Infographics

Racial Bias in Health Care Artificial Intelligence

Published on: September 30, 2021.

Algorithmic predictions accounted for 4.7x

more of the racial disparities in pain relative to standard measures

About this Data Insights

Artificial intelligence (AI) tools such as algorithms are increasingly being used to determine who gets health care. These tools can unintentionally increase the impact of existing racial biases in medicine through the explicit use of race to predict outcomes and risk. Despite evidence that race is not a reliable proxy for genetic differences, how to allocate clinical resources or treatment adherence, using race as a factor has become a common practice when designing clinical algorithms. This new infographic shows how embedding race into health care data and decisions can unintentionally advance racial health disparities and direct more attention or resources to White patients rather than Black, Latino, and other medically underserved patient populations. It also shows how health care AI tools use data that inadvertently capture systemic racism, adding to existing inequities in health care access and status.

This infographic highlights strategies to address bias in algorithms and the potential for AI to support health equity.

Citations

Racial Bias in Clinical Care Algorithms and Examples of Race Correction in Clinical Medicine: Vyas, Darshali A., et al. “Hidden in Plain Sight — Reconsidering the Use of Race Correction in Clinical Algorithms.” New England Journal of Medicine, vol. 383, no. 9, 2020, pp. 874–882., doi:10.1056/nejmms2004740.

Doctors Moving Away from Algorithms with Racial Bias:

eGFR: Inker, Lesley A., et al. “New Creatinine- AND Cystatin C–Based Equations to Estimate GFR without Race.” New England Journal of Medicine, 2021, https://doi.org/10.1056/nejmoa2102953.
The Vaginal Birth after Cesarean: Katie Palmer June, et al. “Researchers Remove Race from a Calculator for Childbirth.” STAT, 3 June 2021, https://www.statnews.com/2021/06/03/vbac-calculator-birth-cesarean/.
STONE Score: American College of Emergency Physicians letter to the House Committee on Ways and Means, December 4, 2020.

NOTE: Get with the Guidelines–Heart Failure Risk Score: American Hearth Association Letter to the House Committee on Ways and Means, September 24, 2020.

Letter states that while race was included as a variable in the publication, Get With The Guidelines® clinical thought leaders elected not to include race in the model programmed into the registry. Thus, race is not included in the heart
failure risk score used by hospitals in Get With The Guidelines® - Heart Failure.

Racial Bias in Other Health Care Algorithms Obermeyer Z, Powers B, Vogeli C, and Mullainathan S. “Dissecting Racial Bias in an Algorithm Used to Manage the Health of Populations.” Science, 36(6464):447-53, October 25, 2019

Using AI to Support Health Equity

Potential to address racial disparities: Pearl, R. “How Ai Can Remedy Racial Disparities in Healthcare.” Forbes Magazine, 15 Feb. 2021, www.forbes.com/sites/robertpearl/2021/02/16/how-ai-can-remedy-racial-disparities-in-healthcare/?sh=793e4c0b30f6.
Experiments have shown some success in AI improving physician’s judgment: Godwin, C. “Can AI Tackle Racial Inequalities in Healthcare?” BBC News, BBC, 6 Feb. 2021, www.bbc.com/news/technology-55940112.
Study on algorithmic approach reducign unexplained pain disparities: Pierson, E., Cutler, D.M., Leskovec, J. et al. An algorithmic approach to reducing unexplained pain disparities in underserved populations. Nat Med 27, 136–140 (2021). https://doi.org/10.1038/s41591...

Strategies to Address Bias in Algorithms

Need more diverse data: Kaushal, Amit, et al. “Health Care AI Systems Are Biased.” Scientific American, Scientific American, 17 Nov. 2020, www.scientificamerican.com/article/health-care-ai-systems-are-biased/.
Incorporate equity into the design of algorithms: Dankwa-Mullan, Irene, et al. "A Proposed Framework on Integrating Health Equity and Racial Justice into the Artificial Intelligence Development Lifecycle." Journal of Health Care for the Poor and Underserved 32, no. 2 (2021): 300-317. doi:10.1353/hpu.2021.0065.
Address lack of diversity: Klawe, Maria. “Why Diversity in Ai Is so Important.” Forbes, Forbes Magazine, 20 July 2020, www.forbes.com/sites/mariaklawe/2020/07/16/why-diversity-in-ai-is-so-important/?sh=2fec45327f2b.

This infographic was reviewed by:

David S. Jones, MD, PhD, Ackerman Professor of the Culture of Medicine, Harvard University
Fay Cobb Payton, PhD, Professor of Information Technology/Analytics at North Carolina State University