Epidemiological Study Designs Applied to Driving Safety: Definitions and Examples
Abstract Four major epidemiological study designs are reviewed: cohort, case-control, case-cohort, and case-crossover. In the medical field, these study designs and their analysis methods are commonly used to estimate the effect of exposure to a disease on an outcome (such as death). The formal epidemiological definition of each design in the medical field is here translated into the context of real-world and naturalistic driving safety studies. For example, instead of an outcome of death, the outcome becomes a crash or other safety-relevant event. Instead of exposure to a disease, the exposure becomes a driver activity such as a secondary task (e.g., talking on a cell phone), a driver impairment (e.g., drunk or drugged), or a driver behavior error (e.g., speeding). The effect size measures of the exposure on the outcome include the rate ratio, risk ratio, and odds ratio.