Medicare Beneficiaries Response to Coverage Gaps Versus Actuarially Equivalent Continuous Coverage for Prescription Drugs

Grant Description: Are Medicare beneficiaries likely to react differently when faced with the doughnut-hole “gap” in Medicare Part D than they would with actuarially equivalent continuous coverage? The researchers challenged the hypothesis that actuarially equivalent, but structurally different cost-sharing arrangements have similar impacts on beneficiaries' prescription drug utilization patterns. They examined whether the relationship between use and benefit structure is sensitive to the overall generosity of insurance coverage. This project built on Stuart's previous HCFO grant assessing the effects of gaps in drug coverage for Medicare beneficiaries with common chronic diseases. That study found that gaps in drug coverage lead to reduced utilization rates and that the effects are magnified for those with common chronic diseases such as diabetes, COPD, and mental illness. This project extended the understanding of how Medicare beneficiaries react to benefit structure, but will be useful to private payers as they search for a cost-sharing formula that contains costs while minimizing disruption in medication regimens. The objective of this project was to provide policymakers with a better understanding of how Medicare beneficiaries behave when faced with alternative cost-sharing structures.

Policy Summary: Perhaps the most controversial aspect of the Medicare Part D prescription drug benefit is the large gap in coverage known as the “doughnut hole.” There is a near universal belief that the doughnut hole reduces utilization of prescription drugs among those affected by it. But all forms of cost sharing do that. An important unanswered question for policymakers is whether the precipitous increase in out-of-pocket cost due to the gap has a fundamentally different impact on demand compared to continuous coverage of equal generosity. In other words, would it matter if Medicare filled the doughnut hole but increased the coinsurance rate by an actuarially equivalent amount? This study compared medication spending patterns for Medicare beneficiaries who experienced gaps in coverage to those with continuous prescription benefits from various sources over the period 1997 through 2003. Unlike prior published studies, the authors controlled for the annual generosity of prescription benefits for both groups thereby permitting conditional estimates of the impact of coverage gaps net of own-price differentials. Consistent with previous research, the authors found that Medicare beneficiaries who experience coverage gaps spend less on drug purchases compared to beneficiaries with continuous coverage and that the effect is strongly associated with gap duration. They also confirmed that the demand for prescription drugs is sensitive to own-price. However, in models containing plan generosity and gap measures (plus other covariates), they found that generosity of coverage dominates gap effects for every type of drug plan and level of drug spending analyzed. There was evidence that gaps independently reduce drug spending, but the impact is very small and statistically significant only for beneficiaries in employer-sponsored plans. Moreover, the authors found no evidence of an independent gap effect for beneficiaries spending above the equivalent of the Part D cap level of $2,250 in 2006 dollars. Another noteworthy finding from the study is that Medicare beneficiaries with high levels of prescription drug spending are more price sensitive than those with lower spending levels. The estimated price elasticity of demand for individuals spending under $2,250 in 2006 dollars (0.34) is within the range reported in the literature, but the estimate for individuals spending over the 2006 cap amount was substantially higher (0.56). These differential elasticities are important from a policy standpoint because they imply that it would cost Medicare more to fill in the doughnut hole than the program would save by raising the coinsurance rate by an actuarially equivalent amount.