Campbell A J, Diep C, Reinken J, McCosh L
J Epidemiol Community Health. 1985 Dec;39(4):337-42. doi: 10.1136/jech.39.4.337.
Between 1977 and 1979 an age stratified sample of people 65 years and over living in the community and in institutions in Gisborne, New Zealand was assessed medically and socially. This sample was followed and reviewed in 1982. At follow up 308 subjects were seen, 227 had died, and 24 had left the area. Factors predicting mortality were assessed. Using a log rank test, factors predicting mortality included age, impaired mental function, functional disability, urinary incontinence, prescribed drugs, pulse pressure, erythrocyte sedimentation rate (ESR), systolic pressure, cardiovascular drugs, and falls. However, a number of these factors increased in prevalence with age. Using a Cox's regression analysis for factors predicting mortality after controlling for age, the following were found to be significant predictors: impaired mental function; functional disability; urinary incontinence; prescribed drugs, ESR and falls. A proportional hazards general linear model showed that the major predictors of mortality in old age were markers of established disease.
1977年至1979年间,对居住在新西兰吉斯伯恩社区及机构中的65岁及以上老年人进行了分层抽样,从医学和社会层面进行了评估。1982年对该样本进行了跟踪和复查。随访时共见到308名受试者,其中227人已死亡,24人离开了该地区。对预测死亡率的因素进行了评估。使用对数秩检验,预测死亡率的因素包括年龄、精神功能受损、功能残疾、尿失禁、处方药、脉压、红细胞沉降率(ESR)、收缩压、心血管药物和跌倒。然而,其中一些因素的患病率随年龄增长而增加。在控制年龄后,使用Cox回归分析预测死亡率的因素,发现以下因素是显著的预测因素:精神功能受损;功能残疾;尿失禁;处方药、ESR和跌倒。比例风险一般线性模型表明,老年死亡率的主要预测因素是已确诊疾病的标志物。