A widely used kidney function test (eGFR) that includes a race-based adjustment systematically gave Black patients healthier-appearing scores, leading to delayed medical care including kidney transplant referrals for over 700 Black patients in the Boston area.
The incident involves the CKD-EPI equation used to estimate kidney function (eGFR) that includes a race multiplier giving Black patients scores boosted by 15.9 percent. A study of 57,000 patients with chronic kidney disease from the Mass General Brigham health system found that one third of Black patients (743 people) would have been classified into more severe kidney disease categories without the race adjustment. Most significantly, 64 Black patients would have qualified for kidney transplant wait lists based on recalculated scores, but none had been referred or evaluated for transplant. The race-based formula was created in 2009 by researchers who added the 'race correction' to account for statistical differences in their data, but the underlying physiological justification remained unclear. The study findings helped convince Mass General Brigham to abandon the race-based formula in June, joining other major hospitals like University of Washington and Vanderbilt. The National Kidney Foundation and American Society of Nephrology formed a task force to evaluate the use of race in kidney testing following growing criticism from physicians and medical students.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Unequal treatment of individuals or groups by AI, often based on race, gender, or other sensitive characteristics, resulting in unfair outcomes and unfair representation of those groups.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed