Clinical Researcher—August 2020 (Volume 34, Issue 7)

PEER REVIEWED

Yaritza Peña; Zachary P. Smith; Kenneth A. Getz, MBA

 

It is well known that the underrepresentation of minority groups in clinical trials decreases the generalizability of clinical trial findings by disguising the potential effects of variation in the pathobiology of disease and race-related differences in drug responses. As a result, several regulatory policy initiatives have focused on developing clinical trial enrollment practices that improve the inclusion of diverse patient subpopulations.

The U.S. Food and Drug Administration (FDA) first released guidance about the importance of studying the effects of products in elderly patients in the 1980s.{1} A decade later, the agency issued a “Guideline for the Study and Evaluation of Gender Differences in the Clinical Evaluation of Drugs” and established the Office of Women’s Health. Despite the progress made as a result of these guidance documents, underrepresentation of racial and ethnic minorities in clinical trials remained highly prevalent.

In 2012, the U.S Congress passed the Food and Drug Administration Safety and Innovation Act (FDASIA) to address ongoing concerns over the lack of diversity and representation in clinical trials. Section 907 of the Act calls for the FDA to improve the inclusion and transparency of clinical trial data representing demographic subgroups.{1,2}  In 2013, a cross-agency task force involving representatives from the Office of the Commissioner, the Center for Biologics Evaluation and Research (CBER), the Center for Drug Evaluation and Research (CDER), and the Center for Devices and Radiological Health (CDRH) found that the FDA’s statutes, regulations, and policies generally provided product sponsors a solid framework for disclosing data on the inclusion of demographic subgroups in their applications.{1}

In 2014, the FDA responded with a new annual publication called “Drug Trial Snapshots.” This publication routinely discloses the extent to which Section 907 of the FDASIA is applied in biomedical research; the print and online versions present the demographic distribution of participants in clinical trials of approved New Molecular Entities (NMEs) for that given year as well as any observed differences in safety and efficacy by demographic subgroup.

Conclusions regarding these differences, however, cannot always be made from the Snapshot reports alone. The data they provide are limited to individual years, thwarting researchers from evaluating trends in participant subgroup demographics.

Aside from FDA recommendations, there are no regulations currently in place that require industry sponsors to include women and minorities in their trials and no programs that provide insight into missing data.{3,4} Perhaps most importantly, current guidance documents do not disclose the information necessary to assess disparities in demographic diversity given individual disease prevalence rates.

What We Need Versus What We Have

More comprehensive data on participant demographic subgroups may aid clinical research professionals in identifying opportunities to improve diversity in their research sites. Specifically, it can help to identify the areas of greatest need, including where demographic subgroup disparities are the greatest, both overall and within specific therapeutic areas or disease conditions.

The information can also be used to assess how participant diversity has changed over time. The availability of results may promote innovations in clinical trial design and avoid duplication of unsuccessful diversity programs or policies, thereby avoiding unnecessary risks to research participants.

To address the need for more comprehensive data and to establish a global baseline measure, in 2019, the Tufts Center for the Study of Drug Development (CSDD)—supported by a research grant from Merck Sharp & Dohme Corp.—conducted a study to address the following objectives:

• Assess the availability and disclosure of participant demographic subgroup data provided by pharmaceutical and biotechnology companies.
• Gather data to inform a baseline assessment of the extent of participant demographic subgroup disparities in the clinical trials of new drug approvals.
• Establish and convey an approach that the FDA, and other stakeholders alike, can apply to improve the value of the Drug Trial Snapshots program and other diversity initiatives.

Since supplemental trials are not required to be reported, this article focuses on disparity in pivotal trial data.

Methods

Tufts CSDD compiled participant demographic subgroup data (i.e., sex, race, ethnicity, age) from pivotal trials supporting all new drugs and biologics approved by the FDA between 2007 and 2017 (n=341). Most of the data were drawn from the FDA website. Tufts CSDD referred to publicly available sources, including ClinicalTrials.gov, medical reviews, and product labeling. Prevalence and incidence data were collected from published sources, including government websites, national health organizations, and peer-reviewed literature.

Tufts CSDD created a summary metric, called the “disparity percentage,” to characterize participant demographic subgroup underrepresentation. This metric is defined as the difference between total actual number of participants by subgroup and the expected level of subgroup representation, divided by the expected level of subgroup participation.

Disease prevalence rates were found in the peer-reviewed literature and public sources for 57% of all approvals. For the remaining 43%, U.S. census data were used as a proxy for the distribution of participant demographic subgroups, as it was assumed that prevalence was distributed proportionately among the population.

Data on 757 pivotal clinical trials and 592,168 study participants were analyzed. An example of the disparity percentage is shown in Figure 1:

Figure 1: Calculating a Disparity Percentage

Disease Condition for Approved Drug:	Peripheral T-Cell Lymphoma
Total Clinical Trial Participants:	788
“Actual” Distribution of Participants Who are Black or of African Descent:	3.7% (29 participants)
Expected or “Predicted” Distribution of Participants Who are Black or of African Descent:	13.5% (106 participants)
Disparity Percentage	-72.6%

Results/Discussion

Data Completeness

While government guidelines mandate that federally funded clinical research to disclose participant demographic data, race/ethnicity data remain incomplete and underreported. Nearly 20% of all drug and biologic approvals between 2007 and 2017 were missing data on participant race for all referenced pivotal trials. More surprisingly, 50% of drug approvals did not include participant ethnicity data on any of their trials (see Table 1).

