Behavioral Science in Clinical Trials: Part 3 — Why Diversity Must Be Built Into Trial Design

A recent study published in Nature examined 341 approved drugs and posed a straightforward question: Do the pivotal phase III trials accurately reflect the racial and ethnic makeup of the US population? The answer was striking. Only 6% of these trials achieved enrollment that reflected the country's demographic diversity. In other words, many drugs approved today are still tested in populations that do not represent the individuals who will ultimately use them, and whose health experiences have not been studied adequately.

The implications of this are significant. If a population is excluded from a clinical trial, we cannot be certain that the results are applicable to them. Potential differences in safety profiles, rate and severity of adverse events, pharmacokinetics, and pharmacodynamics may go undetected. Effectiveness may vary between different populations, and potential risks may be overlooked.

This article—the last in our three-part series on applying behavioral science and service design to improve clinical trials—explores why diversity in clinical trials matters, why progress on this issue has been so slow, and why planning for representative enrollment must begin at the outset, not as an afterthought, but as a core element of trial design and conduct.

Regulatory efforts to improve diversity in clinical trials

In response to long-standing gaps in representation, the FDA issued guidance through the Diversity Action Plans to Improve Enrollment of Participants from Underrepresented Populations in Clinical Studies. Under the guidance, sponsors are asked to submit a Diversity Action Plan comprising three core components. First, they must define enrollment goals broken down by race, ethnicity, sex, and age group. Second, they must provide a clear rationale for those goals, grounded in epidemiology. Third, they must outline specific actions to achieve the enrollment targets.

However, the most significant limitation is that this remains only guidance, meaning sponsors do not face penalties for failing to meet the enrollment targets. If these targets are missed, there is minimal, if any, impact on the approval or commercialization of the drugs.

That said, the broader regulatory direction is clear. Although the FDA currently describes the Diversity Action Plan guidance as non-binding, using terms like “should” or “try” instead of “must,” it is evident that regulators are moving towards a stance where sponsors may ultimately be required to demonstrate that the drugs they want to market to diverse populations in the US were tested in those same populations within the same market.

Designing clinical trials that attract a diverse patient population

Historically, recruitment strategies have focused almost exclusively on statistical power: enrolling enough participants, in aggregate, to achieve a sample size large enough to detect an effect, should one be present. However, many conditions have different incidences and prevalence across racial and ethnic groups, meaning that if a trial population does not reflect those differences, it is fundamentally misaligned with clinical reality. Recruiting a representative population is therefore not an optional ethical add-on; it is a scientific requirement for generating externally valid and generalizable evidence.

Sponsors often see diversity requirements as an added burden on an already challenging recruitment process. While recruitment is difficult even without diversity targets, this perspective overlooks the main issue. Recruiting a diverse patient population isn't just about increasing efforts at the end; it requires a different approach from the outset that tackles the root causes of underrepresentation.

One major barrier is trust. Many individuals from underserved communities distrust clinical research due to historical inequities. Building trust requires visible, sustained investment in people and communities. Patients want to see doctors, investigators, and trial coordinators who look like them, who understand their communities, and who are embedded in their care environments.

Site selection also matters. Clinical trials have traditionally been held at large medical centers, which can be far from where underserved populations live. This makes it hard for them to participate. Asking patients to leave their usual care settings, travel hours to an academic center, and see unfamiliar providers creates a significant barrier to participation.

Sponsors can make trials more accessible by offering remote participation or opening sites at community health centers where underserved patients already receive care. When trials are placed where patients already are, recruitment becomes less about persuasion and more about designing scientifically strong studies that are feasible to deliver in routine care and beneficial to patients.

Protocol design is vital for encouraging participation, especially among underserved communities. Offering flexible schedules, childcare support, transportation assistance, and reimbursements can greatly enhance accessibility. For many, transportation can be the deciding factor for participation. Small adjustments in protocol design can lead to significant improvements in inclusion.

The role of behavioral science in addressing hidden barriers

Behavioral science research can help uncover how people actually think, feel, and decide whether to participate in clinical trials. This includes understanding attitudes toward medical research, perceptions of risk, sources of mistrust, and practical constraints that shape decision-making. Importantly, this work goes beyond simply “talking to patients.” It involves reviewing published evidence on what worked and didn’t work in the past, co-designing and testing different behavioral strategies, framing the intervention, experimenting with the format and timing of its delivery, and choosing recruitment channels to understand what resonates best in specific communities.

Too often, recruitment strategies assume the problem is already known. Minor surface-level changes, such as swapping images in advertisements or translating a flyer, are treated as solutions. These approaches rest on untested assumptions about the true behavioral barrier. When those assumptions are wrong, the result is not just a lack of results but, in some cases, harm, reinforcing stereotypes or deepening mistrust rather than alleviating it.

A more effective approach is to co-design recruitment strategies with the people who have historically been excluded from clinical trials. Community-based participatory research and the use of design-thinking methods allow sponsors to work alongside community members to identify behavioral barriers to participation and co-develop solutions together. The people experiencing the problem are often best positioned to define what would actually make participation feasible and valuable.

Research also offers tools to systematically identify and address logistical barriers. Service design methods can help map the entire patient journey, highlighting where people drop out, the contributing factors, and the most promising opportunities for intervention. Solutions can then be co-developed with patients and tested iteratively, rather than assumed to work in theory.

In summary

The evidence is clear that diversity in clinical trials cannot be solved by last-minute fixes or standalone initiatives added once recruitment is already underway. As this article has shown, meaningful representation is shaped by early study design decisions, including how protocols are developed and whether the study can fit into patients’ daily lives. This includes selecting sites that are geographically accessible and operational beyond standard work hours, providing reimbursement for travel and childcare, and designing protocols that are feasible for both patients and clinicians. It also means choosing investigators and care teams who reflect the communities the study seeks to recruit and considering relevant populations from the outset rather than as an afterthought.

At the same time, the regulatory landscape is shifting. While current FDA diversity requirements remain framed as guidance rather than mandates, the direction of travel is clear.

Addressing diversity in clinical trials is not solely about anticipating future compliance requirements. It is about redesigning the clinical research process so that representativeness is built in from the start. Sponsors who recognize this and act early will not only stay ahead of regulation but also contribute to a more equitable and credible evidence base for medicine.

Olga Elizarova, DDS, MPH, MBA, senior consultant behavioral science, S3 Connected Health