Despite The FDA’s Five-Year Plan, Black Patients Remain Inadequately Represented In Clinical Trials For Drugs

Diverse clinical trial participation is vital to ensuring the safety and effectiveness of medical products for all patients, but few trials have adequate representation from non-White racial and ethnic groups.^1,2 The 2010 Affordable Care Act included requirements for collecting and reporting the demographic data of clinical trial participants. However, in 2013, in furtherance of provisions in the Food and Drug Administration Safety and Innovation Act of 2012, the Food and Drug Administration (FDA) identified widespread “inconsistencies,” particularly around race, in industry’s adherence to these requirements.³

In response, in 2014 the FDA introduced a five-year action plan to “encourage greater participation of diverse populations in clinical trials” and generate “meaningful information on demographic subgroups” for medical decision making.^3,4 Arguably, the centerpiece of the initiative was the creation of Drug Trials Snapshots, a public-facing website posting information on the representativeness of pivotal clinical trials. Snapshots was piloted in 2014 and officially launched in 2015.⁵ Snapshots data allow for evaluation of whether the FDA’s action plan successfully increased the inclusion of racially and ethnically diverse patients in clinical trials and the reporting of race-specific outcomes.

In an effort to assess the impact of the FDA’s 2015 five-year action plan, we analyzed Snapshots data for drugs approved from 2014 to 2021 with race-specific incidence or prevalence data available for the respective indications. We focused on the inclusion of Black patients, for whom health and clinical trial access disparities are most robustly documented.^6–9 We evaluated whether the FDA Snapshots initiative led to an improvement in the inclusion of Black patients in pivotal drug clinical trials and the extent to which this initiative fulfilled the FDA’s goal of providing publicly available race-specific data on the benefits and side effects of these drugs.

Study Data And Methods

Data

We manually extracted data on all newly approved drugs posted on Drug Trials Snapshots from May 1, 2014, to July 2, 2021.⁵ Data included drug name; manufacturer; approval date; indication; National Clinical Trial identifier number; trial start date; numbers of Black and White participants; and racial differences, if any, in benefits and side effects.

Analysis

For each drug’s indications, and to mirror the FDA’s call in the action plan that clinical trial enrollment should mirror prevalence, we compared the share of Black and White patients in the study with the share of disease that Black and White patients shoulder in the population.¹⁰ We did so through a systematic, hierarchical strategy prioritizing federal sources such as the Centers for Disease Control and Prevention over peer-reviewed publications and incidence rates over prevalence rates (see online appendix exhibit A1).¹¹ We excluded indications for which we could not estimate relative disease burden because of a lack of high-quality race-specific data on incidence or prevalence (appendix exhibit A2).¹¹ When incidence or prevalence data by race for an FDA Snapshots drug indication were not available, we considered the inclusion of a broader disease condition that could be used as a reasonable substitute for the more specific indication.

For each drug-indication combination we calculated a representativeness ratio, determined by assessing the ratio of the proportion of Black trial participants to White trial participants relative to the proportion of the disease burden in Black patients relative to White patients in the US population. Values below 1 indicate that Black participants were underrepresented in the clinical trial relative to their disease burden. For our outcome measures we assessed medians and interquartile ranges overall and by trial start date and approval year and tested for interyear differences, using the Kruskal-Wallis test. Each drug-indication combination was assigned its earliest pivotal clinical trial start date. The representativeness ratio over trial start dates was compared for the following periods in relation to the FDA’s action plan: trials starting before the action plan (before August 20, 2013), during the plan’s two-year rollout period (August 20, 2013–August 31, 2015), and during the action plan (September 1, 2015–July 2, 2021). We also calculated the percentage of Snapshots drugs offering race-specific data on benefits and side effects and examined whether there was a trend over time, using the Cochran Armitage test. Statistical tests were performed in SAS version 9.4 with a two-sided $p$ value cutoff of 0.05.

