面向重症监护室血液检查订单的可解释离策略学习 with Side Information（ExOSITO）
摘要
在重症监护室（ICU）为患者订购最少子集的实验室检测往往具有挑战性。医疗团队必须在确保正确信息可用性与减少每项实验室检测的临床负担和成本之间进行权衡。大多数院内环境都会过度订购实验室检测，但正在努力减少对医院资源和环境的负担。本文开发了一种新方法，将离策略学习与特权信息相结合，以识别ICU血液检测的最优订购集合。我们的 approach，EXplainable Off-policy learning with Side Information for ICU blood Test Orders（ExOSITO）为医疗专业人员提供一种可解释的辅助工具，以考虑每个患者的观察和预测未来状态来订购血液检测。我们将此問題 posed as a causal bandit using offline data and a reward function derived from clinically-approved rules; we introduce a novel learning framework that integrates clinical knowledge with observational data to bridge the gap between the optimal and logging policies. 所学得的 policy function 提供可解释的临床信息且在不遗漏任何必需实验室检测的前提下降低成本，优于医生的 policy 和 prior approaches to this practical problem。
引用
@article{arxiv.2504.17277,
title = {ExOSITO: Explainable Off-Policy Learning with Side Information for Intensive Care Unit Blood Test Orders},
author = {Zongliang Ji and Andre Carlos Kajdacsy-Balla Amaral and Anna Goldenberg and Rahul G. Krishnan},
journal= {arXiv preprint arXiv:2504.17277},
year = {2025}
}
备注
Accepted to the Conference on Health, Inference, and Learning (CHIL) 2025