A Clinical Prediction Tool to Guide Treatment of Osteoporosis in the Nursing Home

描述：主要骨质疏松性骨折是常见的，病态的，和昂贵的长期住养老院的居民。尽管这个问题很重要，但目前还没有指导方针来筛查养老院居民的骨质疏松症，也不清楚哪些骨质疏松症药物可以预防长期住院居民的骨折。为了解决这些知识空白，我们提出以下具体目标：1)开发并验证一种使用跌倒和骨折临床危险因素的预测工具，该工具将估算长期住院养老院居民髋部骨折的2年绝对风险和主要骨质疏松性骨折（髋部、肱骨和手腕合并）的2年绝对风险；2)确定骨质疏松药物（即双膦酸盐、降钙素、雌激素、雷洛昔芬和特立帕肽）是否能降低长期住院养老院居民髋部和主要骨质疏松性骨折的发生率；3)确定骨质疏松药物使长期住院居民髋部骨折和主要骨质疏松性骨折发生率降低20%的临床风险阈值。该项目将利用现有数据库，该数据库先前将医疗保险A部分和B部分的索赔数据与药房数据（医疗保险D部分）、临床特征（最小数据集）和设施水平特征（OSCAR）联系起来。使用这个数据库，我们将对2006-2011年间所有参加医疗保险服务收费计划的美国养老院居民进行前瞻性分析，这些居民的住院天数为90天（每年有70万居民）。对于特定目标1，预测工具将完全从所有美国长期居民的临床信息中开发，例如认知和功能状态，非骨质疏松药物使用和近期跌倒史。预测工具将考虑到养老院居民的高死亡率。对于特定的目的2a，我们将通过比较骨质疏松药物“新使用者”与“非使用者”的骨折发生率，并通过倾向评分匹配，来确定骨质疏松药物是否能降低长期住院居民髋部和主要骨质疏松性骨折的发生率。对于特定的目标2b，我们将根据骨折预测工具估计的基线骨折风险，通过检查骨质疏松药物的疗效来确定药物干预的阈值。研究团队在药物流行病学、老年病学、医疗保险索赔数据分析和完成本项目所需的最小数据集特征方面具有丰富的经验。我们的发现将是非常重要的，因为他们将提供第一个指导骨质疏松症的筛查和治疗在养老院设置。我们预计我们的研究结果将很容易为提供者实施，因为骨折预测工具将由所有美国养老院居民已经可用的临床信息组成。从这项研究中获得的知识可能最终导致养老院环境中主要骨质疏松性骨折的发生率降低，从而降低发病率和医疗保健成本。

DESCRIPTION: Major osteoporotic fractures are common, morbid, and costly among long-stay nursing home residents. Despite the importance of this problem, there are no guidelines to screen nursing home residents for osteoporosis, and it is unclear which osteoporosis medications prevent fractures in long-stay residents. To address these gaps of knowledge we propose the following specific aims: 1) develop and validate a prediction tool using clinical risk factors for falls and fracture that will estimate the 2 year absolute risk of hip fracture and the year absolute risk of major osteoporotic fracture (hip, humerus, and wrist combined) in long-stay nursing home residents; 2) determine whether osteoporosis medications (i.e., bisphosphonates, calcitonin, estrogen, raloxifene, and teriparatide) reduce the incidence of hip and major osteoporotic fracture in long-stay nursing home residents; and 3) determine the clinical thresholds of risk at which osteoporosis medications reduce the incidence of hip and major osteoporotic fracture by e 20% in long-stay residents. This project will leverage an existing database that has previously linked claims data from Medicare Parts A and B with pharmacy data (Medicare Part D), clinical characteristics (Minimum Data Set), and facility level characteristics (OSCAR). Using this database we will conduct a prospective analysis on all U.S. nursing home residents enrolled in a Medicare fee-for-service plan and with e 90 day length of stay between the years 2006-2011 (>700,000 residents annually). For specific aim 1, the prediction tools will be developed entirely from clinical information that is available for all U.S long-stay residents, such as cognitive and functional status, non-osteoporosis medication use, and recent history of falls. The prediction tools will account for the high mortality in nursing home residents. For specific aim 2a, we will determine whether osteoporosis medications reduce the incidence of hip and major osteoporotic fracture in long-stay residents by comparing the incidence of fracture among "new users" of an osteoporosis medication with the incidence of fracture in "non-users," matched by propensity scores. For specific aim 2b, we will determine a threshold for pharmacologic intervention by examining the efficacy of osteoporosis drugs according to baseline fracture risk, as estimated by the fracture prediction tool. The research team has experience in pharmacoepidemiology, geriatrics, and analysis of Medicare claims data and Minimum Data Set characteristics necessary to complete this project. Our findings will be highly significant as they will provide the first guidance on screening and treatment of osteoporosis in the nursing home setting. We anticipate that our findings will be easy for providers to implement because the fracture prediction tools will be comprised of clinical information that is already available for all U.S. nursing home residents. Knowledge gained from this study could ultimately result in a decreased rate of major osteoporotic fractures in the nursing home setting, with a subsequent reduction in morbidity and health care costs.