Experts are warning doctors of the pitfalls of bias in artificial intelligence algorithms when determining elements of treatment plans for patients with chronic kidney disease (CKD). CKD is a condition where one or both kidneys function at reduced capacity or do not function at all. The kidneys filter the blood of wastes and toxins which would, otherwise, build up in the body. Doctors have been making use of artificial intelligence (AI) to allow them to make better models and predictions of diseases and their effects.
A critical component of the AI system is the algorithm. The algorithm is a set of rules which are followed by the computer when doing calculations. The systems used by doctors consider complex information such as medical test results and risk factors. But they have been seen exacerbating certain racial bias, such as in the case of Black patients. A recent study warns of one situation where the health of Black patients was often overestimated.
This resulted in those patients receiving less specialized care. Additionally, once these patients are made a lower priority, they end up on a much longer waiting list for a transplant. 56,845 CKD patients were the subject of the study, with 2225 patients who were Black. Doctors used a system that was designed to calculate CKD-EPI, which is used to measure kidney functionality through blood work.
Researchers found 700 Black patients were giving higher health scores even though they showed similar, if not the same symptoms as well as the same risks. These misdiagnosis’s resulted in patients not being referred to kidney specialists or qualifying to be put on a transplant list. Not only did doctors not question the AI system’s results but results for Blacks were inflated by 15.9 percent.
Researchers say these misdiagnoses are part of a long history of Black healthcare being less effective than the healthcare of whites. They suggest that Blacks patients are looked down upon due to socioeconomic status Further research showed these algorithms used false and gross stereotypes.
Have you seen our YouTube channel yet? We update daily with the answers to your kidney health questions.