Background: Mortality from all causes is higher for persons with fewer years of education and for blacks, but it is unknown which diseases contribute most to these disparities. Methods: We estimated cause-specific risks of death from data from the National Health Interview Survey conducted from 1986 through 1994 and from linked vital statistics. Using these risk estimates, we calculated potential years of life lost and potential gains in life expectancy related to specific causes, with stratification according to education level and race. Results: Persons without a high-school education lost 12.8 potential life-years per person in the population, as compared with 3.6 for persons who graduated from high school (ratio, 3.5; P<0.001). Ischemic heart disease contributed most (11.7 percent) to the difference according to education in potential life-years lost (with all cardiovascular diseases accounting for 35.3 percent). All cancers accounted for 26.5 percent, including 7.7 percent due to lung cancer; other lung diseases and pneumonia contributed 10.1 percent of the total, whereas human immunodeficiency virus (HIV) disease accounted for none of the difference according to education. The pattern of disparities according to level of income was similar to that according to level of education. Blacks and whites lost 7.0 and 5.2 potential life-years per person, respectively, as a result of deaths from any cause (ratio, 1.35; P<0.001). Cardiovascular diseases accounted for one third of this disparity, in large part because of hypertension (15.0 percent); HIV disease (11.2 percent) contributed almost as much as ischemic heart disease (5.5 percent), stroke (2.8 percent), and cancer (3.4 percent) combined; trauma and diabetes mellitus accounted for 10.7 percent and 8.5 percent, respectively. Conclusions: Although many conditions contribute to socioeconomic and racial disparities in potential life-years lost, a few conditions account for most of these disparities -- smoking-related diseases in the case of mortality among persons with fewer years of education, and hypertension, HIV, diabetes mellitus, and trauma in the case of mortality among black persons. These findings have important implications for targeting efforts to reduce existing disparities in mortality rates.
