How Reliable AI Chatbots are for Disease Prediction from Patient Complaints?

TL;DR

The paper investigates the reliability of AI chatbots for predicting disease from emergency department patient complaints using few-shot learning and compares them against BERT baselines. It leverages the GED3C gout chief complaint corpus and evaluates GPT-4.0, Claude 3 Opus, and Gemini Ultra 1.0 across two-class and three-class prediction tasks, with Clinical BERT and BERT Base Uncased as additional baselines. The results show GPT-4.0 improves with more training data, Gemini Ultra 1.0 performs best with fewer examples, and Claude 3 Opus remains stable across shot levels; however, none achieve the reliability needed for critical medical decisions, highlighting the necessity of rigorous validation and human oversight. The study emphasizes that AI chatbots should augment, not replace, clinical judgment and calls for further research to enhance reliability in disease prediction from patient complaints.

Abstract

Artificial Intelligence (AI) chatbots leveraging Large Language Models (LLMs) are gaining traction in healthcare for their potential to automate patient interactions and aid clinical decision-making. This study examines the reliability of AI chatbots, specifically GPT 4.0, Claude 3 Opus, and Gemini Ultra 1.0, in predicting diseases from patient complaints in the emergency department. The methodology includes few-shot learning techniques to evaluate the chatbots' effectiveness in disease prediction. We also fine-tune the transformer-based model BERT and compare its performance with the AI chatbots. Results suggest that GPT 4.0 achieves high accuracy with increased few-shot data, while Gemini Ultra 1.0 performs well with fewer examples, and Claude 3 Opus maintains consistent performance. BERT's performance, however, is lower than all the chatbots, indicating limitations due to limited labeled data. Despite the chatbots' varying accuracy, none of them are sufficiently reliable for critical medical decision-making, underscoring the need for rigorous validation and human oversight. This study reflects that while AI chatbots have potential in healthcare, they should complement, not replace, human expertise to ensure patient safety. Further refinement and research are needed to improve AI-based healthcare applications' reliability for disease prediction.

Figures (3)

• Figure 1: Process Flow of Disease Prediction
• Figure 2: A Sample Prompt for 2-Class Prediction
• Figure 3: F1 Score Analysis of Different AI Chatbots