In the last day, two key pieces of research into whether black and ethnic minority (BME) patients are more likely to die from coronavirus have hit the headlines.

We’re going to take a look at what they tell us – and what we still don’t know.

Institute for Fiscal Studies suggests BME patients dying disproportionately

The respected independent think-tank, the Institute for Fiscal Studies (IFS), reports today that deaths among black Caribbean coronavirus patients are 1.7 times higher than white British patients, Pakistani deaths are 2.7 times as high, and black African fatalities 3 times higher.

The findings account for the fact that ethnic minorities are not spread evenly across the population – and that the outbreak has hit areas like London and the West Midlands first and hardest so far.

As FactCheck reported last month, this is important if we’re to understand whether BME patients are at greater risk than their white neighbours.

And the IFS goes a step further by adjusting for the different age profiles of ethnic minority groups.

The researchers conclude that despite being younger on average – which might make them less susceptible overall to the worst effects of Covid-19 – BME people appear more likely to die from coronavirus than white Brits.

They have also attempted to model what will happen once we have data on the ethnicity of people who have died outside hospital, which the Office for National Statistics say they’ll start to release on 7 May.

According to the IFS model, BME patients are still dying at a higher rate than geography and demography would predict, even once we account for the fact that coronavirus victims who’ve died in care homes are overwhelmingly likely to be white. (White residents make up 95 per cent of those living in care homes).

However, there are some important limits to this report that we should bear in mind.

The data they’re using to make comparisons with the wider population is from the 2011 census. It is the best available when it comes to demographic analysis in England and Wales – but it’s nearly ten years old.

The IFS researchers told FactCheck that “for most of the main ethnic groups we consider, the best available evidence (from large surveys) suggests this is not a major problem”. Though they accept that it could be an issue for the smaller ethnic groups they’ve looked at – including the “other” category.

They also acknowledge that when it comes to estimating “excess mortality”, which compares the number of deaths happening with the same time last year, local-level population shares do matter. But until the next census in 2021, there isn’t any more up-to-date alternative available.

Most importantly, we should remember that the epidemic is far from over. The IFS researchers told FactCheck: “Our analysis is based on what has happened already. The adjustment for care home deaths is based on what has been officially attributed to COVID-19 by ONS so far (as of the statistics published on Tuesday).

“It’s certainly true that if the overall profile of deaths changes, the conclusions may change.”

So it may be that BME patients have been disproportionately more likely to die so far in the outbreak – but it’s too early to say what the ethnicity data will show once the epidemic has run its course.

The IFS analysis is the most robust assessment of the impact of coronavirus on ethnic groups in the UK that we’ve seen to date – but we’ll have to wait until the next census (in 2021) and for the crisis to pass before we can properly test their findings.

Imperial College report: black patients ‘may be at increased odds of mortality’

Researchers at Imperial College have looked at the characteristics of 520 coronavirus patients treated across three large London hospitals.

Their most statistically reliable findings were that being older, being male and having underlying conditions (“comorbidities”) put patients at greater risk of dying in hospital.

The researchers also looked at ethnicity, though their conclusions were more tentative.

They say “that there is no crude difference in mortality between the different ethnic groups” – in other words, the proportion of deaths matched the share of the wider population.

But once they adjusted for age and the likelihood of comorbidities among ethnic groups – and the severity of symptoms at the time they were admitted to hospital – black patients “do appear to have increased odds of hospital mortality compared with those of white ethnicity”.

However, the researchers themselves stress that “this finding merits further investigation given its borderline statistical significance” – i.e. more research is needed because the data they have is not conclusive.

They also point out that they could not confirm the ethnicity of 22 per cent of patients in their sample.

Finally, we should remember again that this is a snapshot of one part of the epidemic – people seen in London hospitals between 25 February and 2 April – and its findings may not be applicable to other settings (like care homes, other parts of the country) or to the course of the whole epidemic.

FactCheck verdict

Two key pieces of research have been published in the last day on whether ethnic minority patients in England and Wales are more likely to die from coronavirus – one by the Institute for Fiscal Studies and one by Imperial College.

Both have made strides in terms of adjusting for the different age and illness profiles of ethnic groups. Taking those factors into account, the two reports suggest some ethnic minorities are more likely than white patients to die from coronavirus. Both identify black patients as being at potentially greater risk.

These are important findings that give us some insight into what has happened so far – but we will only be able to test them properly once the epidemic is over and we have up-to-date ethnicity data from the next census, which is slated for 2021.

The Imperial College report says further research is “urgently needed” to investigate whether the pattern it identifies in its sample of London patients is applicable nationally “and if so, to understand the sociodemographic and biologic factors underpinning it”.