Ethnicity, Disability, Religion:  Which groups are at most risk of death involving COVID-19?

As the coronavirus pandemic develops, it is crucial the ONS updates its mortality analysis to ensure the best evidence is available for the right decisions to be taken. We have previously found that those from some BAME backgrounds were at greater risk of coronavirus-related death than those from a White ethnic group. But with more data available, is that still the case? And what impact does religion and disability have on the risk of death involving COVID-19? Ben Humberstone explains more.

Ethnicity has emerged as a key characteristic for understanding why certain people are at more risk of death during the COVID-19 pandemic.

Last month, looking at deaths up to the middle of April, and taking into account age, we concluded that Black males and females were at around four times more risk of death involving COVID-19 than those from a White ethnic group. When taking other socio-demographic characteristics into account, it came down to around two times greater risk. It was a similar picture for people from Bangladeshi and Pakistani ethnic groups. Today we have updated our initial findings using more nuanced methods and with an extra month’s data available – a period which covers the full lockdown.

We’ve been able to do this by producing a linked study that has taken into account what we know about death registrations and what we know about social and economic characteristics, including ethnicity, religion and disability, from the 2011 Census. There are clearly limitations until the census can be updated next year and there is much more work we wish to do. However, we have been able to draw some more detailed conclusions.

Our latest analysis of deaths up to the middle of May shows that Black males are at 3.3 times greater risk of death involving COVID-19 than White males of the same age, while Black females are 2.4 times at greater risk than their White counterparts of the same age. There are a whole range of factors working together here, so what we have subsequently done is adjust for some of those shared characteristics that are at play. These include geographical, demographic and socio-economic factors, and differences in occupational exposure.

In adjusting for some characteristics we are able to explain some of the increased risk when compared to a White person, but there is still a proportion that is unexplained. In fact, our research shows there remains twice the risk for Black males and around one and a half times for Black females. Significant differences also remain for Bangladeshi, Pakistani and Indian men.

This shows that for some ethnicities there is a disproportionate risk of death involving COVID-19. We are planning further research to help understand this difference, including analysis of underlying health conditions.

Religion is another important part of the social make-up of the country we’ve been able to look at by using a similar approach to our ethnicity analysis.

When taking age into account and using the Christian religion as the reference category (as the largest population), those who identified as Muslim at the time of the 2011 Census are 2.5 (males) and 1.9 (females) times at greater risk of a COVID-19-related death than those of Christian religion. Muslims had the greatest risk relative to the Christian population of the same age, although Jews, Hindus, Sikhs and Buddhists also showed a higher risk than the Christian religion.

But when you fully adjust these findings for socio-economic factors, including ethnicity, religion does not appear to be a factor in increased risk for any religious group, other than for the Jewish community.

Our findings show that those who identified as Jewish at the time of the 2011 Census had the highest risk of a death involving COVID-19 compared to the Christian population; Jewish males show twice the risk compared to Christian males, with females at 1.2 times greater risk. Again, we’re unsure what the reasons for this might be at this stage and will undertake further analysis in the weeks and months to come.

We’ve also used census data to help understand the impact of disability on the risk of death involving COVID-19. We’ve found, perhaps not unexpectedly, that those whose daily activities were ‘limited a lot’ at the time of the 2011 Census had a greater rate of death compared to those who were not disabled.

What does all this mean?

There are lots of complex things playing out during the pandemic and your risk of death involving COVID-19 is influenced by a range of factors including where you live, how deprived the area you live is, your ethnicity, the job you do, the health you are in – to name but a few.

But this pandemic is also developing all the time.

In the UK the initial outbreak took hold in London, with high levels of mortality at the end of March and April. This has spread out across the country since then and the type of people exposed to this severe infection will have changed. This will have had an impact on the exposure of different communities to the pandemic.

We have also seen the impact of pre-lockdown – in terms of occupational exposure – and then the impact of lockdown, and how that will have affected people in different ways. We are now coming out of lockdown and we will have to repeat our research again to see if mortality patterns have changed as lockdown has eased.

This is a continually developing picture and the job of the ONS is to continue our research on the impact of ethnicity, religion and disability on COVID-19 related deaths as more data becomes available.