Abstract
Objectives:
There is a need for formal guidance around antiracist research practices to dismantle structural racism, which has been embedded into the design and conduct of health sciences research. Our objective is to review and summarize recommendations from published literature published on antiracist research practices applicable to all health sciences research studies.
Method:
In 2022, we systematically reviewed PubMed, SCOPUS, Excerpta Medica Database, British Nursing Index, PsycINFO, Cumulative Index to Nursing and Allied Health Literature, and Sociological Abstracts using search terms about antiracist research. We found 5362 papers, screened 3962 titles, assessed 171 abstracts for eligibility, and reviewed 103 using a standardized extraction form. We reviewed 11 papers twice to confirm accuracy. We narratively synthesized recommendations relevant to six stages of the research process via combined deductive and inductive qualitative analysis.
Results:
A total of 1407 individual recommendations were grouped into themes and subthemes within each research stage. Project development had the most themes (n = 7) and subthemes (n = 24); dissemination of findings had the fewest (2 themes and 5 subthemes). When designing a project, researchers are encouraged to develop diverse, representative, and empowered research teams; design projects important to marginalized groups; and use conceptual frameworks that build on the work of marginalized scholars. Researchers were encouraged to promote inclusive consent processes, accessible recruitment materials, and recruitment and retention strategies that maximized diversity. When analyzing data, recommendations included involving community partners to ensure appropriate conceptualization and measurement of race, justifying the selection of reference groups, and considering intersectional identities. Recommendations for data interpretation include contextualizing findings, considering structural racism, and improving community involvement to enhance the transparency of data analysis. During dissemination, recommendations highlighted increasing accessibility of research findings.
Conclusions:
We identified a comprehensive list of recommendations describing antiracist research practices. Our synthesis can be used to guide individual research teams, institutional review boards, and grant committees, and to develop educational curricula around antiracist research practices.
Background
Since the racial reckoning in the United States in the summer of 2020, there has been increasing attention to racism across fields and institutions.1,2 These forces have led to reflection on the ways that research has contributed to scientific racism, which the National Human Genome Research Institute defines as “a historical pattern of ideologies that generate pseudoscientific racist beliefs that lead to racial bias and discrimination in science and research.” 3 Race is recognized as a socio-political construct. 4 Nonetheless, several papers have pointed out that even recent research has incorrectly attributed study findings to biological race instead of to structural inequities that account for differential outcomes by race.5,6
To counter scientific racism, intentional actions to promote antiracist research practices are needed to guide the work of all health sciences researchers. Combining the work of varied scholars, we define antiracist research as “research that contributes to the unlearning of internalized belief systems dictated by white supremacy; redistributes power, resources, and privileges to marginalized communities directly harmed by racism; and redresses historical and contemporary injustices.”7–9 Antiracist research aims to ensure that the construct of race is correctly interpreted and that racially and ethnically minoritized communities are engaged throughout the course of a research project. Antiracist research practices do not apply only to health equity-focused research that aims to document or address health inequities, but instead should inform all research on or pertinent to humans.
Some journals have promoted or adopted standards to improve reporting on studies that use race and ethnicity as a variable.9–11 Additionally, the National Institutes of Health has updated guidance to provide specificity when reporting race and national origin. 12 However, there is a lack of actionable, granular guidance to direct researchers in how to incorporate antiracist research principles across stages of a research project.
Scholars from marginalized groups tend to have their scholarship undervalued and under-cited, 13 and they also tend to be the scholars promoting antiracist practices. By reviewing the literature, we aimed to collate, organize, and synthesize published literature on antiracist practices with the goal of providing recommendations to researchers to implement antiracist research practices.
Methods
Our research team was diverse in multiple dimensions; team members’ positionality statements can be found in eMethods of Supplemental Data. We performed our review according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses extension for scoping reviews. A research librarian (S.M.) helped generate and refine the search strategy. Search terms related to ‘racism/anti-racism’ and ‘health science research’ were chosen using a combination of Medical Subject Headings (MeSH) and keywords (Supplementary Table S1). The search was performed in May 2022 in the following electronic databases: PubMed, SCOPUS, Excerpta Medica Database, the British Nursing Index, PsycINFO, and the Cumulative Index to Nursing and Allied Health Literature, and Sociological Abstracts. The papers were filtered by language (English) and publication date (2000–2022). The year 2000 was chosen because the initial draft of the Human Genome Project was released that year, confirming that human DNA is > 99% identical, repudiating any genetic basis for race.
