Skip to main content Skip to navigation

Short Project Descriptions

Show all news items

Tracking the Effect of the COVID-19 Pandemic on Routine Inpatient Hospital Care

Leads: Prof Kate Jolly, Mr Aneel Bhangu (Long-term Conditions), Prof Richard Lilford, Dr Jemma Mytton, Dr Samuel Watson (Meths), Prof Kamlesh Khunti (ARC EM)

Dates: January 2021 - January 2022


Since the beginning of the COVID-19 pandemic there have been concerns regarding the disruptive effects it has had on care for patients with other conditions. There is evidence that admissions for stroke and heart attacks were reduced in Northern Italy, for example, though analysis of ambulance and hospital data in the UK did not show a similar effect.

Various preventive methods have been deployed throughout the pandemic, at varying degrees, in an attempt to control the effects at minimal cost to freedom and the economy. Tracking the effects of the pandemic on health services throughout this period of change is important for both pragmatic and scientific purposes.

Pragmatically, data on non-COVID-19 activity is important for planning purposes since services will have to deal with the shortfall in capacity; for example, reduced diagnostic services for cancer will create a downstream demand. From a theoretical or scientific standpoint, the effect of the pandemic on other services is important because it will inform planning for future epidemics, given that this is unlikely to be the last time that services are disrupted by contagious disease or other catastrophes.

Policy and Practice Partners:

NHS England, NHS Confederation.

Co-Funding Partners:

University Hospitals Birmingham NHS Foundation Trust.

Aims and Objectives:

Our aim is to track the effect of the pandemic on in-patient hospital care over the pandemic in England, specifically to describe the use of services for different categories of patients over time. We then aim to relate those changes to the state of the pandemic at any one time. Many of these are semi-parametric and might include seasonal effects, long term trends, and specific daily effects (Christmas, holidays, etc).


Indicative conditions will be tracked through their associated code, selected to illustrate different organisational factors that may have differential effects - for example, we expect that surgeries where intensive care is often required (e.g. operations for cancer of the oesophagus) will be affected more than those that do not (e.g. breast biopsy/mastectomy). We will examine the difference between activity level from the previous year (2020-21) with the five-year average (2014-18) to generate crude uncertainty bounds.

Following this we will model activities looking at variation over time, for example seasonal effects. Many of these are semi-parametric and might include seasonal effects, long term trends, and specific daily effects (e.g. Christmas, holidays, etc.). We will do this overall, then relate these activities to the changing intensity of the pandemic as manifest in admissions to hospital. We will estimate the effect of different intensity levels of expected decrement in different types of activity allowing for partial pooling of information

Main Results:




Implications for Implementation:

This will help predict the un-requited demand for health care.