Invited commentary: application of case-crossover methods to investigate triggers of preterm birth.

Properties of the case-crossover design have appeal for investigation of acute triggers of preterm birth. Measured and unmeasured time-invariant risk factors are controlled by design, such that maternal race, socioeconomic status, and other personal factors will not confound the exposure-preterm birth association. In this issue of the Journal, Basu et al. (Am J Epidemiol. 2010;172(10):1108-1117) apply the case-crossover approach to assess the short-term relation between ambient apparent temperature and preterm birth. Novel application of the design to preterm birth, a "fatal" event exhibiting dramatic within-subject changes in risk, merits a review of the assumptions underlying the design. Implications of the referent time periods selected and the potential for confounding by seasonal patterns of conception are discussed in this commentary. The provocative associations observed by Basu et al. between high ambient apparent temperature and preterm birth should stimulate follow-up analyses and could ultimately have important public health implications. Future research can also help delineate the relative strengths and weaknesses of different temporal analytic strategies for investigating short-term associations between various exposures and preterm birth.