We examined the relation between the speed of passenger cars and risk
of involvement in a severe crash, in an urban setting, using a case-co
ntrol study. ''Cases'' were 45 vehicles involved in severe crashes in
the Adelaide metropolitan area; we determined their pre-crash speeds u
sing accident reconstruction techniques. For each case, we measured th
e speeds of 10 controls using an amphometer; controls were cars not in
volved in crashes that passed through the crash location at the same t
ime of day, day of week, and season. We found that the risk of involve
ment in a severe crash increased as vehicle speed increased. In partic
ular, within 60 km per hour zones, compared with vehicles traveling at
about the posted limit, vehicles traveling at 75-84 km per hour had a
n odds ratio of 7.8 [95% confidence interval (CI) = 1.4-38.8] for a se
vere crash, whereas vehicles with speeds in excess of 84 km per hour h
ad an odds ratio of 39.0 (95% CI = 9.3-170.5).