A Proposal for a New Design: The Exposure–Crossover Design

Redelmeier proposed the "exposure–crossover design" to assess individual changes due to exposure, rather than average differences as in randomized trials. This design was used to test hypotheses related to motor vehicle accidents, such as the impact of a crash on future crash risk or the influence of becoming a mother on crash risk. It utilized a large database from Ontario, Canada, tracking patients involved in motor vehicle incidents over 5 years. This design was a type of interrupted time series and regression discontinuity design, which were less common in health research but expected to grow with the availability of large databases.