Causal relationships between survival rates, dietary and lifestyle habits, socioeconomic status and physical, mental and social health in elderly urban dwellers in Japan: A chronological study

Abstract

Life expectancy was well known to associate with lifestyle habits, socioeconomic condition, and three health-related dimensions (physical, mental and social health status). However, the causal effect relationship among these variables remains unclear. The purpose of this study was to determine the causal relationships among health and life conditions, socioeconomic status, dietary and lifestyle habits and three health-related dimensions in elderly urban dwellers in Tokyo, Japan. Of 16,462 eligible elderly individuals aged 65 years and older, 13,195 participants responded to the questionnaire survey conducted in September 2001 (a response rate of 80.2%). A follow-up survey was conducted in September 2004 and a total of 8162 survivors were followed until the end of August 2007 through the municipal resident's registry. Exploratory factor analysis defined five latent variables based on the 13 observed variables. From a cross-lagged effects variation model using structural equation modeling, causal relationships were analyzed using longitudinal data obtained at the 2001 and 2004 survey and the number of survival days between 2004 and 2007. After estimating a best-fit model, we discovered that health and life conditions were not determined by current dietary and lifestyle habits, which many studies showed. However, the conditions were more directly affected by three health-related dimensions three years earlier, and indirectly affected by educational attainment and previous annual income as well. The current model suggests that it might be of great importance for elderly individuals to emphasize the maintenance of psychological well being, physical activity, social communication and participation, as well as income, rather than focusing on improvements in diet and health-related lifestyles per se.