Profile and predictors of mortality among free-living elderly subjects: a community-based prospective cohort study from Kerala, India

Objective: There is a paucity of information regarding the mortality profile in community-dwelling elderly as well as the associated risk factors for mortality. The primary objective was to report the 6-year mortality profile in free-living elderly population aged 65 years and above from Ernakulam, Kerala, India via longitudinal follow-up. Methods: The study was a prospective cohort with stratified random cluster sampling. We collected baseline information via house visits using a predefined questionnaire. The subjects were followed up after recruitment for a median of 60 months by phone calls at regular intervals. Calls not answered were followed up by house visits by research personnel. All deaths were verified by reviewing medical records available at subjects' homes. Results: We recruited a total of 1000 participants from 40 individual predesignated clusters (cluster size=25). The median follow-up was 71.8 (70.5-72.8) months and the cumulative follow-up was 5352.16 person years. A total of 262 (26.2%) subjects died during the follow-up period. The mortality rate for the overall cohort was 48.95 per 1000 person years. The corresponding figures for elderly men and women were 54.25 and 45.07 per 1000 person years, respectively. The mortality rate for elderly subjects residing in rural and urban areas was 57.59 and 45.33 per 1000 person years, respectively. Independent predictors of mortality in the study population identified by Cox regression-based analysis were age, sex, orthostatic hypotension, dependence in instrumental activities of daily living and adiposity. Conclusions: Approximately a quarter of community-dwelling elderly individuals died within the next 6 years. Future interventions targeting a reduction in mortality among the elderly need to focus on modifiable risk factors like orthostatic hypotension, adiposity and dependence in instrumental activities of daily living.