Exploring the pathways to longevity: biodemographic, familial and socioeconomic influences on later-life mortality
Valérie Jarry, Université de Montréal
Alain Gagnon, Université de Montréal
Robert R. Bourbeau, Université de Montréal
The close relationship between early childhood conditions and health and mortality outcomes in old age has been extensively studied in both the epidemiological and demographic fields. The channels through which early life is hypothesized to influence mortality are diverse and could be direct or indirect via adult characteristics. Despite the ample evidence on the influence of childhood conditions on overall longevity, less established in the literature is whether this association holds true within long lived families and whether this is a direct effect or an indirect effect. In this paper we investigate the association between biodemographic and socioeconomic factors in early life and mortality after age 40, and through which pathways this effect may operate. In the first part of this study, we examine whether and how the effect of early life conditions influence longevity in the general population as well as in families of centenarians. In the second part, we examine whether the association between childhood conditions and old age mortality can be mediated by the socioeconomic status in adulthood or by marriage. An event-history database that links individuals to their childhood characteristics gathered from the 1901 and 1911 Canadian census records and to their adult characteristics is used. Non-parametric analyses are performed to estimate the effect of early life and adult variables using the Kaplan-Meier estimator as well as gender-specific proportional hazard models with a Gompertz specification of the risk of mortality. We further suggest the use of structural equation modeling (SEM) to identify direct and indirect effects of early life conditions on later life mortality. Overall, this study will contribute to gain insight into the predictors and pathways associated with longevity in addition to deepen our understanding as to whether mortality determinants among the general population may be different from those among long-lived individuals.