Scientific articles, book chapters, commentaries, editorials, or gray literature that provided recommendations on how to conduct health science research using an antiracist approach were eligible for inclusion if indexed in the databases above. We used an expansive definition of “health sciences research” to encompass the study of all aspects of health, disease, and health care. 14 We excluded book reviews; research protocols; articles with only an abstract or conference proceedings; article summaries without additional commentary; papers that did not provide specific recommendations for health sciences research practice; papers solely focused on recommendations for journals, regulatory bodies, or funding agencies; papers that addressed only one specific medical specialty or disease process, or discussed animal studies only. Although racism is a global phenomenon, we aimed to focus this review on recommendations applicable to the conduct of health sciences research in the United States, given the context of this country’s unique history of racism and its impact on science. Therefore, we excluded papers with no U.S.-based authors. This was determined using the location of the affiliations authors self-reported on papers to proxy for authors’ expertise in this sociocultural context. Papers without at least one author reporting a U.S.-based affiliation were considered beyond the scope of this already large corpus of literature.
Data extraction and analysis
Citations identified in the database searches were imported into Covidence systematic review software. 15 Additional papers identified outside of the database search, such as via reference lists of included papers, were added to Covidence manually. Between June and August 2022, two reviewers (T.A. and F.S.) independently screened all papers by title and abstract to identify those that met the inclusion criteria. Discrepancies were resolved through consensus by a 3rd reviewer (K.E.W. or D.M.-W.). Full-text review was completed for studies included after title and abstract review.
For each included paper, any recommendations describing antiracist practices were extracted verbatim into a separate document. Reviewers also noted the article type as classified by the publication source (e.g., review, study type, perspective, commentary, editorial, etc.); where an article type or study method could not be identified, it was listed as a commentary. To ensure data accuracy and mitigate reviewer bias, a random sample of 10% (n = 11) of the papers was extracted twice. While we systematically identified eligible papers, given the heterogeneity of article types, we were unable to systematically grade the quality of the papers, and we did not endeavor to adjudicate epistemic debates across heterogeneous article types. We qualitatively analyzed recommendations across papers and narratively synthesized them.
For this narrative synthesis, each of the extracted recommendation excerpts was coded into one of six predetermined stages of research: (1) project development/study design, (2) participant inclusion, (3) measurement/data collection, (4) data analysis, (5) data reporting/interpretation of results, and (6) dissemination of findings. We started with a deductive analytic approach to assign recommendations to a specific stage of the research process, and then used an inductive approach to qualitatively analyze the recommendations, 16 and group them into common themes and subthemes until consensus was reached (T.A., F.S., K.E.W., and D.M.-W.).
Results
Our initial search identified 5362 titles, which yielded 3962 papers after removal of duplicates. Full text review was conducted on 103 papers that met the inclusion criteria (Fig. 1 and Table 1), from which we extracted 1407 recommendations. Given the heterogeneous nature of the papers, with the vast majority being commentaries or editorials, a standard quality assessment was not possible. Table 2 describes the themes and subthemes by research stage. Almost every recommendation in Table 2 was substantiated by numerous papers from different categories in Table 1. We describe exemplary themes and subthemes below for each stage of the research process.
PRISMA diagram depicting search and included articles. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-analyses.
Studies Included in Review
Themes and Subthemes for Each Step of the Research Process
Stage 1: Project development and study design
This stage encompassed 7 themes and 24 subthemes, focusing on team development and training, and on using conceptual frameworks to guide the project.
Theme: Intentionally constructing a diverse, representative, and empowered research team
Subtheme: Develop institutional infrastructure to support the creation of diverse research teams
Authors offer recommendations to structurally promote the creation of diverse research teams, including by creating measurable objectives and auditing processes. Additional recommendations included creating a diversity task force to keep track of opportunity allocation,17,18 and designing promotion and tenure criteria to increase faculty diversity and retention. 56 Other papers made recommendations for how regulatory agencies such as the Institutional Review Board and Institutional Animal Care & Use Committee could enforce diversity goals for their own workforce and study teams. Researchers could advertise opportunities to minority-serving institutions such as Historically Black Colleges and Universities. 19 Authors emphasized that institutions take responsibility for promoting diversity to shift the onus from marginalized scholars who often undertake extra, uncompensated work.
