MANTEL-HAENSZEL TYPE ESTIMATORS FOR THE COUNTER-MATCHED SAMPLING DESIGN IN NESTED CASE-CONTROL STUDY

We are concerned with a countermatched nested casecontrol study. Assuming the proportional hazards model, the MantelHaenszel estimators of hazard rates are presented in two situations. The proposed estimators can be calculated without estimating the nuisance parameter. Consistent estimators of the variance of the proposed hazard rate estimators are also developed. We compare these estimators to the maximum partial likelihood estimators in the asymptotic variance. The methods are illustrated using the Colorado Plateau uranium miner cohort data.