Abstract
STUDY OBJECTIVE: To examine whether cognitive and psychosocial factors predict mortality once physical health is controlled. DESIGN: A prospective study of community dwelling elderly. Mortality was assessed over a period of 3-4 years after the baseline assessment of predictors. The data were analysed using the Cox proportional hazards model. SETTING: Canberra and Queanbeyan, Australia. PARTICIPANTS: A sample of 897 people aged 70 or over and living in the community, drawn from the compulsory electoral roll. RESULTS: For the sample as a whole, the significant predictors of mortality were male sex, poor physical health, poor cognitive functioning, and low neuroticism. Men had an adjusted relative risk of mortality of 2.5 compared with women. For the male sub-sample, poor self rated health and a poor performance on a speeded cognitive task were significant predictors, while for women, greater disability, low systolic blood pressure, and a low score on a dementia screening test were the strongest predictors. CONCLUSIONS: Mortality was predicted by physical ill health and poor cognitive functioning. Psychosocial factors such as socioeconomic status, psychiatric symptoms, and social support did not add to the prediction of mortality, once sex, physical health, and cognitive functioning were controlled. Mortality among men was more than twice that of women, even when adjusted for other predictors.