Limitations

Our analysis was limited in some cases because of the lack of published race-specific incidence data for several of the disease indications for which companies obtained treatment approvals. This limitation points to a need for companies pursuing indications to both report to the FDA and publicly publish the demographics of the population they hope to serve. Lacking these data, we used federal population-based sources for these data whenever possible, but in some cases, the data were derived from smaller cohorts that we found in peer-reviewed publications. Furthermore, it may be too early to see the full impact of the FDA action plan on clinical trial representativeness. However, we anticipated that by the conclusion of its five-year run, we would have observed some upward trends in trial representativeness. Our study did not include other racial or ethnic groups or vulnerable populations, such as Hispanic, Asian, or older patients, for which additional research is warranted.

Study Results

Drug Trials Snapshots posted 290 drugs (293 drug-indication combinations) from 2014 to 2021. The indications on Snapshots for the drugs of interest spanned a broad range of diseases including autoimmune diseases ($n=16$), cancer ($n=84$), cardiovascular disease ($n=10$), diabetes ($n=6$), dermatology ($n=6$), genetic inherited diseases ($n=17$), gastrointestinal diseases ($n=9$), hematology ($n=6$), infectious diseases ($n=27$), neurology ($n=16$), ophthalmology ($n=7$), psychiatry ($n=11$), pulmonary diseases ($n=5$), and other diseases ($n=12$) that did not fit into these categories. One drug included had three indications, and five drugs had two indications; all other products had a single indication.

Representativeness Of Pivotal Clinical Trials

Disease burden information was available for 225 (232 drug-indication combinations) of the 290 drugs (appendix exhibit A3).¹¹ These 225 drugs constituted our analytic group, and for them the median representativeness ratio was 0.31 (IQR: 0.14–0.69) (exhibits 1 and 2). Black participants were underrepresented in 198 (85.3 percent) trials and in all disease categories except psychiatry (median RR: 2.21; IQR: 0.85–3.07) (exhibit 2), with this exception largely driven by a single drug (Caplyta [lumateperone], RR: 13.67) (data not shown). Black participants were least represented in trials for cardiovascular disease treatments (median RR: 0.19; IQR: 0.12–0.25) (exhibit 2). There was no evidence that the representativeness ratio differed by drug approval year ($p=0.94$). There also was no evidence that the representativeness ratio differed over the periods before the action plan (114 drug indications), during the plan’s two-year rollout period (69 drug indications), and during the action plan (49 drug indications) ($p=0.22$). When we evaluated a narrower (two-year) pre–action plan window (August 20, 2011–August 20, 2013), the results were similar ($p=0.49$) (exhibit 2).

SOURCE Authors’ analysis of data from the Food and Drug Administration’s Drug Trials Snapshots website, May 2014–January 2021. NOTES A total of 34 trials equally or overrepresented Black patients relative to the population. A total of 198 trials underrepresented Black patients relative to the population.

SOURCE Authors’ analysis of data from the Food and Drug Administration’s Drug Trials Snapshots website, May 2014–January 2021. NOTES Representation ratio was calculated by dividing the ratio of Black-to-White representation in clinical trials by the ratio of Black-to-White disease burden in the population. Whiskers represent the interquartile range. The lower bound for “Children” is undefined, as several clinical trials had zero Black enrollees. For “Year of trial start date,” the time periods are specified in the text. For “Year of drug approval,” partial-year data were available for 2014 and 2021.

In addition to lumateperone, several other drugs could be considered outliers (appendix exhibit A5).¹¹ Black participants were underrepresented in trials for Adakveo (crizanlizumab-tmca; RR: 0.35) and Oxbryta (voxelotor; RR: 0.06) for sickle cell disease, which has a uniquely high prevalence in Black Americans. Trials for Pretomanid (pretomanid; RR: 53.71) and Krintafel (tafenoquine; RR: 3.07) for tuberculosis and malaria, respectively, had representativeness ratio values greater than 1. These diseases have low prevalence in the US, and the racial makeup of these trials likely reflects predominantly international participants.

