Annual symposium: Data Science and Health Disparities
March 24, 2023 | 9:00am-3:30pm
Linda K. Paresky Conference Center at Simmons University.
This year’s annual symposium will focus on recent efforts by quantitative scientists to better understand the complex mechanisms that contribute to health disparities and to develop interventions to reduce or eliminate these disparities. Speakers from Harvard T.H. Chan School of Public Health, Brown University, Drexel University, Cornell Medical College, University of North Carolina at Chapel Hill, and the National Institutes of Health will describe advances that have been made as well as challenges that remain.
Welcome and Opening Remarks
Garrett Fitzmaurice, ScD
Professor, Department of Biostatistics, Harvard T. H. Chan School of Public Health
Director, Harvard Catalyst Biostatistics Program
Using Causal Diagrams to Study and Eliminate Racial Health Disparities
Chanelle Howe, PhD
Associate Professor, Doctoral Program Director, Department of Epidemiology,
Brown University School of Public Health
Using Machine Learning to Increase Equity in Healthcare and Public Health
Emma Pierson, PhD
Assistant Professor, Population Health Sciences, Weill Cornell Medical College
Assistant Professor, Computer Science, Jacobs Technion-Cornell Institute at Cornell Tech
Spatially Varying Racial Inequities in Cardiovascular Health: Individual-level risk factors + Neighborhood-level risk factors + Structural racism = Significant Black-White Inequities
Loni Philip Tabb, PhD
Associate Professor, Department of Epidemiology and Biostatistics, Dornsife School of Public Health, Drexel University
Case Studies of Engaged and Participatory Data Science with Rural, Racialized Communities
Leah Frerichs, PhD
Associate Professor, Department of Health Policy and Management, University of North Carolina at Chapel Hill
Health Justice, Critical Science, and the Two-Edged Sword of Data: Structural Problems Require Structural Solutions
Nancy Krieger, PhD
Professor, Social Epidemiology, Department of Social and Behavioral Sciences, Harvard T.H. Chan School of Public Health
A Common Currency of Measurement Utilizing a Cloud Computing Platform to Foster Tech Equity & Advance Health Disparities Research
Deborah Guadalupe Duran, PhD
Senior Advisor, Data Science, Analytics and Systems
National Institute on Minority Health and Health Disparities (NIMHD)
Symposium Discussant Comments
S (“Subu”) V Subramanian, PhD
Professor, Population Health and Geography, Department of Social and Behavioral Sciences, Harvard T.H. Chan School of Public Health
Garrett Fitzmaurice, ScD
Professor, Department of Biostatistics, Harvard T. H. Chan School of Public Health
Director, Harvard Catalyst Biostatistics Program
Abstracts and Bios:
Deborah Guadalupe Duran, PhD, is senior advisor on data science, analytics, and systems for the director of the National Institutes of Minority and Health Disparities at the National Institutes of Health (NIH). She focuses on implicit and explicit biases in data curation, AI designs, algorithm development, and impacts on communities. Duran’s work in AI and ethics strives to increase capacity for underrepresented populations to utilize big data, the cloud, and data science resources. She has developed the ScHARe (Science Collaborative for Health disparities and Artificial intelligence bias Reduction) Research Collaboration Cloud Platform, which focuses on social science data, including social determinants of health, to better understand health disparities and inequities.
Duran works to include social determinants of health in data analytics to better apply diagnostics, treatments, and interventions to reduce health disparities. Prior to this role, she served as director of science policy, scientific planning, and data analytics for more than 20 years. Additionally, she was a principal investigator, a researcher for the National Coalition of Hispanic Health, an advocate for health equity, a professor, and a high school teacher. As an author, Duran has received three Health and Human Services Secretarial Awards and numerous NIH awards.
