Health Care Fairness Impact Lab

Algorithms and Racial Health Equity

Health care practitioners—and the systems they work in—increasingly rely on algorithms to guide decisions. Algorithms can combine a large amount of data and evidence—quickly—to inform better health decisions. Yet, algorithms have also been shown to exacerbate racial, ethnic, and economic disparities in health. 

The Health Care Fairness Impact Lab is a partnership between physicians, economists, and statisticians at the Stanford School of Medicine, Stanford Health Policy and Stanford Health Care (a medium-sized health care center with 2 million outpatient visits a year). Together, they will focus on algorithms used to decide whether to admit patients to the emergency room and whether to refer a patient for chronic kidney disease treatment. These two settings reflect examples of particularly high-stakes decisions where algorithms have the potential to exacerbate or improve health disparities. In the context of emergency care, providers use a combination of judgment and algorithms to make decisions quickly and with limited information. In the context of chronic kidney disease, an established formula for determining access to treatment has been rightfully derided by clinicians and health equity experts as introducing racial bias into treatment decisions–leading to delayed treatment for Black patients. 

By working in partnership, the team will be able to evaluate the impact of algorithms on access to care and health disparities in practical health care settings and across multiple marginalized racial and ethnic groups. They will study new formulas for admitting patients to the emergency room and how chronic kidney disease patients are referred for treatment, evaluating them against different measures of fairness. This research aims to inform decisions at Stanford Health Care and may serve as an input into future research and clinical guidelines in an area of growing attention for healthcare policymakers, clinicians and patients. 


For more information, please contact:

Marika Cusick