Availability Of Data For Black Patients

Of 290 drug approvals listed on Snapshots, the vast majority offered no data on benefits (236, 81.4 percent) or side effects (232, 80.0 percent) for Black patients (exhibit 3). There was no evidence that the availability of this information improved over time ($p=0.70$ for benefits, $p=0.07$ for side effects). Among drugs with race-specific reporting, “no difference” was the most common result for both benefits (53, 98.1 percent) and side effects (53, 91.4 percent). Five drugs reported race-specific differences: One (Eucrisa [crisaborole]) worked better in White than in Black participants, two (Eucrisa and Annovera [segesterone acetate and ethinyl estradiol]) showed increased risks for side effects in White patients compared with Black patients, and three (Entresto [sacubitril/valsartan], Vyleesi [bremelanotide], and Aristada [aripiprazole lauroxil]) had increased risks for side effects in Black patients compared with White patients.

SOURCE Authors’ analysis of data from the Food and Drug Administration’s Drug Trials Snapshots website, May 2014–January 2021. NOTE Partial-year data were available for 2014 and 2021.

Discussion

During the five-year period of the FDA’s action plan to improve clinical trial representativeness, reporting, and transparency, the relative representation of Black participants to White participants in pivotal trials leading to drug approval remained steady. The median representation was one-third the disease burden in the population. Despite the requirement that every pivotal study include race-specific reporting of benefits and risks, only 20 percent of studies provided results on these endpoints for Black patients; this finding was also consistent across the period we assessed.

Our findings suggest that the action plan has not improved the racial representativeness of pivotal drug trials or the availability of race-specific outcomes data. It is hard to know why, but we note that the action plan set no minimum standards for adequate subgroup inclusion; rather, the FDA collaborated with industry and other stakeholders to develop nonbinding recommendations for broadening clinical trial participation, enhancing the quality of demographic data collection and reporting,⁴ and achieving “greater assurance in the safety and effectiveness of the medical products used by a diverse population.”⁴ The FDA’s Drug Trials Snapshots website itself was motivated by the expectation that transparency would help encourage the industry to improve. Unfortunately, there is no evidence that the FDA’s expectations were met. Indeed, FDA officials themselves used Snapshots data in 2015 to uncover important underrepresentation of women in trials of treatments for cardiovascular disease.¹² The FDA also released new guidance in November 2020 that includes additional recommendations for improving trial representativeness. However, as with all FDA guidance, these recommendations are nonbinding.¹³

The FDA should now consider setting minimum thresholds for diverse inclusion, particularly in later-phase clinical trials.

The FDA should now consider setting minimum thresholds for diverse inclusion, particularly in later-phase clinical trials, so that trial participants mirror disease demographics and trials generate adequate data specific to patients from diverse racial and ethnic groups. Although this could require a new congressional regulation to allow the FDA such authority, such a step need not meaningfully impede the conduct of clinical research. During the past twenty years, diverse enrollment in trials funded by the National Cancer Institute, which require sufficient inclusion of minorities for subgroup analyses as a result of the NIH Revitalization Act of 1993, has increased to levels representative of the disease burden in the population overall.¹⁴ Furthermore, industry-sponsored trials conducted at US sites have better representation of Black patients than trials conducted at global sites.¹⁵ Thus, representational requirements for industry, which sponsors most clinical trials for new drug approvals,^16–18 could be met.

