[Case-control study: the importance of selecting the right controls].

Important non-experimental study designs to assess the relation between a possible determinant (e.g. a treatment, a risk factor, a diagnostic or prognostic marker) and a clinical outcome are the follow-up and the case-control study design. In case-control studies, the distribution of the determinant among subjects with the outcome (the 'cases') is compared to the distribution of this determinant in the underlying population from which the patients originate. This is typically estimated in a sample (the 'controls') of that population. The main reason to consider a case-control design instead of a follow-up design is efficiency. A case-control study can be conducted faster and for lower costs, since only the cases and controls need to be studied in detail. We illustrate the case-control study design by an example study into the relation between mobile phone use and the risk of traffic accidents, with a focus on the selection of controls.