EvoClinician: A Self-Evolving Agent for Multi-Turn Medical Diagnosis via Test-Time Evolutionary Learning

TL;DR

EvoClinician is a self-evolving AI agent designed for multi-turn medical diagnosis, capable of iteratively asking questions and ordering tests to improve diagnostic accuracy and efficiency in a realistic, simulated clinical environment.

Contribution

The paper introduces EvoClinician, a novel self-evolving agent that learns diagnostic strategies at test time through a feedback loop, advancing beyond static models and existing continual learning approaches.

Findings

EvoClinician outperforms baseline models in diagnostic accuracy.

The self-evolving approach improves resource efficiency in diagnosis.

EvoClinician adapts strategies effectively during the diagnostic process.

Abstract

Prevailing medical AI operates on an unrealistic ''one-shot'' model, diagnosing from a complete patient file. However, real-world diagnosis is an iterative inquiry where Clinicians sequentially ask questions and order tests to strategically gather information while managing cost and time. To address this, we first propose Med-Inquire, a new benchmark designed to evaluate an agent's ability to perform multi-turn diagnosis. Built upon a dataset of real-world clinical cases, Med-Inquire simulates the diagnostic process by hiding a complete patient file behind specialized Patient and Examination agents. They force the agent to proactively ask questions and order tests to gather information piece by piece. To tackle the challenges posed by Med-Inquire, we then introduce EvoClinician, a self-evolving agent that learns efficient diagnostic strategies at test time. Its core is a ''Diagnose-Grade-Evolve'' loop: an Actor agent attempts a diagnosis; a Process Grader agent performs credit assignment by evaluating each action for both clinical yield and resource efficiency; finally, an Evolver agent uses this feedback to update the Actor's strategy by evolving its prompt and memory. Our experiments show EvoClinician outperforms continual learning baselines and other self-evolving agents like memory agents. The code is available at https://github.com/yf-he/EvoClinician