There has been much concern that more black, Asian and minority ethnic (BAME) people have died with Covid-19 than the proportion of BAME people in the population as a whole. Alongside troubling reports of very high numbers of BAME NHS staff deaths, Public Health England have announced a review into ethnicity as a risk factor during the pandemic.
The Guardian reported that BAME groups accounted for 19% of all Covid-19 deaths in hospitals, yet constitute only 15% of the population of England. According to the latest figures this week, BAME groups now account for 18% of all Covid-19 deaths in hospitals.
On the face of it, this does look disproportionate. But statistically such comparisons are difficult to make. A key problem in the case of Covid-19 is that the pandemic has not spread evenly across the country. For example, London and the Midlands have recorded many more deaths (around 45% of all reported since March 1st), but also have a higher than average proportion of BAME people (more than 40% in the case of London). Given this, we might expect more deaths among BAME groups than the national proportion of BAME people.
While BAME groups are very diverse in terms of demographic profiles and health needs, it may be helpful to estimate an overall proportion of people in BAME groups in the affected populations against which we can compare the reported levels of deaths.
To do this we have weighted regional populations to take account of the uneven distribution of Covid-19 deaths, giving greater weight to more heavily affected areas. This gives us an adjusted figure for the proportion of people in BAME groups in the affected populations.
As the chart shows, on that basis the apparent excess BAME deaths due to Covid-19 disappears – with the proportion of deaths (18%) broadly matching the proportion of BAME people in the affected populations.
But that isn’t the end of the story.
A fuller analysis of the factors explaining Covid-19 deaths will also have to take account of the many other factors we know, or suspect, are linked to the disease.
These will include age – and we know for instance that BAME populations tend to be younger. Given that Covid-19 is a disease predominantly affecting older people, we might expect the rate of deaths in BAME groups to be lower. They will also include underlying health conditions, and we know that heart disease, diabetes and hypertension are all higher in non-white ethnic groups – putting people at greater risk of becoming more severely ill if they are infected with Covid-19. Other factors that will need to be considered include occupation (e.g. whether a job is public-facing or not), home environment, lifestyle, behaviours under lockdown, and income.
This may reveal a specific link with a person’s ethnic group. On the other hand, what it might show is that there are more important – but linked – factors that are predisposing people to succumb to the virus, where someone’s ethnic group is an indicator of a deeper underlying factor – such as income and occupation.
As yet we do not have enough detailed data on the personal circumstances of those who have unfortunately died due to Covid-19, and how they compare with those who survived or did not get infected. However, future studies should be able to untangle the links between these factors and the distribution of Covid-19 deaths across our communities.