Women live longer than men. The process which produces this endpoint has not been described. We do not know whether observed gender differences in longevity reflect a consistent. Proportionately greater hazard throughout the life span for men or whether the sex relative hazard (RH) shifts as a population ages. When a RH is dependent upon length of observation, it is said to be time dependent. Prior epidemiological studies have not described the time structure of the process which results in greater longevity for women. This proposal addresses this gap in knowledge. This study will examine the following questions: 1. Whether the female longevity advantage is due to a constant, proportionately greater hazard in males, or whether female hazards rise as male hazards fall over time. 2. Whether the time dependence of the sex RH varies as a function of behavioral risk variables. A fundamental question in the epidemiology of aging is whether observed sex differences are fixed by biologic characteristics or are alterable by behavior and environment as evidenced by dependence of sex differences on individual risk characteristics. 3. Whether sex differences in survivor effects (reduced associations between baseline risk characteristics and outcomes over time) result from the dependence of the sex RH on time and risk characteristics. These differences could explain known inconsistent findings of epidemiologic studies in populations with different lengths of follow-up. 4. Whether sex differences in survivor effects are explained by sex differences in risk variable change over the follow-up period or by faster deletion of susceptible males. 5. Whether sex differences in the time dependence of risk variable effects explain apparent contradictions between male excess for mortality and female excess for morbidity and functional disability: Do women demonstrate delayed risk variable effects on mortality which result in female excesses of morbidity and functional disability? The examination of this hypotheses is crucial to understanding whether sex difference in time dependence are due to differences in time to death after risk exposure or differences in disease etiology. Hypotheses will be examined in prospective analyses in the Alameda County Study cohort of the Human Population Laboratory, a representative population sample of 6,928 persons selected in 1965 and followed for mortality through 1986. Significantly, this work can be completed without new data collection.
