Why Ethnicity Data Matters for Global Clinical Development

There are encouraging examples of the scientific value of doing so: early genetic studies in the US identified loss-of-function variants in PCSK9 that were considerably more common in African American participants, and were associated with markedly lower LDL cholesterol (Cohen et al., 2005). Follow-up in >3,300 Black participants found that carriers had a 28% lower LDL cholesterol level and an 88% lower risk of coronary heart disease (Cohen et al., 2006). These findings helped validate PCSK9 as a therapeutic target and contributed to the subsequent development of PCSK9 inhibitors now used internationally (Shapiro et al., 2018). It is an example of how research in a population that has historically been underrepresented can demonstrate benefits extending beyond that population.

Advances such as this demonstrate the value of generating evidence in populations most affected by disease. With increasing attention on ethnic representation in clinical trials, it is important to recognise that representation is only one part of the issue. Are ethnicity data collected consistently? Have we considered what those data might tell us? Are they sufficient to inform development and treatment decisions?

Recent work by Brijnath and colleagues highlights this challenge. Their international Delphi consensus on the measurement of ethnicity in clinical trials identified self-identified ethnicity as a critical variable and proposed a minimum set of ethnicity-related information for trialists to consider (Brijnath et al., 2026). This builds on initiatives such as the NIHR INCLUDE Ethnicity Framework, whose opening questions are simple but important: “Who should my trial apply to?” and “Are the groups identified likely to respond in different ways?” (Treweek et al., 2021).

However, ethnicity is still not recorded consistently. Wallace and colleagues reviewed 407 randomised controlled trials published in the NIHR Journals Library between 2016 and 2021. Ethnicity was recorded in 67.3%, but only 9.3% of all reports clearly described both how ethnicity had been recorded and how it was subsequently reported (Wallace et al., 2023). Although these were NIHR-funded UK trials rather than global industry registrational studies, the findings demonstrate that inconsistent handling of ethnicity data remains an issue even across a large body of contemporary randomised research.

Planning for population differences prospectively

Regulators increasingly expect sponsors to consider these questions. FDA guidance issued in December 2025 recommends approaches to enrol populations more representative of those likely to use a medicine, including across racial and ethnic groups, in part so that the influence of participant characteristics on safety and effectiveness can be assessed (FDA, 2025).

However, there is also a practical reason to address these questions early in development. Dedicated bridging or targeted subpopulation studies will sometimes be scientifically necessary, but where differences in exposure, pharmacodynamics, safety or clinical efficacy can reasonably be anticipated, development teams should consider whether the main clinical programme can generate the necessary evidence prospectively rather than identifying the gap during or after regulatory review.

This principle is already reflected in ICH E17 for multi-regional clinical trials. The guideline recommends early consideration of potentially relevant intrinsic and extrinsic factors, their exploration during development, and appropriate collection of such information in confirmatory studies. Well-designed multi-regional trials can support regulatory decision-making across several jurisdictions and may enable more efficient global submissions and earlier access to medicines (ICH, 2017).

Of course, it would be neither feasible nor scientifically appropriate to power every trial for comparisons between every ethnic group. The important question is whether plausible population differences have been anticipated early enough, whether the right data are being collected, and whether the programme is capable of addressing clinically important questions if they arise.

Making ethnicity data useful

So what should this mean in practice?

First, ethnicity should be considered at the design stage rather than treated as another demographic field in the clinical database. Trial teams should ask who the eventual treatment population is, whether there are scientifically plausible sources of population heterogeneity and what information would be required to investigate them.

Ethnicity needs to be collected and reported more consistently. Recent consensus work offers an opportunity to improve practice while recognising that ethnicity categories differ between countries and that ethnicity, race and genetic ancestry are not interchangeable concepts.

Where clinically meaningful differences are plausible, analyses should be planned prospectively. The objective is to generate evidence capable of identifying relevant heterogeneity where it exists, and providing reassurance when it does not.

Finally, development strategies should look beyond individual trials and regulatory submissions. If a medicine is intended for global use, the evidence needed to support its use across those populations should be considered from the outset.

Black History Month rightly prompts us to ask who has historically been represented in medical research. But there is an equally important forward-looking question: who is the evidence we generate today designed to serve?

Collected well and interpreted carefully, ethnicity data can tell us whether findings from a clinical programme are likely to hold across the populations in which the medicine will actually be used.

If the medicines we develop are intended for a global population, our evidence should be capable of travelling with them.