Racial bias skews algorithms widely used to guide care, study finds

The diagnostic tools and algorithms that physicians in the U.S. use to inform their clinical decisions for tens of millions of patients are racially biased, according to a new study. For instance, an algorithm that transplant surgeons use says that kidneys from Black donors are more likely to fail than those from other races. But because Black patients are also likelier to receive an organ from a Black donor, this results in fewer kidneys being available for transplantation. Similarly, an online tool underestimates breast cancer risk for Black and Latinx women, even when other influencing factors are the same. And the discrepancy could mean fewer women of color get screened. “For the medical community, this is an example of structural racism and is one area we can address quickly and decisively,” physician Malika Fair of the Association of American Medical Colleges tells STAT's Sharon Begley.