Myths about diversity in clinical trials reduce return on investment for industry

To the Editor — Misconceptions in drug development often begin at the earliest stages. People enrolled in clinical trials are in general not representative of the population with medical need¹, but the problem is particularly pernicious in early, first-in-human phase 1 studies. This is, in part, a result of stricter eligibility criteria in early studies, and is also due to clustering of phase 1 trial centers in urban locations that are less accessible to patients in rural or underserved areas. Efforts to enhance or mandate enrollment of a diverse population in phase 1 trials are often deferred, in the belief and expectation that later phase 2 and 3 studies can address this need. Yet limiting access to new therapies while they are being tested in phase 1 trials fails to promote inclusive research and is also ethically concerning. In an interview study of 100 adult patients with solid tumors, a significant majority believed that access to investigational new drugs is a fundamental right and yet is also too difficult, citing frequent ineligibility and time-consuming enrollment as barriers to entering these early-phase studies².

Phase 1 studies provide key safety data, can detect preliminary signals of efficacy and, in an era of expedited programs, may even enable accelerated approval of new drugs. The benefits of recruiting patients to a phase 1 trial cannot be reconciled with the longstanding belief in industry that clinical trial participants from under-represented populations need to be included only in the later stages of drug development. The recent accelerated US Food and Drug Administration (FDA) approval of the antibody–drug conjugate sacituzumab govitecan (Trodelvy; Gilead) highlights the dilemma facing drug developers and regulators. Although Black women have mortality rates for triple-negative breast cancer that are 41% higher than those of non-Hispanic white women, FDA approval was based on a study of 108 patients in which only 7% were Black³. Ultimately, a subgroup analysis of the confirmatory phase 3 study revealed an equal clinical benefit for Black patients, but it would have been much better to address efficacy in this patient group at an earlier stage, to determine the true impact of the approval of this medication.

As researchers in industry, we have been concerned to observe the lack of diversity in phase 1 clinical trials of perhaps the most significant development of recent times, vaccines for COVID-19. Forty of the 45 patients in the Moderna mRNA vaccine phase 1 study were white⁴; similarly, the Oxford COVID Vaccine Trial Group reported that the first-in-human study of their adenovirus-based vaccine consisted of “fairly young and healthy volunteers, the majority of whom were white”⁵. There is no doubt that these initial studies were remarkable in their innovation and speed, and later studies explicitly enrolled more diverse populations. But, beyond the moral and scientific imperatives, lack of inclusivity in the earliest phases of drug development can meaningfully impact and impair public trust, particularly among under-represented populations. If a drug is not tested in a representative sample at an early stage, this could lead to hesitancy to take the treatment; indeed, under-represented, minority populations in both the UK and USA saw reduced uptake of the COVID-19 vaccines.

As drugs that we have developed move from early development into later-stage studies, we have personally encountered unsubstantiated fear that diversity may lead to heterogeneous and difficult-to-interpret results. Yet we argue that, to truly understand how medicines work in the real world, clinical trials must be purposefully designed to reflect the populations who will use them. Inclusive research can reveal potential variations in outcomes in subgroups, as well as yielding opportunities to identify unique or specific responses, as observed in a study of the epidermal growth factor receptor (EGFR) inhibitor erlotinib (Tarceva; Genentech, Roche) in non-small-cell lung cancer (NSCLC)⁶. Both women and people of Asian descent had higher response rates than the overall trial population, a difference that was ultimately attributed to the L858R EGFR mutation, which is common in women of Asian descent. The enrollment of a diverse population in this study illuminated a predictive pharmacogenetic marker and uncovered an efficacy signal that has brought benefit to a myriad of patients.

Another myth to bust is that enrolling heterogeneous populations automatically equates to heterogeneity in outcomes. For example, the Surveillance, Epidemiology, and End Results databases indicate that Black patients with colorectal cancer in the real world have a 32% higher mortality risk than white people. Yet recent analysis suggests that this arises from unequal access to care: notably, Black and white patients with colorectal cancer had similar outcomes in a well-controlled clinical trial in which patients received similar healthcare⁷. Artificial intelligence provides similar evidence that a broader, more inclusive patient population will not alter trial outcomes. Trial Pathfinder used the Flatiron Health database to simulate recent large phase 3 studies in NSCLC⁸. When eligibility criteria were either fully relaxed or broadened with a data-driven approach, the number of eligible patients was drastically increased without affecting overall survival hazard ratios.

The specter of financial pragmatism looms large in industry, and we are frequently asked whether delays to recruit diverse participants will incur additional costs. Our first-hand evidence suggests that this is not the case. For example, in two similar studies examining the role of tocilizumab in COVID-19 pneumonia, speed of enrollment in the Empacta⁹ study (16% white) was comparable to that for Covacta¹⁰ (58% white). On the contrary, inclusive research will be a valuable ally to financial budgets. Implementing a diverse recruitment strategy in early development can improve the accuracy of key stop/go decisions, halting costlier pivotal studies on ineffective treatments and leading to longer-term savings. Enrollment of patients from historically under-represented groups will also afford an opportunity to make greater use of community hospitals. As these sites have reduced overhead costs and require fewer resources, this is a cost-effective strategy.

Embracing inclusion and increasing diversity at all stages of clinical trials can create a virtuous loop with tangible return on investment: inclusive trials with diverse patient populations make it easier to attract subjects from a broader pool; the patients enrolled may be more willing to stay in the study and become more engaged; and this, in turn, will increase the scientific and financial robustness of the study. Partnership between industry, clinical research sites, patient advocates and under-represented and excluded communities will be key to achieving these goals (Table 1).

Table 1 Dispelling DEI myths within industry creates benefits

To quote James Baldwin, “Not everything that is faced can be changed, but nothing can be changed until it is faced.” The commitment to facing our misconceptions within biotech is central to delivering the change needed to deliver diverse, equitable and inclusive research.