ABSTRACT
For cohort data, estimation in Cox’s regression model is based on a partial likelihood, which at each observed death or disease occurrence (generically denoted “failure”) compares the covariate values of the failing individual to those of all individuals at risk at that time. Therefore, standard use of Cox regression requires collection of covariate information on all individuals in a cohort even when only a small fraction of them actually fails. This may be very expensive, or even logistically impossible, for large cohorts. Cohort sampling techniques, where covariate information is collected for all individuals who fail (“cases”), but only for a sample of the individuals who do not fail (“controls”), then offer useful alternatives that may drastically reduce the resources that need to be allocated to a study. Further, as most of the statistical information is contained in the cases, such studies may still 330be sufficient to give reliable answers to the questions of interest. There are two important classes of cohort sampling designs: nested case-control sampling and case-cohort sampling. An overview of the two types of sampling designs is given in Chapter 16, and in Chapter 17 a detailed study of case-cohort sampling is provided. In this chapter we focus on nested case-control sampling.
