This statement summarises the current status of the epidemiological approach in studies of drinking and driving. Descriptive epidemiology attempts to define the magnitude and characteristics of the problem, primarily to help to set priorities. It also guides analytical epidemiology, which attempts to identify high-risk groups and/or conditions, and uses the risk factors approach as its principal tool. The work of the Vermont Alcohol Research Center, USA provides an excellent example of how and why findings, based on epidemiological and experimental approaches, should be linked. The convergence between these two approaches has provided the best-established facts about drinking and driving, and will be essential for achieving a complete understanding of this problem. Unfortunately, most of the worldwide epidemiological research on drinking and driving is of much lower quality than that of the Center. Even though significant progress has been made during recent decades, much more progress will be needed and several challenges must be faced. Most epidemiological research still relies on secondary data, such as police reports and coroners' reports, with all their problems and limitations, instead of on primary data, generated under rigorous control by researchers to answer researchers' questions.
Abstract