The FDA could also grant companies contingent approval (for example, accelerated approval) for drugs based on pivotal trials lacking adequate representativeness of the disease in the population and then require confirmatory trials for full regulatory approval to address the initial trials’ shortcomings. If applied in the context of race-specific enrollment objectives, this regulatory step would in effect declare that outcomes observed in White patients can be used as a surrogate, but not a substitute, for outcomes observed in the population of patients with the condition (a population that includes Black patients). Alternatively, the FDA could delay any approval until there is adequately diverse inclusion by monitoring enrollment in real time, similar to its requirement for industry to expand trial enrollment before COVID-19 vaccine approval under the federal Operation Warp Speed effort. Another approach, complementary to those identified above, is for the FDA to generate publicly available equity “report cards” for each pharmaceutical company. This strategy could bring greater transparency and accountability to the industry track record on racial inclusiveness in clinical trials and equitable access to novel treatments.

In fairness, increasing the diversity of clinical trials is no easy task. Numerous complex barriers exist to increasing the diversity and representativeness of trials, including social determinants of health; implicit bias; lack of trust; and a host of other persistent, structural factors. It is therefore critical for the FDA to strengthen its efforts to address the complexity of these barriers and, in turn, prioritize the resources and expertise required to overcome them. Achieving the goals of the action plan will require greater leadership and commitments by the FDA, with additional resources and support for research sponsors to overcome entrenched barriers to diverse representation in trials.

Conclusion

Black patients remain underrepresented in most trials underlying the approval of new drugs. This finding did not differ among trials that were started before or after the FDA’s 2015 five-year action plan was initiated. This plan does not seem to have been associated with improvements in either racial representativeness in these trials or the proportion in which race-specific outcomes are reported. Addressing this shortfall may require further action by the FDA beyond the steps already pursued under its action plan. These findings suggest that the FDA should consider a new approach to improving clinical trial representativeness, reaching beyond reporting- and transparency-centered measures to implement representational requirements.

ACKNOWLEDGMENTS

Angela Green reports receiving personal fees from Clinical Congress Consultants and research funding paid to the institution from Mereo BioPharma. Peter Bach received funding from Arnold Ventures (Grant No. NCI P30 CA008748). The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Bach also reports employment with Delfi Diagnostics; personal fees from the American Society for Health-System Pharmacists, WebMD, Defined Health, JMP Securities, Mercer, United Rheumatology, Foundation Medicine, Grail, Morgan Stanley, NYS Rheumatology Society, Oppenheimer & Co., Cello Health, Oncology Analytics, Anthem, Magellan Health, Kaiser Permanente Institute for Health Policy, Congressional Budget Office, America’s Health Insurance Plans, Geisinger, EQRx, Meyer Cancer Center of Weill Cornell Medicine, and National Pharmaceutical Council; nonfinancial support from the American Society for Health-System Pharmacists, United Rheumatology, Oppenheimer & Co., Oncology Analytics, Kaiser Permanente Institute for Health Policy, Congressional Budget Office, America’s Health Insurance Plans, and Geisinger; other support from Oncology Analytics; and grants from Kaiser Permanente and Arnold Ventures.

