Methods of evaluating the effectiveness of motor vehicle safety technology: Databases and statistical techniques for studies of evolving frontal air bag designs

Automobile manufacturers introduced a variety of vehicle safety technologies during the past 20 years, including frontal air bags, side air bags, electronic stability control, antilock brakes, seat belt pretensioners, and daytime running lights. Unlike drugs and medical devices, whose efficacy is tested in randomized controlled clinical trials before being approved for patient use, new vehicle safety technologies are installed in vehicles in the absence of controlled intervention trials. Automakers perform crash tests using instrumented dummies and other tests, but experiments do not necessarily replicate actual driving conditions. To determine whether vehicle safety technology works as intended, real-world crash and fatality data must be analyzed to compare the incidence of deaths, non-fatal injuries, or crashes in vehicles with and without the technology. Redesigned frontal air bags have been controversial: although there is consistent evidence that fewer children are being killed by newer airbags, some fear these redesigns are compromising protection of large unbelted occupants, primarily men. Using examples from studies of newer frontal air bags, study designs for evaluation of new technology, together with their strengths and drawbacks, will be discussed. Findings concerning newer air bag designs will be presented from unpublished and published analyses of matched-pair cohorts, deaths per registered vehicle, deaths per crash-involved occupant, and non-fatal injuries in towed vehicles. Our research sheds light on the question of whether newer air bags are providing equivalent or improved protection to drivers and right-front passengers, including unbelted men, relative to first-generation air bags.