Medicare Part D and the Use of Hospital Services

It has been hypothesized that Medicare Part D, which provides prescription drug insurance for Medicare beneficiaries, may partially offset government spending on Part A and B services. The research literature broadly supports this hypothesis, but some uncertainty remains regarding the magnitude of the effect. Furthermore, little is known about the extent to which changes in beneficiary health may explain these results. In this paper I use panel data covering the years 1996 through 2010 to investigate whether the relationship between Part D and hospital utilizations is explained by the effect that Part D has on health. To accomplish this I use a set of econometric models to estimate the following four effects: 1) the effect of Part D on hospital utilizations, 2) the net effect of Part D that omits any effect that operates through changes in health, 3) the effect of health on hospital utilizations, and 4) the effect of Part D on health. In all analyses of Part D I provide separate estimates for beneficiaries who had limited or no pre-Part D drug coverage and for beneficiaries who had generous pre-Part D drug coverage. I estimate that Part D reduced the number of inpatient stays by 13% [95% CI: -24%, -2%] and reduced the number of hospital nights by 17% [95% CI: -34%, 1%] among beneficiaries who had limited or no pre-Part D drug coverage. Among beneficiaries who had generous pre-Part D drug coverage, I estimate that Part D increased the number of inpatient stays by 2.5% [95% CI: -12%, 17%] and reduced the number of hospital nights by 22% [95% CI: -41%, -3%]. Results indicate that health does not explain any of the relationship between Part D and the number of inpatient stays, but suggest that health may explain a small part of the relationship between Part D and the number of hospital nights. Finally, I estimate that Part D decreased the number of active health conditions by 8.5% [95% CI: -18%, 1%] among beneficiaries who had limited or no pre-Part D drug coverage and by 9% [95% CI: -17%, -1%] among beneficiaries who had generous pre-Part D drug coverage. Overall, the evidence indicates that Part D had an effect on the health of Medicare beneficiaries, but that this effect on health does not explain a majority of the negative relationship between Part D and the use of hospital services. Medicare Part D and the Use of Hospital Services William J. Parish November 20, 2014