Revaluating proxy measures for analysis of healthcare data and predictive analytics need to account for inherent historical biases in the data.
Developers focusing on risk prediction for managed care programs stumbled upon an unanticipated variable – the impact of existing racial bias in data sets. The inherent bias occurs because of demographics that spend less on healthcare, where cost and healthcare resources consumed are the largest proxy measure.
“There’s growing concern around AI, machine learning, data science, and the risk of automation reinforcing existing biases through the use of algorithms. It was a confluence of what we know is a potential concern,” said Brian Powers, MD, MBA, physician and researcher at Brigham and Women’s Hospital and lead author of the study.