Economics
Pregnant women are involved in more highway accidents

It may be dangerous to drive a car during pregnancy, especially in the first months. This is what different studies carried out in the past 15 years in Quebec and Norway seem to conclude.
"The correlation is statistically strong-in other words far from zero-but the effect of pregnancy on the number of accidents remains small," notes Professor Marc Gaudry, director of Agora Jules Dupuit (AJD), a transportation economics research group at Université de Montréal, who was the first to identify this factor in 1984. Using algorithms that correlate several dozen factors, the researcher observed a statistically significant correlation between the number of pregnancies and the frequency and seriousness of highway accidents. The model developed at that time, called DRAG-1 (standing for highway demand, accident and their severity)-was applied in a second study of this problem in 1991, which led to the same conclusions. Then, in 1997, a team in Norway led by Dr. Lasse Fridstroem also established a significant correlation with even more accurate data. "Based on these results, the Norwegian government launched in 2000 a review of all highway accidents involving women that had taken place in the past 25 years in order to check whether they were pregnant," added Mr. Gaudry. These new Norwegian results are expected to be made available by the Transport Economics Institute (TOI) of Oslo this autumn.

The DRAG-1 model revealed another surprising correlation that had never been suspected before: low alcohol consumption is associated with a lower risk of accident than abstinence. Once again, care has to be taken when making this analysis, since it is the same kind of indirect relationship: when alcohol consumption per person increases in a given population, the number of fatal accidents and the number of deaths per accident both decrease, while the number of less serious accidents increases. "Our model gives us a new way of understanding the problem of highway accidents, by considering the frequency and seriousness of accidents separately, as these two elements are not influenced by the same factors. If it snows, for example, the number of accidents increases, but the accidents are less serious than accidents that occur in good weather. On the other hand, the frequency decreases at higher speeds, because drivers are often more alert, although their accidents are more serious." The algorithms developed by Marc Gaudry show a similar correlation with alcohol consumption, in particular consumption of wine. "Initially, we thought that our model was wrong," he admits. But similar results were obtained with a second version of DRAG, and then by teams in New Zealand, Norway and Germany, as well as by an empirical study in the United States. According to Gaudry, the explanation is as follows: small doses of alcohol may produce a calming effect that reduces aggressiveness and tends to make for slower driving. The problem is to determine what is meant by "small dose." These results are described in a book, published in November 2000 by Elsevier Science, Oxford, entitled "Structural Road Accident Models: the International DRAG Family".

Researcher : Marc Gaudry
Phone:514 343-7284
Funding : Société de l'assurance automobile du Québec (SAAQ), Transport Canada, the Alexander von Humboldt Foundation (AvH) of Germany and the Conseil National de la Recherche Scientifique (CNRS) of France