Objective: To identify minimum criteria to assist the prediction of decline in physical health-related quality of life in the elderly.
Study design and setting: Participants were women from the Australian Longitudinal Study on Women's Health, who responded to three separate Surveys conducted in 1996 (when they were aged 70-75 years), 1999 and 2002. Using data from these Surveys, three categories were generated which described current physical health-related quality of life and future physical decline as measured by the physical component summary score (PCS) of the MOS SF-36 quality of life survey.
Results: Bivariate analyses reported a large number of variables significantly associated with physical decline (P < 0.001), including age, falls, number of diagnoses, symptoms, doctor visits and medications, days spent in hospital, body mass index, living arrangements and social support. Multivariate analyses, using decision tree analysis, identified three items which accurately predicted 76% of the women who would exhibit physical decline according to our definition.
Conclusions: This study identified a number of variables that may be useful in clinical screening for vulnerability to physical decline.