The doughnut hole: it's about medication adherence.

The Medicare Prescription Drug, Improvement, and Modernization Act went into effect on 1 January 2006. Part of this act was the creation of Medicare Part D, a federal program to subsidize the costs of prescription drugs for Medicare beneficiaries. In 2006, the plan required beneficiaries whose total annual drug costs exceeded $2250 to pay 100% of prescription costs until they spent $5100 on medications out of pocket and entered catastrophic coverage. This coverage gap is well-known among Medicare beneficiaries as the “doughnut hole.” Once the patient spent $5100 in annual out-of-pocket costs, the “hole” closed and all medication expenses above that amount were covered. An exception to the coverage gap exists among beneficiaries who have an income below 150% of the poverty level and are eligible for the low-income subsidy (LIS). The LIS pays for all or part of the monthly premium, annual deductible, and drug copayments. Many patients who fall in the doughnut hole try to save money by cutting back on blood pressure and cholesterol medications, thereby increasing the risk for cardiovascular disease and suboptimal outcomes. Studies have shown that up to one third of older adults avoid costs by taking less medication than prescribed, known as “costrelated nonadherence” (1–3). The cost of the medication is an important predictor of cost-related nonadherence, regardless of whether the patient has insurance (4). Medication adherence has been divided into 2 concepts: adherence and persistence. Although conceptually similar, adherence refers to the use of the medication as intended and persistence refers to the maintenance of the medication over time. Medication adherence and persistence are growing concerns to clinicians and payers because of the mounting evidence that medication nonadherence is associated with worse clinical outcomes and higher costs of care, particularly among those with cardiovascular disease (5–7). In this issue, Li and colleagues (8) sought to determine the magnitude of the effect of the Medicare doughnut hole on coverage for antihypertensive and lipid-lowering medications. They postulated that these medications are used to treat asymptomatic cardiovascular conditions and may be prone to cost-related nonadherence during the coverage gap. They studied 4 groups of Medicare beneficiaries who had hypertension or hyperlipidemia. Group 1 had the LIS (no out-of-pocket medication costs), group 2 had generic medication coverage (out-of-pocket brand-name medication costs) during the gap, group 3 had generic and brandname coverage (no out-of-pocket medication costs) during the gap, and group 4 had no medication coverage (all outof-pocket medication costs) during the gap. The authors found that once patients entered the doughnut hole and had to pay out of pocket, the number of antihypertensive or lipid-lowering medication prescriptions per patient decreased compared with similar patients who did not have a coverage gap because of the LIS. This difference in coverage was associated with an increased risk for nonadherence to and nonpersistence of cardiovascular medications. The results were similar among patients at the highest risk for cardiovascular outcomes, including those with diabetes, coronary heart disease, and stroke. Interestingly, the presence of a commercial insurance plan that covered generic prescriptions during the doughnut hole period did not mitigate these effects. Among patients who had a Medicare Part D plan with generic and brand-name coverage during a gap period, the risk for nonadherence was similar to those with the LIS (essentially no decrease in adherence). They also demonstrated that this “gap effect” (decrease in the number of prescriptions) did not occur for medications that treat symptoms, such as pain relievers. Although Li and colleagues studied the effects of the coverage gap on nonadherence, they did not determine the effect on patient outcomes. Many observational studies have evaluated the association between medication adherence and outcomes. Nonadherence to cardiovascular medications has been associated with increased risk for morbidity and mortality (7). For example, nonadherence to statins in the year after hospitalization for myocardial infarction was associated with a 12% to 25% increased risk for mortality (9, 10). Among patients with coronary heart disease, nonadherence to cardiovascular medications is common (11, 12) and is associated with a 10% to 40% relative increase in risk for hospitalizations and a 50% relative increase in risk for mortality (13). Nonadherence contributes to poor blood pressure and cholesterol control. Observational studies have also demonstrated that nonadherence is associated with worse clinical outcomes. What is the solution? Li and colleagues propose a multifactorial approach by all stakeholders, including clinicians, pharmacists, health plans, and government agencies. This approach should help patients identify and adopt low-cost alternatives (including generic drugs) to maintain adherence. Patients will rarely change medications to a less costly option once they are in the doughnut hole; rather, they often split their existing medications or stop filling the costliest medications. This phenomenon may explain why, in Li and colleagues’ study, patients who had generic medication coverage still became nonadherent. If patients initiated generic medications, wouldn’t fewer people enter the gap? Government agencies must also be part of the solution. The original purpose of instituting the gap was to limit the cost of the Part D program, but it has resulted in unintentional consequences (14, 15). Medicare Part D beneficiaries have an overall decrease in the use of generic medications (16). As part of the design of the program, the government is not permitted to negotiate drug prices; as a result, the Centers for Annals of Internal Medicine Editorial