Workshop Examines the Use of Race and Ethnicity in Genomics and Biomedical Research
The National Institute on Minority Health and Health Disparities (NIMHD) and the National Human Genome Research Institute (NHGRI) hosted a two-day workshop in October at the NIH Neuroscience Center Building in Rockville, Maryland, to discuss the use of race and ethnicity data in biomedical and clinical research and their application to minority health and health disparities research. The meeting brought together more than 60 scientists and thought leaders in genetics, medicine, epidemiology, social sciences, and law, as well as representatives from the National Academy of Sciences, Science magazine, and various NIH Institutes.
“Over the past decade, much has changed regarding the scientific understanding of race and ethnicity, especially in the context of genetic and genomic research,” said Dr. Eric Green, director of the National Human Genome Research Institute, noting that the workshop was the first workshop focused on race and ethnicity in genomics and biomedical research to be held at NIH since 2004. “There is a growing need to consider how both genetic and non-biological descriptors of race and ethnicity should be used in biomedical research and to increase minority participation in research.”
The three main objectives of the workshop were to develop a framework and generate recommendations for the appropriate use of race and ethnicity information in biomedical research, identify research questions to advance understanding of self-identified race and ethnicity (SIRE) and ancestry informative markers (AIMs) in genomics and biomedical research, and generate recommendations for how genomics and biomedical research can describe research participants’ diverse backgrounds and experiences in scientifically and socially meaningful ways.
“There are compelling reasons to measure race and ethnicity in biomedical research from both a biological and behavioral perspective,” said Dr. Eliseo Pérez-Stable, director of NIMHD. “Race, in terms of genetic ancestry, is an important factor in advancing knowledge of disease mechanisms and is increasingly becoming a key predictor of variance in clinical outcomes. Race is also a basic demographic factor and a proxy for sociopolitical status that can be used to track inequity in health care.”
In his opening remarks, Pérez-Stable added, “The definition of race by self-identification [SIRE] is seen as the gold standard in research. However, perceived race, or how an individual’s race is defined by society, is often more important than SIRE in determining the effects of race on health. The challenge of how to measure race is exacerbated by the growing numbers of mixed or multiracial individuals in racially and ethnically diverse regions such as Latin America and parts of the United States.”
The first day of the workshop included presentations and discussions on specific aspects of race and ethnicity in research. These sessions covered whether and how SIRE and AIMs should be used in genomics and biomedical research, minority health research and health disparities research, and reporting of genetic variation in research and non-research contexts, such as clinical and direct-to-consumer testing.
On the second day, presentations and discussion focused on facilitating collaboration between genomics and social sciences, utilizing electronic health records in race- and ethnicity-related biomedical research, and the potential roles of the National Academy of Sciences and scientific journals in implementing recommendations from the workshop. The workshop concluded with discussion sessions to formulate topics for discussion to share with the broader scientific community.
Discussion moderator Dr. Chanita Hughes-Halbert of the Medical University of South Carolina remarked that the conference “was an important first step to ensuring that…studies have adequate representation of diverse populations. Limited racial/ethnic diversity in studies is an ongoing issue; this conference outlined strategies that could have a significant impact on improving the participation of these populations in genomic studies and biomedical research.” Presenter Dr. Eric Boerwinkle, from the University of Texas Health Science Center at Houston, said, “Inclusion of diverse groups in biomedical research, especially genomic research, is scientifically and culturally important. Including diverse groups will increase the range of genomic and social variation, which in turn should increase discovery and impact. This workshop was a critical catalyst of a dialogue and closer collaboration between genomic and social scientists.”
At the end of workshop, Green and Pérez-Stable said that many important issues, which have broad significance beyond NIMHD and NHGRI, were defined and will continue to be discussed beyond the workshop.
“With large cohort studies occurring around the world, there are international dimensions to the issues about race and ethnicity that were discussed in the workshop,” said Green. “However, the workshop’s outcomes will not be confined to policy decisions, and there will be many opportunities in the future for novel research. The scientific community is still learning how best to use genomic tools and leverage other resources to improve medicine, and it needs to define the types of research studies that will best inform the future of biomedical science.”
Pérez-Stable said, “I look forward to a time in the future when the scientific community will have access to standardized race and ethnicity data, DNA sequence data, and standardized data on social determinants for diverse populations and can begin to separate out the various biological and non-biological contributors to health outcomes. The workshop leaves me hopeful for the development of better research methodology for the use of race and ethnicity information in biomedical research.”
