The American Journal of Managed Care November 2008 - Special Issue
Primary Care Physicians' Views of Medicare Part D
Objective: To examine physicians' attitudes about the impact of Medicare Part D and how it varied among seniors, particularly Medicare-Medicaid dual-eligible enrollees.
Study Design: Web-based survey of primary care physicians in North Carolina (generous Medicaid formulary) and Florida, Massachusetts, and Texas (restrictive Medicaid formularies).
Methods: Of 5141 eligible primary care physicians, 716 (14%) responded between November 2007 and March 2008. We examined Part D's effects on access overall and for selected populations. We used descriptive and regression analyses to assess physicians' views about Part D's effects on dual-eligible enrollees and how those views differed between North Carolina and the other states. All analyses were weighted for nonresponse.
Results: More respondents had a favorable (48%) than an unfavorable (37%) view of Part D overall, and 55% reported Part D improved access to prescription drugs in general. However, 44% reported access declined for individuals with prior drug coverage, and 64% reported Part D formularies were insufficient for their patients' needs. Nearly half (49%) reported dual-eligible enrollees' access was worse under Part D in 2007 relative to Medicaid before 2006; 63% reported higher administrative burden. Physicians reported Part D lowered dual-eligible enrollees' access and increased providers' burden more in North Carolina than in the 3 restrictive Medicaid states.
Conclusion: Primary care physicians held generally positive but widely varying views of Part D. Respondents expressed concerns about access to prescription drugs under Part D, particularly for dual-eligible enrollees. Improving the transparency and generosity of Part D formulary coverage may improve access.
(Am J Manag Care. 2008;14(11 Spec No.):SP5-SP13)
Physicians in this 4-state survey had heterogeneous views of Medicare Part D.
- A plurality of respondents viewed Part D and its effects on access to prescription drugs positively overall.
- Respondents were concerned about the sufficiency of Part D formularies for their patientsÃ¢â‚¬â„¢ needs and Part DÃ¢â‚¬â„¢s impact on prescription drug access for patients who had coverage previously.
- For dual-eligible enrollees, Part D in 2007 was viewed as less generous and more cumbersome than Medicaid before 2006.
- Views of Part DÃ¢â‚¬â„¢s impact on dual-eligible enrollees were more negative in North Carolina, a state with generous Medicaid drug coverage.
In this article we analyze how physicians viewed the Part D benefit and its effects on their patients’ access to prescription drugs as of late 2007 and early 2008. This is important because the effects of Medicare Part D on seniors’ access to prescription drugs likely depend partly on the availability and design of coverage before and after Part D’s implementation. Prescription drug plans typically rely on formularies to limit expenditures, both directly by altering patients’ cost-sharing arrangements7 and indirectly by shifting physicians’ prescribing habits. Prior to Part D, most seniors who had prescription drug coverage obtained it through their employers, Medicaid, or Medicare supplemental plans. Under Part D, seniors have a wider choice of plans and associated formularies, including Part D plans, Medicare Advantage plans that cover prescription drugs, and employer-sponsored coverage. Part D plans vary across a range of dimensions, including their copayment levels, which drugs they cover, whether any insurance is provided to cover the “donut hole,” and the types of supply-side controls (eg, prior authorization requirements) that they use.8,9 Variety among plans may have offsetting effects on access: it may improve enrollees’ ability to find coverage that matches their needs while complicating the physician’s decision to prescribe the best medication given the patient’s medical needs and drug coverage.
We also examined physicians’ views about the effects of Part D on Medicare-Medicaid dual-eligible enrollees. Dual-eligible enrollees have a disproportionate need for medical services and prescription drugs, lower incomes, and less education than other Medicare beneficiaries on average.10 Before 2006, dual-eligible enrollees had drug coverage through Medicaid. The transition to Part D for dual-eligible enrollees was especially challenging11 because of the administrative complexity of the enrollment process,12,13 Medicare’s policy of random auto-assignment to prescription drug plans,14 and inconsistencies in the coverage of required drug classes.15 Moreover, because the generosity of the Medicaid drug benefit varied substantially across states, dual-eligible enrollees in states with generous Medicaid coverage may have fared relatively worse than those in states with restrictive Medicaid coverage in the transition to Part D.
We surveyed primary care physicians in 4 large, racially diverse states (Florida, Massachusetts, North Carolina, and Texas) that represent distinct regions in the country and vary in the restrictiveness of their Medicaid drug formularies. As all of these states have relatively low Medicaid managed care penetration, their Medicaid formularies apply to most Medicaid enrollees. Prior to 2005, Florida, Massachusetts, and Texas adopted restrictive Medicaid formularies that do not permit pharmacists to request prior authorizations,16-18 whereas North Carolina continues to have a generous Medicaid formulary.19 Of the 26 on-patent drugs we identified as being used to treat 3 common clinical conditions -- hypercholesterolemia, hypertension, and diabetes -- prior authorization was required for 16 in Florida, 23 in Massachusetts, 14 in Texas, and none in North Carolina as of October 2007.
