A chain graph model of inter-relationships among socio-demographic, societal and behavioural factors that affect health over the life course

Fernando Rajulton, University of Western Ontario
Jianlin Niu, University of Western Ontario

A rational health policy needs a shift from focussing on how individuals differ in their 'relative risks' to examine the multicausal nature of health status. This paper aims at understanding the relationships among socio-demographic, societal and behavioural influences that affect health. Using the longitudinal data from the four rounds of the National Population Health Survey, Canada, we identify not only how these relationships vary by age groups, gender, social and immigrant status, but also how they change over the life courses of individuals. Although several studies have examined the behavioural influences on health, the inter-relationships among sociodemographic, social and behavioural influences have not been clearly brought out, possibly because proper statistical techniques have not been available. In this study, we use Chain Graph Models that concentrate on meaningful connections among variables included in the analysis and help identify the changing nature of these relationships over the life course.