Subtheme: Recruit team members with diverse, representative, and intersectional expertise
The literature encouraged research teams to develop strategies to include scholars from diverse sociocultural backgrounds and experiences, with a focus on research team members whose identities matched the proposed study sample. 28 Suggested strategies included using community partnerships and culturally sensitive recruitment tools, 29 and joining networks of underrepresented scientists.20,30,31
Theme: Create antiracist culture within the research team
Subtheme: Value diversity beyond numbers
The literature reflected that antiracist research requires building a work culture that empowers members of the team to share their experiences and identities, 21 and creating leadership opportunities for minoritized research team members.32,33,57 Teams should have clear procedures and corrective measures for reporting prejudice without fear of retaliation. 18 Authors describe the importance of understanding that forms of self-expression (e.g., hair, jewelry, etc.) differ across cultures and should be celebrated, not viewed as unprofessional. 19 This subtheme also encapsulated the concept that racially minoritized team members should not carry the burden of creating the antiracist culture within the research team; it should be a collective effort by all members. 30
Theme: Creation of conceptual framework that includes, acknowledges, and builds on the work of marginalized scholars
Subtheme: Consider how Race, Ethnicity, and Racism Fit into Conceptual Frameworks, and Name Racism Explicitly Where Relevant
The literature suggested that researchers conducting work that relates to race and racism should be prepared to address health inequities rather than simply rediscover that inequities exist.22,68,97 Antiracist conceptual frameworks can promote awareness of racialized identities in society and clarify the intended purpose of collecting data on race/ethnicity.34,97 Antiracist research frameworks for studies assessing disparities should specifically assess for and analyze experiences of racism,69,92 and they should name racism explicitly where relevant and identify its form and any other oppressive systems that might coexist. 9 Scholars should rely on appropriate frameworks throughout the study process, like when assigning reference groups and interpreting results. 57
Stage 2: Participant inclusion
Four themes and 12 subthemes (Table 2) emerged regarding inclusive participant recruitment and retention.
Theme: Promoting inclusive consent processes
Subtheme: Assure that informed consent processes are tailored to the population of interest
Literature noted several ways to tailor consent processes to include traditionally under-represented participants: providing appropriate time for the consent process to address questions and concerns; recognizing the impact of current and past medical abuses (e.g., the Tuskegee Syphilis Study) on decisions about participating in research; 35 and learning about participants’ vocabulary, beliefs, values, and learning styles prior to designing consent forms to ensure transparency and clarity when obtaining informed consent. Engaging immigrants as potential participants might require alleviating concerns about disclosure of immigration status.35,70 Partnering with community-based organizations may address barriers to participation.35,70
Theme: Developing appropriate recruitment and retention strategies
Subtheme: Develop strategies to increase accessibility for engaging the community of interest
Decreasing the time commitment of participants by either conducting the study within the community or offering virtual study visits can aid equitable recruitment. 35 Authors also recommend compensating participants for time, transportation, childcare, and food. 35 To ensure retention, researchers were encouraged to maintain communication throughout the study through accepted modalities like texting or social media. 58
Stage 3: Measurement and data collection
In this stage, four themes and 11 subthemes emerged, with a focus on how to conceive of study variables, particularly race, within the context of society and with the help of community members.
Theme: Conceptualization of race as a study variable
Subtheme: Avoid the fallacy of race-neutral research by including measures of race and what race serves as a proxy for
Literature suggests that researchers should not adopt a race-neutral approach by eliminating race as a variable but rather should collect data on race, acknowledging it is a social construct impacted by racism.36,114 Consulting experts early can ensure clarity of research questions and avoidance of interpreting race as a genetic factor. 29 Papers warned against stratifying findings by race, as this would obscure disparities in health care outcomes. 96
Subtheme: Determine Ways to Specifically Measure Forms of Racism, Beyond Relying on Racial Categories as a Proxy
For studies seeking to report on the associations between race and outcome, researchers recommend identifying approaches to accurately assess different levels of discrimination and quantitatively measuring intersectional systems of oppression.69,81 One approach would be to measure individual experiences of racism, aggregates of structurally racist policies and practices, and downstream effects of structural racism. 8 Another approach would be to use a racism burden score that combines aspects of individual, institutional, cultural, and structural racism to calculate allostatic load. 71 Researchers could also consider measuring the recurrence, severity, and duration of discriminatory experiences, as these factors can contribute to toxic stress22,81 and are upstream causes of racial disparities. 93
Stage 4: Data analysis
Five themes and six subthemes emerged regarding ways to ensure statistical analyses are antiracist.
