be addressed: 1. How do Medicare Part D stand alone plans compare to Federal Employees Health Benefits plans with respect to coverage of prescription drugs? Ho: There is no difference in drug coverage among FEHB prescription drug plans as compared to Medicare Part D plans. 2. How do Medicare Part D stand alone plans compare to Federal Employees Health Benefits plans with respect to cost sharing on prescription drugs? Ho: There is no difference in cost sharing among FEHB prescr iption drug plans as compared to Medicare Part D plans. By the year 2030, the United States will be comprised of 71 million persons over the age of 65 (Centers for Disease Control 2007). For the first time in history, the United States may have more elde rly individuals than working individuals. Many projections indicate that Medicare will not be able to deliver promised benefits to the next generation of retirees without making changes to the progra m (Butler and Moffit 1995). The p olicymakers and health care professionals are interested in

PAGE 17

17 recommendations to address the anticipated needs of older persons. To avoid extreme increases in payroll taxes and other revenues or major cutbacks in services Medicare must explore ways to change the health care syste m to achieve better value for money. The experience of the FEHBP suggests a possible means of accomplishing this objective. Yet, little empirical research has been done to confirm this viewpoint. This dissertation will contribute to the literature on Medicare reform by providing an assessment of an alternative form of prescription drug coverage. It will provide information for third party payers to aid in their discussions surrounding Medicare reform. Furthermore, results of this study may help older persons who are interested in receiving help with a complex network of prescription drug services, but need more information about the potential benefits of services.

PAGE 18

18 CHAPTER 2 REVIEW OF LITERATURE This dissertation compares the Medicare Part D and t he Federal Employee Health Benefits plans with respect to prescription drug coverage. This chapter first provides a general overview of Medicare and the Federal Employee Health Benefits Program. Next, background information is presented on the prescripti on drug component of the two programs, followed by a review of the research on Medicare Part D and the FEHB prescription drug plans. Lastly, the chapter discusses how this research addresses gaps in the literature. Overview of Medicare Medicare was crea ted under Title XVIII of the Social Security Act in 1965, providing Medicare Part A and Medicare Part B for those age 65 and older. In 1972 Medicare coverage was extended to those that were disabled and to individuals with end stage renal disease. Medica re is administered under the Department of Health and H uman Services Coverage under Part A, with automatic enrollment for eligible individuals, includes care rendered in a hospital, skilled nursing facility care and home health care. Part A is financed through payroll taxes. Coverage under Part B is voluntary and includes outpatient services such as physician services, outpatient hospital care, laboratory tests, speech therapy, ambulance services and medical equipment. Part B is financed through federa l appropriations and premiums paid by enrollees. On August 5, 1997, the Balanced Budget Act, which contained a number of Medicare reforms, was signed into law. Specifically, Medicare+Choice or Medicare Part C was created. Up until that time beneficiari es enrolled in Part A and/or B. Now

PAGE 19

19 beneficiaries had the option of enrolling in a private health insurance plan, a Medicare Advantage Plan. Medicare Advantage plans combine Part A and Part B (Medicare Consumer Guide 2008). Medicare Advantage plans incl ude Health Maintenance Organizations (HMOs), Preferred Provider Organizations (PPOs), private fee for service plans and Medicare special needs plans. Medicare is financed through a combination of payroll taxes, general tax revenues, and premiums paid by beneficiaries. The Medicare Hospital Insurance trust fund currently is operating under a deficit and is projected to be depleted by 2019 (Annual Report of the Boards of Trustees 2007). Medicare reform is needed to address this problem. Prescription Dru g Coverage under Medicare Public debate led to the creation of Medicare Part D through the Medicare Prescription Drug, Improvement, and Modernization Act (MMA), which became effective on January 1, 2006. Medicare Part D is provided exclusively through pri vate insurers, with coverage provided either by stand alone prescription drug plans (PDPs) or comprehensive managed care plans called Medicare Advantage (MA) plans (National Health Policy Forum 2004). Enrollees in Medicare Part A or B are eligible for Par t D. Beneficiaries enrolled in MA plans receive drug benefits as a part of their Medicare Advantage plan, while those enrolled in fee for service coverage may enroll in a stand alone prescription drug plan (National Health Policy Forum 2004). In 2006, a bout 22 million seniors enrolled in Medicare Part D (Henry J. Kaiser Family Foundation 2007). Enrollment is voluntary and for those with no coverage it means new access to thousands of prescription drugs. Additionally, other Medicare beneficiaries have t he option to replace their existing drug coverage (e.g. Medigap or

PAGE 20

20 Medicaid) or keep it (e.g. Veterans Administration) as long as it meets the same standards as Medicare Part D coverage (i.e. creditable coverage). As shown in Figure 2 1, by 2008 enrollme nt in Medicare Part D was 25.4 million, or approximately 57% of all Medicare beneficiaries. Drug coverage under Part D varies by plan. Formularies must meet criteria established by the United States Pharmacopeia (United States Pharmacopeial Convention 200 8). For example, plans are required to include drugs within a pre established list of therapeutic categories and classes (National Health Policy Forum 2004). The process by which drugs are selected for coverage is discussed in more detail in Chapter 3. F igure 2 2 shows the standard Medicare drug benefit as of 20 10 The Part D monthly premium was about $39 a month (Kaiser Family Foundation 2009) Cost sharing included a $31 0 deductible, coverage of 75% of drug costs between $310 and $2,83 0, and no covera ge for drug costs $ 2,830 to $ 6,440 (this is the so called 6,440 and are required to only pay the greater of a co payment or coinsurance of 5% ( Kaiser Family Foundation 2009 ) prescription drug benefit. Policymakers wanted to provide extensive drug coverage, but they did not want to give seniors a blank check for utilization. As shown in Figure 2 2 under the Medicare Part D program seniors receive drug coverage until thei r costs reach an amount of $2,83 0. At that time coverage stops and then resumes again when they reach $6,44 0. During the coverage gap seniors must pay out of pocket.

PAGE 21

21 Oberlander ( Medicare budget. A study by the Kaiser Family Foundation (2008) examined the number of beneficiar ies reaching the gap in coverage. Data was extracted from the Intercontinental Marketing Services (IMS) Health Longitudinal Prescription Drug Database. The final sample consisted of 1.9 million Medicare Part D enrollees. Findings revealed that 26% of en rollees reached the coverage gap. Furthermore, older beneficiaries reached the gap sooner. A finding that is particularly important from a policy perspective is that many who reached the coverage gap changed medications or stopped taking their medicine a ltogether. This is problematic because research has shown that when patients stop taking prescribed medication, especially those with chronic conditions, the result is worse health outcomes and a n increase in health care costs ( Hoadley, Hargrave, Cubanski and Neuman 2008). Medicare Part D has been in operation since 2006 and enrollment has grown to nearly 25 million people. Although many are satisfied with their increased drug poses a unique problem. Other areas of concern are enrollment, dual eligibility, the impact on health professionals, and beneficiary choice of plan. And of particular concerns is the adequacy of financing for the program in the future. Overvi ew of the Federal Employees Health Benefits Program Created in 1959, the Federal Employees Health Benefits Program is the largest employer sponsored voluntary health plan in the world (Ruddock 1966). All federal employees who elect FEHBP coverage register during an open enrollment period

PAGE 22

22 (November to December of each year). Once a plan is selected changes are allowed only in the occurrence of a major life event or during the annual open enrollment period There are five types of plans to choose from : Fee For Service non Preferred Provider Organization (PPO) plans, Fee for Service PPO plans, Health Maintenance Organization (HMO) plans, HMO Plans Offering a Point of Service Product, and Consumer Driven Health Plans. Furthermore, enrollees have the option o f selecting a Health Reimbursement Arrangement, Health Savings Account, or a High Deductible Health Plan (HDHP) (FEHBP Home Page 2008) Enrollees in the Fee for Service (FFS) non PPO plans may obtain health services and then upon filing be reimbursed or al low the plan to pay the medical provider directly. The choice of physician is unlimited. Beneficiaries usually pay more for these plans compared to other plans. Fee for Service PPO plans give the beneficiary access to physicians who reduce their charges to the plan (a network of physicians) resulting in decreased out of pocket spending when using a PPO provider. Furthermore, individuals do not have to file claims or paperwork (FEHBP Home Page 2008). On the other hand, Health Maintenance Organizations a re somewhat more restrictive. Providers must be selected from a network of physicians and hospitals in certain geographic areas. Enrollees are cha rged a co pay for physician visits, but there are no deductibles or coinsurance for hospital visits. The adv antages of these plans are that there is no paperwork or billing involved, which results in low out of pocket costs (FEHBP Home Page 2008). Another option, the HMO Plans offering a Point of Service Product, is an alternative for individuals looking for an HMO that allows use of non network providers. These plans are more costly as shown by higher deductibles and

PAGE 23

23 coinsurances. Beneficiaries must also file claims for reimbursement (FEHBP Home Page 2008). Finally, for individuals who desire to have more con trol over their health care costs, the FEHBP offers Consumer Driven Health Plans and Health Savings Accounts. Under Consumer Driven Health Plans a maximum amount is designated to spend on health care and full coverage is provided for preventive care. The se plans require higher cost sharing once a beneficiary exceeds the designated amount. Following the same concept, Health Savings Accounts utilize pre tax funds and require beneficiaries to save for future health care services. Once deposited into an acc ount, the money is not taxed and is used to pay for current health care services (FEHBP Home Page 2008). The FEHBP beneficiaries have many options. Health pl ans offered are varied not only in their provisions, but also in the associated premiums. Under t he FEHBP, the government makes a biweekly contribution, an amount fixed by law at one half the cost of the least expensive option offered by either one of the two government wide plans (Ruddock 1966). Although considered a generous amount to some, there i Individuals can select from a high option or a basic option enrollment based on information obtained from the Office of Personnel Management. This information includes a list of services provided by the plan (e .g. a list of covered drugs, co pays, coinsurance, premiums and deductibles). The majority of enrollees (86%) choose the high option (Ruddock 1966). Each plan offers a range of prescription drug benefits and enrollees can choose from 12 20 p lans, depending on geographical region.

PAGE 24

24 Three plans (Blue Cross Blue Shield, Government Employees Hospital Association, and PacifiCare of California) cover more than half of all FEHBP enrollees (over 4 million) and those plans paid about $3.3 billion for about 65 million prescriptions in 2001 (Dicken, Agarwal, Dirosa et al 2003). Currently, 2008 data for the standard FEHBP Blue Cross/Blue Shield Plan shows that the employee pays $ 1,616 per year f or an individual plan and $ 3,774 per year for a family plan (Blue Cross Blue Shield Service Benefit Plan Brochure 2008). Furthermore, the FEHB plan premiums rose an average of 7% in 2009 (Office of Personnel Management 2008). A GAO study (2007) was conducted in response to the rising premiums seen in FEHBP plan s. Data obtained from the Office of Personnel Management were examined Retirement System (the second largest public purchaser of employee health benefits). Additional survey d ata was obtained from the Kaiser Family Foundation/Health revealed that premium growth slowed down from 12.9% in 2002 to 1.8% in 2007. The top ten FEHBP plans had a range of 0 1 5.5% in regard to premium growth in 2007. Those plans that did increase their premiums reported increases in utilization, cost of health services, and a high proportion of elderly enrollees as the cause, while other plans saw decreased pr emiums due to the provision of less generous coverage. A study by Mueller and colleagues (2005) examined the availability of choices for federal employees in rural areas. Data was derived from the Health Benefit Data File, the Office of Personnel Management Web site, the Research Institute (RUPRI) Medicare Capitation Files. A multivariate analysis was

PAGE 25

25 conducted to explore whether characteristics of rural areas influence enrollment. Results revealed that Blue Cross Blue Shield represen ted 60% of rural enrollment. Furthermore, enrollment in regional plans was higher in urban areas as compared to rural areas. A positive correlation was found between enrollment and the location of the county (i.e. plans were more likely to be found in cen tral urban counties and rural adjacent counties as compared to rural nonadjacent counties). Therefore, highly populated areas attract more plans. The authors concluded that in some rural counties, available choices are few because not all pl ans are exper iencing enrollment (Mueller, McBride, Andrews et al 2005). Prescription Drug Coverage under the Federal Employees Health Benefits Program Under FEHBP, private plans compete to provide health insurance to enrollees. The plans must meet certain criteria to be able to compete. For example, plans must emergency care, treatment of mental conditions and prescription drug coverage. Once plans contract with the FEHBP (through the Office of Personnel Management) benefits and premiums are negotiated. Service provided by plans may vary in location, including local, regional, or national markets. Furthermore, plans may service particular groups such as postal carriers. Each FEH B health plan has its own formulary. The formulary includes a list of generic and brand name drugs that are covered by the health plan. The formulary is developed by a team of pharmacists and physicians that review and update the list periodically (FEHBP Home Page 2008). The process by which drugs are selected for coverage is discussed in more detail in Chapter 3.

PAGE 26

26 Review of Research Literature Medicare Part D The research literature on Medicare Part D is extensive and continues to grow. This section rev iews the research on Medicare Part D related to enrollment, dual eligibility, health professionals, beneficiary choices, expenditures and the effects of drug coverage. Enrollment Medicare Part D enrollment is increasing. By early 2007 approximately 17 mil lion seniors had enrolled in stand alone prescription drug plans (Henry J. Kaiser Family Foundation 2008), although target populations such as low income seniors only enrolled in small numbers. However, policymakers expected over 29 million seniors to enr oll during the first year of the program (Henry J. Kaiser Family Foundation 2006), in order to ensure fair competition among plan providers. Cubanski and Neuman (2006) analyzed the enrollment of seniors in Medicare Part D. Data for the year 2006 were de rived from the Centers for Medicare and Medicaid Services and included information on prescription drug plans (premiums, deductibles, co payments and gap coverage). Researchers determined which firms increased market share within Medicare and examined tre nds by plan type (PDP vs. MA PD) and benefit design (standard vs. enhanced). The final study sample included 20.4 million enrollees and 2,811 Part D plans (1,446 PDPs and 1,365 MA PD plans). Results revealed that 10 companies covered 72% of enrollees. Although numerous plans were offered, enrollees were concentrated in a small number of plans. This concentration of enrollees may be explained by the MA patterns found prior to Medicare Part D implementation. Many of the same MA organizations saw incre ased

PAGE 27

27 enrollment as Medicare Part D beneficiaries were added. Additionally, some organizations merged, such as United and PacifiCare, resulting in a higher concentration of enrollees. On the other hand, the organizations with low enrollment faced concerns about risk aversion, decreased ability to negotiate low drug prices and warnings from Medicare related to cancellation of contracts. For example, the Center for Medicare and Medicaid Services discussed non renewal of those contracts representing less tha n 5,000 enrollees, a total of 13 organizations (Cubanski and Neuman 2006). Of the top 15 plans, 11 were stand alone plans. United AARP Medicare Rx and and emphasizing low premiums to attract enrollees. Thus, more enrollees took advantage of provisions through stand alone PDPs as compared to MA PD plans (Cubanski and Neuman 2006). T he majority of Part D plans did not offer the standard 2006 benefit (i.e. $250 deductible, 25% coinsurance, $2,850 coverage gap and catastrophic coverage once beneficiaries exceed $3,600) and only 17% of enrollees chose the plans that did offer this standard benefit. Fifty two percent of enrollees chose plans that had no deductible and a tiered co payment structure. Furthermore, 30% of enrollees were in plans with low or no deductible, providing coverage of some excluded drugs and offering gap coverage. Lastly, results showed that in regard to the coverage gap, 95% of plans included a coverage gap and 94% enrollees were in plans that include a gap. It is projected that 48% of enrollees will be affected by the gap due to dual eligibility and low income subsidies (9.3 million enrollees).

PAGE 28

28 The authors concluded that Medicare did meet its goals for enrollment of beneficiaries in Medicare Part D (Cubanski and Neuman 2006). The authors further noted that the relatively small number of plans dominating the market may change as other plans reassess themselves to become more competitive. Name recogniti on and low premiums seemed to be the most effective way to attract enrollees, while gap coverage proved to be not as important as expected. This study reflects the many concerns regarding Medicare Part D enrollment. While many of the goals for enrollment have been met, policymakers must continue to monitor the enrollment process to best meet the needs of beneficiaries in the future. Another important policy issue is dual eligibility. Dual Eligibility Dual eligible individuals are Medicare beneficiaries who are also eligible for Medicaid In a report by the Kaiser Family Foundation (2005), dual eligibles were cited as the most costly, sickest, and poorest group within the Medicare population (Nemore 2005). Under Medicaid, beneficiaries had wider access to drugs included on the formulary. Then, when Medicare Part D was enacted, dually eligible individuals were instructed to choose a Medicare prescription drug plan for drug coverage. As a way to meet the growing needs of this group, a low income subsidy was provided under Medicare Part D, which covered premiums, deductibles and the coverage gap. Many dually eligible seniors have complained of decreased access to medications and higher co pays after the switch. For example, under Medicaid if a n individ ual cannot afford a co pay for a prescription he/she will not be denied that medication. On the other hand, under Medicare Part D the individual plans determine which drugs will be included on the formulary. Those drugs may or may not be the drugs that ar e

PAGE 29

29 needed. If a drug is not listed on the Medicare Part D plan formulary the individual must pay out of pocket. A GAO report (2007) assessed the process for enrolling dual eligible individuals and also examined access to drugs by dually eligible individua ls after enrolling in the program. Results showed that the process for enrolling dual eligible individuals was complicated and time consuming (King et al 2007). Administrative burden and the many individuals involved in the process were noted as limitat ions for some seniors. Furthermore, enrollment for many dual eligible individuals may result in time delays for medications. For example, if an individual switches from Medicaid to Medicare Part D, there may be a delay in the time that the pharmacy recei ves their up to date information. During this time delay individuals may be denied access to medications or forced to pay for them out of pocket. GAO concluded that CMS should notify seniors of reimbursement rules and monitor progress (King et al 2007). An article by Basu et al (2008) also examined the effects of Medicare Part D on dual eligible individuals and came to different conclusions. Prescription claims data on over 10,000 dual eligible individuals were obtained from a national pharmacy chain. T he authors used regression models to compare treatment and control groups in terms of drug usage, out of pocket spending, and total drug expenditures. Results revealed that in terms of drug usage there was not a statistically significant difference in tre atment and control groups for the probability of continuing, discontinuing, or initiating a new medication. Additionally, there was no significant difference in out of pocket spending for treatment and control groups. Lastly, total drug expenditures decr eased for both the treatment and control group after Medicare Part D was implemented. The

PAGE 30

30 authors concluded that the implementation of Medicare Part D did not adversely affect dual eligible individuals (Basu, Yin and Alexander 2008). However, studies on Medicare Part D beneficiaries with mental illnesses have shown that issues of dual eligibility have resulted in interrupted treatment, relapses, and increased hospitalizations (Huskamp, Stevenson, Donohue, Newhouse and Keating 2007; Daly and Moran 2007) The healthcare infrastructure is wide and vast. Responding to the need of transitioning over 6 million dual eligible individuals has been challenging. CMS must not only meet the needs of beneficiaries, but there is also the often daunting task of addre ssing the concerns of health professionals. Health Professionals There are extensive reports in the literature on the views of physicians, pharmacists, and other health professionals in regard to Medicare Part D. Increased administrative burden and decre ased patient access to drugs are often mentioned as issues of concern ( Epstein, Rathore Alexander, Ketcham 2008; National Survey of Pharmacists 2006). A study by Epstei attitudes about Medicare Part D. Over 700 primary care physicians completed surveys in North Carolina, Florida, Massachusetts, and Texas. Results revealed that although many physicians (48%) had a favorable vi ew of Medicare Part D overall, 44% reported that access declined for individuals with prior drug coverage. Furthermore, more than half of the physicians (64%) were dissatisfied with Medicare Part D formularies stating that the formularies did not meet pat ients' needs. Finally, 63% reported higher administrative burden after Medicare Part D was implemented. Similarly, in a study by the Kaiser Family Foundation (2006), 802 pharmacists were surveyed. Findings revealed that 91% of pharmacists felt the Medi care Part D

PAGE 31

31 program was too complicated. This was supported by the fact that 97% of the (National Survey of Pharmacists 2006). The American Society of Health System Pha rmacists issued a statement to the Senate Special Committee on Aging Hearing on Medicare Part D. Pharmacists expressed difficulty in corresponding with plans to answer questions, in filling prescriptions for seniors who switched plans toward the end of th e month, in verifying enrollment status, and in reimbursement for emergency prescriptions. Furthermore, pharmacists were concerned that the majority of plans do not allow a 31 day supply of prescriptions or partial fills of controlled substances, which ar e common preferences among seniors (ASHP 2006). In addition to the issues of physicians and pharmacists, various other professionals have presented concerns to be addressed by policymakers. For example, a study by Summer and colleagues (2008) summarized t he responses of counselors, attorneys, program managers, health professionals, and others. These individuals were identified as those who had direct knowledge of Part D beneficiaries' experiences. Their major concerns involved the accuracy of information as well as the ease of use of information related to Medicare Part D plans. Survey respondents recommended that CMS improve communication among organizations, develop a system that incorporates beneficiary drug needs in the selection of a drug plan, and develop a more rigorous monitoring process. Respondents also noted that there were too many drug plan choices, a common criticism of the Medicare Part D program. Beneficiary Choices A growing body of literature focuses on the choices beneficiaries face in selecting plans. Enrollees typically have 45 to 57 plans to choose from, with the plans varying by

PAGE 32

32 premium, co payment and formulary. The guides that were created to facilitate these choices have been criticized as being too complex and not comprehensive enough (Hoadley 2006). A recent article by the Commonwealth Fund (2008) highlighted views for and against standardization of the program. Ideas for standardization include requiring plans to use universal terms to describe benefits, requiring plans to h ave the same rules for cost sharing, and creating universal rules for formulary design (Hoadley 2008). Those in opposition to standardization claim that as the market adjusts itself over time standardization will not be needed. They feel that the program will become less complex and standardization may occur naturally as plans respond to the market (Hoadley 2008). The authors note that beneficiaries support the idea of standardization. Although they enjoy choices many stated that there are simply too ma ny plan options. These concerns reflect the complexity of the Medicare Part D program. Expenditures Medicare Part D expenditures have been a concern since its inception and continue to challenge policymakers. The challenge stems from an attempt to main tain a program that provides high quality services while containing healthcare costs. Projections from the Congressional Budget Office estimate that the Medicare Part D program will cost about $811 billion over the years 2007 2016 (Congressional Budget Of fice 2007). An article by Stuart and colleagues (2005) assessed out of pocket spending by Medicare Part D beneficiaries. The authors used a simulation model to project out of pocket spending for those who enrolled during the first 3 years of the Medicar e Part D program. Data were derived from the 1998 2000 Medicare Current Beneficiary Survey. The study sample was comprised of 1,333 enrollees per capita drug spending. The

PAGE 33

33 sample was projected using the National Health Accounts representing expectations had the Medicare Part D program not been implemented. Three groups were selected for subsidized coverage ( having incomes above 150% of the federal poverty level) an d cohort of potential Part D enrollees with projected 2006 prescription spending above s pending above $5,100. Coverage thresholds were established and out of pocket spending for those expected to enroll was tracked quarterly for two years. Results revealed that average spending estimates for the entire community dwelling Medicare population was $3,081 in 2006 and $3,891 in 2008, while the potential Part D enrollee population spent on average $2,608 in 2006 and $3,230 in 2008. Furthermore, the high spenders mean drug spending was $5,534 in 2006 and 2008, while catastrophic spenders were expe cted to spend on average $9,106 in 2006 and $7,328 in 2008. Overall, over 3 years potential Part D enrollees were expected to pay for about 44% of total drug spending out of pocket. Furthermore, about 40% of potential Part D enrollees were expected to s pend in the doughnut hole, while 15% will reach the they estimate d that quarterly out of pocket payments for enrollees with average prescription drug spending woul d range from $163 in 2006 to $590 in 2007, while the high spender cohort prescription drug spending would range from $401 in 2006 to $1,391 in 2007. The authors conclude d that beneficiaries with high or catastrophic drug spending would be affected the mos t by the Part D benefit structure. High spenders

PAGE 34

34 could pay more than $10,000 out of pocket from 2006 2008. Considering that the median income for those age 65 and older is about $23,000 this could pose serious concerns related to beneficiary health outco mes. The authors recommend ed that future research in this area address these concerns and explore the importance of costly off formulary medicines and their ability to further increase beneficiary out of pocket spending. On the other hand, some studies re port decreases in spending since the implementation of Medicare Part D. One study found that higher plan participation and lower premiums than anticipated resulted in less spending (Goldman and Joyce 2008). Vogt and colleagues analyzed the effects of Med icare Part D on out of pocket spending using 2006 claims data, and found that Medicare Part D decreased out of pocket spending by 16%, while increasing the number of prescriptions filled by 6.5% (Vogt, Joyce, and Goldman 2008). Effect on Outcomes Recent a rticles have begun to examine the effects of prescription drug coverage (or lack of) on health outcomes. Hsu and colleagues examined the consequences of caps on Medicare drug benefits. The authors used a prospective cohort design to compare over 150,000 Medicare+Choice beneficiaries on the basis of economic and clinical outcomes that occurred in 2003 The data were obtained from Kaiser Permanente. Logistic regression was employed to examine costs. Results revealed that about 79% of beneficiaries had a $1,000 cap on their drug benefits and gaps in coverage resulted in negative effects on health outcomes (Hsu, Price, Huang, Brand, et al 2006). Specifically, beneficiaries with no cap on their drug coverage had lower odds of non adherence to their medicati ons. Beneficiaries with the $1,000 cap had higher

PAGE 35

35 odds of high blood pressure, high cholesterol, and high blood glucose levels. Furthermore, beneficiaries with capped benefits had more emergency room visits and hospitalizations (Hsu, Price, Huang, Brand, et al 2006). Similarly, a study by Rosenberg (2007) reviewed the literature on drug coverage provided by Medicare Part D in relation to the outcomes of patients with psychotic disorders. Study findings revealed that for patients with mental illness Med icare Part D coverage was inadequate. This represents about 17% of the total Medicare Part D population. The authors cite lack of continuity of care and insufficient drug coverage as areas for improvement. A similar study found that when drug coverage w as inadequate in the psychiatric population the result was increased emergency room visits, increased hospitalizations and worse health outcomes (Soumerai 1994). In fact, it has been suggested that some Medicare Part D participating plans developed a stra tegy to discourage the enrollment of high cost individuals, such as those with mental illnesses. Psychiatric drugs are very costly, which creates a potential for adverse selection with respect to these individuals. Review of Research Literature Federal Employees Health Benefits Program Little research has been published about the Federal Employees Health Benefits program. Although a search of the literature found no empirical studies, there is a growing body of conceptual papers that focus on the succes s of the Federal Employees Health Benefits program. Discussion of these papers helps to provide background on why this program has been cited as exemplary. Unique features of the program, such as cost saving techniques, are highlighted.

PAGE 36

36 T he adoption o f the Federal Employees Health Benefits program has been recommended as the way to provide insurance coverage in the future (Enthoven 2003; Richmond and Fein 2005). Enthoven (1989) described managed competition as the idea that purchasers could create inc entives for consumers and providers to make better choices given scarce resources. He cited the FEHBP as an example of this type of reformed healthcare system in which organized units are responsible for care resulting in more cost effective choices, imp roved patient satisfaction and better health outcomes. Another example is a book by Richmond and colleagues that discussed current views on universal health insurance (Richmond and Fein 2005). The book cites the FEHBP program as an ideal model for healthc are reform that would allow for an incremental approach to coverage, financing, and enrollment. Similarly, Palmisano and colleagues (2004) cited the Federal Employees Health Benefits program as a good example for health insurance reform. In their article representatives of the American Medical Association discussed proposals to expand healthcare. They noted the FEHB ures as exemplary. Comparison of Medicare Part D and the Federal Employees Health Benefits Program Prescription Drug Coverage Given that the FEHBP has been proposed as a substitute for Medicare, it is important to compare the programs. Only two studies w ere found in the research literature that directly compared the two programs, indicating the need for additional research in this area.

