Nighttime assaults: using a national emergency department monitoring system to predict occurrence, target prevention and plan services

Background: Emergency department (ED) data have the potential to provide critical intelligence on when violence is most likely to occur and the characteristics of those who suffer the greatest health impacts. We use a national experimental ED monitoring system to examine how it could target violence prevention interventions towards at risk communities and optimise acute responses to calendar, holiday and other celebration-related changes in nighttime assaults. Methods: A cross-sectional examination of nighttime assault presentations (6.01 pm to 6.00 am; n = 330,172) over a three-year period (31st March 2008 to 30th March 2011) to English EDs analysing changes by weekday, month, holidays, major sporting events, and demographics of those presenting. Results: Males are at greater risk of assault presentation (adjusted odds ratio [AOR] 3.14, 95% confidence intervals [CIs] 3.11-3.16; P < 0.001); with male: female ratios increasing on more violent nights. Risks peak at age 18 years. Deprived individuals have greater risks of presenting across all ages (AOR 3.87, 95% CIs 3.82-3.92; P < 0.001). Proportions of assaults from deprived communities increase midweek. Female presentations in affluent areas peak aged 20 years. By age 13, females from deprived communities exceed this peak. Presentations peak on Friday and Saturday nights and the eves of public holidays; the largest peak is on New Year's Eve. Assaults increase over summer with a nadir in January. Impacts of annual celebrations without holidays vary. Some (Halloween, Guy Fawkes and St Patrick's nights) see increased assaults while others (St George's and Valentine's Day nights) do not. Home nation World Cup football matches are associated with nearly a three times increase in midweek assault presentation. Other football and rugby events examined show no impact. The 2008 Olympics saw assaults fall. The overall calendar model strongly predicts observed presentations (R-2 = 0.918; P < 0.001). Conclusions: To date, the role of ED data has focused on helping target nightlife police activity. Its utility is much greater; capable of targeting and evaluating multi-agency life course approaches to violence prevention and optimising frontline resources. National ED data are critical for fully engaging health services in the prevention of violence.