Case-control (or case-referent) studies compare patients with a disease or health outcome of interest (a case) with controls who do not have the disease or outcome. Their relevant past exposures are assessed retrospectively, and an estimate of relative risk (odds ratio) determined. This chapter describes the choice of study population, and the key features of selection of cases and determination of their relevant exposures. Selection of controls is a vexed subject, and the chapter considers in detail the associated issues, including matching, loss of controls from the study, and controls who...

Case-control (or case-referent) studies compare patients with a disease or health outcome of interest (a case) with controls who do not have the disease or outcome. Their relevant past exposures are assessed retrospectively, and an estimate of relative risk (odds ratio) determined. This chapter describes the choice of study population, and the key features of selection of cases and determination of their relevant exposures. Selection of controls is a vexed subject, and the chapter considers in detail the associated issues, including matching, loss of controls from the study, and controls who go on to become cases. The chapter concludes with a discussion of study size and statistical power, and the appropriate statistical techniques for estimating and interpreting odds ratio, attributable risk percent, and population attributable risk percent. The problem of interpreting an association as evidence of causation is discussed.