Intended for healthcare professionals


Using injury data for violence prevention

BMJ 2000; 321 doi: (Published 16 December 2000) Cite this as: BMJ 2000;321:1481

Government proposal is an important step towards safer communities

  1. Jonathan P Shepherd, professor of oral and maxillofacial surgery (shepherdjp{at},
  2. Vaseekaran Sivarajasingam, lecturer,
  3. Frederick P Rivara, professor (fpr{at}
  1. Violence Research Group, Department of Oral and Maxillofacial Surgery, University of Wales College of Medicine, Cardiff CF14 4XY
  2. Harborview Injury Prevention and Research Center, University of Washington School of Medicine, 325 Ninth Avenue, Seattle, WA 98104-2499, USA

    The UK government has indicated that its plans to tackle community violence will be based not just on information about offences but also on injury data derived from emergency departments.1 This stems from evidence both nationally and internationally of the substantial extent to which violence that results in injury is not investigated by the police. 2 3

    In the 11 industrialised countries which took part in the 1996 international victimisation survey the median reporting rate of violent offences was 39%, ranging from 18% to 51%.4 In the UK only about 25% to 50% of offences which lead to treatment in emergency departments appear in police records, 5 6 a proportion consistent with the findings of biennial British crime surveys, which allow comparison of householders' accounts of crime with police records. One study in the US found that only 54% of assaults treated in an emergency department were documented by the police.7 A study of emergency department attendees carried out in Bristol showed that assaults on men and assaults occurring in bars, nightclubs, and public streets were less likely to be recorded by the police than assaults on women and assaults in other locations.5 This evidence, together with evidence of differential rates according to day and time of week, explain why local and regional policing and other crime prevention efforts need to focus on injury as well as offending. Systematic comparisons of emergency department and police data have not yet been done in the US, but rates of injury inflicted by intimates were found to be four times higher than rates reported to the police, according to the national crime victimisation survey.8 Importantly, comparisons have shown that even offences in which very serious injury is sustained may not be recorded. One study found that police recording could not be predicted on the basis of injury severity scores.9

    The reasons for this proposal are sixfold.1-3 Firstly, injury data provide a measure of serious violence which is independent of police measures—which are often inaccurate or incomplete. Secondly, recording injuries treated in emergency departments has the potential for largely complete coverage of serious community violence. Thirdly, such recording provides local information of importance to the police—location, time, weapon, type of incident, relationship with attacker—that will help them respond. Fourthly, injury data would provide a new performance indicator of policing at police force level. Fifthly, injury data can include outcome information on the injured victim, which is currently lacking in police reports. Finally, injury data provide a set of measures which are compatible with other data sources—in this case the British crime survey.

    This approach has been piloted and evaluated in South Wales by the Cardiff violence prevention group, a multiagency task force of emergency departments, police, local authority, voluntary sector, and mental health and legal practitioners which is cited in the 1998 Crime and Disorder Act and a recent government review of crime statistics as an example of good practice. 1 10 Non-confidential, aggregate data on hotspots of violence where injury is sustained are shared with the police each month. Among other locations, this has focused policing on particular licensed premises, the principle being to make the management of public alcohol consumption responsive to injury data from local hospitals.

    The proposal to use injury data from emergency departments builds on the legislative framework set out in the 1998 Crime and Disorder Act, which puts health authorities, the police, and local authorities at the centre of community crime prevention. In particular, it gives emergency departments opportunities to share data with other agencies, as well as treat people injured in assault. As well as helping locally, this approach has identified and led to modifications of one of the most commonly used weapons in assaults in the UK, bar glasses.11 Existing emergency department data have already been used to evaluate the effect on violence of urban closed circuit television.12

    This integrated approach has been used in other countries and for other types of injuries. Combined police and emergency department data have been used to develop a comprehensive system for tracking weapon related injuries in Massachusetts.13 The under-reporting of motor vehicle crashes to the police and the different data available from health and police records14 have led to the development of integrated record systems in the US, such as the fatality analysis reporting system.

    What does this mean for emergency department and primary care physicians? Firstly, it emphasises that health care is often the only public service which knows about many violent offences. Secondly, it will mean the routine recording of core data, usually by emergency department reception staff when the victim first attends, and transfer of aggregate data to the police. Recent substantial investment in information technology in emergency departments facilities makes all this possible, but an important prerequisite is confidential patient registration in emergency departments. Evaluations should be focused not just on injury outcomes but also on process outcomes which local communities can adopt to ensure efficient and effective service coordination.

    Every surveillance system has its problems, however. For example, attendance at an emergency department for minor injuries is known to depend on the proximity of the incident to a hospital.15 However, a medical perspective of violence combined with the traditional police perspective has already begun to reap rewards. Based on a sound ethical framework to protect the interests of patients, it is a rational step towards safer and more just communities.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    14. 14.
    15. 15.