It is difficult to distinguish changes in health due to aging from those related to dying, because the two processes are highly related. Some potentially treatable conditions may mistakenly be dismissed as due to old age. The goal of this article was to examine the relationships of aging and of dying to changes in 10 health-related variables: self-rated health, depression, ADLs, IADLs, minimental state examination, body mass index, blocks walked per week, bed days, hospitalization, and walking speed (all coded so that higher values were better). We used longitudinal data from the Cardiovascular Health Study to estimate the changes in the variables associated with 5 years of aging and also in the 5 years before death, controlling for years from death and for age, respectively. All 10 health variables declined as death approached, and most of them also declined with age. The "effect" of the dying process was usually significantly larger than the effect of aging. Large declines in these health measures are probably not due to aging, and should be taken seriously by patients and their providers.