In the United States, a hospital group uses an algorithm to determine the urgency of performing kidney transplants. However, this algorithm based on skin color recently forced hundreds of black patients off the waiting list.
Algorithms in the health field are intended to help decision-making and simplify the work of professionals for better results. A study published in the Journal of General Internal Medicine on October 15, 2020 shows that this is not always the case. According to a team of researchers from Harvard University (USA), a machine learning system has recently reduced the chances of recovery for black patients.
The study directors considered data from the Mass General Brigham Hospital Network on 57,000 patients with kidney failure chronic. In order to better redirect these patients, the professionals use an algorithm whose goal is to decide on the urgency of such an operation according to the patients.
According to the researchers, the system minimized this urgency for 700 black patients . However, this number represents a third of the black patients whose data were included in the study. In other words, their cases should have been considered more serious. In addition, the waiting list should have integrated 64 of these patients for an emergency kidney transplant.
You should know that this is not the first time that an AI-based system discriminates people based on their skin color. In 2019, this was the case with a device that many health systems were using. The latter, prioritizing access to care for chronic diseases, favored white people over black people. However, the device in question did not take skin color into account. Nevertheless, the system evoked by the Harvard researchers is entirely based on this criterion.
Previous cases of discrimination have elicited relatively little response. On the other hand, this latest news has catch the attention of federal lawmakers Americans. Take, for example, Democratic MP Richard Neal, who spoke of the need to reconsider the question of skin color at the level of algorithms. He pointed to the risk of late or inaccurate diagnoses for black and Latino patients. However, this often results in lower quality care.
Richard Neal asked industry companies and the National Institutes of Health (NIH) to study the impact of clinical algorithms on patients. This request is in line with that made to the Department of Health by Senator Elizabeth Warren and her group.
