Ethnicity and COVID-19: investigating the determinants of excess risk in UK Biobank

What is the problem?
People from minority ethnic groups seem to be disproportionally affected by COVID-19, particularly South Asian and Black and African Caribbean communities.

What has been done so far?
National datasets have shown that minority ethnic groups are up to 4 times more likely to die from COVID-19. However, the reason for this increased risk is not known.

What are we going to do?
We will use a large dataset, called UK Biobank, which has been linked to national COVID-19 data. Within this dataset, we will use statistical modelling to examine whether the increased risk in minority ethnic groups is explained by differences in underlying health status, lifestyle behaviours such as physical activity, and environmental factors such as air pollution or measures of social inequality.

Why is this important?
This work will start to unpick why minority ethnic groups may be at increased risk and whether this increased risk is spread equality across the population. For example, is the increased risk explained by a higher burden of other diseases such as heart disease or diabetes or by living in more polluted areas? Do otherwise healthy individuals from minority ethnic groups still have an increased risk? Addressing these question will help inform public health priorities and actions.

Objective
To use UK Biobank to quantify the risk of a positive test for SARS-CoV-2, hospitalisation for COVID-19 complications and death in South Asian (SA) and Black and African Caribbean (BAC) populations compared to white Europeans and to investigate the key mediators and effect modifiers of excess risk.

Methods
We will use an ongoing UK Biobank project already designed to investigate ethnic health in UK Biobank, which includes half a million adults, of which there are over 8000 SAs and 8000 BACs. Through this application we have access to linked COVID-19 data, which will include test results, hospitalisation and mortality. Logistic and survival modelling will be used to quantify the risk of SAs and BACs compared to WEs and to investigate possible mediators and effect modifiers.

Preparatory work
In preparation for this application, we have downloaded over 3000 unique tests results available to date. SAs and BACs were around twice as likely to have confirmed COVID-19 compared to white Europeans, with early evidence that this risk is modified by key markers of health status, including BMI.

Conclusion
This application is designed to provide greater context to the risk of COVID-19 and resulting complications in BACs and SAs which will help tailor public health policy in the future.