PAGE 37

37 T he National Bipartisan Commission on the Future of Medicare was created in 1997 as a part of the Balanced Budget Act. It wa s composed of 17 members, D emocrats and R epublicans, who we re responsible for examining the Medicare program and making recommendations for improvement All recommendations required 11 votes (not a majority vote) to ensure fairness. In 1999, the Co mmission proposed to reform Medicare into a FEHBP type model. The proposal did not receive enough votes to be officially submitted to Congress, but advocates continued to promote the proposal ( National Bipartisan Commission 1999 and Oberlander 2007). A pa per by Cain (1999) discussed the similarities and differences between Medicare and the FEHBP. Both programs were created in the 1960s and are federally underwriting, a ge rating, and waiting periods for preexisting conditions. Furthermore, both programs utilize pharmacy benefit managers in the provision of drug coverage. In regard to differences, Medicare originated from health policy debates that occurred over a long period of time to address unmet health care needs, while the employees. Additionally, Medicare has payment policies with an excessive amount of detail ( e.g. the volume of laws and regulations exceeds 800 pages per million beneficiaries), stemming from the large number of associated interest groups, while the FEHBP mitigate d interest group politics ( e.g. laws are about twenty pages per million beneficiaries a ratio of forty to o ne compared to Medicare ). Another important difference between the two programs is that Medicare uses its negotiating power to set prices, while the FEHBP attempts to regulate the market by setting boundaries.

PAGE 38

38 Specifically, Medicare sets criteria that p lans must meet in order to get paid, whereas the FEHBP negotiates premiums and benefits with the health plans individually. faction ratings. For Medicare, the federal government holds the financial risk compared to the FEHBP, where the government is involved only as an employer (Cain 1999). For example, the governmental finance committee for the FEHBP is the Post Office and C ivil Service Committee of Congress, while the Ways and Means and Finance Committees govern Medicare. The Civil Service committees are not focused on health policy because the cial risk, and consumers have multiple choices (Cain 1999). This study highlighted many of the advantages of the FEHBP as compared to the Medicare Part D program. Other studies have come to similar conclusions even when including in the comparison addit ional large employer health providers. A 2008 study by Yamamoto and colleagues examined the benefits provided by Medicare and FEHBP as compared to large employer plans. Hewitt Associates provided data on the original fee for service Medicare program, the FEHB P Blue Cross/Blue Shield standard nationwide PPO option, and a typical provider organization (PPO) plan. A benefit value was calculated based on an expected value for participants eligible for Medicare. The comparison involved the value of benefits for Medicare beneficiaries with three health care utilization states: healthy beneficiaries, moderately healthy beneficiaries, and beneficiaries with poor health (Yamamoto, Neuman and Strollo 2008).

PAGE 39

39 The data were derived from the year 2007 Results revea led that the average benefit value of Medicare ($10,610) was lower than that of the typical large employer PPO ($12,160) and the FEHBP standard option ($11,780). Furthermore, Medicare was shown to be less generous due to higher cost sharing and lack of an out of pocket limit on services provided under Part B (Yamamoto, Neuman and Strollo 2008). Medicare also provided less drug coverage when compared to the typical large employer PPO and the FEHBP standard option. In regard to total benefit costs, Medicare paid a smaller share (74%) as compared to the typical large employer PPO (85%) and the FEHBP standard option (83%). compared to the typical large employer PPO (73%) and the FEHBP standard option (80%). When comparing individuals based on utilization, study results revealed that Medicare is less generous for low, moderate, and high users as compared to the typical large employer PPO and the FEHBP standard option (Yamamoto Neuman and Strollo 2008). For example, for low cost Medicare beneficiaries the benefit value was $1,470 on average, while it was $1,650 for the typical large employer PPO and $1,690 for the FEHBP standard option. Similarly, for high cost Medicare bene ficiaries the benefit value was $42,440 on average, while it was $47,600 for the typical large employer PPO and $44,860 for the FEHBP standard option (2007 data) The authors concluded that the Medicare Part D program was less generous than either the typ ical large employer plan or the FEHBP standard plan option (Yamamoto, Neuman and Strollo 2008).

PAGE 40

40 Support for these findings came from a study by researchers of the Lewin Group. They compared the drugs listed on the formularies of two Medicare Part D plans the Veterans Administration Program and the Federal Employees Health Benefits Program (Lewin Group 2007). Data were derived for the year 2006 Data (VONA) and included the top 300 drugs (i.e. highest script volumes) used by tho se age 65 and older. Other data were collected from the VA ) pharmacy website, the Centers for Medicare and Medicaid Services, and the Blue Cross Blue Shield Federal Employee Program website. Results revealed that, of the 300 dr ugs, the FEHBP covered the most (284 or 95%), followed by Medicare Part D plans (282 or 94%) and the VA formulary (194 or 65%). Specifically, the Medicare Part D plans covered more of the 132 brand name drugs (128 or 97%) compared to the FEHBP (125 or 95% ) and the VA formulary (56 or 42%). The authors concluded that the FEHBP and Medicare Part D formularies provide broader drug coverage than the VA formulary. However, concerns persist about the adequacy of drug coverage through Medicare Part D formularie s. Additional research is needed to provide a comprehensive comparison of drug coverage between Medicare Part D and the FEHBP. The two studies mentioned above involve a comparison of one FEHBP plan (i.e. BCBS) and one Medicare Part D plan. Although the a nalyses were helpful, these studies are limited in that the FEHBP program is comprised of about 222 plans, while the Medicare Part D program is comprised of over 1,500 plans. This dissertation proposes to address this gap in the literature by comp aring pl ans/formularies that represent 63 70 % of total program enrollment (i.e. 5 FEHBP formularies and 19

PAGE 41

41 Medicare Part D formularie s). Furthermore, this dissertation proposes to analyze over 250 drugs by therapeutic class, another limitation that has not been a ddressed in the literature. Summary Review of the literature revealed that additional research is needed to compare Medicare Part D prescription drug plans and the FEHB prescription drug plans. Researchers have described the Medicare Part D program as e xperimental, thus requiring continual evaluation and revision (Berndt 2004). This dissertation will focus on the following areas: assessment of prescription drug plan formularies and examination of cost sharing. This section explains gaps in the researc h literature in these areas and how this dissertation will fill the gaps in the literature. The three key areas that legislators had to address in the passing of Medicare Part D legislation were financing, delivery, and regulation. Research to date has f ound that delivery of services has been successful thus far, while some beneficiaries remain concerned with the comprehensiveness of those services. Regulation has been adequate, with the exception of a lack of guidelines to monitor progress and enforce r ecommendations for improvement. Financing, however, poses one of the biggest challenges. As the Medicare program faces large increases in cost incurred by beneficiaries in the coming years, possible reforms must be considered. Suggestions to increase t he effectiveness and efficiency of Medicare Part D will be helpful. Studies have shown that cost sharing may be used as a cost containment measure to redirect enrollee use of health services. Insurers may increase cost sharing on certain drugs (e.g. bran d vs. generic) to limit their use. Encouraging the use of generic drugs whenever clinically appropriate will result in cost savings. The use of

PAGE 42

42 generic drugs is encouraged through lower or waived co payments and formulary compliance programs such as step therapy. Generic drug s cost on average, 71% less than brand drugs. For each percentage point increase in the generic utilization rate, Part D drug spending falls by an estimated $12 billion (Price Waterhouse Coopers 2007). This dissertation will includ e an analysis of generic and brand name drug coverage. Previous research also has revealed a need to examine the inclusion or exclusion of certain drugs on the formulary. Medicare Part D and the FEHBP formularies have not been studied in detail. Examin ing formularies is difficult because of the large number of drugs offered (over 1,400) and the lack of a consensus across plans. Drugs vary by dose, strength, and route of administration. Furthermore, prescription drug plan formularies change once a year While additions and deletions usually do not exceed 15 20 drugs, the fact remains that one of the deleted drugs may be viewed as a life saving drug by a consumer. Some have asked what makes a good formulary The Centers for Medicare and Medicaid Servi ces attempted to answer this question by employing the United States Pharmacopeia to determine whether a plan should target the most commonly prescribed drugs or offer a broad and comprehensive list of drugs. To date, researchers have not been in agreemen t in their conclusions, but they do agree that there is room for improvement. Co payments are another area of concern. Studies show that 91% of Medicare Part D plans have tiered cost sharing (Medicare 2006). A prescription drug plan may have several tier s, and the co payment amount depends on in which tier the drug is listed. Studies have shown that Medicare Part D plan tiers range from 1 4 (Bowman,

PAGE 43

43 Rousseau, Silk, and Harrison 2006), while the FEHBP plan tiers range from 1 6 (Heritage Foundation 2003). Examination of the drug co payments and tiers is important when considering beneficiary access. This dissertation will compare co payments among Medicare Part D and the FEHBP prescription drug plans. In conclusion, this dissertation will directly compa re the Federal Employees Health Benefits prescription drug plans to Medicare Part D stand alo ne plans representing the top 63 70 % of plan enrollees. Results will be useful in considerations of future reform of Medicare Part D. Furthermore, this dissertat ion will use 2009 data, which offers an opportunity for an up to date and timely study. Finally, this research will provide help to older persons who are interested in understanding complex benefit structures to assist in their management of medication co sts. The research questions utilized to address the above mentioned gaps in the literature will reflect the perspective of the consumer. Studies have shown that health policy analyses are sensitive to perspective. For example, assessing drug coverage from a plan perspective versus a consumer perspective would lead to very different conclusions. This research utilizes the consumer perspective. Research questions are: 1. How do Medicare stand alone plans compare to Federal Employees Health Benefits plans with respect to coverage of prescription drugs? 2. How do Medicare stand alone plans compare to FEHBP plans with respect to cost sharing on prescription drugs?

PAGE 44

44 Figure 2 1. Estimates of Prescription Drug Coverage a mong Medicare Beneficiaries, 2008

PAGE 45

45 F igure 2 2. Standard Medicare Drug Benefit, 20 10

PAGE 46

46 CHAPTER 3 CONCEPTUAL FRAMEWORK Competition and Regulation in Health Care Markets Although various theories have been proposed to explain the complex decision making process that Medicare Part D and the FE HB prescription drug plans utilize, there is no consensus on one particular theory. However, review of the literature suggests that differences between the two programs may be reflected in differing degrees of regulation (Moran 2000; Paul 2003; Enthoven 1 989; Cain 1999; Merlis 2003; Atherly 2009 ). Advocates of competition, such as Adam Smith, date back to the early 1700s. In the economic literature, a market is defined as an arrangement where buyers meet sellers to exchange goods, services, or information with no government intervention. Prices are set by the laws of supply and demand (Friedman and Friedman 1980). The law of demand states that there is a negative relationship between quantity demanded and price, where quantity demanded is the amount peop le are willing to buy at a certain price. As price decreases the quantity demanded increases. For this research, the market is prescription drugs and beneficiaries are the demanders of those drugs. Furthermore, cost sharing and inclusion of drugs on the formulary are used as a proxy for price. Thus, as shown in Figure 3 1 the demand curve slopes downward. On the other hand, the law of supply states that there is a relationship between quantity supplied and price that is based on how much sellers can offer to the market (Stigler 1966). For this research, prescription drug plans are the suppliers of goods. Although prescription drug plans are the direct suppliers of drugs to beneficiaries, there are many indirect suppliers (e.g. pharmaceutical manufac turers and pharmacy benefit

PAGE 47

47 managers) that are involved in this process. For example, pharmaceutical manufacturers create new drugs, introduce them to the market and then negotiate the price for those drugs with pharmacy benefit managers. Pharmacy benefi t managers are employed by prescription drug plans to negotiat e on their behalf. In Figure 3 2 the supply curve is shown by the upward sloping curve that is, the quantity supplied increases as the price increases. In a competitive market with no gover nment regulation, equilibrium occurs at the point where the demand and suppl y curves intersect (see Figure 3 3 ). In this case, the supply of drugs from a pharmaceutical manufacturer would meet the demand of drugs by beneficiaries. Demand would be express ed through prescription drug plan cost sharing requirements and inclusion of drugs on the formulary. As mentioned, prescription drug plans hire pharmacy benefit managers to negotiate price discounts for drugs with manufactures. The negotiations involve t he receipt of rebates by pharmacy benefit managers on the drugs selected for the formulary. Manufacturers give rebates share. The rebates that are negotiated by the pharmacy benefit manager result in savings that are passed on to the prescription drug plans, which are then passed on to the beneficiaries (Henry J. Kaiser Family Foundation 2005). In Figure 3 3 at the equilibrium price (P*), quantity demanded by benef iciaries is the same as quantity supplied by prescription drug plans (Q*). For this research, cost sharing and inclusion of drugs on the formulary are used as a proxy for price. When prescription drug plans freely enter and exit the market, deviations fr om equilibrium set in motion forces that drive price (i.e. cost sharing and formulary drug inclusion) back to

PAGE 48

48 equilibrium. F or example, as shown in Figure 3 4 if price is above equilibrium, the quantity supplied by prescription drug plans (Q3) exceeds th e quantity demanded by beneficiaries (Q1), resulting in a surplus (Q3 Q1). In such a situation, prescription drug plans will lower the price to eliminate the surplus, inducing beneficiaries to increase quantity demanded. The process continues until price reaches the equilibrium level. Competitive markets with no or little government regulation have certain benefits for consumers. In particular, competition forces sellers to supply the goods and services that are most desired by consumers because otherw ise new firms will enter to better meet consumer preferences. In addition, a relatively more competitive market will have lower equilibrium prices and higher equilibrium quantities than comparable markets that are relatively less competitive. In some cas es however, markets with no government intervention are not optimal, particularly from a social perspective. When examination of the two programs in this research, one of t he major differences between the two programs is government regulation. Specifically, the Medicare Part D program utilizes the United States Pharmacopeia ( USP ) Guidelines which specify that certain drugs should be covered, to regulate their program, whil e the FEHB program does not. The question arises as to whether the use of USP guidelines stemmed from some type of market failure. There are four types of market failure: market power, externalities, public goods, and asymmetric information. First, in some markets, sellers have substantial market power, which means that they can keep prices higher and quantities lower than would be the case in a more competitive market. This type of market failure is the rationale for

PAGE 49

49 antitrust policy. That is, the g overnment intervenes to prohibit business practices that restrict competition and harm consumers. The USP guidelines r equiring certain drugs to be included on a formulary may or may not affect market power. A 2005 report by the Henry J. Kaiser Family Fo undation noted that pharmaceutical manufacturers give rebates based on the inclusion of their in order to increase their market share. Prior to the implementation of the Medicare Part D program, policymakers assured beneficiari es that the Medicare prescription drug plans had sufficient market power to negotiate low drug prices (Frank and Newhouse 2008; Outterson and Kesselheim 2009). However a Families USA report in 2007 compared the drug prices of Medicare Part D plans and VA plans and found that the VA plans provided lower drug prices presumably due to the greater buying power of VA plans. Furthermore, a survey by Ayres, McHenry & Associates found that seniors had concerns about drug manufacturers helping to determine which of their drugs Medicare should and should not cover, but the preferred method to address these concerns was not to create the USP guidelines. Rather, s even out of ten seniors preferred a plan similar to the FEHBP plans (Senior Journal 2004). A second t ype of market failure is externalities, which occur when private consumption or production has external benefits or costs. An example of a positive externality is vaccination against contagious disease. When one person is vaccinated, that individual rece ives a benefit, but so do all the others that might have been infected. An example, of a negative externality is pollution. When one firm pollutes, the effects exten d well beyond the firm itself. Externalities are the rationale for government

PAGE 50

50 regulation in the form of taxes, subsidies, or by using property rights to force firms and individuals to take the spill over impacts of their economic activity into account If the USP Guidelines benefit society as a whole as well as individuals, externalities co uld explain the use of these regulations by Medicare. For example, Medicare Part D included a low income subsidy to supplement the income of those beneficiaries who could not afford d rugs provided by this program However, a 2007 report by the Commonwea lth Fund noted that many of those eligible for the subsidy were not enrolled in Part D and not receiving the subsidy (Summer, Nemore, and Finberg 2007). Furthermore, in 2008, GAO reported that the majority of denials for the low income subsidy (i.e. >50%) were asset and income related (Steinwald, Thorburn, Garvey et al 2008). Additionally, an article by Stuart and colleagues which assessed out of pocket spending by Medicare Part D beneficiaries found that over 3 years potential Part D enrollees were e xpected to pay for about 44% of total drug spending out of pocket (Stuart, Briesacher, Shea, et al 2005). Some studies have found that hospitalizations among beneficiaries with mental illness have actually gone up (Huskamp, Stevenson, Donohue, Newhouse an d Keating 2007; Daly and Moran 2007). On the other hand, Basu and colleagues examined the effects of Medicare Part D on dual eligible individuals and concluded that the implementation of Medicare Part D did not adversely affect dual eligible individuals ( Basu, Yin and Alexander 2008). Public goods are a third type of market failure. Public goods are non excludable that is, it is not possible to provide a good or service to one person without it thereby being available to others and n on rivalro us meaning that the consumption of a good or

PAGE 51

51 service by one person will not prevent others from enjoying it. A classic example is a lighthouse, others cannot be excluded from using it and many persons can use it ividuals are reluctant to invest in public goods, thus they are often provided by the government. This type of market failure does not seem to be related to Medicare Part D as this program is offered exclusively to Medicare beneficiaries. Finally, asymmet ric information may lead to market failure. This situation occurs when one party to a transaction has more or better information than the other party If the information problem is severe the government may intervene either by providing information publi cly or by mandating the provision of information by market participants. Medicare policymakers may have required the use of USP Guidelines to correct for this problem. As noted, market failure provides a rationale for government intervention, with the t ype of preferred intervention depending on the type of market failure. Medicare policymakers may have utilized the USP Guidelines as a solution to market failure expecting prices to be lower and quality to be higher. Or, t he purpose of USP guidelines may have been to prevent adverse selection. Adverse selection, the tendency for people with higher risk to seek health insurance coverage more than those with less risk, is usually a concern whenever people are allowed to choose among plans for health insura nce. The idea is that since asymmetry of information exists the potential for adverse selection exists.

PAGE 52

52 Unfortunately there is no research that examines the relationship of market failure and Medicare Part D. Therefore, the effect of the difference in regulation between Medicare Part D or the FEHBP on coverage or cost sharing is an empirical question If the increased regulation represents an effective response to market failure then the expectation is that Medicare Part D would have more extensive coverage and low er cost sharing than the FEHBP. If, on the other hand, the increased regulation was not due to market failure, the expectation is that FEHBP would have more extensive coverage and lower cost sharing than Medicare Part D. Decision Making Process: Medicare Part D vs. FEHBP T he next step is to compare Medicare Part D and th e FEHBP plans with respect to the process by which health plans decide on the formulary and the level of cost sharing (see Figure 3 5) Medicare Part D plans are govern ed by Congress via the Ways and Means and Finance Committees, while the FEHBP is governed by the Post Office and Civil Service Committee. A n article by Cain 1999 noted that the FEHBP outscore d the Medicare program in the areas of cost containment, innovat ion, customer satisfaction, and regulatory requirements. The area of regulatory requirements is of particular concern for purposes of this dissertation For Medicare Part D, the United States Pharmacopeia (USP) Guidelines play an important role. Specific ally, in section 1860D 4(b)(3)(C)(ii), the Medicare Prescription The Secretary shall request the United States Pharmacopeia to develop, in consultation with pharmaceutical benefit managers and other interested parties, a list of categories and classes that may be used by prescription drug plans under this paragraph and to revise such classification from time to time to reflect changes in therapeutic uses of covered part D

PAGE 53

53 drugs and additions of 2008). covered in certain drug classes (Public Law 2006). The USP is a not for profit organization that re ceives funding through the sale of products and services. The goal of the USP is to ensure quality pharmaceutical care. The USP standards are well known around the world and provide the official public standards for prescription and over the counter medi cations in the United States (Medicare Prescription Drug Benefit 2008) In the case of the FEHBP, Figure 3 5 shows that the Office of Personnel Management (OPM) is the primary oversight agency, with each prescription drug plan choosing its own formulary. However, about 80% of plans employ a pharmacy benefit manager (PBM), which is a third party administrator hired to help mana ge prescription drug benefits. Specifically, PBMs process and pay prescription drug claims, develop and maintain the formulary, p erform drug utilization reviews, contract with pharmacies, and negotiate discounts and rebates with drug manufacturers (Dicken, Agarwal, Dirosa, Kirksey, Rivera Lowitt and White 2003). For the FEHBP, the Office of Personnel Management monitors the PBMs b y negotiating plan benefits, monitoring drug benefit delivery, reviewing customer service reports, and conducting on site visits. Each PBM has a Pharmacy and Therapeutics (P&T) Committee, made up of physicians, pharmacists, and individuals with specialize d clinical expertise that reviews and updates the list of drugs on the formulary for a given plan (FEHBP Home Page 2008). The formulary is designed by the PBM and the plan, but must be submitted to and approved

PAGE 54

54 by the Office of Personnel Management before the final list of covered drugs is offered to FEHBP beneficiaries. It is important to note that, although some Medicare Part D plans also utilize pharmacy benefit managers, a major difference can be found in the decision making process. Unlike the FEHB P, PBMs that serve Medicare must work with the plan to delicate balance among cost, quality, and access. Specifically, PBM negotiations with pharmaceutical manufacturer s have a great impact on selection of drugs for the formulary. PBMs design the formulary to maximize cost savings for the plan and the consumer. Once the final formulary is complete, it is submitted to the Centers for Medicare and Medicaid Services (CMS) for review. Once approved, the final list of covered drugs is offered to Medicare Part D beneficiaries. Hypotheses The use of USP Guidelines by Medicare Part D may represent a response to potential market failure. If so and if the increased regulation c ompared to the FEHBP is an effective response to market failure, then the expectation is that Medicare Part D would have more extensive coverage and lower cost sharing than the FEHBP. If, on the other hand, the increased regulation was not due to market f ailure, the expectation is that the FEHBP would have more extensive coverage and lower cost sharing than Medicare Part D. The objective of this dissertation is to provide empirical evidence that bears on this question. In particular, the dissertation wil l compare drug coverage and cost sharing between Medicare Part D and the FEHBP.

PAGE 55

55 Figure 3 1. The Demand Curve Figure 3 2. The Supply Curve

PAGE 56

56 Figure 3 3 Equilib rium Figure 3 4 Market Equilibrium

PAGE 57

57 Figure 3 5 Diagram of the Medicare Part D & FEHBP Decision Making Processes Congress USP Formulary Guidelines & Criteria for Cost Sharing Established Office of Personnel Mngmt (OPM) PDPs must meet the guidelines PDPs oversee PBMs and report to OPM FEHBP Post Office and Civil Service Comm. Med Part D Ways and Means and Finance Comm. Final Decision by EACH PLAN of what the Formulary and Cost Sharing will be Final Decision by EACH PLAN of what the Formulary and Cost Sharing will be

PAGE 58

58 CHAPTER 4 DATA AND METHODS The objective of this research is to compare Medicare and the Federal Employees Health Benefits Program with respect to prescription drug plans. Secondary data for study were obtained from the Centers for Medicare and Medicaid Services and the Office of Personnel Management. In addition, pr imary data were collected from various health plan websites. The unit of analysis was the drug or the plan, depending on the particular research question being addressed. This chapter describes the data collection process, provides details on constructio n of the analytic dataset, defines study variables, and discusses methods used in the empirical analysis. Prior to commenc ing the study, approval was obtained from the University of Florida Institutional Review Board. Data Collection Medicare Part D Pre scription Drug Plans Data were obtained from the Centers for Medicare and Medicaid Services (CMS). The initial sample of Medicare Part D prescription drug plans consisted of approximately 2,500 prescription drug plans. Medicare Advantage plans were exclu ded because separate data about prescription drug coverage were not available. After this exclusion, there remained approximately 1,893 stand alone prescription drug plans. The percent and cumulative percent of enrollees within each plan were calculated, then the plans were ranked in terms of total enrollment. Beginning with the largest enrollment, the plans representing 75% of total enrollment were selected for further consideration, resulting in a total of 287 stand alone prescription drug plans.

