Cost-Effectiveness of Full Medicare Coverage of Angiotensin-Converting Enzyme Inhibitors for Beneficiaries with Diabetes

Context Economic barriers may contribute to the underutilization of angiotensin-converting enzyme (ACE) inhibitors in patients with diabetes. Contribution This analysis shows that the Medicare program could improve clinical outcomes and save money by providing full coverage of ACE inhibitors with no out-of-pocket expense for patients with diabetes. Clinical and economic benefits would exist even if full coverage only resulted in a modest increase (7.2%) in ACE inhibitor use. Implications Insurers should consider a policy of covering the entire cost of drugs that improve outcomes and lengthen life. The Editors The prevalence of diabetes and its complications is increasing substantially in the United States (1-4). As the population ages and the number of elderly Medicare beneficiaries increases relative to the rest of the population, so will the number of elderly persons with diabetes and the share of national health expenditures allocated to complications of diabetes. Angiotensin-converting enzyme (ACE) inhibitors slow the progression of renal disease (5-12) and reduce cardiac morbidity and mortality (13-15) in individuals with diabetes. Despite ample clinical (5-15) and economic (16-18) evidence of benefit, ACE inhibitors are underused in elderly individuals with diabetes (19-23). A growing body of evidence suggests that drug copayments reduce the use of essential medications in the elderly (24-30). Specifically, elderly individuals with diabetes seem to curtail essential drug use as their drug coverage decreases (26, 31). In 1999, Medicare beneficiaries paid more than 40% of their drug costs out of pocket (32). Under the new Medicare drug benefit, out-of-pocket spending will continue to be substantial (33), causing an ongoing potential barrier to the use of known effective drugs. A policy that bases patients' out-of-pocket copayments on clinical benefit rather than cost of drugs has been proposed to improve value in health care (34). We aimed to assess the cost-effectiveness to Medicare of first-dollar coverage (that is, no cost sharing) of ACE inhibitors for elderly beneficiaries with diabetes. Methods Decision Analytic Model We developed a Markov model simulating the natural history of renal and cardiovascular complications in diabetes and risk reduction due to ACE inhibition (Figure 1). Model outcomes included progression of renal disease, cardiovascular events, life expectancy, quality-adjusted life expectancy, lifetime costs, and incremental cost-effectiveness ratios. Our model builds on previous models of ACE inhibition for diabetic nephropathy (16). However, the primary complication of type 2 diabetes is cardiovascular disease, accounting for more than 60% of deaths; thus, we extend these models by adding a cardiovascular events component based on recent data from the Heart Outcomes Prevention Evaluation (HOPE) trial. The HOPE trial demonstrated that ACE inhibitors improve both renal and cardiovascular outcomes, compared with placebo, in patients with diabetes (13, 14, 35). Figure 1. Markov model of disease progression. CVD A cohort of individuals 65 years of age with diabetes enters the model and transitions through renal disease states and cardiovascular event states with rate of disease progression modified by the use of ACE inhibitors. The time horizon of the analysis is divided into 1-year cycles and the cohort is followed over its lifetime. Initial Population Distribution The cohort was initially distributed across disease states (Table 1) on the basis of epidemiologic data obtained from the 19992000 National Health and Nutrition Examination Survey (NHANES) (36), a nationally representative cross-sectional survey of the U.S. noninstitutionalized civilian population. Among individuals 65 years of age or older with self-reported diabetes, 28.2% had microalbuminuria, 18.3% had macroalbuminuria, 20.3% had previous myocardial infarction, and 14.5% had previous stroke. Table 1. Model Inputs: Disease Prevalence and Progression Likelihood of Events Renal Disease We obtained baseline risk for developing and progressing through the stages of nephropathy (normoalbuminuria, microalbuminuria, macroalbuminuria, and end-stage renal disease [ESRD]) from the placebo groups of trials in patients with type 2 diabetes (5, 7, 37). We obtained risk reductions from ACE inhibitor treatment from clinical trials (5-7) (Table 1). While ACE inhibition may cause regression of diabetic nephropathy (8, 11, 46, 47), we assumed no regression to ensure conservative cost-effectiveness estimates. Cardiovascular Disease We obtained baseline cardiovascular risk and risk reduction with ACE inhibitors from the placebo and treatment groups, respectively, of the subset of individuals with diabetes who were enrolled in the HOPE trial (14) (Table 1). We obtained these rates after calibrating the model to match the composite end point reported by the HOPE trial. Individuals can