Medicare Part D Plan Generosity & Dual-Eligible Nursing Home Residents

描述（由申请人提供）：医疗保险D部分处方药福利极大地改变了处方药在养老院的融资和管理方式，特别是对于三分之二有双重资格获得医疗保险和医疗补助的居民（“双重杀伤”）。虽然医疗补助计划以前为双重资格的疗养院居民提供药物，但这些人现在被随机分配到私人处方药计划（PDP）。PDP在处方集覆盖范围和特定药物的使用管理技术方面有所不同。因此，一些双重资格的疗养院居民被随机分配到PDP，其目前的药物覆盖范围很广，而其他人则被分配到覆盖范围更严格的PDP。目前还没有实证研究的PDP慷慨的双重资格的养老院居民的健康和福祉的影响。为了评估PDP慷慨对药物使用和居民健康结果的影响，我们建议使用一家大型长期护理药房的药物利用数据，并结合2005-2008年期间（D部分实施前一年和D部分实施后三年）最小数据集（MDS）和医疗保险A部分住院索赔的健康结果数据。本申请的具体目的是：1）为随机分配的双重合格疗养院居民构建每个PDP的覆盖范围慷慨度的药物特定措施; 2）检查PDP慷慨度对药物变更和停药的影响; 3）检查PDP慷慨度对健康结局的影响;以及4）评估药物变更和停药对健康结局的影响。目的2和3利用随机化的双重杀伤PDP估计慷慨的药物使用和结果的影响，是不受选择偏差。为了确定药物中断是否是PDP慷慨影响健康结果的机制，目标4使用随机分配给PDP作为结构模型中的工具变量，估计药物变化和停药对健康结果的影响。我们将检查医生、疗养院管理人员和长期护理药房报告的某些PDP覆盖范围限制的七类药物：骨质疏松症药物、血管紧张素受体阻滞剂（ARB）、非典型抗精神病药、抗抑郁药、胆碱酯酶抑制剂、长效阿片类药物和雾化吸入剂。我们将研究可能与这些药物使用中断有关的健康结果，包括骨折、心脏并发症住院、幻觉或妄想、行为、抑郁症状、认知能力、功能状态和疼痛。我们将估计三项回归模型，逻辑回归模型，线性回归模型和两阶段最小二乘回归模型的组合，以解决这些目标。公共卫生相关性：根据医疗保险D部分，老年双重资格的养老院居民-一个具有巨大医疗需求和有限财政资源的人口-被随机分配到PDP，而不考虑其目前药物的覆盖范围。我们的研究将首次全面评估PDP覆盖范围慷慨对双重合格养老院居民的药物使用和健康结果的影响。这项研究的结果将是有用的联邦和州的政策制定者在评估可能的变化，部分D和其他主要利益相关者负责居民护理，如疗养院管理人员，医生谁照顾疗养院居民，长期护理药房。

DESCRIPTION (provided by applicant): The Medicare Part D prescription drug benefit dramatically changed how prescription drugs are financed and administered in nursing homes, particularly for the two-thirds of residents who are dually eligible for Medicare and Medicaid ("dual eligibles"). While Medicaid formerly covered drugs for dual-eligible nursing home residents, these individuals are now randomly assigned to private prescription drug plans (PDP). PDPs vary with respect to formulary coverage and utilization management techniques for specific medications. As a result, some dual-eligible nursing home residents are randomized to PDPs with generous coverage of their current medications and others to PDPs with more restrictive coverage. There have been no empirical studies of the impact of PDP generosity on the health and well-being of dual-eligible nursing home residents. To assess the impact of PDP generosity on medication use and resident health outcomes, we propose to use drug utilization data from a large long term care pharmacy in combination with data on health outcomes from the Minimum Data Set (MDS) and Medicare Part A hospitalization claims for the period 2005-2008 (one year before and three years after Part D implementation). The specific aims of this application are to: 1) construct medication-specific measures of the generosity of coverage for each PDP to which a dual-eligible nursing home resident was randomly assigned; 2) examine the impact of PDP generosity on medication changes and discontinuations; 3) examine the impact of PDP generosity on health outcomes; and 4) assess the impact of medication changes and discontinuations on health outcomes. Aims 2 and 3 take advantage of the randomization of dual eligibles to PDPs to estimate effects of generosity on medication use and outcomes that are uncontaminated by selection bias. To determine if medication disruptions are a mechanism through which PDP generosity affects health outcomes, Aim 4 uses the random assignment to PDPs as an instrumental variable in a structural model estimating the effect of medication changes and discontinuations on health outcomes. We will examine seven classes of medications for which physicians, nursing home administrators, and long term care pharmacies have reported coverage restrictions for some PDPs: osteoporosis medications, angiotensin receptor blockers (ARBs), atypical antipsychotics, antidepressants, cholinesterase inhibitors, long-acting opioids, and nebulized inhalants. We will study health outcomes that could be related to disruptions in use of these medications, including fractures, hospitalization for cardiac complications, hallucinations or delusions, behavior, depressive symptoms, cognitive performance, functional status, and pain. We will estimate a combination of trinomial regression models, logistic regression models, linear regression models, and two-stage least squares regression models to address these aims. PUBLIC HEALTH RELEVANCE: Under Medicare Part D, elderly dual-eligible nursing home residents -- a population with great medical need and limited financial resources -- are randomly assigned to a PDP without consideration of its coverage of their current medications. Our study will provide the first comprehensive assessment of the impact of PDP coverage generosity on medication use and health outcomes for dual-eligible nursing home residents. The results of this study will be useful to federal and state policymakers in assessing possible changes to Part D and to other key stakeholders responsible for resident care, such as nursing home administrators, physicians who care for nursing home residents, and long term care pharmacies.