BACKGROUND It is estimated that by 2020, road traffic crashes will have moved from ninth to third in the world ranking of burden of disease, as measured in disability adjusted life years. The identification of effective strategies for the prevention of road traffic injuries is of global public health importance. Measures aimed at reducing traffic speed are considered essential to preventing road injuries; the use of speed enforcement detection devices (including speed cameras and radar and laser devices) is one such measure. OBJECTIVE To assess whether the use of speed enforcement detection devices (SEDs) reduces the incidence of speeding, road traffic crashes, injuries and deaths. METHODS We searched the Cochrane Injuries Group's Specialised Register, CENTRAL, MEDLINE, EMBASE, Science (and Social Science) Citation Index, TRANSPORT, PsycINFO, CINAHL, EconLit. We searched the websites of road safety and motoring associations, as well as general internet searches. We handsearched selected journals and conference proceedings, and contacted experts in the field. The searches were conducted during May to November 2004. METHODS Randomised controlled trials and controlled before-after studies that assessed the impact of speed enforcement detection devices on speeding, road crashes, injuries and deaths were eligible for inclusion. For studies involving co-interventions, SEDs had to be the major intervention focus of the study to be eligible. METHODS We independently screened search results, assessed studies for inclusion, extracted data and assessed methodological quality. Due to variability between and within included studies, a pooled analysis was not appropriate. RESULTS No randomised controlled trials were identified. Twenty-six studies met the inclusion criteria, of which 22 were controlled before-after trials incorporating a distinct control or comparison group(s) and four were interrupted time series designs with a comparison group(s). Fourteen studies reported speed and crash outcomes, seven reported crash outcomes only and five reported speed outcomes only. All but one study reported an absolute reduction in pre/post average speeds. A pre/post reduction in the proportion of speeding vehicles ranged across studies from 5% to 70% depending on the speed threshold set. Pre/post reductions of 50% to 65% were reported in the proportion of speeding vehicles travelling >15 km/h over the speed limit. Compared with controls, the relative improvement was from 1% to 15% for average speed and from 14% to 65% for percent speeding. All studies reporting crash outcomes reported an absolute pre/post reduction in all crashes and injury related crashes. In the vicinity of camera sites these pre/post reductions ranged from 14% to 72% for all crashes, 8% to 46% for injury crashes, and 40% to 45% for crashes resulting in fatalities or serious injuries. More generalised effects over wider areas showed an absolute pre/post crash reduction ranging from 9% to 35%, 7% to 30% for all injury crashes and 13% to 58% for crashes resulting in fatalities alone, or in combination with serious injuries. The studies of longer duration showed that these positive trends were either maintained or improved with time. Compared with controls, the relative improvement in pre/post crash numbers resulting in any type of injury ranged from 5% to 36%. CONCLUSIONS Despite the methodological limitations of the studies reviewed, the consistency of reported positive reductions in speed and crash outcomes across all studies suggest that SEDs are a promising intervention for reducing the number of road traffic injuries and deaths. More studies of a scientifically rigorous nature are necessary to provide a stronger evidence base that these interventions are worthwhile. There is a need for international harmonisation of data collection methods, including standards on how best to measure speeds and collect crash data, over lengthy intervention and follow-up periods, as well as some consensus as to the expression of outcomes in studies, so that studies can be compared.