Geographical distribution of road traffic deaths in England and Wales: place of accident compared with place of residence.

BACKGROUND There are two sources of information on road traffic mortality in England and Wales: ONS records giving the place of residence and police records giving the place of the accident. Use of the police records has been limited by the lack of an obvious denominator to control for population at risk. This study compares the two measures and explores the implications of using the same population denominator for both. METHODS The number of road traffic deaths occurring in 403 local authority districts in England and Wales during 1995-1999 was compared to the number of deaths to residents in the same period. Both numbers were related to the expected number of deaths to residents and selected environmental risk factors using regression techniques. RESULTS Large differences were found between the number of deaths in each district and the number of deaths to residents. The expected number of deaths to residents was the strongest predictor of both observed totals. The number of deaths in the district and the number divided by expected deaths of residents were highly predictable from road accident risk factors, but the number of deaths to residents and the conventional SMR were not. CONCLUSION Information on the place of residence of road traffic accident fatalities does not show true variations in accident risk. Police records are better for this purpose. The expected number of deaths to residents estimate provides a partial but effective and unbiased control for population effects.