Technology is a powerful tool to advance equitable health care. As machine learning algorithms (MLA) become more common, it is imperative to ensure that these methods do not contribute to inequities through biased predictions or differential accuracy across racial and sex groups. When biases are unchecked, inequities occur that lead to health disparities. A potential contributing factor is the fact that the AI/ML field currently lacks diversity in its researchers and in data from diverse populations. These gaps pose a risk of creating and continuing harmful biases in how AI/ML is used, how algorithms are developed and trained, and how findings are interpreted. Another contributing factor is data. There is a need to ensure that data is representative and techniques such as synthetic data are appropriate for the target population. Preparing AI ready data often includes the need to aggregate data sets, which requires a common currency or standardization of key variables or extensive mapping. National Institutes of Minority and Health Disparities (NIMHD) is striving to foster tech equity and advance health disparities research by supporting (1) the PhenX toolkit, which includes ongoing efforts to develop standardized measures for collecting population demographic characteristics including race and ethnicity, sexual orientation, gender identity, socioeconomic status, and other individual social determinants of health (SDoH) key variables; (2) a cloud platform called ScHARe to facilitate data sharing among health disparity researchers, with an emphasis on social determinants of health (SDOH) data. The platform will provide a centralized location for internal and external experts to develop AI strategies to mitigate bias from data curation, design, algorithm development, training, and implementation; and (3) institutionalize a set of core common data elements for funded health disparity and health outcomes research, which fosters data interoperability from a common currency of measurement to foster tech equity and advance health disparities research.
Leah Frerichs, PhD, is an associate professor in the Department of Health Policy and Management at the University of North Carolina at Chapel Hill. Her research is focused on the intersection of community-based participatory research and systems science to address health disparities, primarily in chronic disease prevention. Frerichs has experience working with diverse communities including American Indian, Latino, and African Americans to develop, implement, and evaluate community-based interventions and policies. Her work has had a major focus on understanding and developing interventions that target social and physical environmental influences on health behaviors in youth who are part of underserved communities.
Frerichs received her PhD from the University of Nebraska Medical Center in health promotion and disease prevention research and completed a postdoctoral fellowship with the Center for Health Equity Research at the University of North Carolina at Chapel Hill. Prior to her doctoral studies, she managed a cancer prevention and control program for tribal communities in the Northern Plains.
Technological advances have increased our ability to generate and use large volumes and varieties of data. However, there remain inequities in who and which communities have power to generate, use, and translate that data into meaningful action. In particular, historically marginalized and racialized communities have not often benefited from data science, and in some instances, have even been harmed. To that end, research approaches that combine principles from asset-based community development and community-based participatory research have potential to redistribute power in the creation and use of data more equitably to communities of color. The goal of this presentation is to provide case studies of applying these approaches with diverse communities, including rural African American youth, Indigenous birthing people, and others. Through these case studies, I will highlight strategies and methodologies applied, successes, and challenges of this type of approach.
Chanelle Howe, PhD, is an associate professor in the Department of Epidemiology at Brown University School of Public Health. There she serves as director of the Epidemiology Doctoral Program and is associate editor for the American Journal of Epidemiology. Howe’s research interests include study design, quantitative methods, causal inference, infectious diseases, and health disparities.
Causal diagrams have been extensively used by epidemiologists to study population health. However, historically there has been a lack of clarity regarding how best to use causal diagrams to study and eliminate racial health disparities that have persisted in the United States (U.S.). This talk will: (1) provide recommendations for using causal diagrams to study and eliminate U.S.-based racial health disparities; (2) discuss the benefits of these recommendations; and (3) demonstrate implementation of these recommendations.
Nancy Krieger, PhD, is professor of social epidemiology and American Cancer Society Clinical Research Professor at the Harvard T.H. Chan School of Public Health and director of the school’s Interdisciplinary Concentration on Women, Gender, and Health. She is an internationally recognized social epidemiologist with a background in biochemistry, philosophy of science, and history of public health. She has participated in social justice, science, and health activism for more than 35 years.
Krieger is one of the Institute for Scientific Information’s most highly cited scientists, a group that includes less than 0.05% of published researchers. Her conceptual and empirical work to understand, analyze, and improve population health and health equity includes developing the ecosocial theory of disease distribution, which addresses embodiment and equity, as well as etiologic research on societal determinants of population health and health inequities, including structural racism and other types of adverse discrimination, and methodologic research to improve monitoring of health inequities.
