Risk Prediction and Etiologic Heterogeneity

Epidemiologic methods such as case-control and cohort studies are used to identify factors related to the risk of a disease. Discovery of new risk factors increases the predictability of the disease by increasing the variance of the predicted risks in the population at risk.  Enhancement of risk predictability also occurs due to the identification of etiologically distinct disease sub-types. In this talk we show how to define etiologic heterogeneity in a way that facilitates the identification of etiologically distinct sub-types, paving the way for optimized risk factor identification and risk prediction. The ideas will be illustrated using data from examples in cancer epidemiology research.