Addressing health inequities in biomedical research

Structural racism is best defined as racism that is deeply embedded into systems and structures such as laws, policies, entrenched practices, and beliefs. As a white person raised in a predominantly white community in suburban Chicago, an REI intensive two-day workshop deepened my understanding of structural racism and our nation’s racialized history. On an almost daily basis, I’m also confronted with examples of the complex, historical, and ongoing assault that the social construct of race has had on our humanity—with devastating effects on the lives and well-being of people of color.

Put more directly by Dr. Rhea Boyd and colleagues in their excellent Health Affairs article: “In short, racism kills. Whether through force, deprivation, or discrimination, it is a fundamental cause of disease and the strange but familiar root of racial health inequities.” As a researcher, who has supported epidemiological and clinical research for the National Institutes of Health (NIH) for more than 20 years, this message hits close to home with a renewed sense of urgency.

The decades-long scholarship about health inequities and their structural root causes, known as the social determinants of health (SDOH), is both rich and compelling. Social determinants of health are “the conditions in which people are born, grow, live, work, and age,” which are “shaped by the distribution of money, power and resources.” (Alderwick and Gottlieb (2019) Meanings and Misunderstandings: A Social Determinants of Health Lexicon for Health Care Systems)

From the murder of George Floyd to multiple city water crises, racial disparities in birth outcomes, maternal mortality, and the disproportionate impact of COVID-19 on people of color, public health practitioners, researchers, funders, and policymakers face a call to action—to deepen their understanding of health inequities and to realign their approaches to research, interventions, and policy.

So, how do we accomplish this? We need a step-by-step action plan.

Step 1: Create lasting change by understanding racial bias

To shift society away from structured racism, Dr. Rhea Boyd provides insight with the first step in our action plan. “First, do no harm, and while you’re doing no harm, learn as much as you can.”

For those of us who participate in the public health research enterprise—as epidemiologists, data scientists, analysts, programmers, data managers, project managers, field technicians, and other research support staff—it is essential to learn about these systemic inequities and the insidious ways in which racial biases play out in research and health care.

An abundance of research and writing on health inequities and the SDOH is available. This 2003 Institute of Medicine report and a 2022 follow-up article in StatNews provide insight into the minimal progress made over the last 20 years. Additionally, this 2017 Lancet article provides a comprehensive overview as does Braveman and Gottlieb in Public Health Reports (2014)

Bibliographies for further reading about health inequities are available from leading scholars such as Sherman James, David Williams, Camara Jones, Nancy Krieger, Diana L Rowley, Paula Braveman, and Dorothy Roberts, to name a few. The National Institute of Minority Health and Health Disparities (NIMHD) provides a range of resources and webinars for researchers and the public via their website and is also a major funding source and advocate for research into health inequities.

Step 2: Integrate social determinants of health into training programs as a core competency

While the study of SDOH has begun to migrate from the margins into the mainstream in contemporary research, this knowledge is not uniform across disciplines. The evidence of the upstream structural causes of racial inequities needs to be formally integrated into public health, data science, and medical training programs as core knowledge—to provide investigators and research teams in government, industry, and academic settings with broader context and a more realistic landscape from which to conceptualize and implement studies and analyze health and well-being.

Step 3: Integrate social determinants of health into biomedical research

The calls from the fields of medicine, data science, and epidemiology are louder for analytical “models that delineate the causal mechanisms connecting structural processes to individual level health outcomes.” See Adkins-Jackson, et al, AJE.

Some of this has been aimed squarely at the field of epidemiology which has been criticized for “producing disproportionately more work documenting individual-level susceptibility” by McCLure et al, AJE versus investigations of more plausible system-level explanations for poor health outcomes which has “served to reinforce, and, post hoc, justify pervasive narratives of biological and cultural inferiority of Black and Brown people.”

Grounding research in an understanding of structural racism will result in more precise analytical models that can better assess causation and association. According to Duke University faculty epidemiologist Whitney Robinson, having better models that accurately account for these structural causes is even more critical when considering that the future of health research will rely heavily on machine learning and artificial intelligence.