We used physician-level prescribing data from Wolters Kluwer Health to identify a sample frame of primary care physicians in the 4 study states. These data distinguished the source of payment among cash, Medicaid, and other insurance but did not separately identify prescriptions paid for by Medicare Part D. To ensure respondents had adequate exposure to Part D and dual-eligible enrollees in clinical practice, we limited the sample frame to physicians whose patients filled a minimum of 30 new (nonrefill) prescriptions overall and 5 new Medicaid prescriptions for hypercholesterolemia, hypertension, and diabetes in the year preceding the study (September 2006 to August 2007). Respondents who reported being enrolled in or having completed a clinical fellowship, practicing as a hospitalist, or not treating both Medicaid and Medicare patients in the last 2 months were ineligible to take the survey.
Survey Instrument and Fielding
We collected detailed information regarding both physician-level characteristics (race, sex, medical specialty, board certification, international medical graduate status, years in practice, and time spent on clinical medicine) and practice-level characteristics (number of physicians in the practice, and the income level, race/ethnicity, and insurance status of patients in the practice). We asked physicians about their overall impressions of Part D and their perceptions of its effects on access to prescription drugs for their Medicare patients in general, among selected subgroups (their sickest, healthiest, and minority patients), and for their patients who did and did not have drug coverage prior to Part D. In addition, physicians were asked about the sufficiency of prescription drug formularies under Medicare Part D and Medicaid. Finally, physicians were asked to compare aspects of prescription drug coverage retrospectively for dual-eligible enrollees under Part D in 2007 with coverage under Medicaid prior to 2006, including their patients’ satisfaction, compliance with and access to medications, and the physicians’ own ability to prescribe preferred medications and administrative burden.
The self-administered online survey was pilot-tested extensively for clarity and breadth, and took an average of 22 minutes to complete. A paper version of the instrument is available on request. It was fielded from November 2007 through March 2008. Invitation letters were sent by first-class mail in 4 overlapping waves. Honoraria for participation ranged from $50 to $100. Up to 3 postcard reminders were sent to nonrespondents with valid addresses.
Of the 5901 physicians in the sample frame, 760 had invalid mailing addresses or were ineligible for the study. Of the remaining 5141 physicians, 716 completed the survey, for an overall response rate of 14%. The response rate was highest among physicians in North Carolina (19%) and lowest among physicians in Texas (11%). Compared with nonrespondents, a greater proportion of respondents were from North Carolina (31% vs 21%) and female (31% vs 27%), and a smaller proportion of respondents were from Texas (30% vs 37%); on average, respondents were less experienced and younger (P <.05 for all comparisons) than nonrespondents (eAppendix Table, available at www.ajmc.com). We calculated weights to balance all of these characteristics except age and experience between respondents and nonrespondents.
We computed descriptive statistics for respondents overall and also compared the responses of physicians from North Carolina (generous Medicaid formulary) with the responses of physicians from Florida, Massachusetts, and Texas (restrictive Medicaid formularies). Unadjusted analyses used t tests for continuous variables, Fisher exact tests for binary variables, and χ2 tests for categorical variables. Multivariate binary and ordered logistic regressions were estimated to account for the effects of differential distributions of respondent characteristics (ie, physician specialty, board certification status, international medical graduate, race, sex, survey wave, years of experience, practice size, time devoted to clinical practice, and distributions of income level, race, and insurance status of patients in practice) between North Carolina and the 3 restrictive Medicaid states combined. We also tested the equality of responses among Florida, Massachusetts, and Texas using regression-based joint Wald tests. To account for potential nonresponse bias, we weighted our analyses using the weights that balance the respondent and nonrespondent samples as closely as possible. Only deidentified data were used in this analysis, and strict measures were undertaken to preserve participants’ confidentiality. This study was approved by the Arizona State University Institutional Review Board.
Table 1 reports respondents’ unadjusted descriptive characteristics. Respondents were predominantly white (54%), male (69%), and board certified (80%). Respondents treated substantial shares of Medicaid (22%), Medicare (33%), and uninsured (11%) patients on average. Most respondents characterized their practices as serving patients primarily of low or medium income, and almost none reported treating primarily high-income patients.
Bivariate analyses found several differences between physicians from North Carolina and physicians from the other 3 states (data not shown). Compared with physicians practicing in the 3 restrictive Medicaid states, physicians in North Carolina were more likely to be white (66% vs 48%) or black/African American (15% vs 5%), to be women (38% vs 29%, P = .04), and to specialize in family practice rather than internal medicine (74% vs 57%). By contrast, fewer physicians in North Carolina were Asian (12% vs 20%) or Hispanic/Latino (1% vs 16%), board certified (72% vs 83%; P = .001), or international medical graduates (19% vs 26%) (P <.001 for all comparisons unless noted otherwise).
Views on Part D’s Effects on Patients
On average, physicians expressed positive views of Part D overall; however, physician views were mixed and varied based on the subpopulation examined (Table 2). More respondents had a positive overall impression of Part D’s prescription drug benefit than a negative one (48% vs 37%, respectively). In total, 55% of respondents reported that Part D improved access to prescription drugs for their Medicare patients in general; far fewer said that patients’ access worsened (20%) or remained unchanged (18%).