In this presentation, I will reflect on: “Health Justice, Critical Science, and the Two-Edged Sword of Data: Structural Problems Require Structural Solutions.” Topics I will briefly address include critical thinking and theory for better science and action, buttressed by several empirical examples, including COVID and the continued contemporary health harms of past injustice, as exemplified by Jim Crow and historical redlining. My premise is that another world is possible, in which health justice exists – and knowing whether or not we are there requires data. This presentation is informed by the ecosocial theory of disease distribution, including its focus on embodiment, agency, and accountability in societal, ecological, and historical context; its distinctions between “biological expressions of injustice” and “unjust interpretations of biology”; and its recognition that anti-essentialist and anti-racist science is essential. Guided by this theory, I will review concepts and methods relevant to construction and use of area-based social metrics for research to advance health justice, including the Index of Concentration at the Extremes for both racialized segregation and racialized economic segregation, used in relation to diverse health outcomes and exposures.
Emma Pierson, PhD, is an assistant professor of computer science at the Jacobs Technion-Cornell Institute at Cornell Tech , and a computer science field member at Cornell University. She holds a secondary joint appointment as an assistant professor of population health sciences at Weill Cornell Medical College. She develops data science and machine learning methods to study inequality and healthcare. Her work has been recognized by best paper, poster, and talk awards, a National Science Foundation (NSF) Career award, a Rhodes Scholarship, Hertz Fellowship, Rising Star in EECS, MIT Technology Review 35 Innovators Under 35, and Forbes 30 Under 30 in Science. Her research has been published at venues including Nature and Nature Medicine, and she has also written for The New York Times, FiveThirtyEight, Wired, and various other publications.
Abstract: Our society remains profoundly unequal. Worse, there is abundant evidence that algorithms can, improperly applied, exacerbate inequality in healthcare and other domains. This talk pursues a more optimistic counterpoint–that data science and machine learning can also be used to illuminate and reduce inequity in healthcare and public health–by presenting vignettes from domains including COVID-19 and cancer risk scores.
Loni Philip Tabb, PhD, is an associate professor of biostatistics in the Department of Epidemiology and Biostatistics at Drexel University’s Dornsife School of Public Health. She is the principle investigator of a five-year Mentored Career Development Award to Promote Faculty Diversity in Biomedical Research (K01) from NIH/NHLBI to support research on “Assessing the spatial heterogeneity in cardiovascular risk factors within and between Blacks and whites”. The research aims to: (1) describe the spatial heterogeneity in cardiovascular health in Blacks, whites, and Black-white inequities, and (2) assess the contributions of individual- and neighborhood-level risk factors to the spatial heterogeneity in cardiovascular health.
She has also collaborated as a co-investigator on several projects funded by the National Institutes of Health, National Science Foundation, Annie E. Casey Foundation, and Sidney Kimmel Cancer Center. Tabb received her PhD in biostatistics from Harvard University where she developed novel statistical methods to address zero inflation in longitudinal count data, applying these methods to environmental health and health disparities research. She obtained her BS and MS in mathematics from Drexel University.
Black-white inequities in cardiovascular health (CVH) pose a significant public health challenge, with these inequities also varying geographically across the U.S. There remains limited evidence of the impact of social determinants of health on these inequities, with a focus on individual- and neighborhood-level risk factors. Additionally, structural racism plays a critical role, but the statistical and epidemiological operationalization of such a construct varies at both the individual- and neighborhood-levels. Using a national population-based cohort from the REasons for Geographic and Racial Differences in Stroke study, we assessed the spatial heterogeneity in Black-white differences in CVH and determined the extent to which individual- and neighborhood-level characteristics explain these inequities. The neighborhood-level measure of the Index of Concentration at the Extremes was our primary measure of interest in capturing residential segregation, and its impact on CVH inequities. We utilized a Bayesian hierarchical statistical framework to fit spatially varying coefficient models. Results showed overall and spatially varying inequities, where Black participants in the REGARDS cohort had significantly poorer CVH. The maps of the state level random effects also highlighted how inequities vary. The evidence produced in this study further highlights the importance of multilevel approaches – at the individual- and neighborhood-levels – that need to be in place to address these geographic and racial differences in CVH.