Themes and Discussion
While the workshop did not result in specific recommendations or consensus positions, sessions on the second day generated topics for discussion meant to stimulate discussion in the genomics and biomedical research community.
I. Collect a variety of data for research involving race and ethnicity by using self-identification as well as genomic, clinical, behavioral, socioeconomic, and sociodemographic variables. Organize data collection around five race and ethnicity dimensions that encompass sociopolitical group history, family history, and genetics: SIRE, race and ethnicity classification by others, reflected appraisal (i.e., perception of how others may classify one’s race and ethnicity), self-identified ancestry, and genetic ancestry.
II. Modify and expand OMB race and ethnicity categories by disaggregating South Asians from other Asians; adding a Middle Eastern/North African (MENA) category and a U.S. nativity category; adding parents’ and grandparents’ SIREs; permitting multiracial categorization; including sociodemographic variables, such as geographical region and urban or rural location; and replacing or adding to traditional categories with questions about the historical racial narrative (e.g., whether any ancestors were held as slaves).
III. Ensure appropriate use of SIRE terminology in genomics and biomedical research by standardizing racial and ethnic terminology (e.g., ethnic groups should not be called “races”); distinguishing between the study of health disparities and minority health research; creating linguistic guidelines for discussing social categories and reporting on population variability and health disparities; using Office of Management and Budget (OMB) race and ethnicity categories primarily for the recruitment of studies, not to report findings; and including information on how race and ethnicity are defined in clinical trial pre-registration and linking it to the final data.
IV. Teach the lay public how to interpret race-associated biomedical genomics by dispelling misconceptions about race-associated disease risk; using nontechnical but accurate language about disease risk when reporting on race-associated genetic data; and clarifying how genetic ancestry can improve health care and that it is not intended to replace culturally and sociopolitically based racial self-identification. In addition, improve the human population genetics education of current and future physicians, and demonstrate the usefulness of ancestry genetic markers in clinical practice.
V. Make collecting and analyzing race and ethnicity data more robust by including diverse race and ethnicity groups in research to make results more generalizable; oversampling minority groups to increase power in data analyses; recruiting individuals who represent the disease load of the general population; broadening race and ethnicity representation in reference groups for existing studies and designing future studies without generalized reference groups; designing studies to examine longitudinal frameworks of racial and ethnic identity; creating data collection and analysis frameworks that can be generalized beyond North American countries; collaborating with researchers who have access to underrepresented participants, social scientists and demographers offering new study design and analysis perspectives, and diverse nonscientific stakeholders, such as organizations representing different racial and ethnic communities, to better engage underrepresented participants; and improve transparency and communication of the research process with communities to increase participation.
VI. Make race and ethnicity data more accessible, searchable, and comparable by creating guidelines to standardize the analysis, tagging, and reporting of data so that methodology, data, and results can be compared across studies with scientific rigor; harmonizing standards for collecting, analyzing, and reporting research and clinical data; and making community-based studies that include race and ethnicity data accessible to study participants. The scientific community, in collaboration with NIH, could provide guidelines for standardizing race and ethnicity measures in electronic health records (EHRs).
VII. Address issues with race and ethnicity measures in research. NIH could:
- Support research addressing the relationship between race and ethnicity and health outcomes, particularly with outreach into communities that do not traditionally participate in genomics or biomedical research
- Facilitate publication of research that involves community engagement, increasing the scientific community’s access to and understanding of this work
- Sponsor demonstration projects to investigate which race and ethnicity measures are most feasible to include in a wide range of research studies
- Review and revise how language in Funding Opportunity Announcements gives instructions about the use of race and ethnicity measures and affects the design of interdisciplinary studies
- Provide guidance for peer review of grant applications and manuscripts that includes variables to measure race and ethnicity by defining consequences for not including diverse populations in NIH-sponsored research and ensuring that
- Researchers appropriately use race and ethnicity categories and terminology,
- Guidelines for inclusion in recruitment are not conflated with guidelines for reporting results, and
- Researchers are not penalized for including diverse study populations, which typically have small numbers of minority participants from each sub-group.
The National Academy of Sciences could:
- Convene a consensus study to define which race and ethnicity variables should be collected for genomic, clinical, and population research. The panel should consist of a diverse group of stakeholders, including policy makers, community groups, and OMB representatives.
- Develop, publish, and disseminate white papers to raise awareness about issues of race and ethnicity in science and health care