According to a recent article published by Casey Ross of Stat News, infusion of bias into decisions about who should receive stepped-up care is a persistent problem even in newer generations of algorithms designed to predict costs. The problem lies not with the algorithms themselves, but how they are applied in a health care system that puts cost management instead of patient management. According to Casey, there are three tactics physicians and data analysts could employ to ameliorate the problem:
- Stop using cost as a proxy for medical need.
- Get better data.
- Link data on need with community services.
UC Berkeley professor Dr. Ziad Obermeyer said, “These small technical choices make the difference between an algorithm that reinforces structural biases in our society, and one that fights against them.”
Read more here.