Falls among the non-institutionalized elderly in northern Minas Gerais, Brazil: prevalence and associated factors

Abstract Objective: To identify the prevalence of falls and associated factors in non-institutionalized elderly persons. Methods: A cross-sectional study featuring a population-based sample of non-institutionalized elderly persons in a city in the north of Minas Gerais was performed. Interviews were conducted in households by trained staff using validated instruments. We investigated the associations between falls and demographic, socioeconomic and health-related factors. After bivariate analysis, the variables associated with falls to a level of 20% were analyzed together using logistic regression, assuming at this stage a significance level of 5%. Results: The studied population was predominantly female, married and with a low educational level. The prevalence of falls was 28.4%. The factors that were associated with falls were: female gender (OR=1.67; 95% CI:1.13 to 2.47); negative self-evaluation of health (OR=1.49; 95% CI: 1.02 to 2.20); impaired functional mobility (Timed Up and Go test >20 seconds) (OR=1.66; 95CI: 1.02-2.74); the occurrence of hospitalization in the previous 12 months (OR=1.82; 95% CI: 1.17 to 2.84); and frailty measured by the Edmonton Frail Scale (OR=1.73; 95% CI: 1.14 to 2.64). Conclusions: The prevalence of falls was high for the population studied and was related to the individual health conditions of the elderly. Resumo Objetivo: Estimar a prevalência de quedas e os fatores associados em idosos não institucionalizados. Métodos: Estudo transversal com amostra de base populacional de idosos não institucionalizados em cidade polo do norte de Minas Gerais. Foram conduzidas entrevistas nos domicílios por equipe especialmente treinada utilizando instrumentos validados. Investigou-se a associação entre a ocorrência de quedas e variáveis demográficas, socioeconômicas e relacionadas à saúde. Após análise bivariada, as variáveis associadas até o nível de 20% foram analisadas conjuntamente por meio de regressão logística, assumindo-se nessa fase o nível de significância de 5%. Resultados: A população avaliada era predominantemente feminina, casada e com baixa escolaridade. A prevalência de quedas foi de 28,4%. Os fatores que se mostraram associados à ocorrência de quedas foram: sexo feminino (OR=1,67; IC95%:1,13-2,47); a autopercepção negativa da saúde (OR=1,49; IC95%:1,02-2,20); comprometimento da mobilidade funcional (teste Timed Up and Go > 20 segundos) (OR=1,66; IC95%:1,02- 2,74); o registro de internação nos 12 meses precedentes (OR