Tackling Ethnic Minority Under-Representation in Clinical Research

This blog article has been prepared by Dr Kamlesh Sheth FFPM.

The article presents the views of the author and not necessarily those of the Faculty of Pharmaceutical Medicine.

AI Disclaimer: This blog was drafted with the assistance of an AI writing tool (Claude, Anthropic). The author provided the overall structure and direction of the piece including the decision to cover current evidence, the impact on patients, the causes of under-representation and the role of each stakeholder group in addressing the issue (industry, academia, and regulatory agencies) and directed several specific additions during drafting. The AI tool was used to draft and edit text for readability and to locate and format supporting references. All AI-generated content was reviewed, fact-checked and edited by the author, who verified the underlying sources and takes full responsibility for the accuracy and views expressed in this piece.

Clinical research has given us medicines, vaccines and devices we rely on today, yet the people who take part in it still do not reflect the population it is meant to serve. Ethnic minority communities are consistently under-represented in UK clinical trials and this is not simply a diversity statistic to note in passing; it affects whether we truly know if a treatment works as well and as safely for everyone. People are consistently more likely to trust health messages and health messengers, who share their own background and language, a finding that offers an important clue to both why representation in research has lagged and how we might close that gap [1].

In this article I look at what the current data tells us, why this matters for patients, what drives the gap and what pharmaceutical physicians, sponsors, regulators and communities can each do to close it, including a recent experience of my own, discussing this very subject on a Hindi-language radio programme [2], which convinced me that representation begins with something as simple as speaking to people in a language and setting they trust.

The scale of the problem: what the data tells us

The evidence has been consistent for years. A systematic review and meta-analysis of thirty UK COVID-19 randomised controlled trials, covering almost one hundred and nineteen thousand participants, found that Black participants made up around 1% of trial populations against an expected 3.3% and Asian participants around 5.8% against an expected 7.5%, when benchmarked against national census data [3]. This was not an isolated finding specific to the pandemic; it reflects a longer-standing pattern across UK research.

Part of the difficulty is that we often cannot even measure the problem properly. A review of one hundred and forty-eight randomised controlled trials funded by the National Institute for Health and Care Research (NIHR) between 2007 and 2017 found that only 60% reported participant ethnicity at all [4]. A more recent analysis of NIHR trial reports between 2016 and 2021 found that only around 9% did what the authors judged to be “a good job” of recording and reporting ethnicity data [5]. Under-representation and under-recording reinforce one another; if ethnicity is not recorded well, the scale of the gap remains invisible and it becomes harder to hold research accountable for closing it.

Under-representation and under-recording reinforce one another; if ethnicity is not recorded well, the scale of the gap remains invisible and it becomes harder to hold research accountable for closing it.

Why representation matters: the impact on patients

This matters clinically, not just statistically. Genetic variation between population groups can influence how a medicine is absorbed, metabolised and tolerated. Differences in the prevalence of certain drug-metabolising enzyme variants across population groups, for example, are well recognised to affect the way some cardiovascular medicines are processed, with implications for how effective standard dosing is for different patients. Conditions such as glucose-6-phosphate dehydrogenase deficiency, which is more common in South Asian and Mediterranean populations, can affect how safely certain drugs can be used. Cardiometabolic risk itself differs by ancestry; South Asian populations are known to develop type 2 diabetes and coronary heart disease at a lower body mass index than White European populations, which is why UK clinical guidance now applies different risk thresholds for this group.

If the people taking part in the trials that generate our evidence base do not reflect this variation, then the dosing guidance, safety data and efficacy claims we rely on may simply be less accurate or less complete for the communities least represented in the data.

This is not only a concern for medicines; it applies equally to the devices we use to guide everyday clinical decisions. The independent NHS Race and Health Observatory’s 2021 rapid review of pulse oximetry found consistent evidence, dating back to 1990, that these devices can give less reliable readings in patients with darker skin pigmentation and it called for stronger validation and further research to close the gap [6]. Subsequent NHS England data, from pulse oximeters issued through the COVID Oximetry @home scheme, put a number on the consequence: readings ran 0.6% to 1.5% higher, on average, in patients with darker skin tones and the proportion of readings that missed a genuinely dangerous oxygen level, a false negative was between 5% and 35% higher in darker-skinned patients, depending on the threshold used [7]. That something as routine as a pulse oximeter can quietly under-detect low oxygen in exactly the patients least represented in its validation studies is a stark illustration of what under-representation can mean in practice, well beyond trials of new medicines alone.

There is also a less tangible, but equally important, consequence. Under-representation and mistrust reinforce each other in both directions: communities that have historically felt overlooked or poorly served by research are, understandably, more hesitant to accept public healthcare intervention, and that hesitancy was clearly visible during the pandemic, where lower trial participation among ethnic minority groups tracked alongside greater vaccine hesitancy in the same communities [3]. Better representation in research is not only a scientific good; it is also how trust in medicine, more broadly, is built or lost.

That something as routine as a pulse oximeter can quietly under-detect low oxygen in exactly the patients least represented in its validation studies is a stark illustration of what under-representation can mean in practice, well beyond trials of new medicines alone.

Understanding the causes

The reasons behind this gap are varied and no single intervention will close it. Language and health literacy are a real barrier: patient information sheets, consent forms and even initial conversations about a trial are still overwhelmingly conducted in a way that can exclude people who are otherwise perfectly willing to take part. Historical experiences – both specific documented failures of research ethics involving minority communities and more everyday experiences of feeling unheard within healthcare systems have left a legacy of understandable caution that is passed down within families and communities.

