Recent advancements in Large Language Models (LLMs) and multi-agent systems have led to the emergence of Agentic AI, promising enhanced medical reasoning capabilities. However, open-ended conversational agents are still vulnerable to two critical failure modes: premature diagnostic handoff and silent clinical hallucinations that may go unnoticed before reaching patients. This work proposes a multi-agent framework to address these issues by replacing the ``LLM-as-a-judge'' routing with deterministic orchestration constraints. The framework includes two safety mechanisms. First, a neuro-symbolic state-tracking gate enforces the completeness of the OLDCARTS clinical protocol (Onset, Location, Duration, Character, Aggravating/Alleviating factors, Radiation, Timing, and Severity) by blocking diagnostic transitions until all required dimensions are collected. Second, an epistemic uncertainty quantification (UQ) gate computes semantic entropy (H) across K=5 independent diagnostic samples to identify and intercept divergent outputs before delivery. We evaluate the system using simulated patient agents powered by the llama-3.1-70b-instruct model on 150 test cases. The full architecture achieves 49.3% diagnostic precision, representing an absolute improvement of 11.3 percentage points over an unconstrained baseline. Additionally, we observe a statistically significant negative correlation (r = -0.181, p < 0.05).
Blogger's Review: This framework effectively combines neuro-symbolic methods with uncertainty quantification, providing crucial safety measures for AI applications in healthcare. It demonstrates the potential of Agentic AI in enhancing diagnostic accuracy, making its practical implications worthy of attention.