PAGE 59

59 Exami nation of the Medicare Part D prescription drug data revealed that there were multiple plans that had the same formulary. Therefore, the plans were collapsed by formulary and ranked in order of total enrollment. Formularies were included in the sample if enrollment was greater than or equal to .5% of total enrollment. This yielded a final study sample of 19 formularies, covering 63% of total enrollment, as shown in Table 4 1 These 19 formularies represented 232 prescription drug plans. The range of e nrollment for the excluded plans/ formularies was 16,697 82,585 enrollees Once enrollment wa s less than .5% of total enroll ment the enrollment numbers became very small and it wa s more likely that the characte ristics of those smaller plans we re not the no rm. For example, the Center for Medicare and Medicaid Services discussed non renewal of those contracts representing a small number of enrollees (Cubanski and Neuman 2006). The organizations with low enrollment faced adverse selection concerns and decrea sed ability to negotiate low drug prices. It is important to note that, although there are over 1,000 prescription drug plans, enrollees typically choose among 45 57 plans, depending on the state in which they live. The CMS data included Prescription Dr ug Plan Formulary and Pharmacy Network Files, with formulary and pharmacy network data for all Medicare Prescription Drug Plans as of January 2009. The file contained six subfiles. The Geographic Locator File included the Medicare Advantage and Prescript ion Drug Plan Region codes and county codes. The Plan Information File included information for each plan (i.e. name, contract ID, plan ID, service area, and plan type). The Formulary File included formulary details for each plan including National Drug Codes (NDCs), tier level, indicators for step therapy, quantity limits, and prior authorization. The Beneficiary Cost File included plan

PAGE 60

60 level cost sharing details for preferred, non preferred and mail order network pharmacies. The Pharmacy Network File included the National Association of Boards of Pharmacy numbers for each network pharmacy including preferred, retail, and mail order indicators. Lastly, the Record Layout File included a diagram of file linkages, in addition to separate data dictionaries for each of the files listed above. The data provided by CMS did not include drug names, generic or brand name status of drugs, or therapeutic classes. The Agency for Health Care Administration in Tallahassee, FL provided information on the drug names an d generic or brand name status of drugs. Agency staff used the National Drugs Codes (approximately 3,000 per formulary) to look up the requested information. The therapeutic class (defined as groups of drugs that are similar in chemistry, method of actio n, and purpose of use) was determined and entered manually using the United States Pharmacopeia Drug Classification System (United States Pharmacopeial Convention 2008). Federal Employees Health Benefits Program Prescription Drug Plans A list of all plan s serving beneficiaries of the Federal Employees Health Benefits Program was obtained from the Office of Personnel Management. The list included the formularies was obtained Formulary 2009; GEHA Benefit Plan Formulary 2009; and NALC Health Benefit Plan Formulary 2009). The initial sample of the Federal Employees Health Benefits Program prescription drug plans consiste d of 222 prescription drug plans. The plans were ranked in terms of total enrollment, then the percent and cumulative percent of enrollees within each plan were calculated. Beginning with the largest enrollment, the plans representing 70% of

PAGE 61

61 total enroll ment were selected for the study, resulting in a total of 5 prescription dr ug plans/formularies, as shown in Table 4 2. The range of enrollment for the excluded plans/formularies was 1 63,346 enrollees Similarly to the case for Medicare Part D, although there are over 200 prescription drug plans, enrollees typically choose from among 12 20 plans depending on the state in which they live. It should be noted that for both Medicare Part D and the FEHBP, enrollment ranged from 80,000 to 2 million enrollees. Construction of Analytic Datasets Prior to the development of datasets, comparison of Medicare and the Federal Employees Health Benefits programs revealed the need for an appropriate benchmark to compare the programs. After reviewing the literature, IMS Health Data was chosen to obtain a list of the top 200 drugs most commonly used by dispensed prescriptions and the top 200 most commonly used drugs by sales in the United States (IMS Health Data 2009). For instances where the drug name was found on bo th lists (i.e. by dispensed prescriptions and by sales) the drug name was only listed once. Additionally, brand name drugs with generic equivalents were added to the list. This means that if a brand name drug was not listed on the formulary, but its gen eric equivalent was, that drug was considered to be listed on the formulary. Categorization in this way refers to therapeutic equivalence. Two drugs are referred to as therapeutically equivalent if, regardless of product name, packaging or price differen ce, they contain the same amount of the relevant active ingredients and may be used interchangeably. This process yielded a final list of 266 drugs (i.e. 134 duplicate drugs were deleted), representing a total of 23 therapeutic classes. This list was fur ther verified

PAGE 62

62 through literature review to ensure that these drugs represented at least 75% of all Medicare expenditures (Soni 2009; Stevenson, Huskamp, Keating and Newhouse 2007; Simoni Wastila, Shaffer and Stuart 2007; Families USA 2001). The data were m erged to comprise three datasets. Table 4 3 lists the variables in each dataset; a later section presents the justification for the variables included. Dataset 1, which was used to make comparisons based on the number/percentage of drugs covered, brand/g eneric status of drugs, and therapeutic class included the following variables: drug name, generic name, drug type, therapeutic class, and then all 24 formulary names were listed. For this dataset the unit of observation was the drug. The first variable, drug name, was an alphabetic variable (i.e. each of the 266 drugs was listed on the rows). The next variable, generic name was also alphabetic and included the generic name of each drug. The third variable, drug type, was numeric (recoded as 0 if it was a generic drug and 1 if it was a brand name drug). Next, therapeutic class was an alphabetic variable which included the therapeutic class name that each drug was classified under. Lastly, the 24 formularies were listed (19 Medicare Part D and 5 FEHBP f ormularies). Each corresponding row consisted of a 0 or 1 for each formulary depending on whether each drug was listed on the formulary or not (i.e. 0 if not listed on the formulary and 1 if listed on the formulary). Dataset 2 was used to perform the re gression analysis and included the following variables: plan name, type of plan, premium, enrollment, copay, coinsurance, tier, and then all top 15 therapeutic classes were listed (i.e. ADHD agents, analgesics, anti cancer agents, antibacterials, anticonvu lsants, antidepressants, antipsychotics, anxiolytics, arthritis agents, blood glucose regulators, blood products/ modifiers/volume

PAGE 63

63 expanders, cardiovascular agents, gastrointestinal agents, hormonal agents, and respiratory tract agents). For this dataset the unit of observation was the plan. The first variable, plan name, was an alphabetic variable (i.e. each of the 227 plans included for this analysis were listed on the rows). The second variable was type of plan. This variable represented whether the plan was a Medicare Part D or FEHBP plan. In each corresponding row the type of plan was noted (i.e. recoded as 1 for Medicare Part D and 0 for the FEHBP). The next variable was premium, which was a continuous number that reflected the amount enrollees p ay to the plan each month to receive benefits. The enrollment variable followed as a continuous variable that represented the number of beneficiaries per plan. Copay was listed as the fifth variable. For each of the 227 plans, the copay or dollar amount that enrollees pay for prescription drugs was noted. The next variable, coinsurance rate, indicated by the percentage that enrollees pay for drugs was listed on the corresponding rows For comparability, the analysis examined coverage only for tiers 1 an d 2 For Medicare Part D plans, t ier 3 represents non formulary drugs and tier 4 r epresents specialty drugs while none of the FEHBP plans have a tier 4 and the definition for tier 3 varies among the FEHBP plans (see T able 4 4). For example, for some of t he FEHBP plans, tier 3 represents multi source brand name drugs (i.e. drugs available from more than one manufacturer and that have at least one generic equivalent alternative available) whereas other FEHBP plans have no tier 3 Additionally, non formula ry drugs and specialty drugs require higher out of pocket costs for beneficiaries as compared to tier 1 and 2 ( generic/brand ) formulary drugs. Therefore, for this research

PAGE 64

64 the comparison between Medicare Part D and the FEHBP cost sharing is limit ed to tie rs 1 (generic drugs) and 2 (brand name drugs). The tier variable was indicated by a number that distinguished tier 1 generic drugs from tier 2 brand name drugs and the copay and coinsurance listed was for tier 1 generic drugs. To account for the change i n copay and coinsurance for tier 2 brand name drugs, the dataset above was copied and pasted below itself. Then the variables copay and coinsurance were updated with new correct entries for each row/plan. That is, the copay and coinsurance changed becaus e tier 2 brand name drugs are different than tier 1 generic drugs. This means that the total number of plans in the first column/row increased to 454 (i.e. 227 plans for tier 1 and 227 plans for tier 2). Finally, the last 15 variables consisted of the t op 15 therapeutic classes. Each row listed the number of drugs within each class that was included in each plan. The third and final dataset 3 used to conduct the analysis of cost sharing included the variables: plan name, tier, cost type (copay or coinsu rance), and cost amount. It is important to note that in the construction of the dataset for the analysis of cost sharing there was a change in sample size. Initially, there were 19 Medicare Part D formularies and 5 FEHBP formularies. These formularies represented 232 Medicare Part D plans and 5 FEHBP plans, respectively, for a total of 237 plans. D ata on cost sharing was missing for the following 14 Medicare Part D plans: Community CCRx Basic (1 plan), Prescription Pathway Bronze (11 plans), BravoRx (1 plan), and Medco Medicare Prescription (1 plan). The deletion of these plans yielded a final total of 218 Medicare Part D plans for the cost sharing analysis. For the FEHBP plans there was no missing data, but examination of cost sharing revealed that t here was variation among plans

PAGE 65

65 depending on whether a plan was for a single person (self) or for a family (self+family). Therefore, a total of 10 FEHBP plans were included in the cost sharing analysis. That is, self and self+family for each of the origi nal 5 plans. This resulted in a final sample of 228 plans for the cost sharing analysis. For this dataset the unit of observation was the plan. The first variable, plan name, was an alphabetic variable (i.e. each of the 227 plans included for this analys is were listed on the rows). The second variable tier was indicated as 1 for tier 1 generic drugs and 2 for tier 2 brand name drugs. The third variable, cost type was indicated for each plan as 1 if copay and 2 if coinsurance. Lastly, the amount or perc entage for each plan was listed on each row, which depended on the tier and whether the plan required a copay or coinsurance. Definition of Study Variables This section provides a brief review of the literature for each of the variables included in this st udy. Specifically, for the regression analysis, the relationship between the dependent and independent variables is described in detail. Some variables were re coded to facilitate analysis. The variables were classified into two main groups by unit of o bservation. For comparisons involving the number/percentage of drugs covered, generic/brand name status of drugs, and therapeutic class, the unit of observation was the drug. For the regression and cost sharing analyses, the unit of observation was the p lan. Tables 4 3 and 4 4 provide additional information. Current standards require prescription drugs to be offered under a health plan policy or contract. For the Medicare program a total of 1,893 stand alone prescription drug plans were offered, while the FEHBP offered 222 plans. The variables in which the unit of observation was the drug were drug name, generic name, drug type,

PAGE 66

66 therapeutic class, and formulary name. The variable drug name is used to identify drugs within formularies. The drug name refers to an owned, proprietary, brand or generic drug. There were 266 total drugs used for analysis in this research, as discussed in a previous section. generic status on a formulary. Factors such as whether a drug is generic or brand may affect whether the drug is included on a formulary. A brand name drug is defined as a trade name and is protected by a patent. A generic drug is defined as a prescription drug that has t he same active ingredient formula as a brand name drug. An active ingredient is any component that provides pharmacological activity or other direct effect in the diagnosis, cure, mitigation, treatment, or prevention of disease, or to affect the structure or any function of the body of man or animals. Generic drugs usually cost less than brand name drugs and are rated by the Food and Drug Administration (FDA) to be as safe and effective as brand name drugs. Drugs that are listed on a formulary are class ified into therapeutic class. A therapeutic class is defined as a group of drugs that are similar in chemistry, method of action, and purpose of use. Most prescription drug plans offer approximately 40 therapeutic classes on their formularies (United Sta tes Pharmacopeia 2008). The number of drugs per therapeutic class is important to beneficiaries because it can be a measure of access or extent of drug coverage ( Elam et al 2005; Hoadley et al 2006; Ketcham and Ngai 2008). The variable formulary name was selected as a key factor in the comparison of Medicare Part D and the FEHBP. Although plans vary by program, similar formularies

PAGE 67

67 may exist among plans. For example, multiple Medicare Part D plans have the same formulary. A formulary is defined as the l ist of prescription drugs covered (paid for) by a given health plan. To define the variables listed in the regression analysis, the unit of observation was the plan. The variables included were: plan name, type of plan, premium, enrollment, copayment, c oinsurance, t ier, and therapeutic classes. The plan name was used as an identifier and referred to the official name given to the plan. The variable type of plan (i.e. FEHBP or Medicare Part D) was also selected as a key descriptor of these plans since s tudies and reports point to differences in regulation between the two programs (Paul 2003; Cain 1999; Merlis 2003; Atherly 2009 ). For example, in a 1999 study by Cain it was noted that Medicare uses its negotiating power to set prices, while the FEHBP att empts to regulate the market by setting boundaries. Other research suggests that differences in regulation may affect the number of drugs offered to beneficiaries (Kipp and Ko 2008). To date, there are no studies that show an empirical examination of whe ther the number of drugs offered will increase or decrease depending on the type of plan. Other studies highlight the importance of plan premiums and enrollment as related to number of drugs in a therapeutic class. Plans w ith higher enrollment increased the number of drugs included on their formularies over time Hoadley and colleagues (2008) analyzed tends among Medicare Part D plans and found that the formularies with the highest enrollment increased the number of drugs offered on their formularies as enrollment increased (Hoadley, Hargrave, Merrell et al 2008) Additionally, premiums have been shown to be associated with the number of drugs included on plan

PAGE 68

68 formularies. For many plans, the amount of the premium that is charged to beneficiaries is re lated to the amount of drug coverage. A recent study showed that plans with higher premiums provided more extensive drug coverage as compared to plans with lower premiums ( Hoadley, Hargrave, Cubanski et al 2006). Additionally, studies show that copayment and coinsurance will also affect the number of drugs as they are modified by plans to contain costs. Results in an article by Gellad et al (2007) showed that plans that offered higher copayments and higher coinsurance also offered more drugs. When Part D plans were compared to plans offered prior to the enactment of Part D, results showed that the greater provision of drugs under Part D resulted in higher copayments. After reviewing the literature several covariates were identified as factors expected to influence the number of drugs per therapeutic class: tier, copayment, coinsurance, enrollment, and premium. Tier was used as a control variable and defined as either tier 1 (preferred generic drugs) or tier 2 (preferred brand name drugs). Preferred dr ug refers to drugs selected for inclusion on a plan formulary because of their effectiveness and cost. To discourage use, non preferred drugs are offered at higher co pay and coinsurance rates meaning consumers must pay a substantial out of pocket amount if they want to purchase a non preferred drug. For comparability, this research focuses on preferred drugs only (see Table 4 4). The preferred or non preferred status of a drug is defined by its tier number. Tier 1 refers to preferred generic drugs and tier 2 refers to preferred brand name drugs. Tier was controlled for because it could increase the likelihood of including drugs within a certain therapeutic class. Specifically, plans

PAGE 69

69 usually provide more tier 1 (generic drugs) as compared to tier 2 (br and name drugs) to curtail cost (Ku 2003; Gorman, Gorman, and Newell 2010). Finally, variables were created to represent the top 15 therapeutic classes. The classes included: ADHD agents, analgesics, anti cancer agents, antibacterials, anticonvulsants, an tidepressants, antipsychotics, anxiolytics, arthritis agents, blood glucose regulators, blood products/ modifiers/volume expanders, cardiovascular agents, gastrointestinal agents, hormonal agents, and respiratory tract agents. Examination of the number of drugs included by each plan for each of these classes allows for a more in depth look at drug coverage between Medicare Part D and the FEHBP plans. Table 4 5 lists the therapeutic classes and the drugs in each therapeutic class included in the analysis. It should be noted that there are other variables that may affect the dependent variable, the number of drugs per therapeutic class. Although data was not available from both programs on variables such as sex, race, and employment status these factors ma y explain differences in the formularies. For example, a study by Smetana and colleagues examined factors that influenced patient willingness to accept a medication change to a unified, restrictive formulary (Smetana, Davis, and Phillips 2004). The sampl e consisted of managed care plan members who had received a prescription for a non formulary medication in the previous 4 months and whose primary care physician approved a conversion to a formulary medication. Results revealed that patient age (OR, 1.03; CI, 1.01 to 1.05) and male gender (OR, 2.00; CI, 1.09 to 3.67) were each significant correlates of conversion (Smetana, Davis, and Phillips 2004). Furthermore, formulary conversion reduced costs beginning 3 months after the

PAGE 70

7 0 conversion date. In determinin g the number of drugs per therapeutic class one must consider that males may have a stronger preference for formulary drugs as compared to non formulary drugs. Therefore, the number of drugs per therapeutic class may be greater for males. On the other ha nd, in a NCHS Data Brief it was noted that women were more likely to use prescription drugs than men. This may be the case, but one must consider whether women prefer formulary drugs more often or non formulary drugs (Gu, Dillon, and Burt 2010). F urthermo re, race has been shown in many studies to have an effect on the kinds and number of drugs on formularies W hite patients have the highest prescription drug use so one may speculate that white patients require a higher number of drugs per therapeutic clas s. Additionally, studies show that medication adherence is lower among other races. A recent longitudinal retrospective cohort study revealed that for the therapeutic class blood glucose agents, black patients were as likely as whites to initiate oral th erapy and fill their first prescription, but experienced higher rates of medication discontinuation (HR: 1.8, 95% CI: 1.2, 2.7) and were less adherent over time (Trinacty, Adams, Soumerai, et al 2009). The authors concluded that racial differences in adhe rence to blood glucose agents persist even with equal access to medication. Therefore, if black and white patients have the same drugs listed on their formularies, black patients may still be less adherent to medication. E mploy ment is another factor that may affect the number of drugs per therapeutic class. The Medicare Part D program consists of mostly retired beneficiaries, while the FEHBP is mostly working age adults. Retired beneficiaries may have less of an ability to pay for non formulary medicatio ns or higher cost sharing medications. Therefore,

PAGE 71

71 retirees (i.e. Medicare Part D beneficiaries) may have more dr ugs on their formularies. Huskamp and colleagues examined the effect of three tier formulary adoption on medication continuation and spending among elderly retirees (Huskamp, Deverka, Landrum, et al 2007). Data w ere derived from four retiree plans that moved from a two tier formulary to a three tier formulary meaning the number and cost of drugs changed for the formularies. Results revealed th at there was only a small effect on medication continuation. A few of the retired patients had gaps in use and discontinued use (Huskamp, Deverka, Landrum, et al 2007). Retire ment status of beneficiaries may impact the number of drugs per therapeutic cla ss. Research Questions To compare Medicare Part D and the FEHBP the following questions and hypotheses were addressed in regard to prescription drugs: Research Question 1. How do Medicare Part D stand alone plans compare to Federal Employees Health Be nefits plans with respect to coverage of prescription drugs? 1A. Does the percent of the 266 drugs on the formulary differ between Medicare Part D and the FEHBP? Ho1A: There is no difference in percentage of drugs on the formulary among FEHB prescription drug plans as compared to Medicare Part D plans. 1B. Does the percent of the 266 drugs on formulary differ by therapeutic class between Medicare Part D and the FEHBP? Ho1B: There is no difference in percentage of drugs covered in the top therapeutic clas ses on the formulary among FEHB prescription drug plans as compared to Medicare Part D plans.

PAGE 72

72 1C. For each of the 266 drugs, how does coverage differ between Medicare Part D and the FEHBP? Ho1C: There is no difference in drug coverage among FEHB prescri ption drug plans as compared to Medicare Part D plans. 1D. For each of the 266 drugs that are covered by formulary, does percent brand differ between Medicare Part D and the FEHBP? Ho1D : There is no difference in the percentage of brand name drugs on the formulary among FEHB prescription drug plans as compared to Medicare Part D plans. 1E. Does the number of the top 266 drugs on formulary differ between Medicare Part D and the FEHBP, after controlling for other factors affecting formulary coverage? Ho1E : There is no difference in the number of the top 266 drugs on formulary among FEHB prescription drug plans as compared to Medicare Part D plans, after controlling for other factors affecting formulary coverage. Research Question 2. How do Medicare Part D stand alone plans compare to Federal Employees Health Benefits plans with respect to cost sharing on prescription drugs? 2A. Among plans with a copay, how does the percentage of plans with generic drugs differ between Medicare Part D and the FEHBP? Ho2A : Among plans with a copay, th ere is no difference in the percentage of plans w ith generic drugs among FEHB prescription drug plans as compared to Medicare Part D plans. 2B. Among plans with a copay, does the percentage of plans with brand name drugs diffe r between Medicare Part D and the FEHBP?

PAGE 73

73 Ho2B: Among plans with a copay, there is no difference in the percentage of plans with brand name drugs among FEHB prescription drug plans as compared to Medicare Part D plans. 2C. Among plans with a copay, does the mean copay differ between Medicare Part D and the FEHBP? Ho2C: Among plans with a copay, the re is no difference in mean copay among FEHB prescription drug plans as compared to Medicare Part D plans. 2D. Among plans with a coinsurance, does the percentag e of plans with generic drugs differ between Medicare Part D and the FEHBP? Ho2D: Among plans with a coinsurance, there is no difference in the percentage of plans with generic drugs among FEHB prescription drug plans as compared to Medicare Part D plans. 2E. Among plans with a coinsurance, does the percentage of plans with brand name drugs differ between Medicare Part D and the FEHBP? Ho2E: Among plans with a coinsurance, there is no difference in the percentage of plans with brand named drugs among FEHB p rescription drug plans as compared to Medicare Part D plans. 2F. Among plans with a coinsurance, does the mean coinsurance rate differ between Medicare Part D and the FEHBP? Ho2F: Among plans with a c oinsurance, there is no difference in mean coinsuranc e among FEHB prescription drug plans as compared to Medicare Part D plans.

PAGE 74

74 Data Analysis Descriptive Analysis Data examination began with the benchmark list of 266 drugs previously identified. Each of the 24 formularies (i.e. Medicare Part D and the FEHBP ) was compa red to the list of 266 drugs. The average number and percentage of drugs included in the formulary were compared across plans. That is, formulary coverage was compared across plans by drug. Then, the percentage of drugs that were brand name a nd generic was compared between Medicare Part D and the FEHBP plans. Next, the percentage of drugs on the formulary was calculated within each therapeutic class and compared between Medicare Part D and the FEHBP prescription drug plans. Table 4 5 lists t he therapeutic classes and all drugs used in the analysis. If there were less than three drugs in a therapeutic class that class was excluded from the analysis. This was necessary because an accurate comparison of Medicare Part D and the FEHBP could not be completed through the examination of a small number of drugs. For example, the therapeutic class anti migraine agents included only one drug. Comparing Medicare Part D and the FEHBP based on coverage of one drug would not be an appropriate comparison. Frequencies, mean values, and standard deviations were used to describe the sample. For comparisons involving the number/percentage of drugs covered, generic/brand name status of drugs and therapeutic class the unit of observation was the drug. For the regression and cost sharing analyses, the unit of observation was the plan. Independent Samples t test In addition to descriptive analyses, a bivariate method of analysis was utilized in which the independent variable x was a binary variable that defined the two groups

PAGE 75

75 compared and the dependent variable y was quantitative. The data was entered into SPSS 17.0 and an independent samples t test was utilized to determine whether there were statistically significant differences between Medicare Part D and th e FEHBP prescription drug plans with respect to percent of drugs covered, percent brand/generic, therapeutic class, and copay. The t test was chosen because the data represented a sample of the larger population of prescription drug plans. As a result, i t was not possible to determine the true standard deviation and calculate a z score. An assumption of the t test is that the observations are normally distributed. To examine the distribution of variables the data were plotted on a histogram and the mean and median were de termined. The mean and the median were approximately equal revealing that the distribution of observations was symmetric. Furthermore, the skewness and kurtosis indicated that the data was normally distributed. Another assumption of the t test is that the variances of the populations to be compared are equal. The equality or non equality of variances was determined for each independent samples t test using the Levene Statistic for the Test of Homogeneity of Variances ( NIST/SEMATECH e Handbook of Statistical Methods 2010) When the test statistic was non significant (p<.05), equal variances was assumed. When the test statistic was significant (p>.05), the t test was performed using a calculation that does not assume equal variances. The procedure for use of the independent samples t test was: y = average percent of drugs covered (or percent brand/generic or therapeutic class or copay)

PAGE 76

76 x = plan (i.e. 0 = FEHBP, 1 = Medicare Part D) In this case, small sample inference was used fo r comparing means because of the small sample sizes used, that is, at least one of the samples ( n 1 or n 2) was less than 20 (Agresti and Finlay 1997). It is assumed that the population distributions were normal. This approach is the same as for large samp les sizes, but the normal distribution is substituted with the t distribution. __________ The confidence i nterval has the form of: ( mean Y 2 mean Y 1 ) t s 1 + s 2 ------n 1 n 2 It was assumed that the two groups had the same _____________________________ _________________ 1 i mean Y 1 2 i mean Y 2 ) ( n 1 1) s 1 + ( n 2 1) s 2 ____________________________ ___ _____ __ _______ n 1 + n 2 2 n 1 + n 2 2 1 i meanY 1 ) is the sum of squared distances of measurements in the first sample from their meanY 1 2 i meanY 2 ) refers to pools information from the two samples to provide a single estimate of variability. The degr ees of freedom for this estimate equal df = n 1 + n 2 2. This equals the total number of observations ( n 1 + n 2) minus the number of parameters estimated in order to error of m eanY 2 meanY 1 simplifies to: ___________ _______ 2 meanY 1 = ___ + ___ __ + __ n 1 n 2 n 1 n 2 The confidence interval for has the form (meanY 2 meanY 1 ) t 2 meanY 1 The t score comes from the t table with df = n 1 + n 2 2, for the desired

PAGE 77

77 confidence level (Agresti and Finlay 1997). Negative Binomial Regression While results of this bivariate analysis provides a direct comparison of drug coverage between Medicare and the Federal Health Employees Benefit programs, the possibility remains that there are other variables that may affect the difference in means. To address this concern, a negative binomial regression analysis was performed. Analyses included a total of 16 regressions. The first regression was an overall model in which the dependent variable was number of drugs per ther apeutic class and therapeutic class is a control variable. The next 15 regressions were separate models for each of the top 15 therapeutic classes in which the dependent variable was the number of drugs per therapeutic class. For example, ADHD agents rep resented therapeutic class 1 and analgesics represented therapeutic class 2. The independent variables represent factors that may affect the number of drugs per therapeutic class and included type of plan (i.e. FEHBP or Medicare Part D), premium, copaymen t, coinsurance, tier, enrollment, and therapeutic class. Number of drugs per therapeutic class = f [Part D/FEHBP; Premium; Copay; Coinsurance; Tier; Enrollment; Therapeutic Class] 0 1 2 3 4 5 6 1 CL 1 15 CL 15 1 CL 1 *F 15 CL 15 *F+ where (ND) is the number of drugs per therapeutic class, (F) is the type of plan, (P) is the premium, (CP) is the copay, (CI) is the coinsurance, (T) is the tier, (E) is the enrollment, (CL) is the therapeutic class, (CL*F) is an i nteraction term and () is the error term.

PAGE 78

78 STATA version 11.0 and TSP version 5.0 software was used to perform the analysis (StataCorp 2009; TSP 2007). Negative binomial regression was chosen as an appropriate method because the dependent variable is a co unt of the number of incidents occurring in a given period of time Traditional linear regression (OLS) assume s a normally distributed outcome variable with equal variances over the range of predictor variables, and may not be optimal for modeling count o utcomes. Furthermore, when trying to estimate using OLS, the homoskedasticity assumption may be violated and results may provide negative predictions with biased coefficients (Wooldridge 2002). A negative binomial distribution assumes a mixture of Poisso n variables which follow a gamma distribution. In this model the likelihood of observing Yi is: f ( yi ) = yi ) ( __ __ 1__ ) ^yi yi is the gamma function and is the same as a Poisson = e^ This procedure uses a maximum likelihood estimation providing an odds measure, which shows the odds that a particular outcome will occur for each independent variable Differences with p value of less than .05 were considered s tatistically significant. Categorical variables included the type of plan (i.e. FEHBP or Medicare Part D) and tier. Continuous variables included the number of drugs covered in a therapeutic class, plan premium, plan enrollment, therapeutic class, copay and coinsurance. As noted previously, results of this research may conclude in three ways. It may be shown that more regulation is associated with a less competitive market. In that case, FEHB prescription drug plans would be expected to provide broader drug coverage as compared to Medicare Part D plans. On the other hand, results may show

PAGE 79

79 that Medicare Part D plans provide broader drug coverage as compared to the FEHB prescription drug plans because of regulation that addresses market failures. Lastly, results may show that there is no difference in prescription drug coverage offered between the two programs.