There are also practical and structural barriers: trial sites are not always located where minority communities live in large numbers, recruitment often happens through channels that do not reach every community equally and the research workforce itself does not yet reflect the population it studies. Finally, simple awareness plays a large role; many people are genuinely unaware that clinical trials exist for their condition, that participation is free, or that trials are subject to rigorous, independent oversight. Addressing that awareness gap was, in fact, the whole premise behind a recent invitation I had to discuss clinical research on a local Hindi-language radio programme, precisely because that audience is rarely reached through conventional research communication channels [2].

The wider regulatory and policy landscape

UK regulators have begun to formalise the expectation that inclusion should be designed in, not bolted on. The Health Research Authority (HRA) and the Medicines and Healthcare products Regulatory Agency (MHRA) have jointly developed draft guidance requiring sponsors and researchers to produce an Inclusion and Diversity Plan when designing a clinical trial or clinical investigation, setting out how the study has been designed to include the people who stand to be affected by its findings [8]. This sits alongside the NIHR’s INCLUDE framework, which since 2020 has offered practical, condition-specific guidance to help trial teams identify under-served groups early and design recruitment and delivery around them, rather than simply reporting a shortfall after the fact [9].

Beyond formal regulation, community faith organisations and local media, remain an underused route into this. My own experience discussing clinical trials in Hindi, on Sabras Radio, was a reminder of how far a conversation in someone’s own language and trusted local outlet can go towards making an unfamiliar subject feel approachable [2]. Patients and communities have a role too – being open to asking their GP or hospital team whether a suitable trial exists, is often the only step standing between interest and participation.

Where representation is decided: industry, academia and regulation

In practice, the decisions that shape representative recruitment are taken at several distinct points in the system and pharmaceutical physicians sit at each one. In industry, those leading clinical development or medical affairs write the protocols and eligibility criteria as well as selection of research sites that determine, from the outset, who a trial can realistically include. In academia, those leading investigator sites shape which populations a study reaches through team composition and community engagement well before a trial opens. Within regulatory agencies, physicians assess whether an Inclusion and Diversity Plan reflects genuine planning or box-ticking at authorisation and given the pulse oximetry findings above, that scrutiny increasingly extends to device validation as well as medicines trials. Recognising where in this chain each of us sits is a useful starting point for understanding representation as something designed in from the beginning, rather than measured and regretted at the end.

the decisions that shape representative recruitment are taken at several distinct points in the system and pharmaceutical physicians sit at each one.

Conclusion

None of this is a single fix and it will not happen quickly. But the direction is clear: regulators are beginning to require better planning for inclusion, researchers have practical frameworks to act on it and pharmaceutical physicians in industry, academia and regulatory agencies alike are well placed to be trusted advocates within their own settings. Representation in research starts with trust and trust starts with understanding that is conveyed, wherever possible, in a language and a setting people already know. The FPM as well as individual pharmaceutical physicians have a part to play in closing this gap.

References

1. Alsan M, Garrick O, Graziani G. Does Diversity Matter for Health? Experimental Evidence from Oakland. American Economic Review 2019;109(12):4071-4111. www.nber.org/papers/w24787 [Accessed 5 August 2026]
2. Sheth K. Interview: clinical research and clinical trials (in Hindi). Sabras Radio, 2026. player.autopod.xyz/1387267 [Accessed 28 July 2026]
3. Murali M, Gumber L, Jethwa H, Ganesh D, Hartmann-Boyce J, Sood H, Zaccardi F, Khunti K. Ethnic minority representation in UK COVID-19 trials: systematic review and meta-analysis. BMC Med 2023;21:111. www.ncbi.nlm.nih.gov/pmc/articles/PMC10049782/ [Accessed 28 July 2026]
4. National Institute for Health and Care Research. Randomised controlled trial participants: Diversity data report. www.nihr.ac.uk/about-us/who-we-are/research-inclusion/randomised-controlled-trial-participants-diversity-data-report [Accessed 5 August 2026]
5. Underrecording and underreporting of participant ethnicity in clinical trials is persistent and is a threat to inclusivity and generalizability. www.sciencedirect.com/science/article/pii/S0895435623002147 [Accessed 28 July 2026]
6. NHS Race and Health Observatory. Pulse Oximetry and Racial Bias: Recommendations for National Healthcare Regulatory and Research Bodies. 2021. nhsrho.org/research/pulse-oximetry-and-racial-bias-recommendations-for-national-healthcare-regulatory-and-research-bodies/ [Accessed 5 August 2026]
7. The impact of skin tone on performance of pulse oximeters used by NHS England COVID Oximetry @home scheme: measurement and diagnostic accuracy study. www.ncbi.nlm.nih.gov/pmc/articles/PMC12801414/ [Accessed 28 July 2026]
8. Health Research Authority, Medicines and Healthcare products Regulatory Agency. HRA and MHRA draft inclusion and diversity guidance. https://www.hra.nhs.uk/about-us/news-updates/guidance-developing-and-submitting-inclusion-and-diversity-plan-second-draft/ [Accessed 28 July 2026]
9. The INCLUDE Ethnicity Framework. https://www.nihr.ac.uk/support-and-services/support-for-delivering-research/improving-inclusion-under-served-groups-clinical-research-guidance-include-project [Accessed 28 July 2026]