Pruning the Path to Optimal Care: Identifying Systematically Suboptimal Medical Decision-Making with Inverse Reinforcement Learning
Machine Learning
2024-11-11 v1 Quantitative Methods
Applications
Computation
Machine Learning
Abstract
In aims to uncover insights into medical decision-making embedded within observational data from clinical settings, we present a novel application of Inverse Reinforcement Learning (IRL) that identifies suboptimal clinician actions based on the actions of their peers. This approach centers two stages of IRL with an intermediate step to prune trajectories displaying behavior that deviates significantly from the consensus. This enables us to effectively identify clinical priorities and values from ICU data containing both optimal and suboptimal clinician decisions. We observe that the benefits of removing suboptimal actions vary by disease and differentially impact certain demographic groups.
Cite
@article{arxiv.2411.05237,
title = {Pruning the Path to Optimal Care: Identifying Systematically Suboptimal Medical Decision-Making with Inverse Reinforcement Learning},
author = {Inko Bovenzi and Adi Carmel and Michael Hu and Rebecca M. Hurwitz and Fiona McBride and Leo Benac and José Roberto Tello Ayala and Finale Doshi-Velez},
journal= {arXiv preprint arXiv:2411.05237},
year = {2024}
}
Comments
13 pages, 4 figures