die of myocardial infarction or stroke or another cardiovascular cause, or they can survive and are then still subject to progression of renal disease and the possibility of subsequent cardiac events. Once an individual has a myocardial infarction or stroke, the risk for subsequent cardiovascular events increases according to event recurrence rates in patients with diabetes (38, 39). Cardiovascular event rates also increase with increasing renal dysfunction, with increased risk obtained from a prospective longitudinal study of patients with type 2 diabetes (40). Noncardiovascular Mortality We based the age-dependent probability of death from causes other than cardiovascular disease or ESRD on year 2000 U.S. life table data (41, 42) multiplied by a standardized mortality ratio for diabetes in the elderly (2). We then applied a proportional hazards (that is, multiplicative) model to remove cardiovascular disease and ESRD mortality (42) from age-based diabetes hazards, because these are modeled separately in our model. We obtained age-dependent ESRD mortality rates from the U.S. Renal Data System (43). Utilities Health state utilities, or measures of value for given health states, can be thought of as quality-of-life weights that are bounded by 1 for perfect health and 0 for death. When utilities are multiplied by the lengths of time individuals spend in their respective health states, the resultant metric is a quality-adjusted life expectancy (measured in quality-adjusted life-years [QALYs]), which reflects both the quantity and quality of remaining years of life (48). We obtained utilities for our study (Table 2) from published studies (49-53). Table 2. Model Inputs: Utilities, Costs, and Discount Rate Costs Because we assessed a Medicare coverage decision, the base-case model took the Medicare perspective, including direct medical costs and future (related and unrelated) health care costs, because all are borne by Medicare. We performed additional analyses from the societal perspective to allow for comparison with other cost-effectiveness analyses. Because of the well-known difficulties in measuring price changes in the medical sector (57, 58), we standardized costs to 2003 U.S. dollars by using the Consumer Price Index for all urban consumers (59). We based all costs, except medication costs, on Medicare claims data (Table 2). Event Costs We obtained Medicare expenditures occurring in the year of an event from the 2001 Medicare Standard Analytic Files, a nationally representative 5% random sample of fee-for-service Medicare beneficiaries, and limited them to beneficiaries 65 years or older with diabetes. We obtained annual expenditures associated with ischemic strokes, myocardial infarctions, and deaths (both cardiovascular and noncardiovascular) from these data. We assigned costs for the year of death on the basis of the costs incurred in the last year of life by beneficiaries who had a myocardial infarction, a stroke, both events, or neither event in the year of their death. We obtained annual costs associated with treating patients with ESRD (including costs of dialysis; transplantation; and other health care, including cardiovascular event care) from the U.S. Renal Data System. The annual costs comprise Medicare payments for individuals with ESRD due to diabetes (54). Ongoing Care Costs We obtained ongoing costs of care (including medical costs of future years of life added) for years in which no discrete event occurred from a diagnostic classification system. This system was developed for the Centers for Medicare & Medicaid Services (CMS) to make risk-adjusted payments to Medicare managed care plans on the basis of patient demographic characteristics and clinical risk factors (55). Many expenditures depend on beneficiary demographic characteristics (age, sex, and percentage of Medicaid enrollment), but incremental costs accrue for disease history (that is, history of stroke, myocardial infarction, and diabetes-related renal failure) (Table 2). We based the annual cost of therapy ($233) on the average wholesale price of lisinopril, a once-daily, off-patent ACE inhibitor (56). Interventions The intervention of interest was Medicare first-dollar coverage of ACE inhibitors (that is, Medicare bears the full drug costs with no beneficiary cost sharing). The comparator was current practice (before the planned 2006 implementation of the new Medicare drug benefit) (60). For current practice, we assign a base rate of ACE inhibitor use from which Medicare accrues the benefits and cost offsets (that is, avoided costly health outcomes), but we assume that Medicare bears none of the costs for drugs in this comparator group. With first-dollar coverage, we very conservatively assume that 100% of the current practice drug costs get shifted to Medicare (that is, all third-party payers currently paying for ACE inhibitors will drop this coverage, effectively shifting all ACE inhibitor costs to Medicare). ACE Inhibitor Use The rates of ACE inhibitor utilization with current practi