Ethnicity and Race

How should ethnicity and race be reported in health datasets?

People’s race and/or ethnicity are part of their personal identity.

They are protected in law - it’s illegal to discriminate based on these factors.

Despite this, we know that in many countries, people in minority racial/ethnic groups receive worse healthcare, and suffer worse outcomes as a result.

In the UK, people in minority ethnic groups have longer life expectancy than white British people, but are more likely to live in poverty and suffer with disability or chronic disease at a younger age.

Many countries do not adequately collect race/ethnicity data in healthcare.

Even countries who do collect data often disagree about how this should be categorised. In the UK, the Office for National Statistics (ONS) advises ethnicity to be recorded differently in England, Scotland, Wales and Northern Ireland - four different approaches in the same country. In the USA the term race is used instead of ethnicity, and many organisations collect only ‘White’, ‘Black / African American’, ‘Asian’, or ‘Other’.

Digital health data and the move towards Artificial Intelligence (AI) promises to dramatically improve healthcare by reducing barriers to diagnosis and treatment. However, this brings with it a real risk of worsening racial/ethnic biases, generating unfairness.

It’s impossible to uncover and correct unfair treatment and inequality in healthcare unless we have accurate race/ethnicity data. It’s also very challenging to compare inequality across countries if we don’t all record race/ethnicity the same way.

If healthcare datasets contained reliable race/ethnicity data, we could make AI tools which better serve all members of society, not just the majority.

Many people are concerned about sharing personal data, for privacy reasons, or in case this is misused to harm them or their communities. Safeguards - such as de-identification, data licensing (legal agreements which outline what users can and cannot do with the data), and compliance with data protection regulations are crucial to protect individuals, demonstrate trustworthy approaches and build public confidence, allowing sensitive data to be collected and stored.

Demonstrating trustworthy approaches and building public approval for this is crucial to ensure we get this right, and that no one is left behind.

Themes we're exploring with our PPIE subcommittee:

  • What do you understand by the term ethnicity?

  • What do you understand by the term race?

  • Do you think both terms are equally important or is one more important than the other?

  • How important do you think it is for countries to agree on terminology and categories for race, ethnicity or both? Why is that?

  • What are your thoughts on the routine collection of ethnicity, race, or both in healthcare data?

    • Is it important? Why?

    • Should it be avoided? Why?

  • How do you think ethnicity should be categorised?

  • How do you think race should be categorised?

  • What do you think of the way the ONS categorises ethnicity in the UK?

  • What are your thoughts on people combining and simplifying ethnicity data into larger categories?

    • For instance, combining all of those responding as ‘Indian’, ‘Pakistani’, ‘Bangladeshi’, ‘Chinese’, and ‘any other Asian background’ into a new larger category such as ‘Asian or Asian British

  • In an ideal world, what would be the best way to capture and categorise information on race and/or ethnicity that can be standardised across the UK and internationally?