A87 Estimating changes in casualty risk per km of travel by age, sex and travel mode in road collisions in Great Britain, 1995-2010

Abstract

Abstract Background The risk of casualty per km of travel in road collisions can be imputed from casualty numbers and distance travelled. The established trend in Great Britain between 1948 and 2013 has been a reduction in risk of fatality of about 4.7 per cent per annum. During the period 1995-2010 the number of fatalities reduced by 49 per cent, corresponding to an annual reduction of about 4.2 per cent, though concurrently driving in the highest risk age groups (young males) also reduced. Using travel and casualty data from this period, we examined the changes in risk by age, sex and travel mode to investigate whether the reduction in aggregate risk was due to the reduction in exposure for this highest risk group. Methods The outcome measures were the risk of fatality by travel mode calculated as number of fatalities per distance travelled in that mode (fatalities per thousand million km of travel). Data on fatal casualties by mode were extracted from STATS19, the national reporting system on personal injury road collisions. Data on exposure per person were extracted from the National Travel Survey (NTS) travel diary reports and grossed up by population size. Variables recorded in and matched from each source were age and sex of casualty and travel mode. Analysis was conducted using negative binomial log-linear regression models to test for differential rates of reduction in fatality risk. Results Fatality rates for all travel modes fell from 8.43 fatalities per 109 km of travel to 3.79 fatalities per 109 km of travel during the period 1995–2010. The reduction in risk was found not differ significantly (α=0.05) across age, sex or travel mode except for the following groups. Those under 10 years of age, who had a greater reduction in risk and those between 15-17 years of age, who had the least reduction. Risk of fatality per km of travel approximately doubled for users of motor cycles, whilst risk of fatality per km of travel was approximately unchanged for users of taxi. Conclusions The reduction in risk of fatality per km of travel observed in aggregate at the national level has affected almost all groups of age, sex and travel mode approximately equally. The reduction in total number of road user fatalities has therefore been strengthened rather than caused by reduced usage in the group with greatest risk.