Abstract
Past epidemiologic studies of coronary disease have shown such variables as elevated blood pressure, elevated serum cholesterol, cigarette smoking, obesity, diabetes and family history of coronary disease to be factors predictive of elevated risk of incidence of this disease in groups and in individuals. Enthusiasm over these positive results should not obscure the fact that these risk factors, taken together, provide an incomplete estimate of the coronary-disease burden of a population and an insensitive prediction of the risks of individuals.Keys and his colleagues1 applied multiple logistic riskfactor equations derived in Framingham to European data and then reversed the process, . . .