The health care system in England has been highly affected by the surge in demand due to patients afflicted by COVID-19. Yet the impact of the pandemic on the care seeking behaviour of patients and thus on Emergency department (ED) services is unknown, especially for non-COVID-19 related emergencies. In this report, we aimed to assess how the reorganisation of hospital care and admission policies to respond to the COVID-19 epidemic affected ED attendances and emergency hospital admissions.

We performed time-series analyses of present year vs historic (2015-2019) trends of ED attendances between March 12 and May 31 at two large central London hospitals part of Imperial College Healthcare NHS Trust (ICHNT) and compared these to regional and national trends. Historic attendances data to ICHNT and publicly available NHS situation reports were used to calibrate time series auto-regressive integrated moving average (ARIMA) forecasting models. We thus predicted the (conterfactual) expected number of ED attendances between March 12 (when the first public health measure leading to lock-down started in England) to May 31, 2020 (when the analysis was censored) at ICHNT, at all acute London Trusts and nationally. The forecasted trends were compared to observed data for the same periods of time. Lastly, we analysed the trends at ICHNT disaggregating by mode of arrival, distance from postcode of patient residence to hospital and primary diagnosis amongst those that were subsequently admitted to hospital and compared these data to an average for the same period of time in the years 2015 to 2019.

During the study period (January 1 to May 31, 2020) there was an overall decrease in ED attendances of 35% at ICHNT, of 50% across all London NHS Trusts and 53% nationally. For ICHNT, the decrease in attendances was mainly amongst those aged younger than 65 and those arriving by their own means (e.g. personal or public transport). Increasing distance (km) from postcode of residence to hospital was a significant predictor of reduced attendances, which could not be explained by weighted (for population numbers) mean index of multiple deprivation. Non-COVID emergency admissions to hospital after March 12 fell by 48% at ICHNT compared to previous years. This was seen across all disease areas, including acute coronary syndromes, stroke and cancer-related emergencies. The overall non-COVID-19 hospitalisation mortality risk did not differ (RR 1.13, 95%CI 0.94-1.37, p=0.19), also in comparison to previous years.

Our findings suggest emergency healthcare seeking to hospitals drastically changed amongst the population within the catchment area of ICHNT. This trend was echoed regionally and nationally, suggesting those suffering a medical emergency may not have attended other (i.e. closer-to-home) hospitals. Furthermore, our time-series analyses showed that, even after COVID-19 cases and deaths decreased (i.e. from early April), non-COVID-19 ED attendances did not increase. The impact of emergency triaging systems (e.g. 111 calls) and alternative (e.g. private hospital, chemist) health services on these trends remains unknown. However, another recent report found increased non-COVID excess deaths in the community, which may be partially explained by people experiencing an emergency and not attending health services at all. Whether those that attended ED services have done so with longer delays from the moment of emergency onset also remains unknown. National analyses into the factors causing reduced attendances to ED services and strategies to revert these negative trends are urgently needed.