Explore the clinical tools that use race to steer care

Every day, physicians use clinical algorithms to make decisions about the patients in their exam rooms. To help weigh a patient’s surgical risk or likelihood of disease, they factor in attributes such as blood pressure, age, weight, surgical history — and in some cases, a patient’s race. Like many clinical researchers, informatician Shyam Visweswaran started learning about those race-based tools in 2020, when a catalyzing New England Journal of Medicine paper laid out 13 common examples in medicine.

“That started me thinking,” said Visweswaran, who helps clinicians implement their experimental algorithms as vice chair of informatics at the University of Pittsburgh. “I wanted to see what the scope of this is, how many such algorithms are out there.” He couldn’t find a resource to answer the question — so he decided to build one. In the most comprehensive list to date, Visweswaran and his colleagues describe 48 clinical tools with race adjustments, three of which were added based on STAT’s reporting.

With Visweswaran’s permission, STAT created its own version that allows users to filter tools by specialty, racial variables, and more.

Race-based algorithms have come under scrutiny because they perpetuate the false idea that race makes people biologically different. The database, which was released in July 2023 and is under review at a journal, highlights another concern: the messy and inconsistent use of race in these tools. Across the calculators, Visweswaran’s group found 49 distinct categories for race and ethnicity. It’s rarely clear how a patient of mixed ancestry might fit into any one of them, who gets to choose the best fit, and how their care might be directed differently as a result.

It’s not a comprehensive list. It doesn’t account for the home-grown tools hospitals often use to support clinical decision-making, or the growing number of algorithms used by health insurers. But it’s a starting point, said Visweswaran, to help keep track as clinicians and health systems begin to reform their approach to race.

“As these clinical guidelines and equations get updated to hopefully fix the issues, the initial versions disappear,” he said. “What is becoming more and more clear is that we need to understand the history.”