NOTES

• 1 Taylor AL, Wright JT Jr. Should ethnicity serve as the basis for clinical trial design? Importance of race/ethnicity in clinical trials: lessons from the African-American Heart Failure Trial (A-HeFT), the African-American Study of Kidney Disease and Hypertension (AASK), and the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Circulation. 2005;112(23):3654–60, discussion 3666. Medline, Google Scholar
• 2 Oh SS, Galanter J, Thakur N, Pino-Yanes M, Barcelo NE, White MJet al. Diversity in clinical and biomedical research: a promise yet to be fulfilled. PLoS Med. 2015;12(12):e1001918. Crossref, Medline, Google Scholar
• 3 Food and Drug Administration. Collection, analysis, and availability of demographic subgroup data for FDA-approved medical products [Internet]. Silver Spring (MD): FDA; 2013 Aug [cited 2022 Jan 13]. Available from: https://www.fda.gov/files/about%20fda/published/Collection--Analysis--and-Availability-of-Demographic-Subgroup-Data-for-FDA-Approved-Medical-Products.pdf Google Scholar
• 4 Food and Drug Administration. FDA action plan to enhance the collection and availability of demographic subgroup data [Internet]. Silver Spring (MD): FDA; 2014 Aug [cited 2022 Jan 13]. Available from: https://www.fda.gov/media/89307/download Google Scholar
• 5 Food and Drug Administration. Drug Trials Snapshots 2015 [Internet]. Silver Spring (MD): FDA; 2015 [last updated 2020; cited 2022 Jan 13]. Available from: https://www.fda.gov/drugs/drug-approvals-and-databases/drug-trials-snapshots Google Scholar
• 6 National Center for Health Statistics. Health, United States, 2015: with special feature on racial and ethnic health disparities [Internet]. Hyattsville (MD): NCHS; 2016 May [last updated 2017 Jun 22; cited 2022 Jan 13]. Available from: https://www.cdc.gov/nchs/data/hus/hus15.pdf Google Scholar
• 7 Bailey ZD, Krieger N, Agénor M, Graves J, Linos N, Bassett MT. Structural racism and health inequities in the USA: evidence and interventions. Lancet. 2017;389(10077):1453–63. Crossref, Medline, Google Scholar
• 8 Zimmerman FJ, Anderson NW. Trends in health equity in the United States by race/ethnicity, sex, and income, 1993–2017. JAMA Netw Open. 2019;2(6):e196386. Crossref, Medline, Google Scholar
• 9 Centers for Disease Control and Prevention. Infant mortality and low birth weight among black and white infants—United States, 1980–2000. MMWR Morb Mortal Wkly Rep. 2002;51(27):589–92. Medline, Google Scholar
• 10 Census Bureau. QuickFacts United States [Internet]. Washington (DC): Census Bureau; 2019 [cited 2022 Jan 13]. Available from: https://www.census.gov/quickfacts/fact/table/US/PST045219 Google Scholar
• 11 To access the appendix, click on the Details tab of the article online.
• 12 Whyte J, Woodcock J, Wang J. Review of the Drug Trials Snapshots Program of the US Food and Drug Administration: women in cardiovascular drug trials. JAMA Intern Med. 2017;177(5):724–7. Crossref, Medline, Google Scholar
• 13 Food and Drug Administration. Enhancing the diversity of clinical trial populations—eligibility criteria, enrollment practices, and trial designs guidance for industry [Internet]. Silver Spring (MD): FDA; 2020 Nov [cited 2022 Jan 13]. Available from: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/enhancing-diversity-clinical-trial-populations-eligibility-criteria-enrollment-practices-and-trial Google Scholar
• 14 Ong MBH. Participation by minority racial, ethnic groups in NCI-funded trials nearly doubles in 20 years. Cancer Letter [serial on the Internet]. 2020 Jun 26 [cited 2022 Jan 13]. Available from: https://cancerletter.com/articles/20200626_1/ Google Scholar
• 15 Food and Drug Administration. 2015–2016 global participation in clinical trials report [Internet]. Silver Spring (MD): FDA; 2017 Jul [cited 2022 Jan 13]. Available from: https://www.fda.gov/files/drugs/published/2015---2016-Global-Clinical-Trials-Report.pdf Google Scholar
• 16 Gresham G, Meinert JL, Gresham AG, Meinert CL. Assessment of trends in the design, accrual, and completion of trials registered in ClinicalTrials.gov by sponsor type, 2000–2019. JAMA Netw Open. 2020;3(8):e2014682. Crossref, Medline, Google Scholar
• 17 Research!America. U.S. investments in medical and health research and development, 2013–2017 [Internet]. Arlington (VA): Research!America; 2018 [2022 Jan 13]. Available from: https://www.researchamerica.org/sites/default/files/Policy_Advocacy/2013-2017InvestmentReportFall2018.pdf Google Scholar
• 18 Buchkowsky SS, Jewesson PJ. Industry sponsorship and authorship of clinical trials over 20 years. Ann Pharmacother. 2004;38(4):579–85. Crossref, Medline, Google Scholar