Learning Diagnostic Reasoning for Decision Support in Toxicology

TL;DR

This paper introduces DeToxR, an RL-finetuned LLM that effectively synthesizes diverse data sources to improve decision support in toxicology, outperforming baseline models and even expert toxicologists.

Contribution

It presents the first RL adaptation of LLMs for emergency toxicology, with a novel data-fusion engine and a clinical reward-based optimization approach.

Findings

DeToxR outperforms unadapted LLMs and supervised baselines.

DeToxR surpasses expert toxicologists in identifying poisons.

The model demonstrates significant potential for high-stakes clinical decision support.

Abstract

Acute poly-substance intoxication requires rapid, life-saving decisions under substantial uncertainty, as clinicians must rely on incomplete ingestion details and nonspecific symptoms. Effective diagnostic reasoning in this chaotic environment requires fusing unstructured, non-medical narratives (e.g. paramedic scene descriptions and unreliable patient self-reports or known histories), with structured medical data like vital signs. While Large Language Models (LLMs) show potential for processing such heterogeneous inputs, they struggle in this setting, often underperforming simple baselines that rely solely on patient histories. To address this, we present DeToxR (Decision-support for Toxicology with Reasoning), the first adaptation of Reinforcement Learning (RL) to emergency toxicology. We design a robust data-fusion engine for multi-label prediction across 14 substance classes based on an LLM finetuned with Group Relative Policy Optimization (GRPO). We optimize the model's reasoning directly using a clinical performance reward. By formulating a multi-label agreement metric as the reward signal, the model is explicitly penalized for missing co-ingested substances and hallucinating absent poisons. Our model significantly outperforms its unadapted base LLM counterpart and supervised baselines. Furthermore, in a clinical validation study, the model indicates a clinical advantage by outperforming an expert toxicologist in identifying the correct poisons (Micro-F1: 0.644 vs. 0.473). These results demonstrate the potential of RL-aligned LLMs to synthesize unstructured pre-clinical narratives and structured medical data for decision support in high-stakes environments.