Evaluating the Challenges of LLMs in Real-world Medical Follow-up: A Comparative Study and An Optimized Framework

TL;DR

This study compares end-to-end LLM-based medical follow-up chatbots with a modular, structured approach, demonstrating that the latter improves dialog stability, accuracy, and efficiency in complex medical tasks.

Contribution

The paper introduces a modular framework with task decomposition and flow control that significantly enhances LLM performance in medical follow-up applications.

Findings

Modular approach reduces dialogue turns by 46.73%.

Significantly lowers token consumption by 80-87.5%.

Improves dialog stability and extraction accuracy.

Abstract

When applied directly in an end-to-end manner to medical follow-up tasks, Large Language Models (LLMs) often suffer from uncontrolled dialog flow and inaccurate information extraction due to the complexity of follow-up forms. To address this limitation, we designed and compared two follow-up chatbot systems: an end-to-end LLM-based system (control group) and a modular pipeline with structured process control (experimental group). Experimental results show that while the end-to-end approach frequently fails on lengthy and complex forms, our modular method-built on task decomposition, semantic clustering, and flow management-substantially improves dialog stability and extraction accuracy. Moreover, it reduces the number of dialogue turns by 46.73% and lowers token consumption by 80% to 87.5%. These findings highlight the necessity of integrating external control mechanisms when deploying LLMs in high-stakes medical follow-up scenarios.