# CDEG: Learning Decision-Critical Evidence for Long-Horizon Diagnostic Agents

arXiv: https://arxiv.org/abs/2608.22899 · HTML: https://mlsanity.com/paper/2608.22899/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 22 of its 37 references that are in the map.
- First posted on arXiv 24 Aug 2026.

## Builds on (most cited first, 22 of 22)

- [What Disease does this Patient Have? A Large-scale Open Domain Question Answering Dataset from Medical Exams](https://mlsanity.com/paper/2009.13081/) (2020)
- [Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models](https://mlsanity.com/paper/2506.05176/) (2025)
- [Self-Alignment Pretraining for Biomedical Entity Representations](https://mlsanity.com/paper/2010.11784/) (2020)
- [Synapse: Trajectory-as-Exemplar Prompting with Memory for Computer Control](https://mlsanity.com/paper/2306.07863/) (2023)
- [How Memory Management Impacts LLM Agents: An Empirical Study of Experience-Following Behavior](https://mlsanity.com/paper/2505.16067/) (2025)
- [MedAgentBench: A Realistic Virtual EHR Environment to Benchmark Medical LLM Agents](https://mlsanity.com/paper/2501.14654/) (2025)
- [MEDDxAgent: A Unified Modular Agent Framework for Explainable Automatic Differential Diagnosis](https://mlsanity.com/paper/2502.19175/) (2025)
- [Evolving Interactive Diagnostic Agents in a Virtual Clinical Environment](https://mlsanity.com/paper/2510.24654/) (2025)
- [GSEM: Graph-based Self-Evolving Memory for Experience Augmented Clinical Reasoning](https://mlsanity.com/paper/2603.22096/) (2026)
- [Evo-MedAgent: Beyond One-Shot Diagnosis with Agents That Remember, Reflect, and Improve](https://mlsanity.com/paper/2604.14475/) (2026)
- [Improving Clinical Diagnosis with Counterfactual Multi-Agent Reasoning](https://mlsanity.com/paper/2603.27820/) (2026)
- [Explaining a machine learning decision to physicians via counterfactuals](https://mlsanity.com/paper/2306.06325/) (2023)
- [MedEinst: Benchmarking the Einstellung Effect in Medical LLMs through Counterfactual Differential Diagnosis](https://mlsanity.com/paper/2601.06636/) (2026)
- [Benchmarking Multi-turn Medical Diagnosis: Hold, Lure, and Self-Correction](https://mlsanity.com/paper/2604.04325/) (2026)
- [DocCHA: Towards LLM-Augmented Interactive Online diagnosis System](https://mlsanity.com/paper/2507.07870/) (2025)
- [Faithfulness vs. Safety: Evaluating LLM Behavior Under Counterfactual Medical Evidence](https://mlsanity.com/paper/2601.11886/) (2026)
- [Strong Reasoning Isn't Enough: Evaluating Evidence Elicitation in Interactive Diagnosis](https://mlsanity.com/paper/2601.19773/) (2026)
- [Joint Optimization of Reasoning and Dual-Memory for Self-Learning Diagnostic Agent](https://mlsanity.com/paper/2604.07269/) (2026)
- [Evaluation of Causal Reasoning for Large Language Models in Contextualized Clinical Scenarios of Laboratory Test Interpretation](https://mlsanity.com/paper/2509.16372/) (2025)
- [Emulating Clinician Cognition via Self-Evolving Deep Clinical Research](https://mlsanity.com/paper/2603.10677/) (2026)
- [DiaLLMs: EHR Enhanced Clinical Conversational System for Clinical Test Recommendation and Diagnosis Prediction](https://mlsanity.com/paper/2506.20059/) (2025)
- [MedExAgent: Training LLM Agents to Ask, Examine, and Diagnose in Noisy Clinical Environments](https://mlsanity.com/paper/2605.07058/) (2026)

## Abstract (arXiv)

Unlike static medical question answering, long-horizon diagnosis captures the sequential nature of clinical practice: evidence is progressively acquired, integrated, and evaluated over multiple rounds of interaction before reaching a final diagnosis. However, existing doctor agents often fail when critical evidence is either not acquired or not adequately incorporated into diagnostic reasoning. Recent agentic approaches attempt to address these failures by reusing historical trajectories or distilled memories. But their diagnostic gains remain constrained because such experience may contain noisy or incidental information and is typically reused without validating which evidence actually drives diagnostic decisions. To address this limitation, we introduce CDEG, a graph-based framework that learns reusable decision-critical evidence from historical diagnostic trajectories. CDEG contrasts successful and failed trajectories from the same case to identify candidate evidence, validates their diagnostic impact through controlled counterfactual interventions, and organizes the resulting diagnosis--evidence--action relations into a structured graph. During inference, CDEG tracks the evolving patient evidence state to retrieve relevant diagnostic relations and selectively guide missing evidence acquisition or overlooked evidence reappraisal. Across in-domain and out-of-distribution benchmarks with multiple doctor agent backbones, CDEG consistently improves diagnostic performance, achieving up to an 11.5% accuracy gain over vanilla agents. These results demonstrate that reliable long-horizon diagnosis requires moving beyond trajectory-level experience reuse toward evidence-level learning of the factors that truly shape clinical decisions.