Theme: Justify the selection of reference groups
Subtheme: Avoid relying on white supremacist principles that treat white participants as the default reference group
Researchers must carefully consider and provide a clear rationale as to how reference groups are chosen.8,60 Literature encourages the avoidance of relying on white supremacist beliefs that treat White participants as the “default reference group,” as this narrative risks pathologizing minoritized communities and reinforces oppressive systems.37,60 One of several suggested alternative approaches was to compare groups to a population average.23,117
Stage 5: Data reporting/interpretation of results
This stage had 4 themes and 12 subthemes emerge regarding how data are interpreted and reported in publications.
Theme: Writing about findings related to racial groups
Subtheme: Reiterate what race is serving as a proxy for when describing research findings to avoid misattribution
When discussing results, authors should not use words that attribute findings to the race of participants, as this incorrectly implies biological differences. Authors should instead explain how racism and systemic inequities contribute to the outcomes.38,72,82 Referring to previously explained conceptual frameworks in discussion sections can help contextualize the findings. 8
Subtheme: Use precise descriptors of racial and ethnic categories
When reporting participants’ demographic information, it is important to clearly state how this information was collected. For example, manuscripts should specify how racial categories were assigned (i.e., self-report vs assigned by others) and the options that were available. 24 Authors should refrain from using terms like “non-White” and “BIPOC” (“Black, Indigenous, and People of Color”) as they do not accurately identify the sample and hinder authors from being able to assess the applicability of data.31,83 Instead, papers discussed that researchers could consider more specific terms like “Jamaican American.” 34
Subtheme: Follow standards for using inclusive language to demonstrate respect for included populations
Efforts should be made to use inclusive language that conveys respect. 25 The term “minoritized” should be used instead of “minority” and “racialized groups” instead of “racial.” 34 Quotation marks can be used when using the term “race” to reiterate that “race” is a social construct. Additionally, “social indicators of health” should be used instead of “social determinants of health” to emphasize that the factors influencing disparities are not fixed. 34
Stage 6: Dissemination of findings
This stage had the least guidance, with 2 themes and 5 subthemes, focused on ensuring that research findings reach the communities of interest.
Theme: Increasing accessibility of research findings
Subtheme: Engage community partners to identify optimal dissemination modalities and engage community partners as coauthors
Researchers were encouraged to engage community partners in identifying optimal and inclusive dissemination modalities,25,58,61 and offering them the opportunity to share their perspectives on the findings in any dissemination work. 60 The literature also stressed that participating community partners should receive authorship and fair compensation. 25
Discussion
Our review found that engaging racially and ethnically minoritized communities in research is critical to ensuring representative, generalizable, high-quality, and impactful science. The literature review also highlighted how research can structurally exclude participants and incorrectly attribute biological interpretations of race and ethnicity. This can result in erroneous conclusions, race-based pseudoscience, or at best, a failure to recognize the structural forces and factors contributing to racial and ethnic differences in study outcomes. By providing this synthesis, we aimed to equip researchers with recommendations for antiracist research practices across all stages of a study that can help mitigate the risk of these issues. We recognize that while researchers may have the will to improve their antiracist research practices, they may require more focused guidance on how to do so. This review identified 103 recent articles describing 1407 recommendations across disciplines, within and outside of health sciences journals, that provide recommendations on how to incorporate antiracist practices into research. The number of citations and recommendations points to the need to consolidate published recommendations for medical researchers to make those more actionable.
The research stage with the most recommendations was the first stage: project development and study design. This points to how critical it is to invest the time at the initiation of a project to ensure that the project has strong antiracist underpinnings. As noted in the included papers, this foundational work goes a long way to ensure that the research team is diverse, the project is relevant and important to marginalized groups, an antiracist culture is created within the research team, the entire research team is trained in preparation for conducting antiracist research, the study design is appropriate, and there is a strong theoretical foundation for the proposed project.
