Rare illnesses with long latency periods are often studied using a case-control design. Adjustment for confounding factors is then often achieved using matching of cases and controls on known confounders. However, the analysis of matched data requires specific statistical methods, and this is not always recognised in published papers.
Niven et al. reviewed 37 matched case-control studies in order to evaluate the proportion studies that used correct statistical methods. Of these 37 studies, only 16 (43%) were analyzed with proper statistical techniques. Correctly analysed studied were more likely to have cases with cancer and cardiovascular disease (10/16 or 63%) than others (5/21 or 24%). They were also more likely to have been published in a journal with higher impact factor. A majority (12/16 or 69%) were in the top 100 as compared to journals with lower impact factor (1/21 or 5%).
The authors conclude that their results raise concern for published findings from matched case-control studies.
Niven DJ, Berthiaume LR, Fick GH, Laupland KB. Matched case-control studies: a review of reported statistical methodology. Clin Epidemiol 2012;4:99–110.