Language Models as Interfaces, Not Oracles: A Hybrid LLM-ML System for Pediatric Appendicitis
Researchers have developed ClaMPAPP, a hybrid system that uses large language models (LLMs) as an interface for clinical decision support, rather than as direct diagnostic engines. This approach separates natural language processing from predictive inference, addressing LLM limitations like prompt sensitivity and output instability. ClaMPAPP extracts structured clinical features from free-text narratives, validates them, and then feeds them into an XGBoost classifier. This hybrid system demonstrated superior diagnostic performance in identifying pediatric appendicitis compared to end-to-end LLM baselines, particularly in minimizing missed cases. AI
IMPACT This hybrid approach offers a more auditable and robust pathway for integrating LLMs into clinical decision support systems, potentially improving diagnostic accuracy and safety.