CDEG: Learning Decision-Critical Evidence for Long-Horizon Diagnostic Agents
2026-08-24 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors focus on improving medical diagnosis that happens over many steps by better using past cases. Unlike past methods that reuse entire past experiences without checking which parts really matter, they created a system called CDEG that learns the important pieces of evidence by comparing successful and failed cases. CDEG organizes this key evidence into a graph and uses it to guide doctors on what information to gather or reconsider during diagnosis. Tests show that this approach improves diagnostic accuracy compared to previous methods. Their work highlights the importance of focusing on critical evidence rather than whole past experiences for better long-term medical decisions.
long-horizon diagnosismedical question answeringdiagnostic reasoningagentic approachescounterfactual interventionsdecision-critical evidencediagnostic trajectoriesevidence reappraisalstructured graphout-of-distribution benchmarks
Authors
Xiwei Dai, Zijie Meng, Zhiting Fan, Yixuan Tang, Ziru Niu, Zuozhu Liu
Abstract
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.