EXTERNAL COMPARISONS FROM NESTED CASE-CONTROL DESIGNS

The nested case-control design, used to sample within cohorts, is usua lly employed for internal comparisons. We propose to use this design f or external comparisons. We-present two probability-weighted estimator s ui the expected number of cases under a given exposure, based an ext ensile rates, for tu;cs versions of the nested case-control design. Th ese estimators are used, along with their variance estimators, to form confidence intervals for standardized mortality ratios. The estimator s are practically unbiased, whereas the naive estimator that treats th e nested case-control sample as a random sample of the cohere is clear ly biased, an estimator from the alternative Cox model-based approach is found to be substantially biased when applied in this context, Comp aring the proposed estimators for nested case-control designs to a cor responding estimator fur the case cohort design, we found that the cor relation between follow-up time and exposure time (that is, the amount of time under the exposure effect) has an impact on which type of des ign is more efficient for external comparisons. A small correlation fa vors the case cohort design and a large correlation the nested case-co ntrol design. We examine empirical properties of these estimators thro ugh computer simulations, using a cohort study of the incidence of sec ond cancer in 2,189 patients with Hodgkin's disease.