The importance of ensuring a strong theoretical foundation is one of two overarching themes identified by our review that crosscut all stages of the research process. At every stage, authors of the included papers encourage researchers to return to the chosen theoretical framework to guide decisions about who to study, how to analyze and interpret data, and how to report findings.
The second concept that crosscut all research project stages was the need to incorporate principles of community engagement into every stage. The literature highlights formal roles for community members, who are people from the population around which the research question revolves, at each step of the research process. Importantly, papers repeatedly mentioned that such individuals should be empowered by traditional research team members as equal partners and compensated adequately for their expertise.
Importantly, we found evidence in the reviewed papers that antiracist research practices do not apply only to health equity-focused research but can also guide all research on or pertinent to humans. For example, in basic science research, engaging community members ensures that the ultimate products of that research are acceptable and understandable to the intended patient population. When it is time to recruit participants into clinical trials or accept a new therapy, community members can help provide their perspectives to the study team or serve as spokespeople to promote equitable and diverse recruitment. However, there may be instances where each recommendation is not relevant for every study design or context; researchers should be thoughtful in applying these recommendations to their own projects where applicable.
Importantly, since running our search, several important articles have been published that directly address our findings. Most notably, the comprehensive report by the National Academies of Sciences, Engineering, and Medicine (NASEM) titled “Rethinking race and ethnicity in biomedical research” 118 was published in 2024. It is notable that our review and the NASEM report both highlight the importance of conceptual models and community engagement, which gives us confidence that the NASEM authors likely reviewed a very similar corpus of literature to develop that guidance.
This study had several limitations. First, our search strategy might have excluded papers from the gray literature (e.g., blog posts, lay publications) that would have provided additional recommendations on this topic. Recommendations published after our literature searches were missed, which is important because this guidance is ever-evolving. Our decision to include only work by authors with at least one U.S.-based affiliation also means we did not include recommendations with only the perspective of researchers in other countries. This intentional decision was made to limit the scope of our review and to acknowledge the unique history of racism in the United States. We recognize that this may limit the applicability of our findings to non-U.S.-based contexts that also experience structural racism. 119 Third, because of the vast heterogeneity of article types, we were not able to grade the level of evidence/bias. However, given the important contributions of authors with lived and professional experience, it was important to include commentaries, editorials, and reviews to capture their recommendations. Lastly, we were unable to evaluate the expertise of the included authors, who likely have varied experience conducting and/or writing about antiracist science.
Conclusions
This narrative review is a first step in developing clear recommendations on best practices for antiracist research at each research stage. As a next step, there is a need for consensus agreement to ensure that the reviewed published literature is truly comprehensive, and to sort out where recommendations may be contradictory or unclear. Thoughtful implementation of these recommendations may help researchers avoid the pitfalls of health equity tourism and toward more deeply grapple with antiracist scientific methods. 8
Authors’ Contributions
K.E.W., T.A., F.S., S.W.-K., S.M., K.A., A.F.D., N.H.-G., and D.M.-W. contributed to the study conceptualization and design, and helped critically review and edit the manuscript. K.E.W., T.A., F.S., S.W.-K., and D.M.-W contributed to article review and extraction, data analysis, and they drafted the manuscript.
Footnotes
Acknowledgments
The authors wish to thank the scholars whose work contributed to this review.
Author Disclosure Statement
The authors have read the journal’s policy, and they have the following competing interests: K.E.W. has been paid as a consultant to Healthy Steps and Thomas Jefferson University Autism Center. She serves as an expert witness for Kline Spector, LLC. K.A. is Founder and Director of Ideas to Impact, Inc., and is a paid consultant for the following organizations: Child Trends, Robert Wood Johnson Foundation, Annie E Casey Foundation, Chapin Hall, Public Research and Evaluation Services, and the Educational Development Center. D.M.-W. is a paid consultant of the Center for Violence Prevention Research and Chiesi, Inc. The remaining authors have no competing financial interests to disclose.
Funding Information
This study was funded by the
Supplemental Material
Abbreviations Used
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
