Research and analysis

Appendix E: Is ethnicity a risk factor for infection or mortality from COVID-19?

Published 3 December 2021

This was published under the 2019 to 2022 Johnson Conservative government

People from ethnic minority groups are more likely to experience various risk factors linked with higher COVID-19 infection rates. And for a few ethnic minority groups, residual disparities compared with white people remain after taking account of known risk factors.

The data shows a significant residual excess risk for people from the Pakistani and Bangladeshi ethnic groups (and to a lesser extent other ethnic minorities) even after adjusting for all known risk factors including occupation and co-morbidities. Taken alongside the PHE work (summarised in the third quarterly report) that showed worse survival for people from the Bangladeshi ethnic group in particular, this indicates that for some ethnic groups, some aspect of their risk is unexplained. But what underlies that excess residual risk is not known – it could be caused by some as yet, unmeasured or unmeasurable risk factors for which ethnicity is a proxy, or it could be partly genetic.

Risk factors

A risk factor (for infection) is something that increases a person’s chance of becoming infected with COVID-19. Such factors include geography, population density, age, deprivation, overcrowding, living in a multigenerational household, certain occupations (in particular those that are public-facing) and lifestyle factors.

According to SAGE, social factors such as poverty and occupation make a large contribution to the greater burden of COVID-19 in ethnic minorities. For example, 24% of households in the Bangladeshi ethnic group were overcrowded (2016 to 2019 combined) and people from the Pakistani ethnic group were over 3 times as likely as White British people to live in the most overall deprived 10% of neighbourhoods.

There are risk factors linked with higher risk of mortality from COVID-19 (factors that increase a person’s chance of dying following COVID-19 infection). Examples of risk factors for mortality following infection include age, sex and (some) underlying health conditions. Some factors, like underlying health conditions, are risk factors for both higher risk of infection and mortality.

Regression analysis

The importance of risk factors can be quantified using multiple regression. This allows analysts to identify the contribution that each risk factor (individually and collectively) is making to the likelihood of infection or death. In the context of COVID-19 mortality, for example, there are wide disparities between ethnic minority groups and the white group. However, when taking account of known risk factors, most of these disparities are reduced significantly or even disappear.

This does not mean that people from ethnic minority groups haven’t been affected by COVID-19 much more than people from the white group. However, it does mean that the known risk factors account for most of the disparities. This enables decision-makers to take these factors into account in planning interventions and communications, or making decisions about prioritisation.

The risk for some ethnic groups may actually increase when some characteristics are taken into account. For example some ethnic groups may have lower incidence of certain diseases that are related to COVID-19 mortality, such as dementia, and as a result when these data are added in the regression model the risk of mortality for some ethnic groups increases.

In looking at the impact of risk factors on different ethnic groups, this regression approach is only applicable when we have data about each person’s COVID-19 status (such as: infected, hospitalised, died of or with COVID-19), their ethnicity, and the presence or absence of a range of risk factors. In the absence of such data, analysts have been resourceful during the pandemic. For example:

• ONS used occupation data from death certificates to identify those occupations with higher COVID-19 mortality rates, and also looked at the ethnic profile of the ‘at risk’ occupations, allowing inferences to be drawn
• ONS subsequently linked death records with the 2011 Census to get estimates of COVID-19 mortality by ethnicity and occupation (albeit the occupation data was old)

Residual (excess) risk

Even after taking account of a range of risk factors, some excess risk still remains for some ethnic groups like the Bangladeshi ethnic group. The second quarterly report summarises research (‘survival analysis’) by PHE showing that, of those infected and testing positive, people from the Bangladeshi, Chinese, Pakistani, Black Other and Indian ethnic groups had an increased risk of death.

It isn’t possible to say for sure why there are some residual disparities in the relative risk of mortality. This could be because of:

• aspects of data quality, such as out of date statistics – for example, ONS’ work to take account of people’s occupations and people’s housing conditions uses data from the 2011 Census
• data on some risk factors not being available – for example, the impact of schools reopening on adults becoming infected by children – over-70s in the Bangladeshi and Pakistani ethnic groups are much more likely to have contact with adults and school age children within the same household, there have been larger increases in the R rate when schools have been opened (other examples include compliance with NPI rules, differences in health seeking behaviours, access to health services, international travel, or the number and intensity of contacts with people who have the infection)
• other risk factors that have not been included in the model – for example, the role of genetics.

Existing research suggests that the inequalities in ethnic groups from COVID-19 are largely explained by underlying social differences. There is some evidence to suggest that genetic differences may also play a role in the disparities. However, a gene cluster identified as a risk factor for severe coronavirus symptoms is carried by approximately 50% of people in South Asia, compared with 16% of people in Europe. This gene cluster is associated with a risk of respiratory failure and may partially explain why people in the Bangladeshi ethnic group have the poorest survival rates, but more research on that is needed.

Researchers at the University of Oxford have identified a gene responsible for doubling the risk of respiratory failure and death from COVID-19 among under 60 year olds ^{[footnote 1] [footnote 2]}. 61% of people with South Asian ancestry carry the higher-risk version of the gene, compared with 16% of those with European ancestry, 2% of those with Afro-Caribbean ancestry and 2% of those with East Asian ancestry. This genetic factor may, in part, explain the higher rates of hospitalisation and mortality among South Asian people

In addition, research from the International Investigator Group for Ethnicity and COVID-19 estimated a 4.1-fold increased risk for COVID-19–related hospitalization and a 2.6-fold increased risk for COVID-19–related death for people with sickle cell disease. It concludes that “Several aspects of sickle cell phenotypes overlap with the pathophysiology of severe COVID-19, which could be relevant mechanisms worthy of further study, as should the directionality of infection and sickle crisis”

Researchers at Newcastle University have identified a gene that is found 3 times as often in people who had COVID-19 but who were asymptomatic as those who developed severe symptoms – so, people with this gene have a degree of protection from severe COVID-19, but still transmit the virus. People in the North and West of Europe are more likely to have this gene than people from other regions of the world.

1. Identification of LZTFL1 as a candidate effector gene at a COVID-19 risk locus, Nature ↩

2. Researchers uncover gene that doubles risk of death from COVID-19, University of Oxford ↩