Invited Commentary: Self-Control Is a Virtue

The case-crossover design, introduced in 1991 by Malcolm Maclure (Am J Epidemiol. 1991;133(2):144-153), provided a precise and powerful tool for studying short-term effects of transient triggering exposures on abrupt outcomes like myocardial infarction. The design is an example of "self-control." One compares case-time intervals that include experiences that came just before the health event with control-time intervals that capture comparable experiences more remote from the event. Methodologists have since tweaked the general approach, recognizing issues that need to be considered to guard against time-driven confounders. I discuss opportunities for possible expansion and further mining of the data from this ingenious design.