PAGE 80

80 Table 4 1. Medicare Part D Plans Selected for Analysis Formulary Name Total Enrollment (as of Jan 2009) Percent of Total Enrollment* Cumul ative Percent Number of Plans per Formulary AARP MedicareRx Preferred 2,716,518 15.6 15.6 31 Advantage Star Plan by RxAmerica 299,956 1.6 17.2 9 Blue Medicare Rx 285,869 1.6 18.8 7 BlueRx 99,729 0.6 19.4 3 BravoRx 82,585 0.5 19.9 3 CIGNA Medicare Rx Plan One 134,285 0.8 20.7 5 Community CCRx Basic 1,041,610 6.2 26.9 26 First Health Part D Premier 277,085 1.6 28.5 8 Health Net Orange 343,495 1.9 30.4 7 HealthSpring Prescription Drug Plan 171,719 0.9 31.3 5 Humana PDP Enhanced or Complete 1,432, 200 8.2 39.5 30 Humana PDP Standard 1,445,988 8.1 47.6 27 Medco Medicare Prescription Plan 211,477 1.2 48.8 2 MedicareBlue Rx Option 3 298,839 1.7 50.5 3 Prescription Pathway Bronze Plan 314,664 1.7 52.2 12 SilverScript Va lue 353,491 2.1 54.3 12 AARP MedicareRx Saver or UnitedHealth Rx Basic 831,943 4.9 59.2 20 WellCare Classic or Signature 476,022 2.6 61.8 20 WellCare Classic or Signature 200,462 1.2 63 2 *Total enrollment across all plans was 17,313,409

PAGE 81

81 Table 4 2. FEHBP Prescription Drug Plans Selected for Analysis Plan Name Total Enrollment (as of 2009) Percent of Total Enrollment* Cumulative Percent Blue Cross and Blue Shield Standard Service Benefit Plan 2,020,621 50.2 50.2 Blue Cross and Blue Shield Basic S ervice Benefit Plan 391,541 9.7 59.9 GEHA Benefit Plan 215,833 5.4 65.3 NALC Health Benefit Plan 95,481 2.4 67.7 American Postal Workers Union Health Plan (CDHP) 81,626 2.0 69.7 *Total enrollment across all plans was 4,026,575 Table 4 3. Da tasets 1, 2, and 3 Definition of Study Variables Variable Definition Label in Dataset Variable Type Unit of Observ ation Dataset 1 Drug Name Refers to an owned, proprietary, brand or generic drug. DN Alphabet ic Drug Generic Name A generic drug is a prescription drug that has the same active ingredient formula as a brand name drug. An active ingredient is any component that provides pharmacological activity or other direct effect in the diagnosis, cure, mitigation, treatment, or prevention of dise ase, or to affect the structure or any function of the body of man or animals. GN Alphabet ic Drug Drug Type A brand name drug has a trade name and is protected by a patent (i.e. can be produced and sold only by the company holding the patent); A generic d rug is a prescription drug that has the same active ingredient formula as a brand name drug. This variable indicates whether a drug is brand or generic. Br 0=Gener ic, 1=Brand Drug

PAGE 82

82 Table 4 3. Continued Variable Definition Label in Dataset Variable Typ e Unit of Observ ation Therapeutic Class Groups of drugs that are similar in chemistry, method of action, and purpose of use CL Alphabet ic Drug Formulary name A formulary is the list of prescription drugs covered (paid for) by a given health plan. This variable indicates whether or not a particular drug is in the formulary for a given plan. Form 0=No, 1=Yes Drug Dataset 2 Plan Name Used as an identifier and refers to the name of the plan PN Alphabet ic Plan Type of Plan This variable indicates whether a plan is either a Medicare Part D plan or a Federal Employees Health Benefits Plan. TP 0=FEHB P, 1=Med Part D Plan Premium The amount a beneficiary pays each month for their health plan coverage. This amount varies depending on the health plan or drug formulary. P Continu ous Plan Enrollment Refers to the number of enrolled individuals in a prescription drug plan. E Continu ous Plan Copayment A copayment is a fixed amount of money paid by a beneficiary when receiving covered services (e.g. prescription s). CP Continu ous Plan Coinsuranc e In health insurance, the insured person and the insurer share the covered procedures under a policy in a specified ratio. For example, the insurer may pay 80% of a pay the remaining 20%. CI Continu ous Plan Tier A tier indicates the level of copayment for a given drug. Plans have multiple tiers, and the copayment/coinsurance amount depends on the tier level. T 1=First Tier, 2=Seco nd Tier Plan

PAGE 83

83 Table 4 3. Continued Va riable Definition Label in Dataset Variable Type Unit of Observ ation Top 15 Therapeutic Class Names Refers to the number of drugs per therapeutic class (i.e. ADHD agents, analgesics, anti cancer agents, antibacterials, anticonvulsants, antidepressants, antipsychotics, anxiolytics, arthritis agents, blood glucose regulators, blood products/ modifiers/volume expanders, cardiovascular agents, gastrointestinal agents, hormonal agents, and respiratory tract agents). ND Continu ous Plan Dataset 3 Plan Name Used as an identifier and refers to the name of the plan PN Alphabet ic Plan Tier A tier indicates the level of copayment for a given drug. Plans have multiple tiers, and the copayment/coinsurance amount depends on the tier level. T 1=First Tier, 2=Seco nd Tier Plan Cost Type Refers to the type of cost sharing applied for drug coverage CT 1=Copa y, 2=Coins urance Plan Cost Amount This number represents either a copay dollar value or a coinsurance percentage. A copayment is a fixed amount of money paid by a beneficiary when receiving covered services (e.g. prescriptions). Coinsurance is a percentage the insured person and the insurer share to cover the cost of drugs. CA Continu ous Plan

PAGE 84

84 Table 4 4. Datasets Tier Comparison: Medicare Part D and the FEHBP P lan Tier 1 Tier 2 Tier 3 Tier 4 All Medicare Part D Plans Generic drugs Preferred brand name drugs Non preferred Brands and Generics Tier 4=Specialty Drugs means those covered drugs that typically cost $500 or more per dose or $6,000 or more per year and have one or more of the following characteristics: (1) complex therapy for complex disease (2)specialized patient training and coordination of care (services, supplies, or devices) required prior to therapy initiation and/or during thera py; (3) unique patient compliance and safety monitoring requirements; (4)unique requirements for handling, shipping and storage; and (5) potential for significant waste due to the high cost of the drug FEHBP BCBS Basic Formulary Self Only Non Postal Gen eric drugs Preferred brand name drugs Non formulary or non preferred brand name drug N/A FEHBP BCBS Basic Formulary Self+Family Non Postal Generic drugs Preferred brand name drugs Non formulary or non preferred brand name drug N/A FEHBP BCBS Standard Formulary Self Only Non Postal Generic drugs Preferred brand name drugs Non formulary or non preferred brand name drug N/A

PAGE 85

85 Table 4 4. Continued Plan Tier 1 Tier 2 Tier 3 Tier 4 FEHBP BCBS Standard Formulary Self+Family N on Postal Generic drugs Preferred brand name drugs Non formulary or non preferred brand name drug N/A FEHBP American Postal Workers Union Health Plan Formulary (CDHP) Self Only Non Postal Generic drugs Preferred brand name drugs (all drugs are covered) N/A N/A FEHBP American Postal Workers Union Health Plan Formulary (CDHP) Self+Family Non Postal Generic drugs Preferred brand name drugs (all drugs are covered) N/A N/A NALC Health Benefit Plan Self Only Non Postal Generic drugs Preferred brand name drugs (all drugs are covered) N/A N/A NALC Health Benefit Plan Self+Family Non Postal Generic drugs Preferred brand name drugs (all drugs are covered) N/A N/A FEHBP GEHA Benefit Plan Formulary Standard Self Only Non Postal Generic drugs Preferred br and name drugs N/A N/A

PAGE 86

86 Table 4 4. Continued Plan Tier 1 Tier 2 Tier 3 Tier 4 FEHBP GEHA Benefit Plan Formulary Standard Self+Family Non Postal Generic drugs Preferred brand name drugs N/A N/A FEHBP GEHA Benefit Plan Formulary High Option Self Only Non Postal Generic drugs Single source preferred brand name drugs (Single source brand name drugs are available from only one manufacturer and are patent protected. No generic equivalent is available) Multi source brand (Multi source brand name dr ugs are available from more than one manufacturer and have at least one generic equivalent alternative available) N/A FEHBP GEHA Benefit Plan Formulary High Option Self +Family Non Postal Generic drugs Single source preferred brand name drugs Multi sourc e brand N/A

PAGE 87

87 Table 4 5. Top Therapeutic Classes and List of All Drugs Drug Name Therapeutic Class Concerta ADHD Agents Focalin XR ADHD Agents Strattera ADHD Agents Vyvanse ADHD Agents Adderall XR ADHD Agents Acetaminophen and codeine Analgesics Fentanyl Analgesics Hydrocodone/Acetaminophen Analgesics Lidoderm Analgesics Naproxen Analgesics Oxycontin Analgesics Oxycodone and Acetaminophen Analgesics Oxycodone hydrochloride ER Analgesics Propoxyphene napsylate and acetaminophen Analgesics S uboxone Analgesics Tramadol hydrochloride Analgesics Celebrex Analgesics Ibuprofen Analgesics Abraxane Anti Cancer Agents Aldara Anti Cancer Agents Alimta Anti Cancer Agents Allopurinol Anti Cancer Agents Aloxi Anti Cancer Agents Arimidex Anti Can cer Agents Avastin Anti Cancer Agents Casodex Anti Cancer Agents Eloxatin Anti Cancer Agents Erbitux Anti Cancer Agents Femara Anti Cancer Agents Gemzar Anti Cancer Agents Herceptin Anti Cancer Agents Neulasta Anti Cancer Agents Neupogen Anti Canc er Agents Pegasys convenience pack Anti Cancer Agents Rituxan Anti Cancer Agents Sandostatin LAR Anti Cancer Agents Tarceva Anti Cancer Agents Taxotere Anti Cancer Agents Temodar Anti Cancer Agents Thalomid Anti Cancer Agents Velcade Anti Cancer Ag ents

PAGE 88

88 Table 4 5. Continued Drug Name Therapeutic Class Xeloda Anti Cancer Agents Zometa Anti Cancer Agents Amoxicillin Antibacterials Amoxicillin TR and clavulanate potassium Antibacterials Avelox Antibacterials Azithromycin Antibacterials Cephalexi n Antibacterials Ciprofloxacin hydrochloride Antibacterials Cubicin Antibacterials Doxycycline hyclate Antibacterials Fluconazole Antibacterials Levaquin Antibacterials Penicillin VK Antibacterials Solodyn Antibacterials Sulfamethoxazole and trimet hoprim Antibacterials Synagis Antibacterials Zosyn Antibacterials Zyvox Antibacterials Depakote Anticonvulsants Depakote ER Anticonvulsants Gabapentin Anticonvulsants Keppra Anticonvulsants Lamictal Anticonvulsants Lamotrigine Anticonvulsants Ly rica Anticonvulsants Topamax Anticonvulsants Aricept Antidementia Agents Exelon Antidementia Agents Namenda Antidementia Agents Amitriptyline hydrochloride Antidepressant Agents Budeprion XL Antidepressant Agents Citalopram hydrobromide Antidepress ant Agents Cymbalta Antidepressant Agents Effexor XR Antidepressant Agents Fluoxetine hydrochloride Antidepressant Agents Lexapro Antidepressant Agents Paroxetine hydrochloride Antidepressant Agents Sertraline hydrochloride Antidepressant Agents Tra zodone hydrochloride Antidepressant Agents Geodon Antipsychotic Agents Invega Antipsychotic Agents

PAGE 89

89 Table 4 5. Continued Drug Name Therapeutic Class Risperdal Antipsychotic Agents Risperdal consta Antipsychotic Agents Risperidone Antipsychotic Agents Seroquel Antipsychotic Agents Zyprexa Antipsychotic Agents Zyprexa Zydis Antipsychotic Agents Abilify Antipsychotic Agents Atripla Antiretroviral/Antiviral Agents Combivir Antiretroviral/Antiviral Agents Epzicom Antiretroviral/Antiviral Agents Kal etra Antiretroviral/Antiviral Agents Norvir Antiretroviral/Antiviral Agents Reyataz Antiretroviral/Antiviral Agents Tamiflu Antiretroviral/Antiviral Agents Truvada Antiretroviral/Antiviral Agents Valtrex Antiretroviral/Antiviral Agents Viread Antiret roviral/Antiviral Agents Alprazolam Anxiolytics Clonazepam Anxiolytics Diazepam Anxiolytics Lorazepam Anxiolytics Enbrel Arthritis Agents Humira Arthritis Agents Meloxicam Arthritis Agents Orencia Arthritis Agents Remicade Arthritis Agents A CTOplus met Blood Glucose Regulators Actos Blood Glucose Regulators Avandia Blood Glucose Regulators Byetta Blood Glucose Regulators Gleevec Blood Glucose Regulators Glyburide Blood Glucose Regulators Humalog Blood Glucose Regulators Janumet Blood G lucose Regulators Januvia Blood Glucose Regulators Lantus Blood Glucose Regulators Lantus Solostar Blood Glucose Regulators Levemir Blood Glucose Regulators Metformin hydrochloride Blood Glucose Regulators NovoLog Blood Glucose Regulators NovoLog F lexPen Blood Glucose Regulators

PAGE 90

90 Table 4 5. Continued Drug Name Therapeutic Class Aggrenox Blood Products/Modifiers/Volume Expanders Angiomax Blood Products/Modifiers/Volume Expanders Aranesp Blood Products/Modifiers/Volume Expanders Carimune NF Blood Products/Modifiers/Volume Expanders Epogen Blood Products/Modifiers/Volume Expanders Gammagard Liquid Blood Products/Modifiers/Volume Expanders Gamunex Blood Products/Modifiers/Volume Expanders Lovenox Blood Products/Modifiers/Volume Expanders Plavix Blood Products/Modifiers/Volume Expanders Procrit Blood Products/Modifiers/Volume Expanders Venofer Blood Products/Modifiers/Volume Expanders Warfarin sodium Blood Products/Modifiers/Volume Expanders Amlodipine besylate Cardiovascular Agents Amlodipi ne besylate and benazepril Cardiovascular Agents Atenolol Cardiovascular Agents Avalide Cardiovascular Agents Avapro Cardiovascular Agents Benicar Cardiovascular Agents Benicar HCT Cardiovascular Agents Caduet Cardiovascular Agents Cartia XT Cardiov ascular Agents Catapres TTS Cardiovascular Agents Clonidine Cardiovascular Agents Coreg CR Cardiovascular Agents Cozaar Cardiovascular Agents Crestor Cardiovascular Agents Digoxin Cardiovascular Agents Diovan Cardiovascular Agents Diovan HCT Cardio vascular Agents Enalapril maleate Cardiovascular Agents Furosemide Cardiovascular Agents

PAGE 91

91 Table 4 5. Continued Drug Name Therapeutic Class Hydrochlorthiazide Cardiovascular Agents Hyzaar Cardiovascular Agents Intergrilin Cardiovascular Agents Isosorb ide mononitrate Cardiovascular Agents Klor Con M20 Cardiovascular Agents Lipitor Cardiovascular Agents Lisinopril Cardiovascular Agents Lisinopril and hydrochlorthiazide Cardiovascular Agents Lotrel Cardiovascular Agents Lovaza Cardiovascular Agents Metoprolol succinate Cardiovascular Agents Metoprolol tartrate Cardiovascular Agents Niaspan Cardiovascular Agents Potassium chloride Cardiovascular Agents Pravastatin sodium Cardiovascular Agents Ramipril Cardiovascular Agents Simvastatin Cardiova scular Agents Toprol XL Cardiovascular Agents Triamterene and hydrochlorizide Cardiovascular Agents Tricor Cardiovascular Agents Verapamil SR Cardiovascular Agents Vytorin Cardiovascular Agents Zetia Cardiovascular Agents Aciphex Gastointestinal Age nts Asacol Gastointestinal Agents Nexium Gastointestinal Agents Omeprazole Gastointestinal Agents Pantoprazole sodium Gastointestinal Agents Prevacid Gastointestinal Agents Prevacid Solutab Gastointestinal Agents Protonix Gastointestinal Agen ts Ranitidine hydrochloride Gastointestinal Agents Androgel 1% Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid)

PAGE 92

92 Table 4 5. Continued Drug Name Therapeutic Class Avodart Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) Levothyroxine sodium Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hor monal Agents, Stimulant/Replacement/Modifying (Thyroid) Levoxyl Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) Mirena Hormonal Agents, Stimulant/Replacement/Modifyi ng (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) NuvaRing Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) Ortho Tri Cyclen Lo Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) Prednisone Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant /Replacement/Modifying (Thyroid)

PAGE 93

93 Table 4 5. Continued Drug Name Therapeutic Class Premarin Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) Sensipar Hormonal Agent s, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) Synthroid Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modif ying (Thyroid) Trinessa 28 Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) Yasmin 28 Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Ho rmonal Agents, Stimulant/Replacement/Modifying (Thyroid) Yaz 28 Hormonal Agents, Stimulant/Replacement/Modifying (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) Zemplar Hormonal Agents, Stimulant/Replacement/Modify ing (Sex Hormones/Modifiers) AND Hormonal Agents, Stimulant/Replacement/Modifying (Thyroid) Actonel Metabolic Bone Disease Agents Alendronate sodium Metabolic Bone Disease Agents

PAGE 94

94 Table 4 5. Continued Drug Name Therapeutic Class Boniva Metabolic Bone Di sease Agents Evista Metabolic Bone Disease Agents Forteo Metabolic Bone Disease Agents Fosamax Metabolic Bone Disease Agents Fosamax Plus D Metabolic Bone Disease Agents Avonex Multiple Sclerosis Agents Betaseron Multiple Sclerosis Agents Copaxone M ultiple Sclerosis Agents Rebif Multiple Sclerosis Agents Carisoprodol Muscle Relaxants Cyclobenzaprine hydrochloride Muscle Relaxants Skelaxin Muscle Relaxants Cosopt Ophthalmic Agents Lucentis Ophthalmic Agents Lumigan Ophthalmic Agents Restasis O phthalmic Agents Xalatan Ophthalmic Agents Advair Diskus Respiratory Tract Agents Albuterol Respiratory Tract Agents Allegra D 12 Hour Respiratory Tract Agents Astelin Respiratory Tract Agents Clarinex Respiratory Tract Agents Combivent Respiratory Tract Agents Fexofenadine hydrochloride Respiratory Tract Agents Flomax Respiratory Tract Agents Flovent HFA Respiratory Tract Agents Fluticasone propionate Respiratory Tract Agents Nasacort AQ Respiratory Tract Agents Nasonex Respiratory Tract Agent s ProAir HFA Respiratory Tract Agents Promethazine hydrochloride Respiratory Tract Agents Proventil HFA Respiratory Tract Agents Pulmicort respules Respiratory Tract Agents Singulair Respiratory Tract Agents Spiriva HandiHaler Respiratory Tract Agent s Symbicort Respiratory Tract Agents Xolair Respiratory Tract Agents Xopenex Respiratory Tract Agents Ambien CR Sedatives/Hypnotics Lunesta Sedatives/Hypnotics Zolpidem tartrate Sedatives/Hypnotics

PAGE 95

95 Table 4 5. Continued Drug Name Therapeutic Class G ardasil Vaccines Prevnar Vaccines RotaTeq Vaccines Varivax Vaccines Zostavax Vaccines

PAGE 96

96 CHAPTER 5 RESULTS This dissertation compares Medicare Part D and the Federal Employee Health Benefits plans with respect to prescription drug coverage. This ch apter is divided into three sections. The first section describes the results of the examination of formulary coverage among plans. The second section provides results of the regression analysis, which examined the relationship between type of plan and t he number of drugs per therapeutic class, holding constant therapeutic class, premium, copayment, coinsurance, tier, and enrollment. Lastly, section three describes results of the analysis on cost sharing. Tables 5 1 and 5 2 give a brief description of th e data. It is important to emphasize that, while in the Federal Employees Health Benefits Program each formulary is associated with only one plan, such is not the case for Medicare Part D. As shown in Table 5 1 nineteen Medicare Part D formularies and f ive Federal Employees Health Benefits Program prescription drug formularies/plans were examined, resulting in a total of 24 formularies However, there were multiple Medicare Part D plans with the same formulary. For example, the formulary AARP MedicareR x Preferred was utilized by 31 Medicare Part D plans (Table 5 2). Overall, the 19 Medicare Part D formularies identified in this analysis were used by a total of 232 plans. The formulary for each plan was compared to the list of 266 top drugs dispensed and sold in the United States. Of the 266 top drugs utilized in the U.S., 197 were brand name drugs and 69 were generic drugs. The 266 drugs were categorized into 23 therapeutic classes for further inquiry.

PAGE 97

97 Formulary Coverage The first research questi on was (1) How do Medicare Part D stand alone plans compare to Federal Employees Health Benefits plans with respect to coverage of prescription drugs? To answer this question first the number and percentage of the top 266 drugs dispensed and sold in the U.S. was calculated for each formulary. The results of the analysis are reported in Table 5 3. Results revealed, that for the AARP MedicareRx Preferred Prescription Drug formulary, which represents the Medicare Part D formulary with the highest enrollment 249 out of 266 (94%) top drugs were covered, while for the Blue Cross and Blue Shield Standard Service Benefit Plan (the FEHBP formulary with the highest enrollment) 252 out of 266 (95%) drugs were on the formulary. On the other hand, for those formular ies representing the least amount of enrollees (about 80,000 beneficiaries each), Medicare formulary, the percentage of drugs covered was 80% and 99%, respectively. Overall, for t he 19 Medicare Part D formularies analyzed, formulary coverage of the top drugs dispensed and sold in the United States ranged from 72 94%, while the range was 85 99% for the 5 FEHBP formularies examined. On average, Medicare Part D plans covered 84% of t he top drugs dispensed and sold in the United States compared to FEHBP plans which covered about 94% on average. To determine whether these differences were statistically significant an independent samples t test was performed. Results revealed that there was a statistically significant difference in drug coverage for the FEHBP prescription drug plans as compared to Medicare Part D plans. As shown in Table 5 4, the FEHBP prescription drug plans yielded a higher average percentage of covered drugs as

PAGE 98

98 compa red to Medicare Part D plans (p<.01) Considering these independent t test results one may conclude that the FEHBP prescription drug plans provided broader coverage of top drugs dispensed and sold in the United States as compared to Medicare Part D plans but as reported later, further analysis using regression methods yielded different results. Examination by therapeutic class is shown in Table 5 5. These are the therapeutic classes associated with the top 266 drugs sold and dispensed in the United Sta tes. F or the 23 therapeutic classes examined, the number of drugs in the therapeutic class ranged from three to forty two drugs. The top three therapeutic classes in regard to top drugs sold and dispensed in the U.S. were cardiovascular agents (42 of 266 top drugs), anti cancer agents (25 of 266 top drugs), and respiratory tract agents (21 of 266 top drugs). The average percentage of drugs covered per therapeutic class ranged from 0 to 100% for Medicare Part D formularies and from 52 to 100% for the FEH BP formularies. The therapeutic class anxiolytics showed the greatest difference in drug coverage between formularies; among Medicare Part D formularies, none of the drugs were covered in this class, while the FEHBP formularies covered on average 95%. In contrast, for some therapeutic classes the average percentage of drugs covered was the same for Medicare Part D and the FEHBP. For example, drugs in the therapeutic class anticonvulsants and antiretroviral/antiviral agents were completely covered by both Medicare Part D plans and the FEHBP plans. Of the 23 therapeutic classes analyzed, both Medicare Part D plans and the FEHBP plans covered at least one drug in each class with the exception of the class anxiolytics.

PAGE 99

99 Analyses using the independent samples t test revealed that there were statistically significant differences in drug coverage for the following 12 of the 23 therapeutic classes: ADHD agents, anxiolytics, arthritis agents, blood glucose agents, blood products/modifiers/volume expanders, cardiov ascular agents, gastrointestinal agents, hormonal agents, metabolic bone disease agents, multiple sclerosis agents, ophthalmic agents, and respiratory tract agents (Table 5 5 ). Specifically, in all these therapeutic classes the FEHBP was shown to provide broader drug coverage (p<.05). For example, Medicare Part D formularies on average covered 87% of drugs in the class cardiovascular agents as compared to the FEHBP formularies, which on average covered 97% (p<.01). Furthermore, on average 95% blood glucos e regulators were covered on Part D formularies, while on average 100% were covered by the FEHBP formularies (p<.01). To further examine formulary coverage Medicare Part D and the Federal Employee Health Benefits programs were compared drug by drug. Tabl e 5 6 shows the top 266 drugs dispensed and sold in the U.S. and the mean percentage of formularies covering each drug. Of the 266 drugs that were examined, 149 were covered by all Medicare Part D formularies compared to 211 covered by all the FEHBP formu laries. The mean percentage comparison between formularies varied drug by drug For example, the mean percentage of Medicare Part D formularies covering the drug Lipitor, a cardiovascular agent, was 79% compared to a mean percentage of 100% for FEHBP for mularies. On the other hand, the mean percentage of Medicare Part D formularies covering the drug Zostavax, a vaccine, was 100% compared to a mean percentage of 40% for FEHBP formularies. Drug coverage within classes also varied. Fo r Medicare

PAGE 100

100 Part D for mularies in the therapeutic class Smoking Cessation Agents, coverage was 79% for Chantix and 95% for Wellbutrin XL compared to the FEHBP formularies which provided coverage of 100% for Chantix and 80% for W ellbutrin XL. Tables 7 9 examine how drug covera ge differed by brand versus generic status. There were two types of comparisons: (1) within formulary: distribution of total drugs covered between brand and generic; and (2) as % of total brand and total generic among top drugs. For example, Table 5 7 sh ows that the first Medicare Part D formulary, AARP Medicare Rx, covered a total of 249 of the top drugs 63 (25%) being generic and 186 (75%) being brand. Moreover, AARP Medicare Rx covered 63 of the total 69 generic drugs (91%) and 186 of the total 197 br and drugs (94%). Focusing first on the distribution of total drugs between generic and brand, for the Medicare Part D formularies, the number of generic drugs covered ranged from 60 to 63 compared to 67 to 68 for the FEHBP formularies, but when calculated as a percentage of total drugs in the formulary, generic drug coverage was about the same for Medicare Part D (25 32%) and the FEHBP (25 30%). Similarly, for brand name drugs, Medicare Part D formularies covered 134 to 185 (68 75%), while the FEHBP formul aries covered 159 to 197 (70 75%). Further analysis using the independent samples t test showed that study groups did not significantly differ with respect to coverage of generic and brand name drugs as a percent of total drugs covered (Table 5 8). For the second comparison, the data was analyzed in an alternative way to shed additional light on differences in generic and brand name status of drug coverage. Independent samples t tests were performed to compare generic drug coverage among generic drugs only and to compare brand name drug coverage among brand name

PAGE 101

101 drugs only. As noted above, for the Medicare Part D formulary AARP MedicareRx Preferred, 63 generic drugs were covered out of a possible 69 generic drugs and 186 brand drugs were covered out of a possible 197 brand drugs (Table 5 7). Results of this analysis revealed statistically significant differences between Medicare Part D and the FEHBP formularies (p<.05). On average, the FEHBP plans covered about 98% of all generic drugs (among generic drugs only) versus about 90% for Medicare Part D plans (Table 5 9). Similarly, a significant difference was shown for the 197 brand name drugs that were listed. Medicare Part D plans covered on average 82% brand name drugs (among brand name drugs only) a s compared to the FEHBP plans that covered on average 93% (p<.05). To determine whether there were other variables that may have affected the difference in means between the number of drugs on the Medicare Part D and the FEHBP formularies a negative bino mial regression analysis was performed. The first regression was an overall model in which the dependent variable was number of drugs per therapeutic class (results are shown in Table 5 10). In addition, fifteen regression models were conducted in which the number of drugs in a given therapeutic class was the dependent variable (results are shown in Tables 5 11 through 5 25). For example, ADHD agents represented therapeutic class 1 and analgesics represented therapeutic class 2 etc The independent var iables included type of plan (i.e. FEHBP or Medicare Part D), premium, copayment, coinsurance, tier, enrollment, and therapeutic class. The regression equation was: 0 1 2 3 4 5 6 1 CL 1 15 CL 15 1 CL 1 *F 15 CL 15 *F+

PAGE 102

102 where (ND) was the number of drugs per therapeutic class, (F) was the type of plan, (P) was the premium, (CP) was the copay, (CI) was the coinsurance, (T) w as the tier, (E) was the enrollment, (CL) was the therapeutic class, (CL*F) was an interaction term and () was the error term. Table 5 10 shows the results of the negative binomial regressions for the overall model. It should be noted that coefficients i n a negative binomial regression must be adjusted for interpretation The coefficients represented by ) were interpreted as the difference between the log of expected counts Consider x0+1 ) x0 ), where represent where x, the predictor variable, is evaluate d at x 0 and x 0 +1 (UCLA Academic Technology Services 2010). T he difference of two logs is equal to the log of their x0+1 ) x0 x0+1 / x0 ), meaning the parameter estimate is the log of the ratio of expected counts (UCLA Academic Technology Services 2010). I ncidence rates are shown as the irr value s in tables 5 10 5 25. Type of plan was positively associated with the number of drugs per therapeutic class while tier was negatively associated. Specifically, Medicare Part D plans provided a greater number of drugs per therapeutic class as compared to the FEHBP plans holding other factors constant However t he significant interaction terms for all classes except for the class anxiolytics, showed that on average, the numb er of drugs for those therapeutic classes was greater for the FEHBP plans as compared to Medicare Part D plans. For example, for the class analgesics the interaction term indicates that on average, the number of d rugs is expected to have a rate .695 time s greater for the FEHBP plans as compared to Medicare Part D plans (in reference to the class ADHD

PAGE 103

103 agents). Similarly, for the class gastrointestinal agents the inter action terms indicates that on a verage the number of drugs is expected to have a rate 678 times greater for the FEHBP plans (see Table 5 10). Thus, i n the overall model Medicare Part D plans provided broader coverage, while the interaction of type of plan and therapeutic class showed that the FEHBP plans provided broader coverage. The ex planation for this seeming contradiction lies in the difference between main and interaction effects. A main effect in the model shows whether there is an overall effect of a variable, after accounting for other variables in the model. An interaction ter m is added with the understanding that the coefficients of the lower order terms are conditional effects instead of main effects. In other words, the effect of one predictor is conditional on the value of the other (Grace Martin 2009). The coefficient of the lower order term is the effect only when the other term in the interaction equals 0 (Grace Martin 2009). A significant interaction term means that the effect of one predictor variable on the dependent variable is different at different values of the other predictor variable (Grace Martin 2000) Adding an interaction term results in the multiplication of the two predictor variables, which changes the interpretation of all coefficients (Grace Martin 2000) For example, consider the equation derived fr om results shown in Table 5 10: ND = 0 + 1 2 3 4 5 6 1 CL 1 15 CL 15 1 CL 1 *F 15 CL 15 *F+ If there were no interaction term, would be interpreted as the unique effect of type of plan on the number of drugs per therapeutic class. Since the interac tion indicates that the effect of type of plan on the number of drugs per therapeutic class is

