The research goal is to identify factors associated with the maintenance of good health and well-being among women as they age. The study examines differences in the life histories and life patterns of women in different age groups and how these factors relate to health outcomes among the age groups or cohorts. The study uses social, sociopsychological, and demographic data from a household interview survey conducted in 1970-71, a 1982 follow-up survey, and computerized medical record data beginning three years prior to the 1970-71 survey and continuing to the present. The study has two interrelated sets of aims. The first set will be achieved using essentially a cross-sectional research design, while the second set uses a longitudinal or panel design. The first set of aims: 1) to examine cohort differences in family history, demographic characteristics, adult role sets, health beliefs, attitudes, and behaviors; 2) to assess the relationship of these variables to health outcomes within each cohort; and 3) to assess the possible independent, additive, or multiplicative nature of the relationship of these variables to health outcome. The second set of aims are: 1) to assess changes in role sets and attributes to roles including continuity of roles and relationships for each cohort in the decade between the surveys; 2) to assess the effect of these roles and relationship changes on health outcomes among the cohorts; 3) to incorporate findings from the analysis for the first set of aims regarding variables influential in health outcomes; and 4) to identify important intervening variables which may modify the health effects of particular independent variables. The study population is a 5 percent sample of female members of large HMO who in 1970 were age 25 or older. This sample of 1,189 women is divided into three age cohorts for the study. The basic approach to the analysis begins with an examination of bivariate relationships and moves towards multivariate assessments. The independent variables include categorical and ordinal level measures, while the dependent variables include ordinal and interval measures. For purposes of examining bivariate relationships which include categorical and ordinal variables, nonparametric statistics will be used. Where interal level dependent variables are examined analysis of variance will be used. When the objective is to determine the separate and combined effects of the independent variables on the dependent variables, multivariate techniques such as log linear models for multidimensional contingency tables and/or multiple regression and discriminant analysis will be used. The study will be the basis for more specialized studies, including later examination of these same subjects as they advance through the life cycle (since, as members of a panel, medical data will continue to be collected on them).