The level of drug approval data completeness showed notable increases in participant representation by sex and age at 96.2% and 91.8%, respectively. The availability of demographic data for pivotal clinical trials showed a similar pattern, with higher completion rates for participant sex (89.7%), age (83.2%), and race (72.8%) and a considerably lower level of availability rate for study participant ethnicity (36.7%).

The availability of participant demographic subgroup data for all 757 pivotal clinical trials approved in the 10-year period was substantially low; only 36.7% had data available on participant ethnicity and 72.8% of trials had data on participant race. The dearth of available ethnicity data represents both the need to enroll more minorities in studies and the need to be more intentional in referencing health disparate populations.

Table 1: Data Transparency in NDAs and BLAs, 2007 to 2017

	NDAs and BLAs with Data Available on Participants (n=341)	% of Total	Pivotal Trials with Data Available on Participants (n=757)	% of Total
Sex	328	96.2%	679	89.7%
Race	282	82.7%	551	72.8%
Ethnicity	171	50.1%	278	36.7%
Age	313	91.8%	630	83.2%

Note: Drug data collected from the FDA website. Pivotal trial data collected from the FDA drug information portal for medical reviews and printed labeling for each approved drug.

Participant Demographic Subgroup Representation

The highest overall levels of underrepresentation were observed among participants of Black or of African descent, with nearly 47,000 fewer participants than expected (see Table 2). “Other” participants (e.g., Native American, Native Alaskan, Native Hawai’ian, or Pacific Islander) and Hispanic or LatinX participants were also under-represented, with 11,641 and 4,669 fewer participants than expected, respectively. Roughly 20,000 fewer women were enrolled in pivotal clinical trials than expected levels. Asian participants were over enrolled by more than 23,000 participants in pivotal trials, a disparity of +148.9%.

Overrepresentation among Asian participants may be due, in part, to market access requirements in key geographies including Japan and China.{5} However, country-specific variation in the characterization of demographic subgroups may also be a contributing factor. Some studies counted participants of Indian descent as Asian while others did not.

Treating minority populations as homogeneous assumes cultural beliefs and experiences are the same, which could potentially influence racial/ethnic stereotypes about patients and implicit biases in research settings.{6} Understanding cultural differences within subpopulations could emend the cycle of participant distrust in clinical research.

Moreover, inconsistent implementation of racial/ethnic classifications negatively impacts participant disparity percentages. Any significant differences found between groups differentially affects the generalizability of clinical research. Disaggregated analyses may increase our ability to understand exposures and health outcomes across subgroups.{7}

Table 2: Subgroup Disparities for Pivotal Trials (n=757)

	Sex		Race and Ethnicity
	Female	Male	White	Black	Asian	Hispanic/LatinX	Other
Total participants	252,586	309,844	346,884	24,612	39,244	32,877	13,612
Distribution of total participants	44.9%	55.1%	75.9%	5.4%	8.6%	7.2%	3.0%
Expected level of participation*	272,616	288,137	305,443	71,226	15,764	37,546	25,253
Expected distribution	48.6%	51.4%	67.1%	15.6%	3.5%	8.2%	5.5%
Difference	-20,030	+21,707	+41,441	-46,614	+23,480	-4,669	-11,641
Disparity percentage	-7.3%	+7.5%	+13.6	-65.4%	+148.9%	-12.4%	-46.1%

*Based on U.S census and disease prevalence.

Wide variation was observed in the disparity percentages for participant demographic subgroups by individual disease condition. Pulmonary/respiratory disease, neurology, and rheumatology require the most attention and remediation, with racial and ethnic disparities observed for more than 80% of the total approvals for these indications (see Table 3). While these diseases disproportionately affect non-white individuals, pivotal trials in these areas had the highest under-representation of Black/African Americans, Hispanic/LatinX and “Other” subgroups.

Black/African American representation in pivotal trials conducted during 2007 through 2017 was considerably low. Based on the analysis of the data available, three times as many Black/African American participants should have been enrolled in clinical trials during the period observed to be adequately represented by disease prevalence rates or by population census figures. Similarly, the Hispanic/LatinX community was highly underrepresented in pivotal trials of investigational oncology treatments. Gastroenterology and rheumatology were the two top therapeutic areas with high levels of Asian participant under-representation.

