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

The document proposed a new study design called the "exposure–crossover design," which focused on assessing individual changes due to exposure rather than average differences. This design was applied to study the impact of motor vehicle accidents on subsequent crash risk using a large database from Ontario, Canada, over a 5-year period with multiple 28-day segments. It was part of the interrupted time series and Regression Discontinuity designs, which were less common in health research but expected to become more prevalent with large databases. The document also discussed the reliability of assessment tools, patient adherence to treatment, and the impact of missing data in clinical studies.