PAGE 104

104 different for different values of therapeutic class, the unique effect of type of plan on the number of drugs per therapeutic class is not limited to but als o depends on the values of and therapeutic class (Grace Martin 2000) The unique effect of type of plan is represented by everything that is multiplied by type of plan in the model: + *therapeutic class. can now be interpreted as the unique ef fect of type of plan on the number of drugs per therapeutic class only when therapeutic class = 0. Because of the interaction, the predicted number of drugs varies depending on the therapeutic class. Another way of saying this is that the slopes of the r egression lines between type of plan and the number of drugs per therapeutic class are different for the different categories of therapeutic class (Grace Martin 2000) indicates how different those slopes are. Interpreting is more difficult (Grace Martin 2000) is the effect of therapeutic class when type of plan = 0. The effect of therapeutic class is + *type of plan, which is different at each of the tw o values of type of plan. For that reason, the best way to explore the effect of therapeutic class is to plug various values of type of plan into the equation to see how the number of drugs per therapeutic class, the dependent variable, changes. Consider for example in Table 5 10 the therapeutic class analgesics and the associated irr coefficients. If w e write the equation for the FEHBP coverage of the therapeutic class analgesics, we find : ND = 0 1 2 3 4 5 6 1 CL 1 1 CL 1 *F + where (ND) i s the number of dru gs per therapeutic class, (F) is the type of plan, (P) is the premium, (CP) is the copay, (CI) is the coinsurance, (T) i s the tie r, (E) is the

PAGE 105

105 enrollment, (CL ) i s the therapeutic class, (CL*F) i s an interaction t erm and () is the error term. In reference to the irr values found in Table 5 10, we would write the equation as: ND= 1.20 + 1.40 *type of plan + .99 *premium + 1.00 *copay + .93 *coinsurance + .98 *tier + 1.02 *enrollment + 3.10 *therapeutic class analgesics + .70 type of plan therapeutic class analgesics The irr or incidence rate ratio of [ 1.40 *type of plan + .70 type of plan *therapeutic class analgesics] represents the effect of type of plan on the num ber of dru gs per the rapeutic class (Grace Martin 2000) For the interaction of therapeutic class analgesics and type of plan FEHBP t ype of plan = 0 and therapeutic class analgesics = 1 so the effect of type of plan is 1.40 *0 + .70 0 *1 = 0 So for thera peutic class analgesics, FEHBP plans as compared to Medicare Part D plans would be expected to provide the same rate for drugs. The fact that all interaction terms in this model were statistically significant (except class anxiolytics) supports the non s ignificant findings for the classes in the 15 separate regressions. Specifically, the differences between and are very small (e.g. between and or and so there is only a small difference shown between the FEHBP and Medicare Part D plans, with respect to drug coverage and most times this difference is non significant when controlling for other factors. For example, consider the equation: Next consider the coefficients i n Table 5 10 for the variables Type of Plan (.34 ) and Interaction Analgesics ( .36 ). The difference is .34 + .36 = .02 For the

PAGE 106

106 variables Type of Plan (.34 ) and Interaction Anti Cancer Agents ( .34 ), the difference is .34 + 34 = 0 and so on In regard to the other factors in the model, enrollment was the single strongest predictor of the number of drugs per therapeutic class. As enrollment increased, the number of drugs per therapeutic class increased. Furthermore, the number of drugs per therapeutic class was greater for tier 2 brand name drugs as compared to tier 1 generic drugs. The coefficients of copayment, coinsurance and premium were not statistically significant. Results of the alternative specification, using each number of drugs in each therapeutic class separately as the dependent variable, indicated that for only one therapeutic class did Medicare Part D plans differ from the FEHBP plans after controlling for enrollment, premium, tier, coinsurance and copayment (Tables 5 11 thr ough 5 25) Analysis of the therapeutic class anxiolytics (Table 5 18) showed that the FEHBP plans provided more drugs, other factors held constant. Additionally, the therapeutic class respiratory tract agents was borderline significant (p=.067) in favor of the FEHBP providing greater coverage. T he remaining therapeutic classes (i.e. ADHD agents, analgesics, antibacterial agents, anti cancer agents, anticonvulsants, antidepressants, antipsychotics, arthritis agents, blood glucose regulators, cardiovascul ar agents, gastrointestinal agents, and hormonal agents) showed no difference in drug coverage among plans when controlling for enrollment, premium, tier, coinsurance and copayment which is consistent with the results from the overall model with interacti on effects.

PAGE 107

107 In regard to specific therapeutic classes, results revealed varying relationships between the dependent and independent variables, as shown in Tables 5 11 through 5 25 and summarized in Table 5 26 In summary, type of plan did not have a stati stically significant effect on the number of drugs in a therapeutic class with the exception of the class anxiolytics. On the other hand, the variable enrollment was positively associated with the number of drugs per therapeutic class for the classes: ant i cancer agents, blood products, cardiovascular agents, gastrointestinal agents, and respiratory tract agents. Coinsurance was a positive and significant predictor of the number of drugs per therapeutic class for th re e classes: ADHD agents, anti cancer ag ents and respiratory tract agents. Interestingly copay was only positive and significant for the classes ADHD agents and anti cancer agents. Similarly, tier was only significant (but negative) for the classes ADHD agents and anti cancer agents. P erhaps the most striking result was that the variable premium was not statistically significant for any of the classes. Thus, i ncreasing the premium had no significant effect on the number of drugs per therapeutic class. Cost Sharing The second research quest ion was (2) How do Medicare Part D stand alone plans compare to Federal Employees Health Benefits plans with respect to cost s haring on prescription drugs? To address this question copay and tier levels were analyzed among formularies. Analysis of plan f ormularies showed that the number of tiers varied depending on the plan. Tier 1 refers to preferred generic drugs, tier 2 refers to preferred brand name drugs, tier 3 refers to non preferred generic and brand name drugs, and tier 4 refers to specialty dru gs (i.e. those covered drugs that meet specific criteria and typically cost $500 or more per dose or $6,000 or more per year). Furthermore, as

PAGE 108

108 shown in Table 5 27, there was wide variation in copay and coinsurance depending on the tier. As the tier numbe r increases usually the drug out of pocket cost increases. For example, the copay is $7 for drugs offered on tier 1 for the Medicare Part D formulary Humana PDP Enhanced as compared to $40 for drugs offered on tier 2. s Blue Shield Basic formulary the copay is $10 for drugs offered on tier 1 compared to $35 for drugs offered on tier 2. Many formularies switch to applying coinsurance rates for tier 3 drugs. All formularies utilize coinsurance rates for tier 4 drugs. T o discourage use, tier 3 and 4 drugs are offered at higher co pay and coinsurance rates ( i.e. consumers must pay a substa ntial out of pocket amount if they want to purchase these drugs). When comparing plans based on whether they provide cost sharing thr ough copayments versus coinsurance, results showed that both Medicare Part D plans and the FEHBP plans utilize fixed dollar copayments more often than coinsurance for tier 1 generic drugs (Table 5 28). On the other hand, for tier 2 brand name drugs Medica re Part D plans were more likely to utilize copays while the FEHBP plans were more likely to utilize coinsurance. For the plans that were shown to utilize copayments, there was a statistically significant difference between Medicare Part D plans and the FE HBP plans (Table 5 29). For the Medicare Part D plans that utilized copayments, the mean copayment for tier 1 generic drugs was $4.53 (range was $0 $8) as compared to the FEHBP plans mean copa yment of $7.67 (range $5 $10) Therefore, the Medicare Par t D plans provided lower mean copays for tier 1 generic drugs as compared to the FEHBP plans (p<.05). On the other hand, the finding for tier 2 brand name drugs was non significant,

PAGE 109

109 meaning there was no difference between the two programs with respect to copay for tier 2 brand name drugs (Table 5 29). For the Medicare Part D plans that utilized coinsurance for tier 1 generic drugs, mean rates were 17% as compared to the FEHBP plan mean rates of 20% (p<.05). For the Medicare Part D plans that utilized co insurance for tier 2 brand name drugs, mean rates were 26% as compared to the FEHBP plan mean rates of 34% (Table 5 30). Furthermore, the FEHB prescription drug plans provided higher mean coinsurance rates as compared to Medicare Part D plans (Table 5 30)

PAGE 110

110 Table 5 1. Overview of Data Total Medicare Part D Plans 232 Medicare Part D Formularies 19 FEHBP Formularies/Plans 5 Total Number of Drugs Examined 266 Total Number of Brand Name Drugs Examined 197 Total Number of Generic Drugs Examined 69 T otal Number of Therapeutic Classes Examined 23 Table 5 2. Plans vs. Formularies Formulary Name Number of Plans per Formulary Medicare Part D Formularies AARP MedicareRx Preferred 31 AARP MedicareRx Saver/UnitedHealth Rx Basic 20 Advantage Star Plan by RxAmerica 9 Blue Medicare Rx 7 BlueRx 3 BravoRx 3 CIGNA Medicare Rx Plan One 5 Community CCRx Basic 26 First Health Part D Premier 8 Health Net Orange 7 HealthSpring Prescription Drug Plan 5 Humana PDP Enhanced/Complete 30 Humana PDP Standar d 27 Medco Medicare Prescription Plan 2 MedicareBlue Rx Option 3 3 Prescription Pathway Bronze Plan 12 SilverScript Value 12 WellCare Classic/Signature Option 1 20 WellCare Classic/Signature Option 2 2 FEHBP Formularies/Plans American Postal Worker s Union Health Plan 1 Blue Cross Blue Shield Basic 1 Blue Cross Blue Shield Standard 1 GEHA Benefit Plan 1 NALC Health Benefit Plan 1

PAGE 111

111 Table 5 3. Formulary Coverage of 266 Top Drugs in the U.S. Formulary TOTAL NUMBER OF DRUGS COVERED (n) PERCENT OF DRUGS COVERED (n/266) Medicare Part D Formularies (N=19) AARP MedicareRx Preferred 249 94% AARP MedicareRx Saver/UnitedHealth Rx Basic 248 93% Advantage Star Plan by RxAmerica 210 79% Blue Medicare Rx Standard 219 82% BlueRx Value 240 90% B ravoRx 212 80% CIGNA Medicare Rx Plan One 230 86% Community CCRx Basic Formulary 219 82% First Health Part D Premier 232 87% Health Net Orange 226 85% Health Spring Prescription Drug Plan 219 82% Humana PDP Enhanced or Complete 232 87% Human a PDP Standard 232 87% Medco Medicare Prescription Plan 214 80% MedicareBlue Rx 220 83% Prescription Pathway Bronze Plan 209 79% SilverScript Value 248 93% WellCare Classic/Signature Option 1 192 72% WellCare Classic/Signature Option 2 196 74% FEHBP Formularies (N=5) Blue Cross Blue Shield Basic 243 91% Blue Cross Blue Shield Standard 252 95% American Postal Workers Union Health Plan 264 99% GEHA Benefit Plan 227 85% NALC Health Benefit Plan 264 99%

PAGE 112

112 Table 5 4. Independent Samples t Te st: Formulary Coverage of 266 Top Drugs in the U.S. MED PART D N=19 (Mean SD) FEHBP N=5 (Mean SD) P Value Percent Drug Coverage 84 6.10 94 5.93 .004 Table 5 5. Independent Samples t Test: Formulary Coverage by Therapeutic Class Therape utic Class Total Number of Drugs in Class Medicare Part D Formulary Coverage (Mean Percent Covered SD) FEHBP Formulary Coverage (Mean Percent Covered SD) P Value ADHD Agents** 5 65 27.36 96 8.94 .001 Analgesics 13 98 3.62 98 3.58 .783 Anti Cancer Agents 25 77 15.50 85 17.07 .353 Antibacterials 16 93 7.18 96 8.50 .432 Anticonvulsants 8 100 0 100 0 N/A Antidementia Agents 3 100 0 93 14.76 .374 Antidepressant Agents 10 98 4.19 98 4.47 .961 Antipsychotic Agents 9 9 8 4.12 93 14.76 .505 Antiretroviral/Antiviral Agents 10 100 0 100 0 N/A Anxiolytic Agents** 4 0 0 95 11.18 .000 Arthritis Agents** 5 93 9.91 100 0 .005 Blood Glucose Regulators** 15 95 6.62 100 0 .004 Blood Products/Modifiers/Volum e Expanders* 12 72 12.93 90 18.22 .019 Cardiovascular Agents** 42 87 8.62 97 4.22 .003 Gastrointestinal Agents** 9 74 19.25 96 9.84 .005 Hormonal Agents* 15 82 9.36 93 8.17 .019 Metabolic Bone Disease Agents** 7 85 11.33 100 0 .000 Multiple Sclerosis Agents* 4 93 11.31 100 0 .021 Muscle Relaxants 3 77 15.76 80 29.91 .860 Opthalmic Agents** 5 72 10.15 100 0 .000 Respiratory Tract Agents* 21 86 9.63 98 4.47 .017 Sedative /Hypnotic Agents 3 84 20.33 93 14.76 .29 2 Vaccines 5 83 7.49 52 46.04 .205

PAGE 113

113 Table 5 6. Independent Samples t Test: Formular y Coverage for each of the 266 Top Drugs in the U.S. Therapeutic Class Drug Name Mean Percent of Medicare Part D Formularies including this Drug on Formulary (N=19) Mean Percent of FEHBP Plans including this Drug on Formulary (N=5) ADHD Agents Concerta 100 100 Focalin XR 74 80 Strattera 68 100 Vyvanse 16 100 Analgesics Acetaminophen and codeine 100 100 Fentanyl 100 100 Hydrocodone/Acetaminophen 100 100 Lidoderm 100 100 Naproxen 100 100 Oxycontin 84 100 Oxycodone and Acetaminophen 100 100 Oxycodone hydrochloride ER 100 100 Propoxyphene napsylate and acetaminophen 100 80 Suboxone 100 100 Tramadol hydrochloride 100 100 Anti Cancer Agents Abraxane 68 60 Aldara 100 100 Alimta 68 60 Allopurinol 100 100 Aloxi 26 80 Arimi dex 100 100 Avastin 63 80 Casodex 100 100 Eloxatin 74 60 Erbitux 58 80 Table 5 5. Continued *Difference between group means significant at .05 confidence level ** Difference between group means significant at .01 confidence level

PAGE 114

114 Table 5 6. Continued Therapeutic Class Drug Name Mean Percent of Medicare Part D Formularies including this Drug on Formulary (N=19) Mean Percent of FEHBP Plans including this Drug on Formulary (N=5) Anti Cancer Agents Femara 100 100 Gemzar 79 60 Herceptin 68 80 Neulasta 89 100 Neupogen 100 100 Pegasys convenience pack 100 100 Rituxan 100 60 Sandostatin LAR 100 100 Tarceva 100 100 Taxotere 74 60 Temodar 0 100 Thalomid 100 100 Velcade 100 60 Xeloda 0 100 Zometa 63 80 Antibacterials Amoxicillin 100 100 Amoxicillin TR and clavulanate potassium 100 100 Avelox 89 80 Azithromycin 100 100 Cephalexin 100 100 Ciprofloxacin hydrochloride 100 100 Cubicin 89 80 Doxycycline hyclate 100 100 Fluconazole 100 100 Levaquin 100 100 Penicillin VK 100 100 Solodyn 63 100 Sulfamethoxazole and trimethoprim 100 100 Synagis 53 100

PAGE 115

115 Table 5 6. Continued Therapeutic Class Drug Name Mean Percent of Medicare Part D Formularies including this Drug on Formulary (N=19) Mean Percent of FEHBP Plans including this Drug on Formulary (N=5) Zosyn 95 80 Zyvox 100 100 Anticonvulsants Depakote 100 100 Depakote ER 100 100 Gabapentin 100 100 Keppra 10 0 100 Lamictal 100 100 Lamotrigine 100 100 Lyrica 100 100 Topamax 100 100 Antidementia Agents Aricept 100 100 Exelon 100 100 Namenda 100 80 Antidepressant Agents Amitriptyline hydrochloride 95 100 Budeprion XL 100 100 Citalopram hydrob romide 100 100 Cymbalta 100 100 Effexor XR 100 100 Fluoxetine hydrochloride 100 100 Lexapro 84 80 Paroxetine hydrochloride 100 100 Sertraline hydrochloride 100 100 Trazodone hydrochloride 100 100 Anti inflammatory Agents Celebrex 89 100 I buprofen 100 100 Antimigraine Agents Imitrex 100 100 Antiparkinson Agents Mirapex 89 100

PAGE 116

116 Table 5 6. Continued Therapeutic Class Drug Name Mean Percent of Medicare Part D Formularies including this Drug on Formulary (N=19) Mean Percent of FEHBP Plans including this Drug on Formulary (N=5) Antipsychotic Agents Geodon 100 100 Invega 100 80 Risperdal 100 100 Risperdal consta 100 80 Risperidone 84 100 Seroquel 100 100 Zyprexa 100 100 Zyprexa Zydis 100 100 Antiretroviral/Antivira l Agents Atripla 100 100 Combivir 100 100 Epzicom 100 100 Kaletra 100 100 Norvir 100 100 Reyataz 100 100 Tamiflu 100 100 Truvada 100 100 Valtrex 100 100 Viread 100 100 Anxiolytics Alprazolam 0 80 Clonazepam 0 100 Diazepam 0 100 Lorazepam 0 100 Arthritis Agents Enbrel 100 100 Humira 100 100 Meloxicam 95 100 Orencia 68 100 Remicade 100 100

PAGE 117

117 Table 5 6. Continued Therapeutic Class Drug Name Mean Percent of Medicare Part D Formularies including this Drug on Formulary (N=19) Mea n Percent of FEHBP Plans including this Drug on Formulary (N=5) Bipolar Agents Abilify 100 80 Blood Glucose Regulators ACTOplus met 95 100 Actos 100 100 Avandia 89 100 Byetta 100 100 Gleevec 100 100 Glyburide 100 100 Humalog 84 100 Janume t 95 100 Januvia 100 100 Lantus 100 100 Lantus Solostar 100 100 Levemir 84 100 Metformin hydrochloride 100 100 NovoLog 89 100 NovoLog FlexPen 89 100 Blood Products/Modifiers/V olume Expanders Aggrenox 100 100 Angiomax 0 60 Aranesp 89 10 0 Carimune NF 58 80 Epogen 74 80 Gammagard Liquid 74 100 Gamunex 74 80 Lovenox 100 100 Plavix 100 100 Procrit 100 100 Venofer 0 80 Warfarin sodium 100 100

PAGE 118

118 Table 5 6. Continued Therapeutic Class Drug Name Mean Percent of Medicare Par t D Formularies including this Drug on Formulary (N=19) Mean Percent of FEHBP Plans including this Drug on Formulary (N=5) Cardiovascular Agents Amlodipine besylate 84 100 Amlodipine besylate and benazepril 84 100 Atenolol 100 100 Avalide 42 80 Avapro 47 80 Benicar 79 100 Benicar HCT 79 100 Caduet 37 80 Cartia XT 100 100 Catapres TTS 95 100 Clonidine 100 100 Coreg CR 100 100 Cozaar 84 100 Crestor 79 80 Digoxin 100 100 Diovan 79 100 Diovan HCT 79 100 Enalapril maleate 100 100 Furosemide 100 100 Hydrochlorthiazide 100 100 Hyzaar 84 100 Intergrilin 0 40 Isosorbide mononitrate 100 100 Klor Con M20 100 100 Lipitor 79 100 Lisinopril 100 100 Lisinopril and hydrochlorthiazide 100 100 Lotrel 95 100 Lovaza 95 100

PAGE 119

119 Table 5 6. Continued Therapeutic Class Drug Name Mean Percent of Medicare Part D Formularies including this Drug on Formulary (N=19) Mean Percent of FEHBP Plans including this Drug on Formulary (N=5) Metoprolol succinate 100 100 Metoprolol tartr ate 100 100 Niaspan 89 100 Potassium chloride 100 100 Pravastatin sodium 100 100 Ramipril 89 100 Simvastatin 100 100 Toprol XL 89 100 Triamterene and hydrochlorizide 100 100 Tricor 100 100 Verapamil SR 100 100 Vytorin 53 100 Zetia 10 0 100 Central Nervous System Agents Adderall XR 68 100 Provigil 100 100 Cystic Fibrosis Agents Pulmozyme 95 100 Erectile Dysfunction Agents Cialis 0 100 Viagra 0 80 Gastrointestinal Agents Aciphex 42 80 Asacol 100 100 Nexium 79 100 Omepr azole 100 100 Pantoprazole sodium 68 100 Prevacid 53 100 Prevacid Solutab 53 80 Protonix 74 100 Ranitidine hydrochloride 100 100

PAGE 120

120 Table 5 6. Continued Therapeutic Class Drug Name Mean Percent of Medicare Part D Formularies including this Drug on Formulary (N=19) Mean Percent of FEHBP Plans including this Drug on Formulary (N=5) Hormonal Agents, Stimulant/Replacem ent/Modifying (Sex Hormones/Modifiers ) AND Hormonal Agents, Stimulant/Replacem ent/Modifying (Thyroid) Androgel 1% 84 100 Avodart 10 0 100 Levothyroxine sodium 100 100 Levoxyl 100 100 Mirena 0 80 NuvaRing 68 100 Ortho Tri Cyclen Lo 58 100 Prednisone 100 100 Premarin 100 100 Sensipar 100 100 Synthroid 100 100 Trinessa 28 100 60 Yasmin 28 58 80 Yaz 28 68 80 Zempl ar 89 100 Metabolic Bone Disease Agents Actonel 79 100 Alendronate sodium 100 100 Boniva 74 100 Evista 100 100 Forteo 100 100 Fosamax 100 100 Fosamax Plus D 42 100 Multiple Sclerosis Agents Avonex 89 100 Betaseron 100 100 Copaxone 100 1 00 Rebif 84 100 Muscle Relaxants Carisoprodol 100 80 Cyclobenzaprine hydrochloride 100 100 Skelaxin 32 60 Ophthalmic Agents Cosopt 95 100 Lucentis 0 100

PAGE 121

121 Table 5 6. Continued Therapeutic Class Drug Name Mean Percent of Medicare Part D Formular ies including this Drug on Formulary (N=19) Mean Percent of FEHBP Plans including this Drug on Formulary (N=5) Lumigan 79 100 Restasis 100 100 Xalatan 84 100 Phosphate Binders Renagel 89 100 Radiography Agents Omnipaque 0 40 Visipaque 0 40 Res piratory Tract Agents Advair Diskus 100 100 Albuterol 100 100 Allegra D 12 Hour 58 80 Astelin 100 100 Clarinex 37 100 Combivent 100 100 Fexofenadine hydrochloride 100 100 Flomax 100 100 Flovent HFA 100 100 Fluticasone propionate 100 100 Nasacort AQ 47 100 Nasonex 89 100 ProAir HFA 89 100 Promethazine hydrochloride 100 100 Proventil HFA 100 100 Pulmicort respules 79 100 Singulair 100 100 Spiriva HandiHaler 89 100 Symbicort 89 100 Xolair 79 80 Xopenex 58 100 Sedatives /Hypnotics Ambien CR 68 100 Lunesta 84 80 Zolpidem tartrate 100 100

PAGE 122

122 Table 5 6. Continued Therapeutic Class Drug Name Mean Percent of Medicare Part D Formularies including this Drug on Formulary (N=19) Mean Percent of FEHBP Plans including this Drug on Formulary (N=5) Smoking Cessation Agents Chantix 79 100 Wellbutrin XL 95 80 Transplant Agents CellCept 100 100 Prograf 100 100 Urinary Tract Antispasmodics Detrol LA 89 100 Vesicare 95 100 Vaccines Gardasil 100 60 Prevnar 16 60 RotaT eq 100 40 Varivax 100 60 Zostavax 100 40 Vitamins Folic Acid 0 60 Vitamin D 0 60 Table 5 7. Formulary Coverage of Prescription Drugs: Brand versus Generic by Formulary Formulary Total Number of Generic Drugs Covered (ng) Total Number of Brand Nam e Drugs Covered (nb) Total Numb er of Drugs Cover ed (n) Generic Drugs as Percent of Total Drugs Covered (ng/n) Brand Drugs as Percent of Total Drugs Covered (nb/n) Generic Drugs Covered by this Formulary as a Percent of All Generic Drugs (ng/69) Brand Dru gs Covered by this Formulary as a Percent of All Brand Drugs (nb/197) Medicare Part D Formularies (N=19) AARP MedicareRx Preferred 63 186 249 25% 75% 91% 94% AARP MedicareRx Saver/UnitedH ealth Rx Basic 63 185 248 25% 75% 91% 94%

PAGE 123

123 Table 5 7. Continued Formulary Total Number of Generic Drugs Covered (ng) Total Number of Brand Name Drugs Covered (nb) Total Numb er of Drugs Cover ed (n) Generic Drugs as Percent of Total Drugs Covered (ng/n) Brand Drugs as Percent of Total Drugs Covered (nb/n) Generic Drug s Covered by this Formulary as a Percent of All Generic Drugs (ng/69) Brand Drugs Covered by this Formulary as a Percent of All Brand Drugs (nb/197) Advantage Star Plan by RxAmerica 62 148 210 30% 70% 90% 75% Blue Medicare Rx Standard 62 157 219 28% 72% 90% 80% BlueRx Value 63 177 240 26% 74% 91% 90% BravoRx 63 149 212 30% 70% 91% 76% CIGNA Medicare Rx Plan One 60 170 230 26% 74% 87% 86% Community CCRx Basic 62 157 219 28% 72% 90% 80% First Health Part D Premier 62 170 232 27% 73% 90% 86% Health Net Orange 62 164 226 27% 73% 90% 83% HealthSpring Prescription Drug Plan 60 159 219 27% 73% 87% 81% Formulary Total Number of Generic Drugs Covered (ng) Total Number of Brand Name Drugs Covered (nb) Total Numb er of Drugs Cover ed (n) Generic Drugs as Percent of Total Drugs Covered (ng/n) Brand Drugs as Percent of Total Drugs Covered (nb/n) Generic Drugs Covered by this Formulary as a Percent of All Generic Drugs (ng/69) Brand Drugs Covered by this Formulary as a Percent of All Brand D rugs (nb/197)

PAGE 124

124 Table 5 7. Continued Formulary Total Number of Generic Drugs Covered (ng) Total Number of Brand Name Drugs Covered (nb) Total Numb er of Drugs Cover ed (n) Generic Drugs as Percent of Total Drugs Covered (ng/n) Brand Drugs as Percent of To tal Drugs Covered (nb/n) Generic Drugs Covered by this Formulary as a Percent of All Generic Drugs (ng/69) Brand Drugs Covered by this Formulary as a Percent of All Brand Drugs (nb/197) Humana PDP Enhanced/Co mplete 62 170 232 27% 73% 90% 86% Humana PDP Standard 62 170 232 27% 73% 90% 86% Medco Medicare Prescription Plan 63 151 214 29% 71% 91% 77% MedicareBlue Rx Option 62 158 220 28% 72% 90% 80% Prescription Pathway Bronze Plan 62 147 209 30% 70% 90% 75% SilverScript Value 63 185 248 25% 75% 91% 94% WellCare Classic/Signat ure Option1 60 132 192 31% 69% 87% 67%

PAGE 125

125 Table 5 7. Continued Formulary Total Number of Generic Drugs Covered (ng) Total Number of Brand Name Drugs Covered (nb) Total Numb er of Drugs Cover ed (n) Generic Drugs as Perce nt of Total Drugs Covered (ng/n) Brand Drugs as Percent of Total Drugs Covered (nb/n) Generic Drugs Covered by this Formulary as a Percent of All Generic Drugs (ng/69) Brand Drugs Covered by this Formulary as a Percent of All Brand Drugs (nb/197) WellCar e Classic/Signat ure Option 2 62 134 196 32% 68% 90% 68% FEHBP Formularies (N=5) American Postal Workers Union Health Plan 67 197 264 25% 75% 97% 100% Formulary Total Number of Generic Drugs Covered (ng) Total Number of Brand Name Dru gs Covered (nb) Total Numb er of Drugs Cover ed (n) Generic Drugs as Percent of Total Drugs Covered (ng/n) Brand Drugs as Percent of Total Drugs Covered (nb/n) Generic Drugs Covered by this Formulary as a Percent of All Generic Drugs (ng/69) Brand Drugs Co vered by this Formulary as a Percent of All Brand Drugs (nb/197) Blue Cross Blue Shield Basic 68 175 243 28% 72% 99% 89% Blue Cross Blue Shield Standard 68 184 252 27% 73% 99% 93% GEHA Benefit Plan 68 159 227 30% 70% 99% 81% NALC Health Benefit Plan 67 197 264 25% 75% 97% 100%

