Dissecting racial bias in an algorithm used to manage the health of populations


I am still wrapping my head around this one. It boils down to using an algorithm that looks at “income inequalities” and “healthcare costs”.

I am not sure how the hospitals are falling for the snake oil of “artificial intelligence”.

Perhaps the administrators are “naturally stupid”.

The AI in the hospitals requires various signals from the embedded sensors, real-time monitoring and the works. You can’t have a “centralised” algorithm running all over the place (or the district, for example) because it is going incredibly skew the results. I am pained by these clickbait headlines because otherwise, then, it is going to incredibly slow the rollout of better products.

(This aside- I came across this study in WSJ). It seems that magazines and journals have an impeccable marketing department. The headline, of course, could be titled differently, but I followed it through because the algorithm was “accused” of racial bias).

Health systems rely on commercial prediction algorithms to identify and help patients with complex health needs. We show that a widely used algorithm, typical of this industry-wide approach and affecting millions of patients, exhibits significant racial bias: At a given risk score, Black patients are considerably sicker than White patients, as evidenced by signs of uncontrolled illnesses. Remedying this disparity would increase the percentage of Black patients receiving additional help from 17.7 to 46.5%. The bias arises because the algorithm predicts health care costs rather than illness, but unequal access to care means that we spend less money caring for Black patients than for White patients. Thus, despite health care cost appearing to be an effective proxy for health by some measures of predictive accuracy, large racial biases arise.

via Dissecting racial bias in an algorithm used to manage the health of populations | Science