The impact of COVID-19 and associated public health restrictions on trends in police-recorded violence in an English police force area

Abstract

Abstract Background The COVID-19 pandemic, and associated public health measures, had a marked impact on a number of health and wellbeing outcomes, including alcohol use and violence. Current literature presents a mixed view of the impact of the pandemic on violence trends. The current study utilises police offence data from a region of northern England to examine the impact of lockdowns, and subsequent relaxation of restrictions, on trends in violent offences. Methods Time series analyses using seasonal auto-regressive integrated moving average (SARIMA) modelling was used to investigate the impacts of the COVID-19 public health measures on weekly offence trends from April 1 2018 to March 20 2021. Additionally, pre-pandemic data were used to forecast expected trends had the pandemic not occurred. These expected trends were then compared to actual data to determine if the average levels of violence were outside the forecasted expectations. Overall violence and six subtypes (violence with and without injury, sexual offences including rape, domestic violence, and alcohol-related violent offences) were examined. Results Overall, the observed trend in police recorded violent offences demonstrated fluctuating patterns in line with commencement and easing of public health restrictions. That is, offence numbers declined during lockdowns and increased after relaxation of restrictions. However, the majority of observed values fell within the expected range. This broad pattern was also found for subtypes of violent offences. Conclusions While violent crime trends demonstrated fluctuations with lockdowns, and subsequent easing of restrictions, these changes were not demonstrably larger than expected trends within this English region, suggesting that a sustained amplification in violence was not observed within this data. However, it is important to acknowledge the high levels of violence reported in this region across the study period, which should be used as a key driver for investing in long-term approaches to violence prevention. Given the extent of unreported violence generally, and that victims/survivors may come into contact with other support services (without reporting to the police), it is vital that policy and practice decisions take a holistic approach, considering a broad range of data sources.

Keywords

• Violence
• Public health
• Prevention
• Trends
• ARIMA modelling
• Covid-19 pandemic