PAGE 126

126 Table 5 8. Independent Samples t Test: Formulary Coverage Brand versus Generic of Total Drugs Covered MED PART D n=19 (Mean SD) FEHBP n=5 (Mean SD) P Value Percent Brand Name Drugs Covered out of total drugs cover ed 72 2.07 73 2.12 .458 Percent Generic Drugs Covered out of total drugs covered 28 2.07 27 2.12 .458 Table 5 9. Independent Samples t Test: Formulary Coverage Brand versus Generic as % of All Brand and Generic Drugs MED PART D n=19 (Mean SD) FEHBP n=5 (Mean SD) P Value Percent Brand Name Drugs Covered 82 8.00 93 8.02 .015 Percent Generic Drugs Covered 90 1.34 98 1.10 .000 Table 5 10. Negative Binomial Regression Predicting Number of Drugs per Therapeutic Class a mong Medicare Part D and the FEHBP Variable irr SE z p value Intercept 1.20128 1.20128 .129187 9.29877 .000 Type of Plan*** .340841 1.40612 .090507 3.76592 .000 Premium .357216E 03 .999642 .996970E 03 .358302 .720 Copay .662928E 03 1.00066 .544193E 03 1.21819 .223 Coinsurance .069158 .933179 .066941 1.03312 .302 Tier* .017108 .983037 .784080E 02 2.18187 .029 Enrollment*** .021756 1.02199 .510407E 02 4.26242 .000 Analgesics*** 1.13318 3.10551 .130873 8.65858 .000 Anti Cancer Agents*** 1.53474 4.64011 .125799 12.2000 .000 Antibacterials*** 1.24873 3.48591 .128903 9.68736 .000 Anticonvulsants*** .647 861 1.91144 .138627 4.67340 .000 Antidepressants*** .850632 2.34112 .134918 6.30483 .000 Antipsychotic Agents*** .693394 2.00049 .137720 5.03481 .000 Anxiolytics 7.55466 5.24E 04 19.6139 .385168 .700 Arthritis Agents .068611 1.07101 .151928 .451601 .652 Blood Glucose Regulators*** 1.23424 3.43577 .129136 9.55768 .000 Blood Products/Modifiers/Volu me Expanders*** .789408 2.202 09 .134114 5.88611 .000 Cardiovascular Agents*** 2.22288 9.23388 .121593 18.2813 .000 Gastrointestinal Agents*** .625634 1.86943 .137715 4.54296 .000 Hormonal Agents*** 1.13528 3.11204 .1300 55 8.72921 .000 Respiratory Tract Agents*** 1.54368 4.68179 .125922 12.2590 .000

PAGE 127

127 Table 5 10. Continued Variable irr SE z p value Interaction Analgesics *** .363843 .695000 .100933 3.60480 .000 Interaction Anti Cancer Agents *** .340175 .711645 .097236 3.49846 .000 Interaction Antibacterials ** .308413 .734611 .09 9618 3.09596 .002 Interaction Anticonvulsants*** .340163 .711654 .106880 3.18266 .001 Interaction Antidepressants** .330022 .718907 .104092 3.17049 .002 Interaction Antipsychotic Agents** .307468 .735306 .106286 2.89282 .004 Interaction Anxiolytics 7.64641 4.78E 04 19.6136 .389851 .697 Interaction Arthritis Agents** .348412 .708843 .117163 2.97373 .003 Interaction Blood Glucose Regula tors*** .345687 .707734 .099725 3.46642 .001 Interaction Blood Products/Modifiers/Volu me Expanders*** .367880 .692200 .103651 3.54921 .000 Interaction Cardiovascular Agents** .358287 .698872 .093990 3.81198 .000 Interaction Gastrointestinal Agents** .388108 .678339 .106250 3.65276 .000 Interaction Hormonal Agents .373899 .688046 .100414 3.72356 .000 Interaction Respiratory Tract Agents*** .399111 .670916 .097227 4.10494 .000 value, factor change in the dependent variable given a one unit increase in the predictor. Categorical variables were dummy coded: type of plan (1, Medicare Part D; 2, FEHBP), tier (1,preferred generic drugs; 2, preferred brand drugs), and therapeutic class. R = .96, p< .001. *p< .05 **p< .01 ***p< .001 Table 5 11. Negative Binomial Regression Predicting Number of Drugs for ADHD Agents Variable irr SE z p value Intercept 1.23792 1.23792 .579085 2.13771 .033 Type of Plan .077802 .925148 .110629 .703273 .482 Premium .449939E 03 .999550 .126559E 02 .355518 .722 Copay* .030958 1.03144 .928376E 02 3.33468 .001 Coinsurance* 1.70386 5.49511 .592478 2.87582 .004 Tier* 1.84892 .157407 .142749 12.9523 .000 Enrollment .049508 1.05075 .051446 .962328 .336 *Variable is significan t at .05 confidence level

PAGE 128

128 Table 5 12. Negative Binomial Regression Predicting Number of Drugs for Analgesics Table 5 13. Negative Binomial Regression Predicting Number of Drugs for Anticancer Agents V ariable irr SE z p value Intercept 2.36938 2.36938 .172053 13.7712 .000 Type of Plan .821828E 02 1.00825 .036474 .225318 .822 Premium .202492E 03 1.00020 .394189E 03 .513693 .607 Copay* .72275 4E 02 1.00725 .164767E 02 4.38652 .000 Coinsurance* .512864 1.67006 .202123 2.53739 .011 Tier* .085236 .918296 .023811 3.57970 .000 Enrollment* .044644 1.04566 .015483 2.88352 .004 *Variable is significant at .05 confidence level Table 5 14. Negative Binomial Regression Predicting Number of Drugs for Antibacterials Variable irr SE z p value Intercept 2.49227 2.49227 .197514 12.6182 .000 Type of Plan .033287 1.03385 .042582 .781712 .434 Premium .413186E 03 1.00041 .459103E 03 .899985 .368 Copay .128296E 02 1.00128 .188903E 02 .679162 .497 Coinsurance .041466 1.04234 .233062 .177919 .859 Tier .013611 .986481 .027222 .500002 .617 Enrollment .016341 1.01648 .017791 .918490 .358 *Varia ble is significant at .05 confidence level Table 5 15. Negative Binomial Regression Predicting Number of Drugs for Anticonvulsants Variable irr SE z p value Intercept 2.07944 2.07944 .274730 7.56905 .000 Type of Plan .152065E 28 1 .058193 .261313E 27 .999 Premium .162636E 31 1 .648816E 03 .250666E 28 .999 Copay .371930E 30 1 .261608E 0 2 .142171E 27 .999 Coinsurance .336054E 28 1 .322232 .104289E 27 .999 Tier .456440E 29 1 .037649 .121236E 27 .999 Enrollment .260431E 28 1 .024797 .105025E 26 .999 *Variable is significant at .05 confidence level Variable irr SE z p value Intercept 2.57037 2.57037 .218586 11.7591 .000 Type of Plan .025352 .974966 .045 741 .554255 .579 Premium .242869E 03 .999757 .519303E 03 .467682 .640 Copay .109139E 02 .998909 .208861E 02 .522544 .601 Coinsurance .139494 .869798 .257390 .541955 .588 Tier .014674 1 .01478 .030071 .487989 .626 Enrollment .195950E 02 1.00196 .019752 .099205 .921 *Variable is significant at .05 confidence level

PAGE 129

129 Table 5 16. Negative Binomial Regression Predicting Number of Drugs for Antidepressants Variable irr SE z p value Intercept 2.26997 2.26997 .246992 9.19045 .000 Type of Plan .011072 1.011134 .052659 .210254 .833 Premium .644786E 04 1.000064 .584028E 03 .110403 .912 Copay .371952E 03 0.999628 .235272E 02 .158095 .874 Coinsurance .028856 0.971556 .289978 .099513 .921 Tier .443259E 02 1.004442 .033852 .130942 .896 Enrollment .171901E 02 1.00172 .022282 .077149 .939 Variable is significant at .05 confidence level Table 5 17. Negative Binomial Regression Predicting Number of Drugs for Antipsychotic Table 5 18. Negative Binomial Regression Predicting Number of Drugs for Anxiolytics Variable irr SE Z p value Intercept .447114 .447114 1.02663 .435518 .663 Type of Plan* 2.05698 0.127839 .122602 16.7778 .000 Premium .386747E 03 1.000387 .103258E 02 .374545 .708 Copay .104781E 02 0.998953 .013676 .076614 .939 Coinsurance .173947 0.840341 1.24582 .139625 .889 Tier .015653 1.015776 .162323 .096435 .923 Enrollment .421970E 02 0.995789 .096363 .043790 .965 *Variable is significant at .05 confidence level Table 5 19. Negative Binomial Regression Predicting Number of Drugs for Arthritis Agents Variable irr SE Z p value Intercept 1.26990 1.26990 .354064 3.58664 .000 Type of Plan .763167E 02 0.992397 .074596 .102307 .919 Premium .187646E 03 1.000188 .816594E 03 .229791 .818 Copay .342083E 02 1.003427 .339484E 02 1.00765 .314 Coinsurance .153342 1.165724 .417966 .366876 .714 Tier .037654 0.963046 .049021 .768121 .442 Enrollment .021555 1.021789 .031935 .674965 .500 *Variable is significant at .05 confidence level Variable irr SE Z p value Intercept 2.22265 2.22265 .263842 8.42419 .000 Type of Plan .020014 1.020216 .057006 .351083 .726 Premium .313291E 03 1.000313 .626068E 03 .500410 .617 Copay .272798E 02 0.997276 .253748E 02 1.07508 .282 Coinsurance .470700 0.624565 .314871 1.49490 .135 Tier .040382 1.041208 .036623 1.10264 .270 Enrollment .237244E 02 0.99763 .023797 .099695 .921 Variable is significant at .05 confidence level

PAGE 130

130 Table 5 20. Negative Binomial Regression Predicting Number of Drugs for Blood Glucose Regulators Variable irr SE Z p value Intercept 2.43298 2.43298 .203258 11.9700 .000 Type of Plan .564821E 02 1.005664 .042890 .131690 .895 Premium .941978E 04 1.000094 .471124E 03 .199943 .842 Copay .122969E 02 1.00123 .193690E 02 .634877 .526 Coinsurance .193206 1.213133 .237932 .812022 .417 Tier .017576 0.982578 .027835 .631441 .528 Enrollment .019012 1.019194 .018337 1.03680 .300 Variable is significant at .05 confidence level Table 5 21. Negative Binomial Regression Predicting Number of Drugs for Blood Products/Modifiers/Volume Expanders Variable irr SE Z p value Intercept 1.79314 1.79314 .250631 7.15450 .000 Type of Plan .019804 0.980391 .051836 .382052 .702 Premium .625153E 03 1.000625 .549474E 03 1.13773 .255 Copay .109515E 02 0.998905 .243907E 02 .449001 .653 Coinsurance .214497 0.806947 .300938 .712763 .476 Tier .016951 1.017095 .035100 .482938 .629 Enrollment* .042820 1.04375 .022666 1.88916 .059 *Variable is significant at .05 confidence level Table 5 22. Negative Binomial Regression Predicting Number of Drugs for Cardiovascular Agents Variable irr SE Z p value Intercept 3.43159 3.43159 .124658 27.5281 .000 Type of Plan .018014 0.982147 .026110 .689940 .490 Premium .125224E 03 1.000125 .287702E 03 .435256 .663 Copay .342248E 03 0.99 9658 .120154E 02 .284841 .776 Coinsurance .170040 0.843631 .148279 1.14676 .251 Tier .838469E 02 1.00842 .017322 .484039 .628 Enrollment* .022490 1.022745 .011259 1.99740 .046 *Varia ble is significant at .05 confidence level Table 5 23. Negative Binomial Regression Predicting Number of Drugs for Gastrointestinal Agents Variable irr SE Z p value Intercept 1.66373 1.66373 .277372 5.99819 .000 Type of Plan .051670 0.949642 .057120 .904597 .366 Premium .167734E 03 1.000168 .627876E 03 .267146 .789 Copay .192391E 02 0.998078 .271797E 02 .707848 .479 Coinsurance .544265 0.580268 .336764 1.61616 .106 Tier .034763 1.035374 .039293 .884717 .376 Enrollment .043489 1.044449 .025091 1.73327 .083 *Variable is significant at .05 confidence level

PAGE 131

131 Table 5 24. Negative Binomial Regression Predicting Number of Drugs for Hormonal Agents Variable irr SE Z p value Intercept 2.33629 2.33629 .214833 10.8749 .000 Type of Plan .040644 0.960171 .044537 .912584 .361 Premium .142946E 03 1.000143 .491782E 03 .290669 .771 Copay .174050E 03 0.999826 .208432E 02 .083504 .933 Coinsurance .305238 0.736948 .257694 1.18450 .236 Tier .010715 1.010773 .030122 .355731 .722 Enrollment .024813 1.025123 .019424 1.27745 .201 *Variable is significant at .05 confidence level Table 5 25. Negative Binomial Regression Predicting Number of Drugs for Respiratory Tract Agents Variable irr SE Z p value Intercept 2.74634 2.74634 .178137 15.4170 .000 Type of Plan .067012 0.935184 .036483 1.83677 .066 Premium .107815E 03 0.999892 .410563E 03 .262602 .793 Copay .167628E 02 0.998325 .173596E 02 .965621 .334 Coinsurance* .497231 0.608212 .214851 2.31431 .021 Tier .031025 1.031511 .025099 1.23608 .216 Enrollment .027526 1.027908 .016125 1.70707 .088 *Variable is significant at .05 confidence level Table 5 26. Summary of Negative Binomial Regressions Predicting Number of Drugs for Top 15 Therapeutic Classes Type of Plan Premium Copay Coinsurance Tier Enrollment ADHD Agents .078 .450E 03 .031* 1.704* 1.849* .050 Analgesics .025 .243E 03 .109E 02 .139 .015 .196E 02 Anti cancer .822E 02 .202E 03 .723E 02* .513* .085* .045* Antibacterials .033 .413E 03 .128E 02 .041 .014 .016 Anticonvuls ants .152E 28 .162E 31 .372E 30 .336E 28 .456E 29 .26E 02 Antidepressants .011 .645E 04 .372E 03 .029 .443E 02 .172E 02 Antipsychotics .020 .313E 03 .273E 02 .471 .040 .24E 02 Anxiolytics 2.06* .387E 03 .10 5E 02 .174 .016 .422E 02 Arthritis Agents .763E 02 .188E 03 .342E 02 .153 .038 .022 Blood Glucose .564E 02 .942E 04 .123E 02 .193 .018 .019 Blood Products .020 .625E 03 .110E 02 .214 .017 .043* Cardiovascula r .018 .125E 03 .342E 03 .170 .838E 02 .022* Gastrointestinal .052 .168E 03 .192E 02 .544 .035 .043* Hormonal Ag. .041 .143 03 .174E 03 .305 .011 .025 Respiratory Ag. .067 .108E 03 .168E 02 .497* .031 .0 28* *Variable is significant at .05 confidence level

PAGE 132

132 Table 5 27. Cost Sharing Analysis among Medicare Part D (N=19) and FEHBP (N=5) Formularies Formulary Tiers Cost Sharing Medicare Part D Formularies AARP MedicareRx Preferred 4 $7 $38 $66.2 5 98 .33 AARP MedicareRx Saver 4 $5 $22 $49.45 98 .25 Advantage Star Plan by RxAmerica 4 $4.75 5.50 .25 .25 .45 Blue Medicare 4 $2 9 $33 45 $60 75 .25 .30 BlueRx 4 $4 5 $25 $35 $55 65.25 .33 .25 .33 CIGNA Medicar e Rx Plan One 4 $2 2.50 $25 33 $72.50 81 .25 First Health Part D Premier 4 $7 $27 30 $51 64 .33 Health Net Orange 5 $2 5 $30 44 90 .25 .33 .25 .33 HealthSpring Prescription Drug Plan 2 .25 .25 Humana PDP Enhanced 4 $ 7 $40 $70 .33 Humana PDP Standard 3 .10 .15 .25 .40 .48 MedicareBlue Rx 4 $8 $43 $73 .30

PAGE 133

133 Table 5 27. Continued Formulary Tiers Cost Sharing SilverScript Value 4 $8 $32.50 39.25 $98 .25 WellCare Classic/Signature Opt ion 1 4 $0 $30 41 $67 92 .25 .33 WellCare Classic/Signature Option 2 4 $0 $36 39 $79 85 .25 .33 FEHBP Formularies American Postal Workers Union Health Plan 2 $8 .25 Blue Cross Blue Shield Basic 3 $10 $35 .50 Blue Cross Blu e Shield Standard 3 .20 .30 .30 GEHA Benefit Plan 2 $5 .5 NALC 2 .20 .30 Table 5 28. Cost Sharing among Top Drugs in the U.S. Tier Medicare Part D Total number of plans with copay (n=218) FEHBP Total number of plans with copay (n=10) Tier 1 Generic Drugs 181 or 83% 6 or 60% Tier 2 Brand Name Drugs 147 or 67% 2 or 20% Table 5 29. Independent Samples t Test Comparison of Mean Copay among Plans with Copay by Plan Type Copay Medicare Part D Plans (Mean SD) Copay FEHBP Plans ( Mean SD) P Value Tier 1 Generic Drugs (Medicare Part D Plans n=181 and FEHBP Plans n=6) $4.53 3.05 $7.67 2.25 .014 Tier 2 Brand Name Drugs (Medicare Part D Plans n=147 and FEHBP Plans n=2) $35.06 6.32 $35.00 0 .901

PAGE 134

134 Table 5 30. Indep endent Samples t Test Comparison of Mean Coinsurance among Plans with Coinsurance by Plan Type Coinsurance Medicare Part D Plans (Mean SD) Coinsurance FEHBP Plans (Mean SD) P Value Tier 1 Generic Drugs (Medicare Part D Plans n=37 and FEHBP P lans n=4) 17% .046 20% 0 .000 Tier 2 Brand Name Drugs (Medicare Part D Plans n=71 and FEHBP Plans n=8) 26% .020 34% .103 .066

PAGE 135

135 CHAPTER 6 SUMMARY, DISCUSSION, AND CONCLUSION Medicare Part D and the Federal Employees Health Benefits Program both provid e services to millions of enrollees, yet their method s of operations differ. Specifically, Medicare Part D is run through extensive government regulations (approximately 2,400 pages), while the FEHBP runs by laws that comprise ab out twenty pages (Ca in 1999). The Medicare program has been criticized for its lack of provision of services. For example, nine out of ten Medicare enrollees purchase supplemental coverage (Francis 2009). S ome advocate for reforming Medicare to look more like the FEHBP pro gram cost control, lack of fraud/abuse, and protection against interest group politics ( National Bipartisan Commission 1999; Francis 2003; and Oberlander 2007) On the other hand, the Medicare Part D program has exceeded the expectations of many. Premiums and costs are one third below initial predictions (Francis 2009). Additionally, Medicare does not require prior plan approval for nonemergency hospitalization while the FEHBP does. Furthermore, Medicare offers greater access to non preferred providers. Currently, controversy exists as to which program is preferable particularly given the challenges faced by a growing elderly population in the Unit ed States This research aims to compare the two programs with respect to prescription drug coverage and cost sharing. Formulary Coverage Before discussing the findings on formulary coverage it is important to briefly discuss the two different methods that were used for analysis. This d iscussion is particularly important given the independent t tests, which found that the FEHBP plans

PAGE 136

136 provided broader drug coverage whereas findings of the regression analysis showed that there was only a small difference between the two programs and that difference was only statistically significant for the classes anxiolytics and respiratory agents, for which the FEHBP provided broader drug coverage. The independent t test was used for testing hypothes e s related to a single factor while the regression analysis was used for testing hypothes e s related to multiple factors The question arises as to which method is preferable; t he answer depends on the rese arch objectives and questions. Multiple regression provides an advantage because it takes into accou nt the effect of multiple predictor variables simultaneously. That is, the effect of a single predictor variable is estimated assuming that all other predic tor variables are held constant On the other hand, the independent t tests answer the bottom line question of which program provides broader drug coverage when all factors affecting coverage are allowed to vary The first research question addressed in this study was how do Medicare Part D stand alone plans compare to Federal Employees Health Benefit s plans with respect to coverage of prescription drugs? The null hypothesis was that there would be no difference between the plans U sing the independent samples t test for the 19 Medicare Part D formularies analyzed, formulary coverage of the top drug s dispensed and sold in the United States ranged from 72 94% (average 84%), while the range was 85 99% (average 94%) for the 5 FEHBP formularies examined (p<.01). Overall, t he independent sample t test f indings indicate that the FEHBP plans provided broad er drug coverage as compared to Medicare Part D plans.

PAGE 137

137 On the other hand, the regression results indicate that, once other factors and interaction effects were taken into account the programs were shown to be about the same in terms of coverage In oth er words, the bivariate analysis results disappeared in multivariate analysis. T his difference in findings using the two different analytical methods may be useful to different groups For example, c onsumers may be interested in the actual number and kin d of drugs on their formulary so they may find the independent sample t test results useful Other consumers may find the factors copay and coinsurance to be important so they may focus on regression results. Additionally, health plan providers may be mo re interested in how factors like premium, copay, coinsurance, tier and enrollment affect drug coverage. As health plan providers make complex decisions on which drugs to include on their formulary they may need to take the multiple factors used in the r egression analysis into consideration When examined by therapeutic class, independent samples t test s revealed that the FEHBP provide d broader drug coverage in 12 of the 23 therapeutic classes: ADHD agents, anxiolytics, arthritis agents, blood glucose ag ents, blood products/modifiers/ volume expanders, cardiovascular agents, gastrointestinal agents, hormonal agents, metabolic bone disease agents, multiple sclerosis agents, ophthalmic agents, and respiratory tract agents (p<.05). Additionally, results of the t tests revealed that for one of the top three therapeutic classes in terms of drugs sold and dispensed in the U.S., anti cancer agents, there was no difference found between the two programs in regard to drug coverage Yet, for the other top classes, cardiovascular agents and respiratory agents, the FEHBP was shown to provide broader coverage.

PAGE 138

138 Moreover, the therapeutic class anxiolytics, which was composed entirely of benzodiazepine drugs (i.e. alprazolam, clonazepam, diazepam, and lorazepam), showe d the greatest difference in drug coverage between formularies. Results of the independent samples t test revealed that among Medicare Part D formularies, none of the drugs were covered in this class, while the FEHBP formularies covered on average 95% (p <.05). This finding may be explained by the difference in enrollee population characteristics of the two programs (i.e. the majority of Medicare beneficiaries are older vs. the FEHBP which includes working age adults and those 65 and older). For example, some studies show that anxiolytics are not recommended for use in the elderly ( Pontillo, Lang, and Stein 2002), while other studies recommend their use, but only with caution (Merck Manual of Geriatrics 2010). Specifically, short/intermediate acting anxi olytics like alprazolam, lorazepam, and oxazepam are recommended for use, while longer acting anxiolytics like diazepam or clonazepam are not recommended for use in the elderly. Additionally, the dosage is usually lower for elderly patients as compared to younger patients. Anxiolytics are classified as short or long acting based on their onset and duration of action. The decision to prescribe or not prescribe them is also based on the associated side effects and addictive properties of the drugs. Possib le side effects include sedation and disorientation. For this reason some have attributed these drugs to the cause of falls and fractures in the elderly. In fact, Medicare decision makers have cited these factors as reason to exclude these drugs from the ir formularies (Bambauer, Sabin, and Soumerai 2005; United States Pharmacopeial Convention 2008). On the other hand, recent studies revealed that exclusion of anxiolytics from formularies may decrease use, but may not result in decreased fracture

PAGE 139

139 risk (Br iesacher et al 2010; Wagner et al 2007). Considering the lack of consensus on the use of anxiolytics, it could be possible that the difference between programs for this e not to include anxiolytics on their formularies. Yang and colleagues suggest that future studies that examine anxiolytic use in Medicare beneficiaries should focus on age, sex, and racial ethnic differences among beneficiaries ( Yang et al 2008). To prov ide more explanation regarding the therapeutic classes in which differences between programs were found, it is helpful to examine certain drugs. Specifically, the drugs that some FEHBP plans offer that none of the Medicare Part D plans offer were : fo lic a cid mirena, venofer, l ucentis o mnipaque v isipaque angiomax, i ntegrilin, cialis, viagra, t emodar, xeloda and vitamin D. Review of the literature suggests that they provide less coverage, but that provision of these drugs was not necessary for the Medicare ( mostly over 65 years of age ) population. For example, folic acid may not be included on Medicare Part D formularies because it is primarily found in many foods and usually pre scribed to pregnant women ( Office on Specifically, it is a water soluble B vitamin and is added to many cold cereals, flour, breads, pasta, bakery items, cookies, and crackers. It is also found in many vegetables and is used for pre venting and treating folic acid deficiency, as well as anemia (Office Pregnant women take folic acid to prevent miscarriage and birth defects With this in mind, it is understandable why folic acid was not included on Medicare Part D formularies.

