RESUMO
BACKGROUND: The aim of this study was to better evaluate the role of alcohol drinking in fatalities linked to road traffic accidents. METHODS: The data of accidents were collected by a French official agency from police records, including many variables, among which was a blood alcohol test. They were analyzed in a descriptive way and toward a logistic regression. This exhaustive database comprised all of the 500,961 accidents with casualties that involved less than three vehicles (28,506 fatal accidents) recorded in France during a 52 month period (September 1995 to December 1999). The results of the alcohol tests were known in 78.7 of the drivers. RESULTS: The blood alcohol concentration was over the legal limit (0.50 g/L in France) in 9.8% of the accidents with casualties overall. Considering only fatal accidents, the rate of positive alcohol test in drivers was approximately 31.5%. This rate varied depending on the period and the type of accident, raising up to 71.2% in single-vehicle accidents (loss of control) at night during the weekend. The percentage of positive alcohol tests also dramatically increased following the number of fatalities per accident (87.5% in single-vehicle accidents during weekend nights involving three or more killed). The logistic regression in single-vehicle accident shows that the higher odds ratios concern the positive blood alcohol test (OR = 4.19), clearly overwhelming the other precipitating factors of accidents (age of driver, meteorological conditions, time of day, and other factors). CONCLUSIONS: Drinking alcohol before driving is a well known factor of accidents. We clearly demonstrate here that it is the main factor leading to deaths linked to road traffic accidents in France. The results are strengthened, and some analyses are allowed, by the exceptional features of our database. The authors emphasize the need for prevention measures.
