Prevalence study: Main illicit psychoactive substances among all drivers involved in fatal road crashes in France : DRUID - Driving under the Influence of Drugs, Alcohol and Medicines. 6th Framework Program

INTRODUCTION. In 1999, in France, before considering change in the drug legislation, the government wished to know more on the effect of illicit drugs on the risk of road crashes. A systematic screening of illicit drug on all drivers involved in fatal crashes, was hence made compulsory. We estimate prevalences of alcohol, cannabis, amphetamines, cocaine, and opiates. We provide them for drivers of all vehicles types, as well as for car drivers only. We also provide estimates of prevalences in the general population through the careful selection of a control group.
MATERIAL AND METHOD. A total of 10,728 drivers were involved in fatal road crashes in France between October 2001 and September 2003. After exclusion of drivers with missing data on alcohol, cannabis, age, or with age below 18, some 10,519 drivers remain in the dataset. The subset of car drivers contains 7,455 of them. Drivers involved in fatal crashes include killed drivers, injured and non-injured drivers. […]
RESULTS. In all drivers involved in fatal crashes, alcohol prevalence is estimated at 25.9%, THC prevalence at 6.9%, amphetamines at 0.6%, cocaine at 0.5% and opiates at 0.9%. prevalences of alcohol and cannabis are highest in moped users. Prevalences are higher in men than in women, at all ages. In men, prevalence of cannabis is higher at younger age, whereas in women prevalence of alcohol is highest in the 35-49 year old group. Among drivers who are positive to cannabis, 47% are (also) positive to alcohol. Prevalence of cannabis is 8.7% in drivers responsible for the crash and 3.8% in nonresponsible ones. Prevalence of alcohol is 35.6% in drivers responsible for the crash and 9.4% in non-responsible ones. In the control group, prevalence of cannabis is 2.8%, prevalence of alcohol is 5.0%.
DISCUSSION. For all the substances except cannabis, the commonly defined DRUID thresholds are higher than the French thresholds in use in 2001-2003. We hence very likely miss a high proportion of those between the two thresholds. This means that the estimated prevalences should be considered as lower bounds.
Prevalence of alcohol in our control group using the French threshold (2.7%) is close to the prevalence of the whole driving population, estimated in another study (2.5%); this supports the choice we made for the control group and hence the generalisation of its prevalences to the whole driving population.
Prevalences of alcohol and cannabis are higher in the drivers responsible for the fatal crashes compared to the non-responsible drivers or to the control group. It is only a descriptive and univariate analysis; this will be better analysed in the responsibility study.