Offline reinforcement learning with uncertainty for treatment strategies in sepsis

TL;DR

This paper introduces a reinforcement learning approach for sepsis treatment that estimates optimal strategies, assesses confidence, and suggests multiple options, addressing data bias and personalization challenges in critical care.

Contribution

The study presents a novel reinforcement learning framework that incorporates uncertainty estimation and bias mitigation for personalized sepsis treatment recommendations.

Findings

Reinforcement learning can identify treatment strategies with confidence levels.

Bias against aggressive treatment is present and can be mitigated.

Multiple treatment options can be recommended rather than a single policy.

Abstract

Guideline-based treatment for sepsis and septic shock is difficult because sepsis is a disparate range of life-threatening organ dysfunctions whose pathophysiology is not fully understood. Early intervention in sepsis is crucial for patient outcome, yet those interventions have adverse effects and are frequently overadministered. Greater personalization is necessary, as no single action is suitable for all patients. We present a novel application of reinforcement learning in which we identify optimal recommendations for sepsis treatment from data, estimate their confidence level, and identify treatment options infrequently observed in training data. Rather than a single recommendation, our method can present several treatment options. We examine learned policies and discover that reinforcement learning is biased against aggressive intervention due to the confounding relationship between mortality and level of treatment received. We mitigate this bias using subspace learning, and develop methodology that can yield more accurate learning policies across healthcare applications.