The protective effect of child restraint and the relative safety of front and rear seating in a population where children often travel unrestrained was assessed in a population based case-control study. The cases were all 129 children aged 0-11 years injured as car passengers in a motor vehicle accident who contacted, during 1996, one of the two major children's hospitals in Athens; emergency cases are accepted by the two hospitals on alternate days throughout the year, thus generating a random sample of children injured as car passengers. The prevalence of the studied exposures in the study base was estimated from an inspection survey comprising a random sample of 191 children of the same age who travelled in passenger cars. The survey was conducted by medical staff from our centre in collaboration with the road traffic police. Odds ratios (ORs) were calculated after adjustment for confounding factors through the Mantel-Haenszel procedure. The OR for injury was 3.3 among unrestrained children compared with restrained children (comparison essentially limited to children aged 0-4 years) and 5.0 for children seated in the front compared with those seated in the rear (comparison essentially limited among unrestrained children). Protective effect estimates derived from this analytical study suggest that in Greece about two thirds of all childhood injuries from car crashes could have been avoided through the regular use of a proper child restraint. The data also indicate that, in the absence of a child restraint system, a rear seating position conveys substantial protection and could explain the low mortality of children as car passengers in Greece, a country which is characterised by a high overall road traffic mortality as well as a high childhood accident mortality.

Keywords: child car restraints; seating positions; protective effects; road traffic injuries