PAGE 140

140 Similarly, it is recommended that the drug venofer, which r eplenishes depleted stores of iron, be used with caution in the elderly. The elderly may be more sensitive to its side effects of hypotension and nausea (Rx List the Internet Drug Index Venofer 2010) It is also recommended that the drug mirena, a birth control agent be used with caution in the elderly due to their increased sensitivity to its effects. Common side effects include a cne, back pain, breast pain or tenderness, cha nges in menstrual bleeding changes in sex drive, dizziness, lighthe adedness, bleeding, headache, nausea, vomiting and weight gain (Drugs.com Mirena 2010) Review of this side effect profile or the low if an y, number of Medicare beneficiaries taking the drug to prevent pregnancy may have been the reason why Medicare decision makers decided not to include this drug. Next, the exclusion of the drug lucentis may have been due to 2007 reports that the drug is lin ked to stroke in elderly patients. Results of the safety study, (SAILOR) showed that the risk for stroke was significantly higher in patients receiving the recommended dose of lucentis (0.5 mg) compared with a 0.3 mg dose (1.2% vs 0.3%; P = .02) at an ave rage follow up of 230 days (Waknine 2007) Furthermore, those with a history of stroke appeared to be at increased risk for subsequent stroke Lucentis, which treats (wet) age related macu lar degeneration is still widely used among elderly patients, but there is reason to be cautious with its use ( The Eye Digest 2009) Both drugso mnipaque and v isipaque are used as radiographic contrast mediums administered by intravascular injection to allow ra diographic visualization of internal structures. Since these drugs are highly excreted by the kidney it is recommended that

PAGE 141

141 they be used with caution in the elderly (Drugs.com Omnipaque 2010; Drugs.com Visipaque 2010) Those with impaired renal function are at gre ater risk and elderly patients are more likely to have decreased renal function. On the other hand, literature review shows that there is no clear explanation why the other excluded drugs were left out. For example, angiomax is used as a blood thinner for patients who aim to block the formation of clots in their bloodstream. It is recommended that the drug be used with caution in the elderly because they may be more sensitive to its effects, such as bleeding (Rx List the Internet Drug Index Angiomax 2010) Similarly, integrilin keeps the platelets in the bloodstream from coagulating (clotting) to prevent unwanted clots that can occur with certain heart or blood vessel conditions (Rx List the Internet Drug Index Integrilin 2010) Caution is advised with t his drug for the same reasons as described for angiomax. Although it is advised that these drugs be used with caution in the elderly that advice holds true for the other drugs in the same class (e.g. plavix, warfarin) that are covered by Medicare Part D plans (Rx List the Internet Drug Index Plavix 2010; Rx List the Internet Drug Index Warfarin 2010) Therefore, it is unclear why these drugs were excluded from Medicare Part D plans. Additionally, some have questioned why cialis and viagra were excluded. These drugs are used for erectile dysfunction (ED), which primarily plagues senior men. According to the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) in older men, ED usually has a physical cause, such as disease, injury, or side effects of drugs (NIDDK 2010) Furthermore, the American Urological Association has stated that "By far, the most important cause of the development of ED is the presence

PAGE 142

142 of illnesses like high blood pressure, diabetes mellitus, high cholesterol leve ls and cardiovascular disease ( American Urological Association Education and Research 2010). Incidence has been shown to increase with age (i.e. about 5% of 40 year old m en and between 15 and 25% of 65 year old men experience ED). Various senior advocat e groups have expressed concern related to the exclusion of cialis and viagra from Medicare Part D formularies (Senior Journal 2006) They claim that many senior men find relief through these drugs, yet the drugs still remain excluded. Furthermore, for th e drugs temodar and xeloda, which are used to treat brain tumors it is interesting that they are covered by Medicare Part B, but not Medicare Part D (Mahay 2009; Nelson 2010) hospital or clinic a nd some self administered prescription drugs. Temodar and xeloda are self administered drugs. Also, due to the fact that these drugs are included in the protected class of anticancer (antineoplastic) agents one may speculate that they would be included on the Medicare Part D formularies. It is unclear why Part D plans excluded these two drugs from their formularies. Lastly, findings revealed that none of the Medicare Part D plans covered vitamin D. It is unclear why the plans decided not to cover vitamin D. A study by Mosekilde 2005 provided a review that summarized knowledge on vitamin D status in the elderly (Mosekilde 2005) The author concluded that studies suggest fortification of food cannot provide sufficient vitamin D to the elderly. Furthermor e, vitamin D insufficiency was noted to be related to a number of other disorders frequently observed among the elderly, such as breast, prostate and colon cancers, type 2 diabetes, and cardiovascular disorders, yet causality has not been established (Mose kilde 2005) In regard to

PAGE 143

143 treatment, it was suggested t on with calcium be used for high blood pressure in elderly women (Mosekilde 2005) In summary, some of the drugs excluded by all Medicare Part D plans can be justified by supporting literature cautioning use of the drugs in the elderly. During the final decision making process plans reserve the right to choose which drugs to include on their formularies. This review has shown clinical reasons may acco unt for some of the differences in coverage between Medicare Part D and FEHBP plans. There are few studies that compare the FEHBP and Medicare Part D with respect to coverage within therapeutic classes. But what can be found in the literature is informa tion on how well these programs individually cover specific therapeutic classes. For example, in a study by Bowman et al 2006, formulary drug coverage among Medicare Part D plans was examined within the therapeutic class anti cancer agents. The authors f ound that the majority of cancer drugs were covered by almost all Medicare Part D plans. Furthermore, Gellad et al 2007, examined Medicare Part D plan coverage of angiotensin receptor blockers which are categorized under the class cardiovascular agents. Results showed that all Medicare Part D prescription drug plan formularies included at least one angiotensin receptor blocker, while 35% of plans covered all seven listed in the study. Consistent with previous research this study also found that Medicare Part D plans covered a large number of anti cancer and cardiovascular drugs. Moreover, a recent JAMA article reported results of a study on Medicare Part D plan coverage of drugs in the classes gastrointestinal agents, cardiovascular agents, respiratory ag ents, antidepressants, blood glucose regulators, and analgesics (Tseng et

PAGE 144

144 al 2007). Results revealed that the greatest coverage was found for cardiovascular agents (85 90% of plans). On the other hand, the authors concluded that less than half of the dru gs examined (34 out of 75) were widely covered by Medicare Part D plans. Broad coverage of formulary drugs has also been shown within therapeutic classes in the FEHBP. For example, a 2003 GAO report found that FEHBP enrollees generally had unrestricte d access to prescription drugs. Furthermore, formularies were not considered to be overly restrictive, which was determined by the finding that they included drugs in most major therapeutic categories (Dicken et al 2003) With respect to coverage of brand vs. generic drugs, r esults of the independent samples t test revealed that the FEHBP plans covered about 98% of generic drugs (among generic drugs only) versus about 90% for Medicare Part D plans. Additionally, FEHBP plans covered on average 93% brand na me drugs (among brand name drugs only) as compared to Medicare Part D plans which covered on averag e 82% brand name drugs (p<.05). On the other hand, further examination using the t test comparison showed that generic and brand name drug coverage (i.e. ge neric/brand coverage out of total drug coverage) was about the same for Medicare Part D and the FEHBP. In contrast, a study by the Lewin Group 2007 revealed Medicare Part D plans covered more of the 132 brand name drugs (128 or 97%) compared to the FEHBP (125 or 95%) although no statistical tests were performed and thus one is unable to determine if there was a statistically significant difference between the two groups Another study by Tseng 2007 showed that Medicare Part D plans covered 90% of gene ric drugs This discrepancy in findings in the literature and the results of this study

PAGE 145

145 may be due to the cross sectional nature of the study design. More accurate measures of generic and brand name drug coverage may be shown in longitudinal studies (Jun g 2010). The final set of analyses to address formulary coverage differences between the two programs used negative binomial regression analysis. In the overall regression model which included all therapeutic classes results indicated that after controlling for premium, tier, enrollment, copay, coinsurance and therapeutic class, Medicare Part D plans provided greater drug coverage while the interaction of type of plan and therapeutic class showed that the FEHBP plans provided broader coverage; whe n both similar number of drugs being covered by Medicare Part D and FEHBP. To shed further light on this finding separate regression analyses were conducted with the de pendent variable defined as number of drugs in a given therapeutic class. The only two classes that showed a statistically significant difference with respect to type of plan were the classes anxiolytics and respiratory agents, both of which revealed the FEHBP provided broader drug coverage. Thus the results in the separate regressions by therapeutic class are consistent with the results from the overall regression in general, no difference in formulary coverage between the two plans A 2010 GAO study o f specialty drugs for FEHB P /Medicare Part D beneficiaries examined reasons why some drugs were included on the formulary and why others were not included. Results revealed that Medicare Part D plans considered limited ability to negotiate price concession s with manufacturers, low utilization for some drugs, and CMS USP guided formulary requirements as barriers to inclusion of drugs on

PAGE 146

146 formulary (GAO 2010). Plan sponsors noted that they are limited in the leverage they have when negotiating prices for drug s. Normally, when negotiating a drug price, plan therapeutic alternative. However, many specialty drugs belong to one of the six classes of clinical concern for which the U SP guidelines state Part D plan sponsors must include all or substantially all drugs on their formularies. Therefore, that normal negotiating leverage is removed (GAO 2010). As mentioned previously in the conce ptual framework, regulation could possibly a ffect the inclusion of drugs on a f ormulary. As for other control variables premium did not have a statistically significant effect on the number of drugs Review of the literature revealed that premiums are heavily dependent on the degree of cost shar ing with deductibles hav ing the greatest impact on premiums. Perhaps including deductibles in the regression analyses would have resulted in significant effects for the variable premium (Gorman, Gorman, and Newell 2010). A recent examination of New Yor k health plans showed that changes in deductibles and cost sharing resulted in premium changes of as much as 50% (Gorman, Gorman, and Newell 2010). Other studies have shown that while copay and coinsurance are used to deter enrollees from seeking services premiums do not directly affect the number of services utilized. Instead, premiums affect the amount and type of services utilized (Brook et al 1984; Manning et al 1988). Future research which considers deductibles may yield more significant premium re lated results. The regression anal ysis did fin d a positive association between the number of drugs per therapeutic class and copay for the classes ADHD agents and anti cancer agents. A n increase in copay result ed in an increase in the number of drugs in these

PAGE 147

147 classes. Additionally, coinsurance was a significant predictor of the number of drugs per therapeutic class for the classes: ADHD agents, anti cancer agents and respiratory tract agents. T he number of ADHD and anti cancer drugs increased as coinsur ance increased while the number of respiratory drugs decreased as coinsurance increased. This study has shown the importance of taking individual therapeutic classes into consideration in the interpretation of findings on cost sharing. A study by Avaler e Health that examined Medicare Part D plan cost sharing for cancer drugs noted that most Medicare plans place cancer drugs on higher tiers and that the coinsurance maximum is 33% (Murphy et al 2008). Earlier results of the regression analysis revealed th at a n increase in copay resulted in an increase in the number of anti cancer agents. Therefore, both enrollees and plans can benefit from this type of benefit structure. Beneficiaries pay more through higher copays and coinsurance, yet they receive more drugs as a result. Similarly, plans provide more drugs, yet beneficiaries use less (Huskamp 2003). A more extensive discussion on copay and coinsurance can be found later in the section labeled cost sharing. For the next independent variable, tier, a si gnificant difference was found for the classes ADHD agents and anti cancer agents. For both classes, the number of drugs per therapeutic class was greater for tier 2 brand name drugs as compared to tier 1 generic drugs. It is interesting that there were more brand name drugs listed on the formularies vs. generic drugs considering that generic drug promotion is often used as a cost containment measure (Tseng 2007). Lastly, in the overall model e nrollment was the single strongest predictor of the number o f drugs per therapeutic class. As enrollment increased, the number of drugs

PAGE 148

148 per therapeutic class increased. Specifically, an increase was shown in the classes: anti cancer agents, blood products/modifiers/volume expanders, cardiovascular agents, gastroi ntestinal agents, and respiratory tract agents. Cost Sharing Plans utilize cost sharing as a way to control costs. The goal is to decrease utilization by promoting the use of more cost effective therapies and to discourage use of unnecessary services ( O ffice o f Technology Assessment 1993) Out of pocket costs among health plan beneficiaries have been increasing in recent years. In relation to prescription drug coverage, beneficiaries may incur out of pocket costs in the form of fees that exceed the amo unt of reimbursement allowed by the health plan or payment for drugs received during the waiting period of eligibility for coverage, experimental treatments, or drugs purchased that are not listed on plan formularies (Office of Technology Assessment 1993). The most common out of pocket costs among beneficiaries are deductible s copay s (a fixed amount of money paid by a beneficiary when receiving covered services) or coinsurance (a percentage of covered procedures). This research focused on copay and coins urance. Out of pocket costs vary depending on whether a health plan charges a copay or coinsurance. The second aim of this research was to address the question of how Medicare Part D stand alone plans compare to Federal Employees Health Benefits plans w ith respect to cost sharing on prescription drugs Results of this research were consistent with the literature in that there wa s wide variation in copay and coinsurance depending on the tier. Cost sharing usually increase d by tier from lowest to highest starting with generic then brand then non formulary drugs.

PAGE 149

149 Furthermore, results revealed that both Medicare Part D plans and the FEHBP plans utilize fixed dollar copayments more often than coinsurance for tier 1 generic drugs. But for tier 2 brand name drugs t he FEHBP plans utilize d coinsurance more often. Studies have shown that coinsurance is more effective in controlling cost for brand name drugs. Findings from the independent samples t test revealed that for the Medicare Part D plans that utilize d copayments, the mean copayment for tier 1 generic drugs was $4.53 as compared to the FEHBP plans mean copayment of $7.67, a difference of $3.14 (p<.05). Furthermore, there was a small, but significant difference related to coinsurance. For tier 1 gener ic drugs, Medicare Part D plans coinsurance mean rates were 17% as compared to the FEHBP plan mean rates of 20% (p<.01). Additionally, for tier 2 brand name drugs, mean rates were 26% as compared to the FEHBP plan mean rates of 34% (p=.06). Considering that there were statistically significant differences found, the question nsurance rate From a consumer perspective lower cost sharing is better. Furthermore, studies have shown that when copay increases, utilization decreases and negative outcomes may result. Gaynor et al found that raising the amount of cost sharing wa s not a very effective method for controlling cost (Gaynor, Li, and Vogt 2007). The authors commented that inc reased cost sharing resulted in decreased utilization of health services, but the resulting savings w ere offset by an increase in outpatient service use. On the other hand, a study by Gruber found that increasing coinsurance rates result ed in decreased he alth care utilization without a negative effect on h ealth outcomes

PAGE 150

150 (Gruber 2006). Other studies clearly show that as cost sharing increase s prescription drug utilization decrease s (Joyce et al 2002; Goldman et al 2004; Huskamp et al 2003; Soumerai, Ross D egnan, and Gortmaker 1987; Harris, Stergachis, and Ried 1990; Johnson et al 1997; Tamblyn et al 2001; Motheral and Fairman 2001) Studies also have show n that the effect of cost sharing is greatly impacted by income. For example, the Kaiser Commission c onducted a study on Medicaid and the uninsured to examine the effects of increasing cost sharing to curtail costs (Ku 2003). This study was interesting considering the fact that concern exists as to whether those with low incomes will be negatively affect ed by increased cost sharing. The Commission concluded that as copayments increased utilization of prescription drugs decreased among low income Medicaid beneficiaries, and there was concern about resulting negative health outcomes. With respect to employer based health insurance beneficiaries, a study of the New York health insurance market found that typical generic drugs ha d a $10 copay, while brand name drugs carried a $30 copay (Gorman, Gorman, and Newell 2010). These numbers were a bit high er, but comparable to findings in this research of $4 $8 for generic drugs and $35 for brand name drugs. Gorman et al also reported that within the New York health care market a cost sharing benefit of $10 for generic drugs and $50 for brand name drugs re sulted in a savings of 12% (Gorman, Gorman, and Newell 2010). This is interesting considering that these amounts are larger than the average copay amounts revealed in this research. Correspondingly, a Kaiser Family Foundation report yielded findings of a verage employer sponsored copays of $11 per prescription in 2006 (Kaiser Family Foundation and the Health Research and Educational Trust 2006).

PAGE 151

151 In terms of coinsurance, the most common employer based coinsur ance rate is 20 percent. This research has sho wn that there is no one size fits all in the determination of appropriate cost sharing levels. The goal for plan providers is to encourage beneficiaries to purchase less expensive yet effective drugs. Since different approaches to financing and health ca re delivery exist, it is a challenge to compare cost sharing coverage options. Some have suggested that standardization may be a solution for easier comparison of plans. For example, standardizing out of pocket maximums may be a good start. There are le ssons to be glea n ed here for both Medicare Part D a nd the FEHBP. Limitations Several limitations should be considered in interpreting the findings of this research. First, data was not available on the demographics of enrollees within each plan. Facto rs such as age, income sex, race, and employment status may affect results. Age was not controlled for in this study. The FEHBP population likely consist s of younger enrollees as compared to the Medicare population and those age 65 and older have been f ound to enroll in different plans as compared to younger enrollees (Francis 2009). However, younger FEHBP enrollees are found in mostly HMO plans, while many of those age 65 and older enroll in PPO plans (e.g. the Blue Cross Standard Plan) (Francis 2009). Demographics may also affect cost sharing witho ut of pocket costs vary ing with income, type of service, or patient characteristics (Brook et al 1984; Manning et al 1988). These factors also may affect administrative costs and h ealth costs tend to ris e

PAGE 152

152 with age. In particular, e nrollee income may have an impact on prescription drug coverage. For example, recent changes in Medicare Part D charge those with higher incomes a higher premium. Currently, the FEHBP charges all enrollees the same dollar amo unt regardless of income. Therefore, income is likely to be an important predictor in the difference between the two programs. Although data was not available from both programs on sex, race, and employment status one can speculate that these factors may also explain many of the differences in the formularies. Secondly, caution should be used in the interpretation of results due to the cross sectional nature of the data. Plans are subject to change over time. The data used in this resea rch represents the year 2009 Therefore, this data set might not capture the full impact of drug coverage today although it is unlikely that prescribing trends have changed dramatically in a one year period Thirdly, not all Medicare Part D and FEHBP plans were include d in the study. Analyses included the top 63 70% of prescription drug plans among the FEHBP and Medicare Part D programs. However, previous studies only compare d a maximum of 3 Medicare Part D/ FEHBP prescription drug plans (Lewin 2007 ; Yamamoto 2008; Bow man 2006). Lastly, caution in the interpretation of these results stems from the fact that these two programs may be overlapping in their e ffect on one another. For example, Americans who work pay for Medicare as wages ar e deducted from each paycheck. T herefore, federal employees, or FEHBP beneficiaries contribute to the cost of Medicare through payroll deductions. On the other hand, when an enrollee has both FEHBP and Medicare coverage, Medicare is the primary payer. Over 1 million of the FEHBP enroll ees are Medicare Part D enrollees meaning Medicare is contributing to the

PAGE 153

153 finance of the FEHBP. Specifically, a total of one third of the FEHBP are age 65 and older and make up about one half of the FEHBP spending (Francis 2009). The possibility exists t hat changes in Medicare Part D could result in corresponding changes in the FEHBP. Conclusion In conclusion, there are differences in prescription drug coverage and cost sharing among plans within and between the Medicare Part D and FEHB program. The nature of these differences depends on the method of analysis, perspective, and therapeutic class examined. Analysis using independent t tests revealed that the FEHBP plans provided broader prescription drug coverage as compared to Medicare Part D plans. However, further regression analysis taking into account the effect of various factors affecting coverage showed that there wa s no difference between the two programs with respect to drug coverage. Consumers, policymakers, and health insurance providers may be interested in reviewing all results to determine appropriate conclusions in the examination of plans. In regard to c ost sharing benefit structures among both p rograms we re wide and varied. Average copays for generic drugs were $4.53 for Medicare Part D plans and $7.67 for the FEHBP plans. Additionally, generic drug coinsurance rates were 17% for Medicare Part D plans and 20% for the FEHBP plans. Both Medicare Part D and the FEHBP provide prescription drug coverage in a variety of the top therap eutic classes of drugs dispensed and sold in the United States. Lessons in formulary decision making and cost sharing are derived through comparison of the two programs.

PAGE 154

154 Policy Implications Health care reform is one of the natio time Among the topic s of interest the scope and depth of health insurance benefits are of particular importance. Specifically, the provision of prescription drugs can affect access and quality among beneficiaries. Considering the projected depletion of funds to support Medicare in the coming years Medicare policymakers can use this research to make more informed decisions related to reform. Similarly, the FEHB program will need to make adjustments in the coming years to keep up with the growing eld erly population. Comparisons to Medicare and other employer based health insurance programs will help to identify best practices among plans in the provision of the top drugs dispensed and sold in the U.S. Policymakers can take a closer look at the benef it structure of those plans providing the greatest number of top drugs in addition to plans with the lowest and highest cost sharing levels. Clinicians can also benefit from results of this research. A 2007 study by Tseng and colleagues addressed clinic ian concerns about variations in Medicare plan formularies (Tseng 2007) and concluded that future research that identified within a class, which drugs were widely covered would greatly aid clinicians in the reduction of their administrative burden. Furth ermore, the authors stated that such information would lower the risk that Medicare beneficiaries are prescribed non formulary drugs. Moreover, special attention would be given to those classes with no widely covered drugs, in this case the anxiolytics. Tseng and colleagues suggested that this information be provided to clinicians in the form of a Web site, personal digital

PAGE 155

155 assistant based tool, or e prescribing software to aid in decision making in the selection of medications (Tseng 2007). In summary, t he importance of learning from the lessons of the FEHBP and the Medicare Part D programs should be emphasized. Ignoring these lessons can result in wasteful spending throughout the health care system. It has been stated that Medicare is financed primaril y by persons who do not directly benefit from it until in their future. This becomes important when we consider the ratio of our growing elderly population to the working population. Comparing the coverage and cost sharing between these two programs invo lves many issues that warrant further exploration. Future Research This research has opened the door to many areas for future research. Important areas for future research are the role of demographic factors in prescription drug coverage, the market behav ior of prescription drug plans, and health outcomes associated with drug coverage. Specifically, future comparison of the two programs should include an analysis of the relationship of drug coverage/cost sharing and demographic factors such as income and g eographic location. Studies show that the cost of medical care varies by area and physician practice patterns vary by area. This may result in drug coverage that varies by area. Furthermore, many studies suggest that lower income people may be affected negatively by cost sharing ( ). In regard to market behavior, it would be interesting to explore the role of market share in drug coverage. In short, prescription drug plans hire phar macy benefit managers to negotiate price discounts for drugs with manufactures. The negotiations

PAGE 156

156 involve the receipt of rebates by pharmacy benefit managers on the drugs selected for the formulary. Manufacturers give rebates based on the inclusion of the ir drugs on a the pharmacy benefit manager result in savings that are passed on to the prescription drug plans, which are then passed on to the beneficiaries (Henry J. Kaiser Family Foundation 2005). Exploring how market share affects access and quality of drugs provided to beneficiaries would be beneficial to consumers. Lastly, while this research provided an examination of drug coverage and cost sharing, it did not explore whether broader drug coverage led to better health outcomes or whether harm resulted from the provision of less drug coverage. Studies show that substantial cost sharing will lead to adverse outcomes for some patients resulting from forgone care (Khan, Kaestner, and Lin 2008). The question remains as to how prescription drug coverage and cost sharing affect health outcomes. In the future, research that examines the link between Medicare Part D and the FEHBP drug coverage and common health outcom es (e.g. blood pressure level, cholesterol level, glucose/hemoglobin A1C) may be beneficial to consumers and plans.

PAGE 157

157 LIST OF REFERENCES Agresti A, Finlay B. Statistical Methods for the Social Sciences. 3rd ed. New Jersey: Prentice Hall; 1997. AMCP (Aca demy of Managed Care Pharmacy). AMCP Guide to pharmaceutical payment methods, 2009 update. JMCP. 2009;15(6):S1 61. http://www.amcp.org/data/jmcp/1002.pdf Accessed August 7 2009. Ame rican Urological Association Education and Research. Non Surgical Management of Erectile Dysfunction (ED) 2010. http://www.urologyhealth.org/adult/index.cfm?cat=11&topic=174 Accessed August 5, 2010. Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds. Washington, D.C. 2007. Ce nters for Medicare and Medicaid Services Web Site. www.cms.hhs.gov/ReportsTrustFunds/ Accessed May 7, 2009. SCHIP: recent state experiences. Kaiser Commission on Medicaid and the Uninsured. 2005. www.kff.org/ medicaid /7322.cfm Accessed August 10 2010. ASHP (American Society of Health System Pharmacists). Statement for the record submitted to the Senate Special Committee on Aging h earing on meeting the challenges of Medicare Part D implementation. 2006. www.ashp.org/emplibrary/Statementforthe20%RecordSpecialCommitteeAging.pdf Accessed D ecember 19, 2008. Atherly A, Florence C, Thorpe KE. Market structure and plan characteristics in the Federal Employees Health Benefits Program. Proceedings of the Annual meeting of the Economics of Population Health: Inaugural Conference of the American Society of Health Economists. 2006. http://www.allacademic.com/meta/p93446_index.html Accessed April 6, 2009. Bambauer KZ, Sabin JE, Soumerai SB. The exclusion of benzodiazepine coverage in Medicare : simple steps for avoiding a public health crisis Psychiatr Serv. 2005;56:1143 1146. Basu A, Yin W, Alexander GC. The impact of Medicare Part D on Medicare Medicaid dual eligible beneficiaries' prescription utilization and expenditures. NBER Working Paper Series. 2008. Vol. w14413. http://ssrn.com/abstract=1288412 Accessed April 6, 2009.

PAGE 158

158 Berndt ER. Uniqu e issues raised by drug benefit design. Health Aff. 2004;23:103 107. Blue Cross Blue Shield Formulary: Basic and Standard Version. Caremark Web Site. http://www2.caremark.com/fepstandardretail/ .com Accessed January 14, 2010. Blue Cross Blue Shield Servic e Benefit Plan Brochure. 2008. Federal Employees Program Blue Cross Blue Shield W eb Site. http://www.fepblue.org Accessed February 11, 2009. Bowman J, Rousseau A, Silk D, Harrison C. Access to cancer drugs In Medicare Part D: formulary placement and bene ficiary cost sharing in 2006 Health Aff. 2006 ;25: 1240 1248. Briesacher BA, Soumerai SB, Field TS et al. Medicare Part D's exclusion of benzodiazepines and fracture risk in nursin g homes Arch Intern Med. 2010;170:693 698. Brook RH, Ware JE, Rogers WH et al. The effect of coinsurance on the health of adults. Results from the RAND Health Insurance Experiment. Santa Mo nica, CA: RAND Corporation, 1984. Report No.: R 3055 HHS. ISBN 0 8330 0614 2 Butler SM. A new policy framework for health care markets. Health Aff (Millwood). 2004; 23:22 24. Butler SM, Moffit RE. The FEHBP as a model for a new Medicare program. Health Aff. 1995;14(4):47 61. Cain HP. Moving Medicare to the FEHBP model, or how to make an elephant f ly. Health Aff 1999;18:25 39. Catlin A, Cowan C, Hartman M et al. National health spending in 2006: A year of change for prescription drugs Health Aff. 2008;27(1): 14 29. Cannon MF. Health care needs a dose of competition. The Cato Institute. 2005. http://www.cato.org/pub_display.php?pub_id=5070 Accessed June 8, 2009. Centers for Disease Control and Pr evention and Merck Company Foundation. The state of aging and health in America 2007 Whitehouse Station, NJ. 2007. www.cdc.gov/ aging /pdf/saha_ 2007 .pdf Accessed June 8, 2009. Centers for Medica re and Medicaid Services. Office of the Actuary, National Health Statistics Group, 2006 National Health Care Expenditures Data 2008. http://www.cms.gov/nationalhealthexpenddata/01_ove rview.asp ? Accessed June 9, 2009. http://www.cbo.gov/budget/factsheets/2007b /medicare.pdf Accessed August 1, 2007

PAGE 159

159 Cubanski J, Neuman P. Status report on Medicare Part D enrollment in 2006: analysis of plan specific market share and coverage. Health Aff. 2006;26:w1 w12. Daly R, Moran M. Dual eligibles faced serious health problems in shift to Part D Psychiatr News 2007;42(10):1 58. Daschle T, Greenberger SS, Lambrew JM. Critical What We Can Do About the Health Care Crisis. 1st ed. New York (NY): Thomas Dunne Books; 2008. Dav is K, Schoen C. Creating consensus on coverage choices. Health Aff Web Exclusive. 2003; w3:199 211. http://content.healthaffairs.org/cgi/reprint/hlthaff.w3.199v1 ? Accessed Augus t 10, 2010. Dicken J, Agarwal R, Dirosa R, Kirksey B, Rivera Lowitt C, White A Federal Plans, Enrollees, and Pharmacies. US Government Accountability Office; 2003. GAO 03 196. Dicken J, Dirosa R, Federal Employees Health Benefits Program: Premiums Continue to Rise, but Rate of Growth Has Recently Slowed US Government Accoun tability Office; 2007. GAO 07 873T. Dicken J, Peterson K, Bogart G et al Retiree Health Benefits: Majority of Sponsors Continued to Offer Prescription Drug Coverage and Chose the Retoree Drug Subsidy US Government A ccountability Office; 2007. GAO 07 572. Dicken J, Dirosa R, Lee D, Shuman S, Walker T, Weber M. Federal Emp loyees Health Benefits Program: Enrollee Cost Sharing for Selected Specialty Prescription Drugs. US Government Accountability Office; 2009. GAO 09 517R Dorn S, Meyer JA. Health coverage for laid off workers: searching for common ground. Economic and Soc ial Research Institute. 2002; No 3:1 4. http://www.esresearch.org/newsletter/may/alert3.htm Accessed August 9, 2010. Drugs.com Omnipaque. 2010. Drugs.com Web Site http://www.drugs.com/pro /omnipaque injection.html Accessed August 5, 2010. Drugs.com Visipaque. 2010. Drugs.com Web Site http://www.drugs.com/pro/visipaque.html Accessed August 5, 2010. Drugs.com Mirena. 2010. Drugs.com Web Site http://www.drugs.com/cdi/mirena iud.html Acces sed August 5, 2010. Elam L, Murawski MM, Childs S, Vanable JW. Patient safety forum: do state Medicaid preferred drug lists affect patient safety? Psychiatr Serv. 2005;56:1012 1016.

