Race-Based Disease Scores: Good for One, Not for The Other
(Thursday, June 22, 2023)
About a year ago, the National Kidney Foundation updated the formula for calculating the estimated Glomerular Filtration Rate (eGFR) by eliminating the race-criteria because it was found that the race-based algorithm incorporating GFR is delaying diagnosis and treatment of worsening chronic kidney disease (CKD). In a report published last week in the Journal of the American Medical Association (JAMA), the same is not true for colorectal cancer in which case eliminating the race-based formula may result in worse prediction accuracy for racial and ethnic minority groups that may lead to inappropriate care recommendations that ultimately contribute to health disparities in managing colorectal cancer.
Race- and ethnicity-based algorithms are being widely reviewed to address non-intentional bias and related health disparities. But these two cases emphasize that one logic does not fit all cases. It is necessary to evaluate the scoring formulas for each specific disease scenario before any changes to established scoring or calculation systems are made. In both diseases, colorectal cancer and CKD, race and ethnicity of the patients have frequently been cited as important factors for assessment of the seriousness of the disease. But as the two reports present, what might be true for one case, i.e., CKD where the race-based algorithm was found to lead to erroneous conclusions regarding the seriousness of the disease, the opposite was found to be true for colorectal cancer.
Race-based algorithms are even more important for software and Artificial Intelligence (AI) based digital health devices. Several discrepancies have been reported for imaging- and questionnaire-based devices where the training of the device with non-diverse data leads to errors in the conclusions. The above studies point out that this question is not as simple as one may be inclined to argue. It needs systematic and vigorous evaluation in each case to justify inclusion or elimination of race-based algorithms.
Dr. Mukesh Kumar
Founder & CEO, FDAMap
Linkedin: Mukesh Kumar, PhD, RAC