The case-crossover study is a relatively unknown way of identifying short-term transient risk factors for acute-onset diseases. In patients with the disease of interest, the frequency of exposure to a certain risk factor is compared between two time periods. If the exposure is more common in the period directly preceding disease onset than in an earlier period, the control period, it is likely that the exposure contributes to the development of the disease. The problem of confounding is minimized in case-crossover studies since the patient serves as his or her own control. A potential disadvantage is that sufficient biological knowledge of the clinical picture is needed to provide a good estimate of the risk period. As administrative databases are now commonly used for research purposes, future use of case-crossover methods is likely to increase. We illustrate the case-crossover study with the question of whether antibiotic use increases the risk of unwanted pregnancy in women who use the contraceptive pill.