Objective:To examine the association between generosity of drug coverage and essential cardiovascular medication use among retired seniors. Study Design: Retrospective analysis of the 1997 to 2000 Medicare Current Beneficiary Survey, a nationally representative survey of the Medicare population.
Methods:The study examined community-dwelling fee-for-service Medicare beneficiaries aged 65 years or older with retiree health insurance and with coronary heart disease and hyper-lipidemia (n = 1220) or congestive heart failure (n = 1147). Generosity of drug coverage was defined as the percentage of the beneficiary’s annual drug expenditures paid by the employer. Dependent variables were any statin use for the group with coronary heart disease and hyperlipi-demia and any angiotensin-converting enzyme inhibitor or angiotensin receptor blocker use in the group with congestive heart failure. Logistic regression analyses estimated the adjusted odds of essential medication use by generosity category in each disease group. We estimated the extent to which medication use would change if generosity levels moved to those under standard Medicare Part D levels.
Results: The overall prevalence of statin use was 64.1%, and that of angiotensin-converting enzyme inhibitor or angiotensin receptor blocker use was 50.0%. In both disease groups, retirees in the less generous drug coverage categories had significantly lower adjusted odds of use than retirees with the most generous drug benefits (ie, covering >76% of annual drug expenditures). Overall, the shift to a standard Medicare Part D structure would result in mean declines of 8.4% in statin use and 5.2% in angiotensin-converting enzyme inhibitor or angiotensin receptor blocker use. Retirees with the most generous drug coverage face twice the mean decline in drug use.
Conclusion: Retirees who already have generous drug benefits from their employers may be placed at risk for decreased utilization of effective medications due to any future scaling back of retiree drug coverage.
(Am J Manag Care. 2007;13:425-431)
With a double-digit growth in pharmaceutical expenditures, many employers have increased prescription cost-sharing for retirees in recent years. The new Medicare Part D has created additional (albeit unintended) incentives for terminating or scaling back retiree drug coverage.
One third of the Medicare beneficiaries now have drug coverage through employer-sponsored health plans, and most of these are offered by their own or their spouse's former employer or union. Most of these beneficiaries receive more generous coverage under these plans than they would under the new Medicare prescription drug coverage (Part D) which began January 1, 2006.1 Competitive labor markets have provided sufficient incentive in the past few decades for employers to voluntarily offer retiree health benefits with generous drug coverage to workers and their families.2 However, because an accounting rule change in the early 1990s requiring employers to include retiree health benefit liabilities on their financial statements, coupled with a double-digit growth in healthcare costs (particularly for pharmaceuticals), many employers have been overhauling their retiree health and drug benefits to control, reduce, or eliminate their retiree health benefit liabilities.3 With the enactment of the Medicare Prescription Drug Improvement and Modernization Act, additional (albeit unintended) incentives for terminating or scaling back retiree drug coverage have been created. First, the existence of Medicare Part D as an alternative source of drug coverage for retirees would reduce political resistance for employers considering dropping coverage altogether. In fact, the Congressional Budget Office4 estimated that 17% of retirees were likely to lose their employer-sponsored drug benefits; more recent estimates from the Department of Health and Human Services suggest that it might be double that.5 Second, in light of the ongoing trends of reductions in generosity of retiree drug coverage due to business, accounting, and cost factors,3 employers planning to continue offering benefits may now have Medicare Part D as a concrete example of a "floor" or minimum level of generosity for retiree drug benefits. This is because the US Congress has built in financial incentives in terms of direct subsidies of up to $88 billion to employers for maintaining their retiree drug coverage. To receive a tax-free subsidy (equal to 28% of total drug costs between $250 and $5000 for each covered retiree), the employer would need to offer employment-based retiree drug coverage that is at least actuarially equivalent to the standard Medicare Part D coverage. Hence, employers planning to continue offering drug coverage could over time reduce the generosity of their benefits all the way to Medicare Part D levels and still receive the benefit of the federal subsidies.
Regardless of whether employers will drop coverage to standard Medicare Part D levels or not, recent surveys suggest that employers are scaling back prescription coverage through increased drug cost-sharing for retirees and are likely to continue doing so during the long term.3,6,7 The question of how substantial increases in out-of-pocket shares for prescription expenditures could potentially affect medication use, particularly among Medicare beneficiaries with retiree health insurance, has not been studied in detail. This study takes advantage of the existing variation in retiree drug benefits to examine the potential effect of significant changes in prescription cost-sharing on use of essential chronic medications among retired seniors. To reduce the complexity in testing this relationship, we specifically examine the association between the generosity of drug coverage and use of 2 effective cardiovascular medications selected from well-established clinical guidelines,8-10 namely, (1) statin use among retirees with coronary heart disease (CHD) and hyperlipidemia and (2) angiotensin-converting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB) use among retirees with congestive heart failure (CHF).
The data source used for the study was the 1997 to 2000 Medicare Current Beneficiary Survey (MCBS) linked with Medicare claims. The MCBS is a nationally representative survey of Medicare beneficiaries conducted under the auspices of the Centers for Medicare & Medicaid Services. It includes detailed information on beneficiaries' demographics, health and functional status, health insurance coverage, and annual use and spending for all healthcare services, including prescription drugs. The health insurance section of the survey captures the source of supplemental insurance (eg, current employer, former employer, self-purchased, Medicaid, and others) and whether these plans offer prescription drug coverage. The prescription drug use section captures drug names during the thrice-yearly interviews of respondents who are instructed to keep a medication log, save insurance receipts, and show the interviewers all of their medication containers.
The sample frame consists of community-dwelling fee-for-service Medicare beneficiaries aged 65 years or older with supplemental health insurance from their own or their spouse's former employer from 1997 to 2000. We limited the sample to fee-for-service beneficiaries because availability of Medicare claims was necessary to identify disease conditions. From the
1997 to 2000 pooled cross sections, we selected 2 groups for analysis based on 1 or more inpatient or outpatient claims for (1) CHD and hyperlipidemia (n = 1220) and (2) CHF (n = 1147) (an appendix containing the International Classification of Diseases, Ninth Revision, Clinical Modification and Current Procedural Terminology codes is available online at www.ajmc.com). For the primary analysis, we excluded Medicare retirees with additional drug coverage from sources other than an employer (153 retirees from the group with CHD and hyperlipidemia and 167 retirees from the group with CHF). However, secondary analysis without excluding this group indicated that our results are not sensitive to this sample exclusion criterion, and results only for the primary sample are presented.
Our dependent variable was an indicator of the presence of any use of the essential cardiovascular medications. The indicator represented any statin use for the group with CHD and hyperlipidemia and any ACE inhibitor or ARB use in the group with CHF.
The main independent variable of interest was the generosity of drug coverage. This variable was defined as the percentage of the beneficiary's annual drug expenditures paid by the employer. We grouped beneficiaries into the following 4 generosity categories: 0%, 1% to 50%, 51% to 75%, and 76% to 100%. The 0% category primarily represents beneficiaries with no retiree drug benefits but also includes a small group with such limited drug benefits that none of their drug spending was paid by the employer.
Control variables included socioeconomic and demographic factors such as age, sex, education, income, metropolitan status, and geographic region of residence. We also controlled for smoking status and whether the beneficiary had ever visited a cardiologist. In addition, we included the following clinical characteristics: numbers of physician visits and hospitalizations for the specific study condition, presence of contraindication to the essential medication, cardiovascular risk factors likely to increase essential medication use, number of chronic conditions other than cardiovascular disease, and self-reported health status (the details are available in an appendix from the author). Survey year of observation was also used as a control.