In this study we will examine recent patterns of multiple cause mortality in the U.S. for different race and sex cohorts. Of special interest will be identification of the different roles various conditions play in causing death at advanced age. We will be particularly interested in evaluating the nature of those patterns among the oldest-old population (i.e., those 85+), determining race and sex differences and establishing how those patterns differ from those at younger ages (e.g., 65 to 74 and 75 to 84). In these studies we will have the advantage of time series (i.e., 18 years or more) of mulitple cause mortality data which will be long enough to identify major cohort differences in risks for major chronic diseases--and to isolate those differences from secular trends. In order to analyze the multiple cause mortality data we will construct biologically motivated analytic models. Such models will be designed to deal with such phenomena as dependent competing risks, the interaction of multiple medical conditions and limitations in causing death, the effects of basic processes of biological senesence on the mortality processes, and the possible consequence of age related biological bounds on mortality processes. These models will be evaluated using a variety of demographic and epidemiological data. Of particular importance in constructing and evaluating our models of mortality processes at advanced ages is the fact that, in addition to having a multiple cause mortality time series lengthy enough to describe cohort differences, we will also have the advantage of recent nationally representative surveys with longitudinal follow-ups of individuals. In these surveys, for different components of the U.S. elderly population (i.e., the disabled community population, the 1982-1984 NLTC survey; the general community population, the 1984 to 1986 LSOA; the institutional population, the 1985-1987 NNHS follow-up) we will have large nationally representative surveys (with large samples of the oldest-old) for whom multiple chronic diseases and conditions are reported while the subjects are alive, and for whom mortality follow-up is conducted. In addition we will have detailed longitudinally followed community populations with extensive and frequent assessment of physiological variables. These additional data sets (i.e., the national surveys and the longitudinal populations) will help us both substantively (by helping us to directly examine the morbid and disability processes leading to death and the medical conditions that emerge in those processes) and methodologically (by providing detailed longitudinal data on the health and physiological status of elderly and extreme elderly persons which can be used to evaluate the assumptions and predicted outcomes of biologically motivated models designed to analyze the multiple cause mortality time series).
Showing the most recent 10 out of 54 publications
