Road traffic accidents (RTAs) are the leading cause of mortality in adolescents and young adults in all regions of the world and rank among the top five causes of death (Feachem et al. 1992; World Bank 1993). Traffic-related fatality accounts for 10% of all deaths in the 5-44 year age group in developing countries and is increasingly becoming an important public health problem (Downing et al. 1991). In Kenya, deaths from RTAs increased by 43% over the last 10 years, and alcohol is suspected to be an important contributing factor. There are, however, no specific drink driving countermeasures or legislation setting a blood alcohol concentration (BAC) limit for drivers, and the prevalence of drinking and driving is unknown. Roadside surveys have been used widely for collecting BAC data from the driving population since the Grand Rapids Study in 1964 (Borkenstein et al. 1964). A standard roadside survey requires specific facilities, a high level of organization and a significant amount of resources. The study design involves sampling drivers from the main traffic stream into specially prepared, highly visible off-road survey sites fitted with a portacabin or a designated research vehicle where interviews and breathtests can be done in privacy, and is most suited for highways in rural locations with ample parking spaces. Permanent police roadblocks, usually placed along inter-urban roads, may also be used as sampling sites, but they present serious biases, in terms of their locations and vehicle selection criteria. Alternatively, a method of sampling drivers stopping at a red traffic signal in a metropolitan area has been reported (McLean and Holubowycz 1981). The procedure does not involve working with the police, but it presumes the availability of a network of signalized intersections in the city, and is therefore not suitable for urban centres (like Eldoret) without such facilities. This paper describes how traffic police checks were used in conjunction with a roadside survey in Eldoret, a town with a population of 178,000, located in western Kenya. The purpose of the study was to obtain drivers' BAC data that can be used as reference in formulating drink driving policy in Kenya, as well as to assess the feasibility of random breath testing as a strategy for implementing such a policy.
Abstract