It is proposed to use a large existing data base describing 34,198 California Seventh-day Adventists to investigate the effects of health habits on survival. Forty-five percent of the subjects in this special population are lacto-ovo vegetarian with many of those remaining having habits typical of other Americans. Forty-one percent of subjects became members at a variety of ages during adult life, this usually being accompanied by great changes in health habits. Thus the population provides the unique advantage of a wide variation in current health habits, so increasing statistical power. In additional there is a varying duration of these dietary and other habits amongst similar aged subjects. Thirteen years of mortality follow-up will be available (1978-90), following collection of baseline exposure data (1974-76). The projected number of deaths during this period is approximately 6,544. The exposure data have been carefully verified for the purpose of previous studies of cancer and ischemic heart disease. Its proposed use for aging and survival analyses is thus extremely cost effective. A major aim is to establish whether particular dietary habits, physical inactivity, obesity levels or psychosocial attributes predict survival. Lengthened survival in the population may result from an increased lifespan, representing an effect on the aging process. This is often also accompanied by delayed expression of common chronic disease, hence not affecting compression of mortality. On the other hand we hypothesize that other lifestyles will not affect the aging process but may delay common chronic diseases, hence producing compression of mortality. We will compare the effects of certain health habits on survival between the two sexes. Similarly we will investigate with special interest exposure-survival associations in the very old (> 80 years) and contrast these with effects in younger persons. Many causes of mortality have long induction periods. We hypothesize that a longer duration of exposure is a separate and important modifier of the effect of current habits but expect that later age of first exposure will have less effect on survival. Analyses will largely use product-limit non-parametric survival estimators and the Cox proportional hazards model.
Showing the most recent 10 out of 11 publications
