Abstract
BACKGROUND: Studies on falls in older adults have mainly been conducted in high income countries. Scant, if any, information exists on risk factors for falls in the older population of sub-Saharan African countries. METHODS: A cross-sectional survey and a 12-month follow-up study were conducted to determine risk factors for falls in a representative multi-ethnic sample of 837 randomly selected ambulant community-dwelling subjects aged ≥65 years in three suburbs of Cape Town, South Africa. Logistic regression models were fitted to determine the association between (1) falls and (2) recurrent falls occurring during follow-up and their potential socio-demographic, self-reported medical conditions and physical assessment predictors. RESULTS: Prevalence rates of 26.4 % for falls and 11 % for recurrent falls at baseline and 21.9 % for falls and 6.3 % for recurrent falls during follow-up. In both prospective analyses of falls and recurrent falls, history of previous falls, dizziness/vertigo, ethnicity (white or mixed ancestry vs black African) were significant predictors. However, poor cognitive score was a significant predictor in the falls analysis, and marital status (unmarried vs married) and increased time to perform the timed Up and Go test in the recurrent fall analysis but not in both. Other than the timed Up and Go test in recurrent falls analysis, physical assessment test outcomes were not significant predictors of falls. CONCLUSION: Our study provides simple criteria based on demographic characteristics, medical and physical assessments to identify older persons at increased risk of falls. History taking remains an important part of medical practice in the determination of a risk of falls in older patients. Physical assessment using tools validated in developed country populations may not produce results needed to predict a risk of falls in a different setting.