Home Department: Data Science
Mentor: Sherri Rose (Health Policy)

"Towards a Rigorous Framework for Analyzing the Social Impact of Healthcare Algorithms"

Healthcare AI adoption is increasing rapidly, with a reported 71% of hospitals using predictive AI systems in 2024 compared to 66% in 2023. Most such systems have not been rigorously evaluated for potentially negative consequences prior to or following deployment, at the risk of harming patients and exacerbating health inequities. With algorithmic evaluation methods needed swiftly, Peter’s project will use rigorous causal inference techniques (e.g., combining double robust machine learning and microsimulation modeling) to analyze the impact of several healthcare algorithms on various healthcare metrics, including access, outcomes, quality, and costs, contributing to a developing framework for such analyses.