case-control study

The Centre for Evidence-Based Medicine at the University of Oxford describes a case-control study as one that takes patients with a certain outcome or disease, together with an appropriate group of controls without it, and examines what they were exposed to beforehand. The direction is the defining feature. A cohort study records an exposure and waits to see what happens; a case-control study begins with what happened and reaches back for the exposure, using recall or existing records.

That reversal is what makes the design useful. The centre lists it as quick and cheap, and as the only feasible method for very rare disorders or for those with a long lag between exposure and outcome, since a cohort study of a rare disease would need to follow enormous numbers of people for years to accumulate enough cases. It also needs fewer participants than a cross-sectional study to answer the same question.

The costs come from the same reversal. Establishing exposure by recall or records is open to recall bias, where people who already have a diagnosis search their history differently from people who do not. Selecting the control group is described as difficult, and a poorly chosen one distorts the comparison from the start. Confounders remain, as in any observational design. The centre places case-control studies below cohort studies in the hierarchy of evidence, because starting from the outcome makes the order of events harder to establish, and below randomised controlled trials, where allocation by chance removes the problem at the design stage.