Among 7,634 Montreal drivers, interviewed in 1973-74 by telephone about medical and related factors, and about driving habits, 347 had had at least one accident (the 'index accident') causing injury or death, in the period 1 January 1973 to 31 March 1976. These 'cases' were compared with 347 controls, one for each case, closely matched for sex, age, reported annual mileage and type of driver's permit. No control had an accident causing injury or death.

There was no suggestion of differences between cases and controls in answers to nine questions concerning exercise, headaches, physical defects, diabetes, tranquillisers, prescribed drugs, other medicaments, whether employed at the time of interview, and whether they ever drove after drinking. The use of seat belts on the highways was slightly less common in cases than in controls, but there were no differences in usage in towns. Cases included higher proportions (a) who reported smoking while driving, (b) who worked irregular shifts, and (c) who were overweight.

Compared with the chance of a control suffering any accident, a case had a relative risk greater than 2 of having an accident in addition to the index accident in the same period.