Medical Dialogue Generation via Intuitive-then-Analytical Differential Diagnosis

TL;DR

IADDx addresses the gap in medical dialogue generation by explicitly modeling the differential diagnosis through an intuitive-then-analytic process and a diagnosis-oriented graph. It uses retrieval-augmented reasoning to refine potential diseases and guide knowledge retrieval and response generation via a Fusion-in-Decoder framework. Across MedDG and KaMed, IADDx achieves superior automatic and human evaluations and offers interpretable diagnostic paths and intermediate results. This approach enhances trust and practicality in clinical and patient-facing medical dialogue systems.

Abstract

Medical dialogue systems have attracted growing research attention as they have the potential to provide rapid diagnoses, treatment plans, and health consultations. In medical dialogues, a proper diagnosis is crucial as it establishes the foundation for future consultations. Clinicians typically employ both intuitive and analytic reasoning to formulate a differential diagnosis. This reasoning process hypothesizes and verifies a variety of possible diseases and strives to generate a comprehensive and rigorous diagnosis. However, recent studies on medical dialogue generation have overlooked the significance of modeling a differential diagnosis, which hinders the practical application of these systems. To address the above issue, we propose a medical dialogue generation framework with the Intuitive-then-Analytic Differential Diagnosis (IADDx). Our method starts with a differential diagnosis via retrieval-based intuitive association and subsequently refines it through a graph-enhanced analytic procedure. The resulting differential diagnosis is then used to retrieve medical knowledge and guide response generation. Experimental results on two datasets validate the efficacy of our method. Besides, we demonstrate how our framework assists both clinicians and patients in understanding the diagnostic process, for instance, by producing intermediate results and graph-based diagnosis paths.

Figures (2)

• Figure 1: An example of differential diagnosis in a medical dialogue, which contains intuitive and analytic reasoning.
• Figure 2: Left: The architecture of Differential Diagnosis, which includes the intuitive association stage and the analytic refinement stage. The multi-disease classifier in Stage 2 generates a refined diagnosis to guide response generation. Right: The structure of Response Generation. The diagnosis combined with the dialogue acts are used to retrieve relevant knowledge.