Introduction

The Chief Medical Officer (CMO) for England asked a subgroup of NERVTAG (the New and Emerging Respiratory Virus Threats Advisory Group), led by the University of Oxford, to develop a risk prediction model for COVID-19 in line with the emerging evidence.

QCovid® assesses factors such as age, sex at birth, ethnic category, body mass index (BMI), and specific health conditions and/or treatments to estimate the risk of a person catching coronavirus and becoming seriously unwell and not surviving.

The QCovid® model used for PRA was based on data collected during the 97-day period of the first wave of the pandemic in the UK (January – April 2020), but it has since been tested with new data and continues to accurately predict outcomes. The model has been independently validated by the Office for National Statistics (ONS) and shown that it performs well in identifying people at high risk from coronavirus.

DHSC commissioned NHS Digital to use the University of Oxford’s QCovid® model to identify additional people to be added to the Shielded Patient List (SPL).

To identify people who might be high risk, NHS Digital conducted a COVID-19 Population Risk Assessment (PRA) using the QCovid® model to calculate risk assessment results for people in England using nationally held data. Individuals whose risk assessment results were higher than an agreed threshold were added to the SPL as a precautionary measure and informed in writing. The letter explained why they had been added to the SPL and what they could do to be removed from it. It also made clear that shielding advice was advisory and not the law. Clinicians were encouraged to review these additions using the COVID-19 Clinical Risk Assessment Tool (also based on QCovid®) and remove any patients which they did not think should be on the SPL.

All adults were considered for the PRA except for:

• people who were already on the SPL
• people who were previously on the SPL but were removed by a clinician involved in their care
• people who had already been assessed individually by a clinician and identified as low risk or medium risk
• people whose combined risk factors could not possibly meet the agreed threshold to be considered high risk, to minimise the number of records processed in line with best practice
• people coded with Down’s Syndrome in their clinical record who had not already been identified through the national SPL - this is due to Down’s Syndrome already being included in the national SPL ruleset and a known coding issue

This report supports transparency by publishing socio-demographic and risk threshold distributions of people who were risk assessed, as well as a summary on the use of default values in the QCovid® model for missing data, or data outside minimum-maximum thresholds.

Further information on COVID-19 Risk Assessment, Shielded Patient List and the external validation of the QCovid® model by ONS are available at the links below.

Coronavirus (COVID-19) risk assessment

Shielded Patient List

External validation by ONS

The QCovid® model was not developed to consider all factors which may determine an individual’s risk, specifically:

• an individual’s behaviour (for example hand washing, wearing face coverings and visiting friends or family)
• occupation
• infection rates
• local and national lockdown measures

The data and analysis presented in this report therefore only describe the socio-demographic attributes of people assessed by QCovid®.

This report shows distributions for two cohorts of people. These are the high risk cohort (from the QCovid® model) and the total PRA cohort assessed by the QCovid® model, defined as:

• the 17 million people identified as potentially high risk (henceforth referred to as the total PRA cohort) based on the COVID-19 PRA
• the 1.5 million people identified as high risk from the COVID-19 PRA (henceforth referred to as the high risk cohort) based on thresholds set by England’s Chief Medical Officer (CMO) and senior clinicians

Key observations of the analysis described within this report include:

• The identification of people who are high risk based on the Absolute Risk criteria of the COVID-19 Population Risk Assessment appear to be skewed towards males and those in older age groups. Comparatively, those identified based on the Relative Risk criteria appear to be more skewed towards females and those in younger age groups.
• Almost 28% of people identified in the high risk cohort have missing data or data outside the minimum-maximum range, specifically for ethnic category and BMI values.
• The age-sex at birth distribution for both cohorts appear to be slightly skewed towards females. The total PRA cohort appears to be balanced for most age groups compared to the age group distribution for the high risk cohort.
• The age group distribution of ethnic categories is skewed towards young adults in the total PRA cohort and younger to middle-age groups in the high risk cohort for all categories apart from those who are White British or White Irish.
• People aged between 40 and 74 who are obese, based on their recorded BMI values, appear to be highly represented in both cohorts, with a higher proportion of obese people in the high risk cohort compared to the total PRA cohort.