NURSING HOME POLICY VIA HIERARCHICAL DURATION MODELS

通过分层持续时间模型制定疗养院政策

• 批准号: 3372877
• 负责人: CARL N MORRIS
• 金额: $ 13.31万
• 依托单位: HARVARD MEDICAL SCHOOL
• 依托单位国家: 美国
• 财政年份: 1992
• 资助国家: 美国
• 起止时间: 1992-07-01 至 1994-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/3372877
• 关键词: health care cost /financing; health care service utilization; mathematical model; mental health; model design /development; nursing homes; statistics /biometry

The project objective is to develop statistical methods to estimate duration (survival) models, and to apply these methods to nursing home and mental health duration data. In spite of increasing interest in nursing home utilization among the elderly, there have not been corresponding improvements in our knowledge of how long people stay in nursing homes. In particular, elderly people who enter nursing homes as private patients face the risk of spending-down their assets until they become eligible for Medicaid, and it is not known how great this risk is. The first aim is to develop and validate on several data sets hierarchical models incorporating linear multiple regression terms that will provide estimates of important policy variables, while accounting for commonly encountered sampling problems. These sampling problems include censoring and length-biased sampling. The hierarchical modeling framework is important because it enables fitting individual-level models, permits heterogeneity across individuals, and results in more general (mixture) duration distributions. The other two aims are to apply the models developed to improve estimates of nursing home length of stay and the probability of spend-down. The nursing home data set includes all residents in thirty-six San Diego nursing homes during a year and a half period during the early 1980s. As second major focus will be an estimating length of stay of inpatients in mental health facilities, using data from Massachusetts during 1985-1992. The methods developed will be applicable to other health data sets.

该项目的目标是制定统计方法， 持续时间（生存）模型，并将这些方法应用于养老院和 心理健康数据。 尽管人们对护理的兴趣越来越大 在老年人的家庭利用率方面， 我们对人们在养老院呆多久的了解有所改善。 在 特别是作为私人病人进入养老院的老年人， 他们的资产，直到他们成为合格的风险 医疗补助，目前尚不清楚这种风险有多大。 第一个目标是在多个数据集上分层开发和验证 包含线性多元回归项的模型， 对重要政策变量的估计，同时考虑到 遇到了采样问题。 这些抽样问题包括审查 和长度偏差采样。 分层建模框架是 重要的是，因为它能够拟合个人层次的模型，允许 个体间的异质性，并导致更普遍的（混合） 持续时间分布 另外两个目标是应用开发的模型来改进估计 养老院的住院时间和花费的可能性。 的 养老院数据集包括圣地亚哥36个城市的所有居民 在80年代初的一年半时间里， 作为 第二个主要重点将是估计住院病人的住院时间， 精神卫生设施，使用1985-1992年马萨诸塞州的数据。 开发的方法将适用于其他健康数据集。