# Constructing and Evaluating Clinical Reasoning Trajectories for Medical Agent

arXiv: https://arxiv.org/abs/2609.05090 · HTML: https://mlsanity.com/paper/2609.05090/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Clinical and biomedical applications, one of 17 directions in LLM reasoning.
- Builds on 34 of its 55 references that are in the map.
- First posted on arXiv 4 Sep 2026.

## Placed in

- LLM reasoning / [Clinical and biomedical applications](https://mlsanity.com/lr/clinical-and-biomedical-applications/)

## Builds on (most cited first, 30 of 34)

- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [Tree of Thoughts: Deliberate Problem Solving with Large Language Models](https://mlsanity.com/paper/2305.10601/) (2023)
- [Voyager: An Open-Ended Embodied Agent with Large Language Models](https://mlsanity.com/paper/2305.16291/) (2023)
- [Large Language Models Encode Clinical Knowledge](https://mlsanity.com/paper/2212.13138/) (2022)
- [PubMedQA: A Dataset for Biomedical Research Question Answering](https://mlsanity.com/paper/1909.06146/) (2019)
- [AgentBench: Evaluating LLMs as Agents](https://mlsanity.com/paper/2308.03688/) (2023)
- [What Makes Good In-Context Examples for GPT-3?](https://mlsanity.com/paper/2101.06804/) (2021)
- [Learning To Retrieve Prompts for In-Context Learning](https://mlsanity.com/paper/2112.08633/) (2021)
- [MedMCQA : A Large-scale Multi-Subject Multi-Choice Dataset for Medical domain Question Answering](https://mlsanity.com/paper/2203.14371/) (2022)
- [Selective Annotation Makes Language Models Better Few-Shot Learners](https://mlsanity.com/paper/2209.01975/) (2022)
- [MedXpertQA: Benchmarking Expert-Level Medical Reasoning and Understanding](https://mlsanity.com/paper/2501.18362/) (2025)
- [Many-Shot In-Context Learning](https://mlsanity.com/paper/2404.11018/) (2024)
- [MDAgents: An Adaptive Collaboration of LLMs for Medical Decision-Making](https://mlsanity.com/paper/2404.15155/) (2024)
- [Self-Adaptive In-Context Learning: An Information Compression Perspective for In-Context Example Selection and Ordering](https://mlsanity.com/paper/2212.10375/) (2022)
- [A Survey of Large Language Models in Medicine: Progress, Application, and Challenge](https://mlsanity.com/paper/2311.05112/) (2023)
- [Self-Discover: Large Language Models Self-Compose Reasoning Structures](https://mlsanity.com/paper/2402.03620/) (2024)
- [A Survey of LLM-based Agents in Medicine: How far are we from Baymax?](https://mlsanity.com/paper/2502.11211/) (2025)
- [Rationale-Guided Retrieval Augmented Generation for Medical Question Answering](https://mlsanity.com/paper/2411.00300/) (2024)
- [ClinicalLab: Aligning Agents for Multi-Departmental Clinical Diagnostics in the Real World](https://mlsanity.com/paper/2406.13890/) (2024)
- [KG4Diagnosis: A Hierarchical Multi-Agent LLM Framework with Knowledge Graph Enhancement for Medical Diagnosis](https://mlsanity.com/paper/2412.16833/) (2024)
- [Automatic Evaluation of Healthcare LLMs Beyond Question-Answering](https://mlsanity.com/paper/2502.06666/) (2025)
- [An Empirical Evaluation of Prompting Strategies for Large Language Models in Zero-Shot Clinical Natural Language Processing](https://mlsanity.com/paper/2309.08008/) (2023)
- [MedReflect: Teaching Medical LLMs to Self-Improve via Reflective Correction](https://mlsanity.com/paper/2510.03687/) (2025)
- [Chain-of-Though (CoT) prompting strategies for medical error detection and correction](https://mlsanity.com/paper/2406.09103/) (2024)
- [MediEval: A Unified Medical Benchmark for Patient-Contextual and Knowledge-Grounded Reasoning in LLMs](https://mlsanity.com/paper/2512.20822/) (2025)
- [ClinicalAgents: Multi-Agent Orchestration for Clinical Decision Making with Dual-Memory](https://mlsanity.com/paper/2603.26182/) (2026)
- [Benchmarking Multi-turn Medical Diagnosis: Hold, Lure, and Self-Correction](https://mlsanity.com/paper/2604.04325/) (2026)

## Abstract (arXiv)

Evaluation of medical artificial intelligence agents remains predominantly answer-centric, assessing only the correctness of final outputs while overlooking the quality of intermediate reasoning. In clinical settings, however, a correct answer reached through fabricated evidence or incoherent logic is as dangerous as an incorrect one. We propose MedTraj, a framework that treats reasoning trajectories as critical objects for construction, evaluation, and optimization. The pipeline generates structured multi-step clinical reasoning chains from medical diagnosis sources. Each trajectory is then parsed into clinical observations, evidence, numbered reasoning steps, and a final conclusion, and scored across five quality dimensions: coherence, evidence support, hallucination, completeness, and traceability. Controlled error injection introduces targeted faults into otherwise correct trajectories to establish causal links between specific reasoning failures and measurable quality degradation. Building on this, step-level filtering based on marginal contribution identifies which individual reasoning steps drive or undermine trajectory quality. Finally, quality-weighted context learning feeds trajectory evaluations back into the model at inference time, allowing it to learn from both strong and weak reasoning demonstrations. Experiments across CareQA, PubMedQA, and CECMed demonstrate that trajectory context consistently improves reasoning coherence, with gains of +0.029 to +0.041 over a zero-shot baseline. On the CECMed severity assessment task, quality-weighted context nearly doubles the correctness over the zero-shot baseline while cutting the hallucination ratio by 87%. Marginal-contribution analysis further shows that a small minority (15.3%) of reasoning steps carry most of the quality signal, and that extending chains beyond four steps yields diminishing returns.