Published: 24 February 2022
Identifying patients at high risk from COVID-19
At the outbreak of the COVID-19 pandemic, the NHS urgently needed a way to identify people whose health was most at risk if they caught coronavirus.
The Chief Medical Officer for England commissioned a team of leading researchers and clinicians from around the UK to create a way of predicting patient groups most at risk of serious outcomes from COVID-19 infection.
Funded by the NIHR, the research team, comprising experts from 12 institutions, collaborated to develop a new approach to identifying patients’ risk at a population level. Led by Professor Julia Hippisley-Cox, Professor of Clinical Epidemiology and General Practice at the University of Oxford, the team began their project with an analysis of a pre-existing database of more than 8 million people aged 19–100 years. The information included pseudonymised GP records, hospital records, COVID-19 test results and death registrations corresponding to the first wave of the pandemic (late January to April 2020).
The researchers planned to identify which combinations of health and personal factors put patients at greater risk of hospital admission or dying from COVID-19 infection, so they considered characteristics such as age, gender, ethnicity and body mass index (BMI). They also looked at the effects of certain treatments and medical conditions, including cardiovascular disease, diabetes, respiratory disease and cancer.
Using its findings, the team developed a clinical risk prediction model, QCovid®. It then tested how well this model predicted hospital admissions or deaths from COVID-19 infection using the pre-existing health records of a separate set of people.
QCovid® performed well in predicting patients’ outcomes, with those identified by the model to be in the top 5% for predicted risk of death accounting for approximately 76% of actual COVID-19 deaths during the study period, and people in the top 20% accounting for 94% of COVID-19 deaths. The study’s results have been published in the British Medical Journal.
Our population-based risk model has broken new ground by identifying the patients at highest risk of COVID-related death and hospital admission, so that the NHS can target resources to the most vulnerable and those most likely to benefit. It really demonstrates the value of having a cross-partnership team of multiple specialities in delivering innovative research and improvements for the healthcare system.”
Professor Julia Hippisley-Cox, Professor of Clinical Epidemiology and General Practice at the University of Oxford
The model has been independently validated by the Office for National Statistics (ONS), confirming that the model performs in the ‘excellent’ range and can accurately identify patients at highest risk from COVID-19. ONS validation is considered the gold standard of evidence and assurance, and QCovid® is thought to be the only COVID-19 risk prediction model in the world to meet this standard. The validation study’s results were published in The Lancet Digital Health.
The QCovid® model was also validated for use in Scotland using Scottish population data (EAVE II, results published in Thorax), and in Wales using the SAIL Databank (published in the International Journal of Population Data Science). Ronan Lyons, Professor of Public Health at Swansea University and contributor to the QCovid® research, commented that: “The validation of QCovid® in Wales has helped enormously in informing policy responses to those at greatest risk."
Extending COVID-19 risk assessment to all adults
At the beginning of 2021, people on the existing Shielded Patient List (SPL) - which comprises people identified as being at high risk of dying from COVID-19 based on having a single underlying disease - were prioritised for COVID-19 vaccination, in line with guidance from the Joint Committee on Vaccination and Immunisation.
Less than a year after the NIHR-funded research began, the risk prediction model was used by NHS Digital to develop the COVID-19 Population Risk Assessment. In this new assessment, NHS Digital applied the QCovid® model to NHS patient data in England, to identify people not already included on the SPL in England who might be at high risk of dying from COVID-19 infection.
Using the same assessment of combined factors, including age, BMI, specific health conditions and treatments, the COVID-19 Population Risk Assessment identified a further 1.7 million high-risk people who were added to the SPL and advised to shield. This included 820,000 adults aged 19–69 years who were prioritised for vaccination (as the over-70s in the identified high-risk group had already been prioritised). Automatically adding at-risk patients to the SPL ensured that as many patients were protected as quickly as possible.
For the first time, we are able to go even further in protecting the most vulnerable in our communities. This model is a tribute to our health and technology researchers. The model’s data-driven approach to medical risk assessment will help the NHS identify further individuals who may be at high risk from COVID-19 due to a combination of personal and health factors.
Dr Jenny Harries, former Deputy Chief Medical Officer for England and Chief Executive of the Health Security Agency
NHS Digital also used QCovid® to produce the COVID-19 Clinical Risk Assessment Tool, to help clinicians review individual patients’ risk level and add or remove them from the SPL as required.
The adoption of this risk assessment model by the NHS will play an important role in supporting clinicians and patients with conversations about COVID-19 and enable decisions to be made with a greater understanding of personal risk.
Professor Andrew Goddard, President of the Royal College of Physicians
Although QCovid® has been specifically designed to inform UK health policy and interventions to manage COVID-19 related risks, the research team have suggested that it could be implemented by other countries following their own local validation.
The QCovid® research team won the Royal College of General Practitioner’s Overall COVID research paper of the year award for its academic publication in the BMJ. The research team and wider group of collaborators who developed the COVID-19 Population Risk Assessment - which included representatives from the Department of Health and Social Care, NHS Digital and NHS England - also collectively won awards, such as the Florence Nightingale Award for Excellence in Healthcare Data Analytics.
Identifying COVID-19 risk after vaccination
As the COVID-19 vaccination programme was developed and rolled out to the adult UK population in 2021, there were also concerns that the vaccine may not effectively protect some patient groups. This would leave them at greater risk from COVID-19 infection and subsequent hospital admission or death.
To help identify and urgently protect those most at risk, the QCovid® team were commissioned by the Chief Medical Officer on behalf of the UK government to develop new risk scores to predict people’s risk of hospital admission or dying from COVID-19 after receiving either one or two doses of vaccine. They analysed a sample of over 6.9 million vaccinated adults, of whom 5.2 million had received both vaccine doses, which was representative of the UK population.
The research, published in the BMJ, recorded 1,929 COVID-19-related hospital admissions and 2,031 COVID-19 deaths within the sample, of which 71 admissions and 81 deaths occurred at least 14 days after the second vaccine dose.
The team’s findings indicated that people receiving treatment for cancer or autoimmune disorders, care home residents and those with HIV/AIDS or neurological disorders were among those who remained at higher risk of hospitalisation or death from COVID-19 after one or two vaccine doses. The study did not distinguish between the type of vaccine received.
This enormous national study of over 5 million people vaccinated with two doses across the UK has found that a small minority of people remain at risk of COVID-19 hospitalisation and death. Our model helps to identify those who remain most at risk post-vaccination.
Professor Aziz Sheikh, Professor of Primary Care Research and Development at the University of Edinburgh and a member of the QCovid® research team
NHS Digital subsequently updated the Clinical Risk Assessment Tool to include this new evidence.
Together the QCovid® model developed by NIHR researchers and the risk assessment tools developed by NHS Digital and its collaborators have helped protect the patients most at risk during the COVID-19 pandemic.
- The research team is led by the University of Oxford and includes researchers from the universities of Cambridge, Edinburgh, Swansea, Leicester, Nottingham and Liverpool with the London School of Hygiene & Tropical Medicine, Queen’s University Belfast, Queen Mary University of London, University College London, the Department of Health and Social Care, NHS Digital and NHS England.
- The COVID-19 Population Risk Assessment team included representatives from the Department of Health and Social Care, NHS Digital, NHS England, the Office for National Statistics, Public Health England, University of Oxford, New and Emerging Respiratory Virus Threats Advisory Group (NERVTAG), Oxford University Innovation and the Winton Centre for Risk and Evidence Communication.
- More information about this study is available on the NIHR’s Funding & Awards website.
- Read more Making a Difference stories.