Wolinsky F D, Johnson R L, Stump T E
Department of Medicine, Indiana University School of Medicine, Indianapolis, USA.
Gerontologist. 1995 Apr;35(2):150-61. doi: 10.1093/geront/35.2.150.
Mortality risks over an 8-year period are assessed among the 7,527 older adults interviewed as part of the Longitudinal Study on Aging. Using a modified version of the behavioral model, hierarchical methods are used to sequentially introduce the predisposing and enabling characteristics, disease history, disabilities and functional limitations, perceived health, and health services utilization both at and after baseline. Based on their partial r statistics (shown in parentheses), the proportional hazard analyses identify the mean annual number of hospital episodes after baseline (.13), age (.07), female gender (-.05), non-kin social supports (-.03), body mass (-.03), and having a history of diabetes (.03) as the six most salient predictors. These six variables account for 80% of the overall model fit.
在作为老龄化纵向研究一部分接受访谈的7527名老年人中，评估了8年期间的死亡风险。使用行为模型的修改版本，采用分层方法依次引入基线时及基线后个体的易患因素和促成因素特征、疾病史、残疾和功能限制、自我感知健康以及医疗服务利用情况。基于其偏r统计量（括号内显示），比例风险分析确定基线后每年平均住院次数（0.13）、年龄（0.07）、女性性别（-0.05）、非亲属社会支持（-0.03）、体重（-0.03）以及有糖尿病史（0.03）为六个最显著的预测因素。这六个变量占整个模型拟合度的80%。
