In the 1990s, a well-intentioned government policy aimed at addressing health disparities inadvertently paved the way for problematic race-based algorithms in medicine. The National Institutes of Health’s requirement to collect and report racial data in funded research revealed stark racial divides in health outcomes. However, this quantification also enabled the development of algorithms that misused race as a health risk factor.Researchers, new to handling race data, often categorized subjects simplistically as Black, Hispanic, or Asian, overlooking complex ancestries within these groups. Many still considered race a biological explanation for differences, rather than a social construct with weak genetic ties.
The kidney function estimation algorithm exemplifies this issue. Based on higher average creatinine levels in Black people, researchers introduced a race adjustment. They attributed this to greater muscle mass in Black individuals, an assertion rooted in slavery-era stereotypes and supported by thin scientific evidence.
Similar oversimplifications occurred across medical specialties. In pulmonology, the long-held belief that Black people have lower normal lung function led to overlooked chronic lung diseases. Obstetrics saw echoes of unfounded notions about Black women’s pelvises influencing birth risk calculations.
These race-based algorithms have had real consequences. In lung function tests, adjustments for race have delayed disability payments for Black workers and led to missed or underestimated respiratory diseases.
As medicine grapples with its history of racial bias, efforts are underway to reevaluate and remove race from clinical algorithms. The challenge lies in addressing health disparities without perpetuating harmful stereotypes or overlooking the complex socioeconomic factors that influence health outcomes.
See “How race became ubiquitous in medical decision-making tools” (September 4, 2024)
