Black, Hispanic populations underrepresented in trials supporting FDA cancer drug approvals
Landmark trials that led to FDA oncology drug approvals during the past decade frequently underrepresented black and Hispanic populations, according to results of a database study published in JAMA Oncology.
Researchers also observed poor reporting of race in these trials.
“Our findings show that the science might not be applicable to the population that’s going to receive the medications,” Jonathan M. Loree, MD, assistant professor in the department of medicine, division of medical oncology, at University of British Columbia, said in a press release. “If patients are going to be receiving this drug, we need to know that it’s going to work for them with the same effectiveness that’s seen in the trial.”
Loree and colleagues, along with researchers from The University of Texas MD Anderson Cancer Center, Fred Hutchinson Cancer Center and Baylor University, investigated race reporting and representation in trials that served as the basis for FDA approval of oncology drugs between July 2008 and June 2018. The researchers searched PubMed and clinicaltrials.gov to acquire primary reports of trials, and reviewed FDA archives to identify approvals. They determined the U.S. population-based cancer estimates by race using the SEER and U.S. Census databases.
The percentage of trials reporting race and the racial breakdown of trial populations served as primary endpoints. Race subgroup analyses reporting and disparities between race representation in trials and in the U.S. population served as secondary outcomes.
The researchers identified 232 clinical trials that led to 204 FDA drug approvals over the past decade (73.5% full and 26.4% accelerated) for patients with solid tumors and hematological cancers. The approvals applied to 108 drugs with various indications. Subsequent analyses excluded two trials that had no available report.
The remaining 230 trials included a total of 112,293 participants. Of these trials, 145 (63%) reported on a minimum of one race, 18 (7.8%) reported on the four major races in the United States (white, Asian, black and Hispanic) and 58 (25.2%) included at least one subgroup analysis by race.
One-hundred-forty-four trials (62.6%) reported white race, 110 (47.8%) reported Asian race, 88 (38.2%) reported black race and 23 (10%) reported Hispanic race.
During the study period, researchers observed little change in the proportion of trials reporting race. Between July 2008 and June 2013, 45 trials (56.6%) reported race vs. 100 trials (67.1%) from July 2013 to June 2018 (OR = 1.63; 95% CI, 0.93-2.87). The proportion of trials that reported race subgroup analyses also increased modestly (16.6% vs. 30.2%; OR = 2.26; 95% CI, 1.16-4.67).
Whites represented 76.3% of trial participants, followed by Asians (18.3%), blacks (3.1%) and Hispanics (6.1%). The proportion of Hispanic participants increased slightly in the second half vs. the first half of the study period (6.7% vs. 5.3%; OR = 1.27; 95% CI, 1.08-1.49), whereas the proportion of black participants declined (2.9% vs. 3.6% OR = 0.81; 95% CI, 0.73-0.9).
Relative to their proportion of cancer incidence in the United States, blacks (22% of expected) and Hispanics (44% of expected) were underrepresented compared with whites (98% of expected) and Asians (438% of expected).
Trials related to targeted therapy (72.6%) and chemotherapy (78.9%) had a greater likelihood of reporting race than studies of other drug modalities. Likewise, solid tumor trials (69.9%) appeared more likely than hematology trials (48.6%) to report race.
Global recruitment may have affected the various racial and ethnic mixes, researchers acknowledged. They also noted that race representation should be increased in combination with further study on the biological factors contributing to racial differences.
Loree said the findings also are significant in Canada, where the same trials often serve as the basis for drug approvals.
“One thing particularly relevant to the Canadian context is that we weren’t able to analyze the participation of Native Americans in trials because there were only 13 patients reported out of a total of 112,000 patients,” Loree said. “That’s shocking and definitely shows an area where improvement is needed.” – by Jennifer Byrne
Disclosures: The researchers report no relevant financial disclosures.