Original article
Interpretation and bias in case-crossover studies

https://doi.org/10.1016/S0895-4356(97)00196-0Get rights and content

Abstract

The case-crossover design is an innovative epidemiologic technique with distinct strengths and limitations. We review the fundamental logic of this self-matching non-randomized design and direct attention to 15 concerns related to the available data, unavailable data, analytic technique, quantitative statistics, and etiologic model. Implications for each concern are discussed in the context of a recent report on whether cellular telephone calls are associated with an increased risk of a motor vehicle collision. We suggest that an understanding of the case-crossover design may help investigators explore selected questions in behavioral medical research.

References (31)

  • TA Louis et al.

    Crossover and self-contained designs in clinical research

    N Engl J Med

    (1984)
  • BW Brown

    The crossover experiment for clinical trials

    Biometrics

    (1980)
  • DT Campbell et al.
  • MA Mittleman et al.

    Triggering of acute myocardial infarction by heavy physical exertion

    N Engl J Med

    (1993)
  • MA Mittleman et al.

    Triggering of acute myocardial infarction onset by episodes of anger

    Circulation

    (1995)
  • Cited by (62)

    • Matters of the Heart: Case-Crossover Analysis of Myocardial Infarction on Valentine's Day

      2021, CJC Open
      Citation Excerpt :

      Case-crossover studies are suitable for transient exposures such as Valentine's Day, and for acute outcomes, such as myocardial infarction.5-7 The case-crossover study is a variant of the case-control study.8,9 In a case-crossover study, the day of the event is matched with other nearby days, and the characteristics of case days are compared with those of control days.

    • Full moon and traffic accident-related emergency ambulance transport: A nationwide case-crossover study

      2018, Science of the Total Environment
      Citation Excerpt :

      Second, although our study demonstrates a significant increase in the risk of traffic accidents on full moon nights, there may be concerns regarding the interpretation of the magnitude of the observed effect. Given that most car occupants and motorcyclists are only likely to look at a full moon for a very transient period of time, distractions due to a full moon should result in only a small rise in fatal crashes (Redelmeier and Tibshirani, 1997). Further, a full moon is inert, distant, and familiar, suggesting that it should have less of an effect on road accidents compared to a combination of other roadway distractions (Klauer et al., 2014).

    • Climate and the eye: Case-crossover analysis of retinal detachment after exposure to ambient heat

      2017, Environmental Research
      Citation Excerpt :

      Case days are afterwards compared with control days for key differences in temperature exposure. Because patients with retinal detachment are their own controls, adjustment for unmeasured confounders such as age, socioeconomic status, myopia or history of intraocular interventions is automatic, as long as these do not vary between case and control days over time (Levy et al., 2001; Maclure and Mittleman, 2000; Redelmeier and Tibshirani, 1997). In this study, we defined case days as the date of treatment for retinal detachment, and used a time-stratified approach to select control days.

    • The exposure-crossover design is a new method for studying sustained changes in recurrent events

      2013, Journal of Clinical Epidemiology
      Citation Excerpt :

      Crossover rules are not usually a problem unless the intervention needs to be applied repeatedly. Faulty data are always a concern, particularly if the outcome event is ascertained by patient recall rather than routine well-ordered monitoring [57]. Inappropriate statistical analyses for repeated outcomes are always a concern, although the graphical display of aggregate rates can help avoid major mistakes.

    View all citing articles on Scopus
    View full text