Table 3: Top Therapeutic Areas with Participant Demographic Disparities

Subgroup	Therapeutic Area	Approved Drugs which Underrepresent Demographic (>20%)	Average Disparity Percentage per Drug
Black/African American	Pulmonary/respiratory diseases	100%	-80%
	Rheumatology	100%	-80%
	Neurology	88%	-70%
Asian	Gastroenterology	100%	-86%
	Rheumatology	83%	-46%
Hispanic/LatinX	Oncology	93%	-63%
	Neurology	85%	-54%
	Pulmonary/respiratory diseases	80%	-51%
Other Racial Identities	Neurology	89%	-72%
	Pulmonary/respiratory diseases	86%	-72%
	Immunology	100%	-71%

Conclusion

Findings from the Tufts CSDD study highlight not only the need to improve transparency and reporting of clinical trial participant demographic data, but also the high level of participant subgroup under-representation in FDA-regulated pivotal trials during the past 11 years.

Developing trust between study participants and clinical research professionals begins with improvements in transparency and disclosure. The results of this study indicate efforts to improve participant diversity have not been broadly successful and more needs to be done.

This study has its limitations. The analysis is based on publicly available data. As a result, the findings may underestimate participant subgroup diversity levels. It is likely that sponsor companies collected but did not report participant demographics for some of their trials; further emphasizing the need for disclosure and reporting in the industry.

The results do not include an assessment of drug development programs that failed to receive FDA approval. Additionally, Tufts CSDD relied on U.S. census data to determine the expected or predicted level of population demographic representation when disease-specific prevalence rates were unknown. Future research will look to apply country-specific population census data and other study exclusion criteria to improve the accuracy of diversity assessment.

Low levels of trust, poor access, study participation burden, low education, and lack of clinical trial awareness are among the many barriers that contribute to minority under-representation in clinical research. Poor disclosure and transparency have contributed to public distrust.{8} Improvements in data reporting and completeness on participant demographic diversity will not only go far in improving public trust, they will also play a key role in guiding the clinical research enterprise in addressing the under-representation of participants by race and ethnicity.{9}

Authors’ Notes

Data collection began in 2018 and continued into 2019. While more current data are available now, these were not available at the time our data collection was completed and were out-of-scope for the project being conducted. Tufts plans to periodically update the dataset with more current data. In calculating the impact of FDASIA, we see little evidence of change over time for the years leading up to and after 2012, but in time an examination of this topic may make up its own paper.

Kenneth A. Getz reports an educational grant from the Investigator-Initiated Studies Program of Merck Sharp & Dohme Corp. during the conduct of the study.

References

1. U.S. Food and Drug Administration. 2019. Enhancing the Diversity of Clinical Trial Populations—Eligibility Criteria, Enrollment Practices, and Trial Designs (Draft Guidance for Industry). https://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/default.htm
2. U.S. Department of Health and Human Services. 2016. Final Rule—Clinical Trials Registration and Results Information Submission. Federal Register 81(183). https://s3.amazonaws.com/public-inspection.federalregister.gov/2016-22129.pdf
3. Diaz C. 2003. Increasing Diversity in Clinical Trials. Best Practices, Health Disparity Symposium. Bethesda, Md.
4. Getz K, Faden L. 2008. Racial Disparities Among Clinical Research Investigators. Am J Ther 15(1):3–11.
5. Comer B. 2020. Expanding the Tent: Improving Trial Participation Among Under-Represented Patient Populations. In Vivo. https://invivo.pharmaintelligence.informa.com/IV124476/Expanding-The-Tent-Improving-Trial-Participation-Among-UnderRepresented-Patient-Populations
6. Pérez-Stable E J, El-Toukhy, S. 2018. Communicating with Diverse Patients: How Patient and Clinician Factors Affect Disparities. Patient Ed and Counsel 101(12):2186–94. https://doi.org/10.1016/j.pec.2018.08.021
7. Islam NS, Khan S, Kwon S, Jang D, Ro M, Trinh-Shevrin C. 2010. Methodological Issues in the Collection, Analysis, and Reporting of Granular Data in Asian American Populations: Historical Challenges and Potential Solutions. J Health Care Poor Underserved 21(4):1354–81. doi:10.1353/hpu.2010.0939
8. Smirnoff M, Wilets I, Ragin DF, Adams R, Holohan J, Rhodes R, Winkel G, Ricci EM, Clesca C, Richardson LD. 2018. A Paradigm for Understanding Trust and Mistrust in Medical Research: The Community VOICES Study. AJOB Empirical Bioethics 9(1):39–47. https://doi.org/10.1080/23294515.2018.1432718
9. Joshi M, Bhardwaj P. 2018. Impact of Data Transparency: Scientific Publications. Persps in Clin Res 9(1):31–6. https://doi.org/10.4103/picr.PICR_104_17

Yaritza Peña (yari.pena@tufts.edu) was a Research Analyst with the Tufts Center for the Study of Drug Development in Boston, Mass. at the time of the study detailed in this article, and is continuing to pursue a Masters at Tufts while working as a Project Manager at Takeda Pharmaceuticals in Cambridge, Mass.

Zachary P. Smith (zachary.smith605922@tufts.edu) is a Research Analyst with the Tufts Center for the Study of Drug Development, Tufts University School of Medicine, in Boston, Mass.

Kenneth A. Getz, MBA, (kenneth.getz@tufts.edu) is Deputy Director and Professor at the Tufts Center for the Study of Drug Development, Tufts University School of Medicine, and Founder and Board Chair of the Center for Information and Study on Clinical Research Participation.