Case-control studies in pediatric epidemiology: parent surrogates and potential pitfalls of inaccurate and selective recall.

The case-control study is quite popular as a study design for exploring associations between risk factors and disease in pediatric epidemiology. Since data concerning exposures to the child are often collected through interviews with parents or other surrogates, researchers should be aware of the opportunities for bias due to inaccurate or incomplete recall. Methods which exist for the control of this problem are presented. These include: the selection of control groups with childhood conditions of similar etiologic uncertainty as the disease being studied; collecting exposure data from both parents; collection of data from children where possible; diligent interviewer training; reviewing clinical records; and use of validity scales. Strengths and weaknesses of these strategies are discussed.