This study investigated whether social and/or psychologic factors help to predict older dialysis patients' continued survival. A stratified (by race and sex) random sample of patients aged 60+ years was selected from the ESRD Network census of all patients in that age category residing in a single southeastern state (Georgia) and receiving chronic dialysis as of November 1987; personal interviews with patients were completed in 1988. This analysis includes 287 patients (mean age, 69 years) receiving outpatient hemodialysis for whom primary cause of renal failure and functional status data were complete. Patient tracking and vital statistics data determined that 49% of the sample survived as of October 31, 1990. Study variables included demographic, dialysis, health status, social situation, and psychologic outlook variables reported at the patients' 1988 interviews. Log rank tests showed univariate associations between patients' continued survival and race/gender, recovery time following dialysis treatments, cardiovascular co-morbidity, exercise activity score, freedom from health limitation of daily activity, functional status, leisure activity score, self-rated health status, overall life satisfaction, depression, and public religiosity. The Cox proportional hazards model was fit to the data, with continued survival from the time of the 1988 interview as the dependent variable. There was a significantly increased mortality risk for white men relative to the other race/gender groups and for patients reporting severely impaired functional status at the 1988 interview. With functional status in the model, no other social or psychologic variables were significant predictors of mortality.(ABSTRACT TRUNCATED AT 250 WORDS)