Comparisons among countries with regard to the level of alcohol involvement in fatal crashes are not simple or straightforward. Officially reported alcohol related crash rates are subject to major differences in measurement and reporting methodology, which can make comparisons inaccurate. The reported rates are also prone to error. It is highly unlikely that some of these reported rates are accurate reflections of what the rates would be if measured using methods similar to those used in the United States. For example, the officially reported rate in Sweden of 3.3% is based on police reports of alcohol involvement. Because of the nature of duties of police officers at the scene of fatal crashes, they frequently are not in a position to judge or may not have the knowledge or experience to determine whether alcohol was involved. Autopsies carried out on all fatally injured drivers in Sweden find a rate of 18% alcohol involvement (Laurell 1999). This discrepancy illustrates some of the serious reporting and measurement problems that may distort alcohol-related fatality rates and make comparisons across countries difficult and possibly misleading.
Further work is needed to collect and interpret information on laws, alcohol-related crashes, and data quality. It is possible that a computational algorithm could be constructed to correct for some of the differences in definitions of alcohol-related fatalities. This sort of correction, however, probably would not overcome some of the error and underreporting that undoubtedly occurs.