Inadequate demographic representation calls trial generalizability into question

Neither premarketing nor postmarketing studies used by the Food and Drug Administration (FDA) to evaluate new cancer therapies adequately represent cancer patients in the U.S., a cross sectional study revealed.

In an analysis of 77 premarketing studies and 56 postmarketing studies, Black patients in particular were underrepresented in both sets of studies, as were older adults, despite the fact that older adults are at greater risk for cancer, senior author Gary Gross, MD, Yale School of Medicine, New Haven, Connecticut, and colleagues reported in JAMA Network Open.

In contrast, women were adequately represented in both pre- and post-marketing studies while Asian patients were in fact overrepresented, investigators noted.

Discerning the safety and efficacy of therapeutics for the wider population requires that the clinical trial participants be representative of the patient population, Gross and colleagues noted. “Inadequate demographic representation in premarketing and postmarketing studies calls into question the generalizability of cancer clinical trial results to clinical practice as well as the FDA’s ability to encourage or enforce demographic representation in clinical studies,” they argued.

Between 2012 and 2016, the FDA approved a total of 45 different cancer therapies, including both novel drugs and biologics, Gross and colleagues found.

They calculated the percentage of trials reporting sex, age, and race or ethnicity, as well as the participation to prevalence ratios (PPRs), which were determined by dividing the percentage of a particular population among study participants by the percentage of the particular population in the disease population, they explained. Underrepresentation was defined as a PPR less than 0.8.

Postmarketing studies were less likely than premarketing studies to report the sex of participants (75% vs 100%P<0.001), the investigators pointed out. Conversely, there was no difference between post and pre-marketing studies when it came to reporting either the proportion of older adults in a given study or their ethnicity.

The study authors found that, at a mean PPR of 0.91 (95% CI, 0.90-0.91) in premarketing studies and a mean PPR of 1.00 (95% CI, 1.00-1.01), women were adequately represented in both types of studies across all approved therapeutics. Similarly, White patients were adequately represented in both premarketing studies (mean PPR 0.95 [95% CI, 0.94-0.85]) and postmarketing studies (mean PPR 0.92 [95% CI, 0.91-0.93]).

In contrast, at a mean PPR of 0.73 (95% CI, 0.72-0.74) in premarketing studies and a mean PPR of 0.75 (05% CI, 0.75-0.76) in postmarketing studies, older adults were underrepresented in both types of studies across a number different cancer types, the authors found. Black patients were also underrepresented in both premarketing (mean PPR 0.32 [95% CI, 0.31-0.32]) and postmarketing studies (mean PPR 0.21 [95% CI, 0.20-0.22]), again across a wide range of different cancer types.

Perhaps surprisingly, Asian patients were overrepresented in both pre- and post-marketing studies at a mean PPR of 3.99 (95% CI, 3.99-4.00) and 4.26 (95% CI, 4.26-4.27), respectively. The authors hypothesized that this overrepresentation may simply reflect the fact that several studies in the sample analyzed were performed in Asian countries.

Obstacles to Black patient participation in industry-sponsored trials include factors such as transportation, child care and access to health care — and, the study authors added, Black patients often carry a disproportionate burden of comorbidities, which often excludes them from clinical trial eligibility.

“Furthermore, there is a well-documented and justified mistrust of research institutions among Black patients stemming from a long history of exploitation of Back patients by medical research,” the authors pointed out.

As for the low numbers of older patients in both pre and post-marketing studies, the authors again pointed to exclusion criteria based on comorbidities or the use of concomitant medications as barriers leading to underrepresentation.

Given that exclusion based on comorbidities or concomitant medication use disproportionately impacts older adults and Black patients, the study authors argued that these criteria “should be included only when necessary.”

“Ultimately, these findings suggest that the FDA must do more to improve the transparency of clinical trial demographic data and to ensure that women, older adults, and racial/ethnic minority groups are adequately represented in premarketing and postmarketing studies of novel cancer therapeutics,” the study authors concluded.

Commenting on the findings, Joseph Unger, PhD, Fred Hutchinson Cancer Research Center, Seattle, Washington, agreed with the authors that the clinical trial system in the U.S. is beset by many obstacles, which together generate clinical trial cohorts that are “distinctly nonrepresentative” of the U.S. population as a whole. However, he argued, rectifying the problem is not so easy.

“Full representation of demographic and socioeconomic groups is clearly desirable, for reasons of patient equity and scientific inquiry,” he wrote. “However, given the challenges of achieving this, the potential trade-offs should be considered. If fully ascribed to, the requirement to achieve fixed demographic enrollment targets could in fact delay new drug development, as trials struggle to achieve required accrual, especially among older patients. If so, this well-intentioned strategy could end up hurting the very populations it aims to help by delaying access to novel therapeutics…

“As [Gross] et al suggest, a prospective strategy that better targets treatment settings with diverse cancer populations will improve demographic and socioeconomic representativeness and limit the potential for harmful delays in the development of new treatments,” he concluded. “At the same time, continued and focused efforts to eliminate structural, clinical, physician, and patient barriers to enrollment are necessary to allow patients of all backgrounds to more easily participate. By such comprehensive means, the United States may finally begin to recognize the ideal of an open, inclusive clinical trial system readily accessible to all patients with cancer.”

  1. Neither premarketing nor postmarketing studies used by the FDA to evaluate new cancer therapies adequately represent cancer patients in the U.S.

  2. This inadequate representation calls into question the generalizability of clinical trial results to the underrepresented groups, including Black patients and older adults.

Pam Harrison, Contributing Writer, BreakingMED™

Gross reported receiving funding from Genentech and Johnson and Johnson as well as travel and speaking fees from Flatiron Health.

Unger had no conflicts of interest to declare.

Cat ID: 332

Topic ID: 78,332,730,332,935,192,150,151,725,925