|
Abstract:
|
Nested case-control study design is popular in epidemiological practice due to its cost-effectiveness. Typically, for all the cases and the selected (matched) controls, complete information is collected, while the information on other members in the study cohort is limited. The ordinary conditional logistic regression is widely used, however, its efficiency can be of concern. We present a maximum likelihood estimator under the nested case-control sampling in the context of Cox proportional hazards model. In particular, we are interested in estimating the effect of a time-dependent exposure that have a variable number of repeated measurements in cases and their matched controls. The method is applied to data from a nested case-control study of hormones and breast cancer.
|