PAGE 160

160 Enthoven AC. Effective management of competition in the FEHBP. Health Aff (Millwood) 1989;8(3):33 50. Enthoven AC. The fortune 500 model for healthcare: is now the time to change? J Health Politics, Pol Law 2003;27(1):37 48. Epstein AJ Rathore SS Alexander GC Ketcham JD Primary care physicians' views of Medicare Part D. Am J Manag Care. 2008;14(11 Suppl):SP5 13. The Eye Digest Lucentis (ranibizumab) FDA approved new drug for Wet Macular Degeneration. University of Illinois Eye & Ear Infirmary Chicago, IL 60612. 2009. http://www.agingeye.net/mainnews/lucentis.php# Accessed August 3, 2010. Families USA. Enough to make you sick: prescription drug prices for the elderly. 2001. Publication No. 01 103. http://www.familiesusa.org/assets/pdfs/Enough to Make You Sick.pdf Accessed March 11, 2009. F amilies USA. No bargain: Medicare drug plans deliver high prices. 2007. Publication No. 07 101. http://www.familiesusa.org/assets/pdfs/no bargain medicare drug.pdf Access ed December 11, 2009. FEHBP Home Page. Insurance programs: types of plans. 2008. US Office of Personnel Management Web Site. http://www.opm.gov/insure/ health/planinfo/types.asp Accessed June 10, 2009. Francis W. pp269 303 in Helms RB, Health Policy Reform: Competition and Controls 1st ed. Washington, D.C.: AEI Press; 1993. Fr ancis W. The FEHBP as a model for Medicare reform: separating fact from fiction. 2003. The Heritage Foundation Web Site. http://www.heritage.org/Research/Reports/2003/08/The FEHBP as a Model for Medicare Reform Separating Fact from Fiction Accessed January 9 2009. Francis W. Using the Federal Employees' model: nine tests for rational Medicare reform. 2003. The Heritage Foundation Web Site. http://www.heritage.org/Research/Reports/2003/08/Using th e Federal Employees Model Nine Tests for Rational Medicare Reform Accessed January 9 2009. Francis W. Putting Medicare Consumers in Charge: Lessons from the FEHBP 1st ed. Washington, D.C.: AEI Press; 2009 Frank RG, Newhouse JP. Should drug prices be negotiated under Part D of Medicare? and if so, how? Health Aff. 2008;27(1):33 43.

PAGE 161

161 Friedman M, Friedman R. Free to Choose: A Personal Statement. Florida: Harcourt Inc. 1980. Fuchs BC. Increasing health insurance coverage through an extended Federal Emplo yees Health Benefits Program Inquiry 2001;38(2):177 192. GAO ( Government Accountability Office). Medicare P art D : Spending, Beneficiary Cost Sharing, and Cost Containment Efforts for High Cost Drugs Eligible for a Specialty Tier. 2010. GAO 10 242. Gaynor M, Li J, Vogt WB. Substitution, spending offsets, prescription drug benefit design. Forum for Health Econo mics & Policy. 2007;10(2):Art. 4. GEHA Benefit Plan Formulary. 2009. GEHA Web Site. http://www.geha.com/prescriptions/preferred_prescriptions.html Accessed J anuary 13 2010. Gellad WF, Huskamp HA, Phillips KA, Haas JS. Angiotensin receptor blockers on the formularies of Medicare drug plans. J Gen Intern Med. 2007; 22 :1172 1175. Goldman DP, Joyce GF, Escarce JJ et al. Pharmacy benefits and the use of drugs by the chronically ill. JAMA 2004;291:2344 2350. Goldman DP, Joyce GF. Medicare Part D: A successful start with room for improvement, JAMA 2008;299:1954 1955. 2010. Uni ted Hospital Fund Web Site. h ttp://www.uhfnyc.org/assets/795 Accessed August 10, 2010. Grace Martin K. StatNews #40. Interpreting interactions in regression. 2000. T he Cornell Statistical Consulting Unit Web Site. http://www.cscu.cornell .edu/news/statn ews/stnews40 .pdf Accessed August 10, 2010. Grace Martin, K. Linear regression analysis testing and dropping interaction terms in regression and ANOVA 2009. http://ezinearticles.com/?Linear Regression Analysis Testing and Dropping Interaction Terms in Regression and ANOVA&id=2074991 Accessed August 10, 2010. Gruber J. The role of consumer copayments for health care: lessons from the RAND health insurance experiment and beyond. 2006. Henry J. Kaiser Family Foundation Web Site. http://www.kff.org/insuranc e/upload/7566.pdf Accessed August 10 2010 Gu Q, Dillon CF, Burt VL. Prescription drug use continues to increase: U.S. prescription drug data for 2007 2008. NCH S data brief, N o 42. Hyattsville, MD: National Center for Health Statistics. 2010.

PAGE 162

162 Hargrav e E, Hoadley J, Cubanski J, Neuman T. Medicare prescription drug plans in 2009 and key changes since 2006: summary of findings. 2009. Henry J. Kaiser Family Foundation Web Site. ht tp://www.kff.org/medicare/upload/7917.pdf Accessed May 9 2010. Harris BL, Stergachis A, Ried LD. The effect of drug copayments on utilization and cost of pharmaceuticals in a health maintenance organization. Medical Care 1990;28: 907 917. Henry J. Ka iser Family Foundation. Follow the pill: understanding the US commercial pharmaceutical supply chain. The Health Strategies Consultancy LLC. 2005. Henry J. Kaiser Family Foundation Web Site. http://www.kff.org/rxdrugs/upload/Follow The Pill Understanding the U S Commercial Pharmaceutical Supply Chain Report.pdf Accessed September 3, 2009. Henry J. Kaiser Family Foundatio n. Medicare: Medicare prescription drug coverage enrollment update. 2006. Henry J. Kaiser Family Foundation Web Site. http://www.sen.ca.gov/health/KFF_MEDICARE_ENROLLMENT_UPDAT E.pdf Accessed December 5, 2009. Henry J. Kaiser Family Foundation. 2007. Fact sheet: the Medicare prescription drug benefit. 2007. Henry J. Kaiser Family Foundation Web Site. http//www.ktf. org/medicare/upload/7044 06.pdf Accessed March 7, 2008. Henr y J. Kaiser Family Foundation. 2007. Medicare Chartpack: Overview of Medicare Part D Organizations, Plans and Benefits By Enrollment in 2006 and 2007. Henry J. Kaiser Family Foundation. Prescription drug trends. 2008. Henry J. Kaiser Family Foundation Web Site. http://www.kff.org/rxdrugs/upload/3057_07.pdf Accessed August 11 2010. Heritage Foundation. Patients' right to choose a better plan should drive Medicare reform. 2003. Heritage Foundation Web Site. http://new.heritage.org/Press/NewsReleases/nr022703.cfm Accessed August 1 2010. Hoadley J. Medicare Part D: Simplifying the program and improving the value of Information for beneficiaries. 2008. The Common wealth Fund Web Site. http://www.commonwealthfund.org/usr_doc/Hoadley_MedicarePartD_1118_ib.pdf Accessed January 11, 2009. Hoadley J, Hargrave E, Cubanski J, Neuman T. An in depth examination of formularies and other features of Medicare drug plans. 2006. Henry J. Kaiser Family Foundation Web Site. http://www.kff.org/medicare/upload/7489.pd f Accessed January 5 2009.

PAGE 163

163 Hoadley J, Hargrave E, Cubanski J, Neuman T. The Medicare Part D Coverage Gap: Costs and Consequences in 2007. 2008. Henry J. Kaiser Family Foundation publication 7811. Hoadley J, Hargrave E, Merrell K, Cubanski J and Neuman T Medicare Part D 2008 Data Spotlight: Formularies 2008. Henry J. Kaiser Family Foundation Web Site. http://www.k ff.org/medicare/upload/7734.pdf Accessed October 5 201 0. Hoff J. Medicare Private Contracting: Paternalism or Autonomy? American Enterprise Institute. 1998. Hsu J, Price M, Huang J et al. Unintended consequences of caps on Medicare drug benefits. N Engl J Med. 2006;354(22):2349 2359. Health Insurance Premiums and Cost Sharing: Findings from the Research on Low Income Populations. Kaiser Commission on Medicaid and the Uninsured (March 2003). Huskamp HA, Deverka PA, Epstein AM, Epstein RS, McGuigan KA, Frank RG. The effect of incentiv e based formularies on prescription drug use and spending. N Engl J Med. 2 003;349(23):2224 2232. Huskamp HA, Deverka PA, Landrum MB, Epstein RS, McGuigan KA. The effect of three tier formulary adoption on medication continuation and spending among elderly retirees. Health Serv Res. 2007; 42 (5) :1926 1942. Huskamp HA, Stevenson DG, Donohue JM, Newhouse JP, Keating NL. Coverage and prior authorization of psychotropic drugs under Medicare Part D. Psychiatr Serv 2007;58(3):308 310. http://ps.psychiatryonline.org/cgi/content/full/58/3/308 Accessed August 11, 2010. IMS Health Data. 2009. Pharmacy Times Web Site. http://www.pharmacytimes.com/issue/pharmacy/2009/2009 05/RxFocusTop200Drugs 0509 and IMS Web Site http://www.imshealth.com/portal/site/imshealth/menuitem.a4 6c6d4df3db4b3d88f611019 418c22a/?vgnextoid=85f4a56216a10210VgnVCM100000ed152ca2RCRD&cpsextcurrc hannel=1 Accessed January 5, 2010. Johnson RE, Goodman MJ, Hornbrook MC, Eldredge MB. The impact of increasing patient prescription drug cost sharing on thera peutic classes of drugs received and on the health status of elderly HMO members. Health Serv Res 1997;32(1):103 122. Joyce GF, Escarce JJ, Solomon MD, Goldman DP. Employer drug benefit plans and spending on prescription drugs. JAMA. 2002; 288(14):1733 17 39.

PAGE 164

164 Jung C. Impact of Medicare Part D on Generic Drug Utilization in the Long term Care Facilities. The H. John Heinz III College Carnegie Mellon University, Pittsburgh, PA 15213 Committee: Rema Padman (Chair), Shamena Anwar, Martin Gaynor. 2010 http://www.heinz.cmu.edu/research/377full.pdf Kaiser Family Foundation. Medicare Part D Fact Sh eet: Kaiser Family Foundation. August 2008. Henry J. Kaiser Family Foundation Web Site. www.kff.org. Access ed January 5, 2010. Kaiser Family Foundation. The Medicare Prescription Drug Benefit Fact Sheet November 2009. Publication 7044 10 Henry J. Kaiser Family Foundation Web Site. www.k ff.org Accessed October 5, 2010. Kaiser Family Foundation and Health Research and Educational Trust Survey of employer sponsored health b enefits, 2002 2006. 2006. Henry J. Kaiser Family Foundation Web Site. http://www.kff.org/insurance/75 27/upload/7527.pdf Accessed August 12 2010. Ketcham JD, Ngai JK. How similar are states' Medicaid preferred drug lists? The American Journal of Managed Care. 2008;14(11):SP46 SP52. http://www.ajmc.com/issue/managed care/2008/2008 11 vol14 n11SP/Nov08 3706pSP46 SP52 Accessed June 5, 2010. Khan N, Kaestner R, Lin SJ. Effect of prescription drug coverage on health of the elderly. HSR: Health Services Res earch 43:5, Part I 2008 King KM, Katz R, Achman L, Poppe S. Medicare Part D: Enrolling New Dual Eligible Beneficiaries in Prescription Drug Plans. 2007. GAO 07 824T. Government Accountability Office Web Site. http://www.gao.gov/new.items/d07824t.pdf Accessed December 3 2008. Kipp RA, Ko C. Medicare Part D administrator survey: potential cost impacts resulting from CMS guidance on special protections for six protected drug clas sifications and section 176 of the Medicare Improvements for Patients and Provides Act of 2008. 2008. AMCP Web Site. http://www.amcp.org/cfr/waSys/f.cfc?method=getLinkAsset& id=4979E963 Accessed August 5, 2009. Ku L. Charging the poor more for health care: cost sharing in Medicaid 2003. Washington: Center on Budget and Policy Priorities Web Site. http://www.cbp p.org/archiveSite/5 7 03health.pdf Accessed April 5, 2010. Ku L, Wachino V. The Effect of Increased Cost Sharing in Medicaid: A Summary of

PAGE 165

16 5 Landsman PB, Teutsch SM. Pharmacy benef its and the use of drugs by the chronically ill. JAMA. 2004; 291(19 ):2344 2350. Lewin Group. 2007. Comparison of VA National Formulary and Formularies of the Highest Enrollment Plans in Medicare Part D and the Federal Employee Health Benefit Program. Lope rt R, Rosenbaum S. What is fair? choice, fairness, and transparency in access to prescription medicines in the United States and Australia. The Journal of Law, Medicine & Ethics 2007;35(4 ):643 656. Lundy K Janes S. Community Health Nursing: Caring for the P Pp. 159 162. 2009. http://books.google.com/books?id=fj5dInclgw0C&printsec=frontcover&source=gbs_navli nks_s Accessed June 3, 2009. Mahay, H. Oral Chemotherapy: Patient Advantages and Challenges. 2009. Pharmacy Times Web Site http://www.pharmacytimes.com/issue/pharmacy/2009/August2 009/CounselingChemoth erapy 0809 Accessed August 5, 2010. Manning WG, Newhouse JP, Duan N et al. Health Insurance and the Demand for Medi cal Care. Evidence from a Randomized Experiment. Santa Monica, CA: RAND Corporation; 1988. Report R 3476 HHS. McBride T. et al. An Analysis of Availability of Medicare+Choice, Commercial HMO, and FEHBP Plans in Rural Areas: Implications for Medicare Ref orm. 2003. Rural Policy Research Institute Rural Policy Web Site. www.rupri.org/ruralHealth/publications/PB2003 5.pdf Accessed December 5, 2008. Medicare Consumer Guide. 2008. Medicare Consumer Guide Web Site. http://www.medicareconsumerguide.com/ Accessed December 11, 2008. Medicare Payment Advisory Commission. 2002 Survey of physicians about the Medicare program. 2002. Medicare Payment Advisory Commission Web Site. www.medpac.gov Accessed December 5, 2008. Medicare Payment Advisory Commission. Report to the Congress: increasing the value of Medicare. 2006. Medicare Payment A dvisory Commission Web Site. http://www.medpac.gov/documents/Jun06_EntireReport.pdf Accessed March 5, 2008 Medicare Prescription Drug Benefit. Summary of t he USP Approach and Methodology to the Model Guidelines Version 3.0. 2008. USP Web Site. http://www.usp.org/pdf/EN/mmg/modelGuidelinesApproachMethodology .pdf Accessed April 7 2009

PAGE 166

166 Merck Manual of Geriatrics. Generalized anxiety disorder. 2010. Merck Web Site. http://www.merck.com/mkgr/mmg/sec4/ch34/ch34b.jsp Accessed A ugust 6 2010. Merlis M. The Federal Employees Health Benefits Program: Program Design, Recent Performance, and Implications for Medicare Reform. Menlo Park, Calif: The Henry J. Kaiser Family Foundation; 2003. Publication #6081. Moon M. Medicare, Mod ernization and FEHBP. 2002. The Urban Institute Web Site. http://www.urban.org/publications/900489.html Accessed May 7, 2009. market d can the common interests of the for profit pharmaceutical and managed care industries hold the regulators at bay? Health Aff. 2000;19(2):63 77. Mosekilde L. Vitamin D and the elderly. Clin Endocrinol. 2005;62(3):265 281. Motheral BR, Fairm an KA. (2001) Effect of a Three Tier Prescription Co Murphy L, Sadownik S, Ford C, Barton V, Schmidt C and Barber S. Avalere Health. 2008. Cost Sharing for Cancer Patients in Medicare. American Cancer Society Cancer Action Network. Mueller KJ, McBride TD, Andrews C, Fraser R, Xu L. The Federal Employees Health Benefits Program: A model for competition in rural America? The Journal of Rural Health. 2005;21(2):105 113. Murphy L, Sadownik S, Ford C, Barton V, Schmidt C, Barber S. Cost Sharing for Cancer Patients in Medicare, 2009. 2008. Avalere Health. American Cancer Society Web Site. http://action.acscan.org/ site/DocServer/Avalere_ACS_CAN_Part_D_study_final_version _with_Sarah_s_.pdf/482831466?docID=10441&verID=1 Accessed January 10, 2010. NALC Health Benefit Plan Formulary 2010 FIND 2009. http://www.nalc.org/depart/hbp/Benefits/BROCHURE/index.html Accessed April 2010 National Bipartisan Commission on the Future of Medicare. Building a better Medicare for today and tomorrow. 1999. Medicare Commission Web Site. http://medicare.commission.gov/medicare/bbmtt31599.html Accessed June 5, 2009. National Health Policy Forum. The Basics: Medicare Modernization Act of 2003: Overview of Part D. 2004. The George Washingto n University.Washington DC. National Survey of Pharmacists, findings on Medicare Part D. 2006. The Henry J. Kaiser Family Foundation Web Site. www.kff.org/kaiserpolls/7555.cfm Accessed December 19 2008.

PAGE 167

167 Nelson D. Medicare Part D and Temodar. 2010. eHow Web Site http://www.ehow.com/facts_6900733_medicare part temodar.html Accessed August 4, 2010. Nemore P. Medicare Part D: Issues for dual eligibles on the eve of implementation. 2005. Henry J. Kaiser Family Foundation Web Site. http://www.kff.org/medicaid/7119.cfm Accessed January 5, 2009. NIDDK Erectile D ysfunction. The National Institute of Diabetes and Digestive and Kidney Diseases National Institutes of Health 2010. http://kidney.niddk.nih.gov/kudiseases/pubs/impotence/index.htm Accessed August 16 2010. NIST/SEMATECH e Handbook of Statistical Methods 2010. NIST Web Site. http://www.itl.nist.gov/div898/handbook/eda/section3/eda35a.html Accessed June 5, 2010. Oberlander J. Is premium support the right medicine for Medicare? Health Aff. 2000; 19(5):84 99. Oberlander J. Through the look ing glass: the politics of the Medicare prescription drug, improvement, and Modernization Act. Journal of Health Politics, Policy and Law. 2007;32(2):187 219. Office of Personnel Management. OPM Announces Open Season for Health Benefits, Dental and Vision Insurance, and Flexible Spending Accounts 2008. OPM Web Site http://www.opm.gov/news/opm announces o pen season for health benefits dental and vision insurance and flexible spending accounts,1435.aspx Accessed January 5, 2009. Office of Technology Assessment, U.S. Congress. Benefit Design in Health Care Reform: Background Paper Patient Cost Sharing. 199 3. OTA BP H 112 http://www.fas.org/ota/reports/9310.pdf Accessed August 15, 2010. Office on Women's Health. Folic Acid FAQ. Department of Health and Human Services. 2010. http://www.womenshealth.gov/faq/folic acid.cfm Accessed August 15, 2010. Outterson K, Kesselheim AS. How Medicare could get better prices on prescription drugs. Health Aff. 2009;28(5):w832 w841. Palmisano DJ, Emmons DW, Wozniak GD. Expanding insurance coverage through tax credits, consumer choice, and market enhancements: the American Medical Association proposal for health insurance reform. JAMA. 2004 ; 291(18):2237 2242. Paul R. Paying dearly for free prescription d rugs. 2003. http://www.lewrockwell.com/paul/paul133.html Accessed April 12, 2009.

PAGE 168

168 Author. Payment on Pharmaceutical and Other Medical Use. Medical Care year;39(12):1293 1304. Peck B. Private Insurance Plans and Medicare: The Disappointing History. Public Pontillo DC, Lang AJ, Stein MB Management and treatment of anxiety disorders in the older patient. Clinical Geriatrics. 2002;10(10):38 49. Price Waterhouse Coopers. Medicare Part D: An Assessment of Plan Performance and Potential Savings. 2007. Pharmaceutical Care Management Association Web Site. http://pcmanet.org/assets/2008 03 25_Research_PwC%20Medicare%20Savings%20and%20Generics%20Report%20200 6%20(Jan%202007).pdf Access ed April 7, 2009. Public Law 110 275. 2006. Enacted by Congress, includes a section (176) on formulary requirements with respect to certain categories or classes of drugs Richmond JB, Fein R. The Health Care Mess: How We Got Into It and What It Will Tak e to Get Out, 307 pp, ed. Cambridge, Mass, Harvard University Press. 2005. Rosenberg JM. Overview of Medicare Part D prescription drug benefit: potential implications for patients with psychotic disorders. Amer J of Health System Pharm 2007;64.2:S18 S23 Ruddock AE. Federal employees health benefits program. I. History and future of the federal program -1964. Am J Public Health Nations Health. 1966;56(1):50 53. Rx List the Internet Drug Index Venofer. 2010. Rx Lis t Web Site http://www.rxlist.com/venofer drug.html Accessed August 5, 2010. Rx List the Internet Drug Index Plavix. 2010. Rx List Web Site http:/ /www.rxlist.com/plavix drug.html Accessed August 5, 2010. Rx List the Internet Drug Index Warfarin. 2010. Rx List Web Site http://www.rxlist.com/coumadin drug.html Accessed August 5, 2010. Rx List the Internet Drug Index Angiomax. 2010. Rx List Web Site http://www.rxlist.com/angiomax drug.html Accessed August 5, 2010. Rx List the Internet Drug Index Integrilin. 2010. Rx List Web Site http://www.rxlist.com/integrilin drug.html Accessed August 5, 2010. Senior Journal. Erectile dysfunction drugs being dropped by Medicare drug plans. 2006. Senior Journal Web Site http://seniorjournal.com/NEWS/MedicareDrugCards/6 10 18 ErectileDysfunctionDrugs.html Accessed August 16, 2010.

PAGE 169

169 Senior Journal. Poll says senior voters want Medicare drug plan choices, competition. 2004. Senior Journal Web Site. http://seniorjournal.com/NEWS/Medicare/4 11 11WhatSeniorsWant.html Accessed October 4, 2009. Simoni Wastila L, Shaffer T, Stuart B. A national comparison of prescription drug expenditures by Medicare beneficiaries living in the community and long term care facility settings. 2007. University of Maryland, Baltimore, Peter Lamy Center on Drug Therapy and Aging. U.S. Depart ment of Health and Human Services Web Site http://aspe.hhs.gov/daltcp/reports/2007/pdnatcom.html Accessed March 8, 2009. Smetana GW, Davis RB, Phillips RS. Factors that influence patie nt response to requests to change to a unified restrictive formulary, a prospective study. J Gen Intern Med. 2004; 19 (12) :1212 1219. Soni A. The top five therapeutic classes of outpatient prescription drugs ranked by total expense for adults age 18 and old er in the U.S. civilian noninstitutionalized population, 2006 Statistical Brief #232. 2009. Agency for Healthcare Research and Quality Web Site. http://www.meps.ah rq.gov/mepsweb/data_files/publications/st232/stat232.pdf Accessed March 11, 2009. Soumerai SB, Ross Degnan AD, Gortmaker S. Payment restrictions for prescription drugs under Medicaid: effects on therapy, cost and equity. N Engl J of Med. 1987;317:550 556 Soumerai SB, McLaughlin TJ, Ross Degnan D et al. Effects of a limit on Medicaid drug reimbursement benefits on the use of psychotrophic agents and acute mental health services by patients with schizophrenia. N Engl J Med. 1994; 331:650 5. StataCorp. 2 009. Stata Statistical Software: Release 11 College Station, TX: StataCorp LP. Steinwald B, Thorburn P, Garvey C, Long D, Tholandi M, Wedum E. Medicare Part D Low Income Subsidy : Assets and Income Are Both Important in Subsidy Denials, and Access to State and Manufacturer Drug Programs Is Uneven. 2008. GAO 08 824 http://www.gao.gov/htext/d08824.html Accessed D ecember 7 2009. Stevenson DG, Huskamp HA, Keating NL, Newhouse JP. Medicare Part D and nursing home residents. J Am Geriatr Soc. 2007;55:1115 1125. Stigler GJ. The Theory of Price. 3d ed. New York: Macmillan; 1966. Econlibrary Web Site. www.econlib.org Accessed April 2, 2009. Stuart B, Briesacher BA, Shea DG, Cooper B, Baysac FS, Limcangco MR. Riding the rollercoaster: the ups and downs in out of pocket spending under the standard Medicare drug benefit. Health Aff 2005;24(4):1022 1031.

PAGE 170

170 Summer L, Nemore P, Finberg J. Improving the Medicare Part D Program for the Most Vulnerable Beneficiaries, The Commonwealth Fund, May 2007. Summer L, Nemore P, Finberg J. Medicare Part D: How Do Vulnerable Beneficiaries Fare? The Commonwealth Fund Publication #1127, Vol.35. www.commonwealthfund.org Accessed December 12, 2008. Tamblyn R, Laprise R, Hanley JA et al. Adverse events associated with prescription drug cost sharing among poor and elderly persons. JAMA. 2001;285(4):421 429. Trinacty CM, Adams AS, Soumerai SB et al. Racial differences in long term adherence to oral antidiabetic drug therapy: a longitudinal cohort study. Health Serv Res. 2009; 9 :24. Tseng CW, Mangione CM, Br ook RH et al. Identifying Widely Covered Drugs and Drug Coverage Variation Among Medicare Part D Formularies. JAMA. 2007;297 (23):2596 2602. United States Pharmacopeial Convention. Medicare Prescription Drug Benefit: Summary of USP Approach and Methodol ogy to the Model Guidelines Version 4.0: Drug Categories and Classes in Part D. 2008. USP Web Site. http://www.usp.org/hqi/mmg/revisions.html Accessed June 8, 2010. Vogt WE, Joyce GF, Goldman DP. Effects of Medicare Part D on Pharmacy Use and Out of Pocket Spending. Pittsburgh, PA: RAND Corporation. 2008. Wagner AK, Ross Degnan D, Gurwitz JH et al. Effect of New York State regulatory action on benzodiazepine prescribing and hip fracture rates. Ann als of Internal Medicine 2007;146:96 103. Waknine Y. Lucentis Linked to Stroke Risk in Elderly. Medscape Today 2007. Medscape Today Web Site http://www.medscape.com/viewarticle/551479 Accessed Au gust 6, 2010. Wooldridge JM. Econometric Analysis of Cross Section and Panel Data 2ed. State: The MIT Press; 2002 Yamamoto D, Neuman T, Strollo MK. Medicare Issue Brief: How Does the Benefit Value of Medicare Compare to the Benefit Value of Typical Larg e Employer Plans? The Henry J. Kaiser Family Foundation, Publication #7768. 2008. Yang HK, Simoni Wastila L, Zuckerman IH, Stuart B. Benzodiazepine use and expenditures for Medicare beneficiaries and the implications of Medicare Part D exclusions. Psychi atr Serv. 2008;59:384 391.

PAGE 171

171 BIOGRAPHICAL SKETCH Annesha White Lovett graduated magna cum laude from Florida A & M University with a Doctor of Pharmacy Degree in April 2001. Realizing her passion for outcomes research after taking a related course, she de cided to pursue a Master of Science degree with a focus in p harmacoeconomics, wh ich she obtained in July 2003. She received her Doctor of Philosophy degree from the University of Florida in the fall of 2010. She has worked on various projects with the Fl orida Medicaid program including her thesis entitled, The Economic Burden of Hyperphosphatemia Related End Stage Renal Disease in Florida Medicaid Patients. She has also in terned at the Government Accountability Office, which resulted in a publication ent Benefits: Effects of Using Pharmacy Benefit Managers on Health Plans, Enrollees, and Pharmacies. Dr. Lovett is a recipient of the South Florida Society of Health System Pharmacists Student Achievement Award, the Association of Black Health System Pharmacists Student Achievement Award, the Florida Society of Health System Pharmacists Student Scholarship Award, the I nternational S ociety of P harmacoeconomics and O utcomes R esearch Distinguished Service Award and the Ph armaceutic al R esearch and M anufacturers of A merica Foundation Fellowship. She is also an active member of the A merican S ociety of H ealth Systems P harmacists the A merican Ph armacist s A ssociation and the International Society of Pharmacoeconomics and Outcomes Resea rch Dr. Lovett prides herself in believing that she can truly make a difference in