Kitchener Martin, Bostrom Alan, Harrington Charlene
Department of Social and Behavioral Sciences, University of California, San Francisco 94118, USA.
Inquiry. 2004 Summer;41(2):189-202. doi: 10.5034/inquiryjrnl_41.2.189.
This paper draws from a rich longitudinal California data set to analyze the scope and nature of nursing home closures between 1997 and 2001, and to present a Cox proportionate hazards model of the risks of closure that arise from a range of facility and market characteristics. When compared with the sample total of 1,482 facilities operating in the baseline year of 1997, only 56 facilities closed through 2001, involving the loss of 3.8% of facilities and 2,915 beds (2.3%). The multivariate Cox model of factors associated with closure reports that: 1) hospital-based facilities are 600% more likely to close than are free-standing homes; 2) reducing bed size by one standard deviation (52 beds) increases the risk of closure by 460%; 3) facilities with losses of 5% or worse are more than twice as likely to close; and 4) a one-standard deviation increase in the spare bed capacity measure of county competition raises the risk of facility closure by 140%.
本文借鉴了加利福尼亚州丰富的纵向数据集,以分析1997年至2001年期间养老院关闭的范围和性质,并提出一个Cox比例风险模型,用于分析一系列机构和市场特征所引发的关闭风险。与1997年基准年运营的1482家机构的样本总数相比,到2001年仅有56家机构关闭,占机构总数的3.8%,损失床位2915张(占2.3%)。与关闭相关因素的多变量Cox模型表明:1)医院附属机构关闭的可能性比独立式养老院高600%;2)床位规模减少一个标准差(52张床位)会使关闭风险增加460%;3)亏损5%或更严重的机构关闭的可能性是其他机构的两倍多;4)县内竞争的备用床位容量指标每增加一个标准差,机构关闭风险就会提高140%。