Medicare beneficiaries with hypertension in our sample who transitioned from no drug coverage to Part D plans had increased odds of using any antihypertensive medication as well as increased counts of antihypertensive medications post-Part D. These findings point to an important public health benefit from Part D and likely translate into improvements in blood pressure control and overall health.
We are not aware of prior studies that have specifically studied the effect of Medicare Part D on prescription medication use by patients with hypertension during the first 2 years of the program. Previous studies found Part D increased overall drug use 6–22 percent and reduced out-of-pocket spending 13–23 percent (Ketcham and Simon 2008
; Yin et al. 2008
;). The effect is consistently higher among those with no prior prescription coverage; researchers have found that this group increased its overall medication use by 11–37 percent, and its overall pharmacy spending by as much as 74 percent (Schneeweiss et al. 2009
; Zhang et al. 2009
We find less of an impact of Part D for those with limited prior drug coverage. There were no improvements in the likelihood of using any antihypertensive post-Part D among the sample with prior prescription coverage, and only minimal improvements in annual counts of medications. This lack of improvement may represent a ceiling effect, as the U.S.$150-cap and U.S.$350-cap groups had relatively high usage to begin with.
Our findings also suggest that Part D may encourage the use of more costly agents that, on average, may not improve health outcomes. While Part D is associated with an increased likelihood of patients' without prior coverage using an antihypertensive, it is also associated with these patients being more likely to initiate treatment with more expensive ARBs over ACEs. It is unlikely that these differences can be accounted for by differences among the study groups in rates of ACE-related side effects (the main indication for using ARB over ACE). Because ARBs are available as brand-name drugs only, while ACEs are primarily lower priced generics, these use patterns have important cost implications. This outcome is consistent with the finding of a recent study that noted Part D was associated with less use of generic drugs (Zhang et al. 2008
There are some potential limitations to our study. First, our results are based on drugs purchased at network pharmacies, but we believe any bias from missing claims is negligible, for several reasons. Our study design, with longitudinal data and a control group, should guard against this bias. Members who filled prescriptions in the network received a 15 percent discount from the plan's negotiated prices that are already lower than retail price. In addition, network pharmacies were numerous (around 58,000 nationwide) and included almost all local pharmacies. It is possible that we might not have observed all prescriptions filled by members through a U.S.$4 program introduced by Wal-Mart in November 2006; however, the use of these programs was limited in 2007, and any decrease in filled hypertension claims post-Part D would, if anything, underestimate the Part D effects. Second, our results might not be generalizable because our study members were continuously enrolled in MA-PD products offered by a single insurer in western Pennsylvania. In addition, our samples were more likely to be white with relatively high median income compared with national averages, and their higher incomes may make them less sensitive to changes in out-of-pocket-costs than other Medicare beneficiaries. Finally, although we used pre/post-design and propensity scores to control for health status differences between comparison groups, we acknowledge the potential for unobserved differences in the health status of our subjects as an explanation for our results. In addition, some of the measures used for deriving the propensity score are measured at the community level and therefore there is the potential for substantial error in deriving a propensity score with those measures.
In conclusion, Part D increased medication use among seniors with hypertension who previously lacked drug coverage. In addition, Part D appears to have increased the use of ARBs more than less expensive alternative medications. If these results are reproducible, a more careful design of drug coverage advocating for use of the most cost-effective drugs may improve the efficiency of the U.S. health care system. This is especially important with the passage of national health care reform (Patient Protection and Affordable Care Act), in which the Part D benefit will become more generous as the coverage gap is eliminated.