In his PhD studies in the 1980s Jonathan Shepherd matched police data with hospital emergency department (ED) records of people injured in violence. Startlingly, three quarters of violent offences which led to ED treatment were not represented in police records. This finding, since confirmed in several Western countries, led him to conceive, develop and, in Cardiff University’s violence research group which he founded, evaluate new ways to measure and prevent violence. These innovations have established violence as a problem which can be understood and tackled from a public health as well as a crime and justice perspective. Here, in a year dominated by COVID-19, he summarises the new insights gathered in his group’s 20th annual report on violence in England and Wales.
On the face of it, social distancing should decrease injury caused by violence, especially in a country where fists and feet are most commonly used as weapons. Outside the home that is. Inside the home, where people are confined during lockdowns and where weapons – knives for example – are freely available, the story might be very different.
The National Violence Surveillance Network (NVSN) of hospitals, established by Cardiff University’s violence research group in 2000, provides a unique perspective of the impact of COVID-19 restrictions and their easing. In contrast with previous years, the other reliable measure of violence, the Crime Survey for England and Wales, has not been possible because it relies on face-to-face interviews. Police records, long recognised by the Office for National Statistics as sensitive to a range of reporting and recording biases, represent a measure of police activity rather than a reliable violence measure. ED data, a valid and reliable measure of physical harm, reflect violence serious enough to result in hospital treatment and are not a measure of psychological and social violence harms.
To serve these measurement and prevention purposes, new items needed to be included in the standard ED data set, on violence location, perpetrator numbers (more than one suggests gang violence) and their relationship with the injured, and weapons. First introduced in Cardiff in the mid-1990s, these “Cardiff Model” data were subsequently codified and published by NHS Digital and then included in the new Emergency Care Data Set implemented in every ED in England in 2017.
The principal findings, from 133 EDs in Wales and England in 2020 were:
- 119,111 people attended emergency units in England and Wales for treatment of violence-related injuries; 56,653 and 193,922 fewer attendances compared to 2019 (a 32% reduction) and 2010 (a 62% reduction).
- Every COVID-19 related national lockdown, tightening and and easing of restrictions had a significant effect on violence levels.
- Violent injury among males and females declined by 33% and 29% respectively in 2020, compared to 2019; the biggest falls since NVSN records began in 2001.
- Serious violence affecting all age groups decreased in 2020 compared to the previous year; falls among children (0–10-year-olds; down 66%), adolescents (11–17-year-olds; down 37%), young adults (18–30-year-olds; down 36%), those aged 51 years and over (down 30%) and those aged 31-50 years (down 23%).
- Those at highest risk of violence-related injury in 2020 were males (2.7 per 1,000 population: more than twice the risk for females) and those aged 18-30 (4.7 per 1,000 population).
Detailed examination in Cardiff showed that an overall reduction reflected a substantial fall in violence outside the home, not in the home where levels did not change. Importantly, people who would be expected to be injured outside the home, including with a weapon, were not injured after lockdowns were imposed. This says something about the comparative safety of the average home compared with the safety of environments outside, especially night-time economies where most violent injury is sustained. Despite the availability of weapons in the home, they were used comparatively less often than outside.
The policy implications include that violence prevention needs to continue to be focussed on night-time economies; pubs, clubs and bars, and the streets where they are concentrated. Effective solutions already identified – real-time public space CCTV surveillance, limitations on alcohol availability, minimum alcohol prices, purposeful data sharing and analysis, and precisely targeted policing, for example, need constant attention. As trends in 2020 demonstrate, violence can increase as rapidly as it decreases.
The substantial reduction in numbers of injured children treated in EDs associated with COVID restrictions also underlines the comparative safety of the home and suggests that their physical safety outside can be improved further.
Multi-agency prevention, if it’s evidence based, has a great deal to offer. As the response to the COVID epidemic has taught us, this has its challenges and needs to be led from the top of government. The “Public Health Duty” currently before parliament for specific agencies, including the NHS, to collaborate to prevent serious violence is an important next step. Encouragingly, the UK government was persuaded to introduce this new legislation on the basis of its impact assessment which largely relied on evaluations of the effectiveness and cost benefit of the Cardiff Model – the use of specific ED data to locate and target temporal and geographic concentrations of violence.
This approach is needed more widely too, for example in the United States where, in Georgia, only 10% of violence leading to ED treatment is known to police. In the wake of unrest in Ferguson, Missouri following the shooting of Michael Brown by a police officer there, in 2014 President Obama set up his Task Force on 21st Century Policing. This acknowledged that joint problem solving reduces crime and that “Keeping the community and all key stakeholders informed…can build trust and increase collaboration.” Recommendations included the use of multidisciplinary teams and the appointment of new task forces. But little has changed, as further lethal policing and the Black Lives Matter movement have demonstrated.
The Cardiff Model, first implemented in 2001 and refined since, provides a template for this collaborative problem solving and the U.S. Centres for Disease Control and Prevention has published its toolkit to drive implementation. The violence prevention boards at the centre of the Model, replicated and expanded in the 18 new violence reduction units in England and Wales, act on ED as well as police data, and engender interagency accountability for violence prevention, including police accountability.
Professor Jonathan Shepherd, Crime and Security Research Institute, Cardiff University ShepherdJP@cardiff.ac.uk