Step 4: Adopt SDOH best practices and recommendations

The NIMHD Minority Health and Health Disparities Research Framework provides an “evolving” model for conceptualizing and conducting research into health disparities by depicting the multidimensional aspects of individual, interpersonal, community, and societal levels of influence alongside biological, behavioral, physical/built environment, sociocultural, and health system domains. In 2020, an SDOH assessments collection was added to the PhenX Toolkit (consensus measures for phenotypes and exposures) to help standardize and improve data collection and facilitate harmonization of SDOH data across studies.

At a September 2022 NIMHD virtual conference, SDOH presenters, including researcher Paula Braveman, cogently outlined the rationale for incorporating the “upstream” causes of racial disparities into research design. Expert after expert echoed the groundwater metaphor—that instead of focusing only on individual biomedical and behavioral factors, we need to also consider the “long and complex social chains that give rise to health and health disparities.” We need to study more than just “fragments of the causal chain” and consider the upstream factors, which often go unmeasured and unaddressed.

Epidemiologist Scarlett Lin Gomez, along with other researchers, also outlined various measures of structural racism that have been used in models including redlining and racial bias in mortgage lending, residential segregation and housing discrimination, gentrification, persistent poverty, and the footprint of Jim Crow. See Alson et al, Health Equity, 2021, Groos et al, J Health Disparities Res Practice, 2018, and the RESPOND study at UCSF for more examples.

Step 5: Utilize (or build) publicly available SDOH databases for research

A critical tool in health disparities research are databases that ingest and house SDOH data from various public sources to facilitate standardization and integration of SDOH variables as well as hypothesis generation. The Agency for Health Research Quality (AHRQ) has released a beta version of the Social Determinants of Health Database, which contains community-level SDOH variables across various domains from multiple public sources, linkable at the county, zip code, and census track level.

Researchers at the University of California in San Francisco developed the Health Atlas, an interactive map of more than 100 variables to explore neighborhood-level characteristics to see how they relate to health at the population level.

Additionally, the UCSF Dream Lab created the California Neighborhoods Data System (CNDS), which provides “an integrated system of small area-level measures of socioeconomic and built environments for California, which can be linked to individual-level geocoded records.” This data system studies the impact of neighborhood characteristics (e.g., crime, walkability, access to healthy food, air pollution, and more) on cancer incidence and outcomes in populations.

Step 6: Use a team science approach to solve complex problems

As recommended by UCSF researchers during the September 2022 NIMHD virtual conference, given the complexity of the causal chain, health disparities research requires a team science approach leveraging expertise from social sciences, epidemiology, genetic and molecular epidemiology, data sciences, bioinformatics, biostatistics, environmental epidemiology, geospatial, GIS analytics, demography, and community engagement. This interdisciplinary approach will lead to greater integrity in the research design, enriched soil for innovation, and broader, more impactful analyses.

Answering the call to action as health scientists

The groundwater metaphor is especially relevant for the field of biomedical research and the way in which racial disparities in health outcomes have been examined and interpreted historically, with the lens trained on individual-level factors and the broader societal and structural context left out of scope.

While there is growing awareness of structural racism and its impact on health, the recommendations from experts in the field summarized here are practical and attainable steps for the research community to take to confront structural racism and contribute in more meaningful ways to the body of knowledge about health disparities.

As a researcher, I feel fortunate to work at ICF, where diversity, equity, and inclusion are part of ICF’s history and ongoing commitment, where team science is already championed, where training and professional growth are integrated into the ICF culture and infrastructure, technology is modern and human centered, and innovation is by design.

ICF recently partnered with Amazon Web Services (AWS) on an Envision Engineering engagement to build a platform called the Research Analysis Accelerator (RAA)—to help researchers accelerate their work by shortening their time to science.

Data ingested into the platform will be aligned with FAIR principles and be readied for predictive modelling beyond traditional statistical tests. Our initial use-case focuses on climate change, creating a more effective tool for studying its impact on health and disease outcomes, and creating forecasts to predict future health outcomes.

The next phase for ICF’s RAA platform will be to integrate additional data domains such as SDOH variables into this platform to add further dimensionality. This will enable researchers to investigate climate-related health disparities, which are predicted to increase as temperatures become warmer.

A paradigm shift in biomedical research is in motion and ICF is well positioned to play a part in the push toward achieving equitable access to the social determinants of health that contribute to good health and well-being.