Combining data to identify ethnic COVID-19 inequalities
31 January 2022
Just a few weeks into the first COVID-19 wave in the UK, reports identified a higher risk of deaths among ethnic minorities. The creation of the Public Health Research Database (PHRD) – a combination of GP and hospital medical information and 2011 Census records - confirmed these, and allowed evidence-driven policy changes to be to protect the most vulnerable members of our community.
Just a few weeks into the first COVID-19 wave in the UK, reports in the press identified a higher risk of deaths among ethnic minorities. In time, it became increasingly apparent that people from Black and South Asian ethnic backgrounds were disproportionately affected by the pandemic compared to White people.
It was unclear to what extent this was due to pre-existing differences between ethnic groups in the UK, such as health, environmental exposure and lifestyle behaviours, such as obesity and smoking. Without fine-grained data on how these factors relate to an individual’s risk of COVID-19, it is difficult to fully understand their impact.
Researchers at the University of Leicester, funded by Health Data Research UK, worked with the Office for National Statistics to analyse an unprecedented data resource to shed new light on health inequalities in ethnic minorities. This research showed that socio-demographic and economic factors – such as age, living environment, and occupation – and medical history could explain part of the difference, but they are not the whole story.
How this was made possible
Medical records do not regularly hold information on ethnicity, religion, socioeconomic status and many other factors that influence health outcomes. Without this information, research studies either cannot include these factors or must rely on data at a population level, comparing different regions for instance.
The Data and Connectivity programme jointly led by Health Data Research UK was created to connect UK health data to support and accelerate research on COVID-19. This programme supported the creation of the Public Health Research Database (PHRD) – a combination of GP and hospital medical information and 2011 Census records. The database includes 29 million anonymised records of adults in England and it is the first time that administrative and health data have been combined.
By linking the records of individuals from these different sources, researchers can get a much more detailed picture of the impact of COVID-19. The database is held securely in a ‘Trusted Research Environment’, meaning that only accredited researchers can access the data to work on research projects for the public good.
Ethnicity differences in COVID-19 deaths
This invaluable resource was analysed by Prof Kamlesh Khunti, from the University of Leicester and Chair of the SAGE Ethnicity subgroup, his colleague Prof Tom Yates, and the Office for National Statistics.
Using the linked data from the PHRD, the researchers could show that all ethnic minority groups had a higher risk of COVID-19-related death compared to the White British population during the first wave of the pandemic. However, during the second wave from September 2020, the risk of COVID-19 death remained particularly high for people from Pakistani and Bangladeshi backgrounds but not for people from Black ethnic groups. This higher risk remained even after accounting for factors like age, population density, region, occupational exposure and pre-pandemic health.
The researchers highlighted that the relative improvement for people from Black ethnic backgrounds shows that inequalities can be addressed. This could have been due to widespread coverage of racial disparities during the first wave, leading to behavioural changes. However, the continued higher mortality rate in people with Bangladeshi and Pakistani ethnicities indicated an urgent need for a public health campaign and policy response.
Further data from 12 million people with body mass index (BMI) in their records showed that obesity has a much greater impact on the risk of dying from COVID-19 for minority ethnic groups. The greatest risk was for people of South Asian ethnicity, where an adult with a BMI of 27 had the equivalent risk of a White person with a BMI of 40. Although the reasons for this were not clear, it emphasised the need for public health messaging to reduce obesity, especially among ethnic minorities.
Ethnicity differences in vaccine uptake
Linking the PHRD with vaccination records also revealed differences in ethnicity and COVID-19 vaccine uptake. In a study of six million people aged 70 and over in England, the team found that people of Black ethnic backgrounds were about five times more likely to be unvaccinated than White British people as of 15 March 2021, even accounting for other factors. Low uptake was also seen in people from Pakistani backgrounds, who were 3.5 times more likely to be unvaccinated.
While previous research had identified that vaccination rates tend to be lower among certain ethnic groups, this study examined the impact of wider factors, such as individual socioeconomic status and disability. While the researchers found some differences based on these factors, they had a smaller effect than ethnicity or religion.
As the vaccination rollout continued, a later study gave the first insights into vaccine inequalities among younger adults. This confirmed that ethnicity, religion and deprivation were major factors at all ages, although their influence varied by age. These results showed the importance of analysing data by age group to better understand how to address vaccine uptake inequalities.
Combining data from different sources has given us an unprecedented insight into health inequalities in the UK. By identifying the importance of some factors and our lack of understanding of others, this work has underlined the urgency of research to understand why these disparities exist and how they can be addressed through public health or community engagement programmes.