The Guardian published over the last few weeks several articles about the report produced by the Office for National Statistics (ONS) which highlights the higher mortality of Black, Asian or minority ethnic (BAME) people compared to whites. Firstly, let’s look at the key findings:
- Black people are more than four times (females are 4.3, males 4.2) more likely to die from coronavirus than white people
- Similar findings for other ethnicities like Bangladeshi, Pakistani or Indian
We’re writing this as a summary that provides more context to a worrying headline.
As we’ve mentioned in the previous posts, such research, which can be classified as a cross-sectional observational study, can’t explain the cause of this racial disparity. However, the findings are valid and do signal an important problem which needs to be understood and resolved in order to reduce mortality rates.
ONS went one step further and adjusted their statistics for the factors which are most likely to be correlated with a higher death rate:
- geographical factors
- levels of deprivation
- measures of health
- measures of socio-economic status
They explain how they used logistic regression to adjust for all the variables. There’s obviously no reason to doubt the competence of the data scientists working for ONS. However, one thing to point out is that the whole analysis relies on various datasets collected during the 2011 census, so the distribution may have changed. We know that inequality has been constantly increasing over the recent years. Also, even though the number of people who have tragically died in the UK is relatively high, the total number of deaths included in this report is 17,866, with 14,781 white and 1,022 black. Strictly from a statistical point of view, this number can impact the relevance of any correlations observed.
Coming back to the results after the adjustment, the ONS found that black males were 1.93 times and black females were 1.89 times more likely to die from COVID-19 than their white counterparts. The original article in The Guardian mentions these figures but focuses more on the continued disparity rather than the reduction from 4x to less than 2x. This is not insignificant and it provides a very clear indication that, while the genetic factor may play a part, poverty and other socio-economic factors are some of the main drivers of poor health in general and COVID morbidity in particular.
|Gender||Overall death rates vs white||Adjusted death rates|
Having said that, we must stress that it’s important to consider all the factors and see what the best measures are for premature death to be prevented for all ethnicities. As we wrote in our very first post, more knowledge it better than less knowledge.
As to the current hypotheses trying to explain these differences, The Guardian did post a very informative video which highlights a few additional likely factors:
- Vitamin D: It is known that many black people are deficient in vitamin D because of lower levels of exposure to sunlight after a specific latitude (like UK or some northern regions of the US) and because pigmentation reduces vitamin D production in the skin. The NHS in the UK recommends that all adults take vitamin D supplements during the autumn and winter. Vitamin D is believed to lower the risk of cardio-vascular disease, cancer and diabetes.
- Job types: the proportion of BAME workers is higher for certain jobs that involve exposure to many people. Two examples are bus drivers and couriers.
- Accommodation: this is included in the adjustment done by the ONS but it’s worthwhile pointing our that BAME people are more likely to live in crowded houses or estates, which increases the risk of coronavirus transmission.