The Anticipatory Effects of Medicare Part D on Drug Utilization

描述（由申请人提供）：2003年，《医疗保险处方药、改进和现代化法案》（MMA）签署成为法律，该法案扩大了医疗保险计划，将处方药的自愿保险纳入其中。该计划被称为D部分，于2006年1月1日实施。鉴于该项目的主要目标之一是扩大处方药的可及性，评估对药物利用的影响至关重要。然而，确定这种影响是复杂的，因为D部分是在2003年12月MMA通过时宣布的，比该计划实施早了两年。前瞻性的个人可能已经预测到未来的价格变化带来的补贴覆盖和改变他们的药品支出之前，D部分生效。本项目将估算2003年D部公布后的因果需求反应。这个项目有三个含义。首先，这项调查量化了在政策实施之前宣布政策的效果，以及对计划评估的影响。在项目评估文献中，政策公告能够对重要结果产生实际影响（与其实施效果分开）的观点在很大程度上尚未得到探讨。其次，该项目提供了在毒品需求背景下前瞻性行为的第一次测试。在许多其他情况下，基于每个时期的相对价格，优化个人对当前和未来消费的权衡。然而，在确定当前药物消费水平时，个人是否会对未来价格的预期作出反应尚不清楚。第三，该项目查明以前对d部分利用效果的估计可能存在很大的偏差，因为有一种预期的反应；这些研究——从政策实施前一到两年的使用变化来确定项目效果——将错误地描述d部分的净效果。实证方法通过比较d部分宣布前后慢性药物与急性药物的使用情况，来确定是否存在因果预期反应。因此，他们的需求可能只对目前的价格有反应，而长期吸毒可能对预期的未来价格下降有消极或积极的反应。该项目还使用了其他群体（如老年人和近老年人）对公告的暴露程度的变化来确定影响。最后，该项目产生了第一个D部分的治疗效果利用估计，其中考虑了方案前预期反应。

DESCRIPTION (provided by applicant): In 2003, the Medicare Prescription Drug, Improvement, and Modernization Act (MMA) was signed into law, which expanded the Medicare program to include voluntary insurance coverage for prescription drugs. This program, known as Part D, was implemented on January 1, 2006. Given that one of the program's primary aims was to expand access to prescription drugs, evaluating the impact on drug utilization is of first-order importance. However, identifying this effect is complicated by the fact that Part D was announced with the passage of the MMA in December 2003, two years before the program was implemented. Forward looking individuals may have anticipated future price changes brought about by subsidized coverage and changed their drug spending before Part D took effect. This project will estimate the causal demand response to the announcement of Part D in 2003. There are three implications of this project. First, this investigation quantifies the effect of announcing a policy before its implementation, with consequences for program evaluation. The idea that policy announcements can have real effects (separate from their implementation effect) on important outcomes has been largely unexplored in the program evaluation literature. Second, the project provides the first test of forward-looking behavior in the context of drug demand. In many other contexts, optimizing individuals trade-off present and future consumption based on the relative prices in each period. However, it is not known whether individuals are responsive to expectations of future prices when determining current drug consumption levels. Third, the project identifies a potentially large source of bias in the previous estimates of the utilization effect of Part D. To the extent that there was an anticipatory response, these studies- which identify the program effect from changes in utilization relative to the year or two just prior to the policy's implementation-- will misstate the net effect of Part D. The empirical approach identifies whether there is a causal anticipatory response by comparing utilization for chronic drugs relative to acute drugs before and after the announcement of Part D. This strategy hinges on the observation that acute drugs treat illnesses that are largely unpredictable, thus their demand is likely to respond only to current prices, whereas chronic drug use may respond negatively or positively to anticipated future price reductions. This project also uses variation in exposure to the announcement across other groups such as the elderly and near-elderly to identify the effect. Finally, this project produces the first treatment effect utilization estimates of Part D that take into account a pre-program anticipatory response. PUBLIC HEALTH RELEVANCE: This project seeks to estimate the causal anticipatory effects of Part D on drug utilization. Estimating the utilization effect of Part D is important for evaluating whether the primary aims of the policy-- expanding access to prescription drugs-- were achieved and for assessing the program's future financial solvency. Accounting for the combined effect of the anticipatory response and the response at implementation will produce a more accurate estimate of the program effect than previous studies that have considered the implementation effect alone.