Analysis of Case-Control Data

Following a presentation of the main steps in the analysis of case-control data starting with editing and cleaning data and selecting continuous predictors, this chapter discusses some of the challenges related to the analysis of case-control data. It deals with common strategies of dealing with confounding such as stratified analysis, Mantel-Haenszel estimates, and multivariate analyses. The uses of these analytic approaches are also discussed in assessing interaction or effect modification. The second part of the chapter presents bivariate and multivariate methods, such as the conditional logistic regression method for the analysis of matched data. The final section presents a framework for model building in case-control analyses and discusses the interpretation and presentation of analysis results.