Accuracy and Consistency of LLMs in the Registered Dietitian Exam: The Impact of Prompt Engineering and Knowledge Retrieval

TL;DR

This study evaluates the accuracy and consistency of state-of-the-art LLMs on the Registered Dietitian exam, analyzing how different prompting techniques and knowledge retrieval methods affect their performance in nutrition queries.

Contribution

It introduces a comprehensive evaluation of LLMs on nutrition questions, assessing the impact of advanced prompting techniques and retrieval methods for the first time.

Findings

GPT-4o with CoT-SC outperforms other models in accuracy.

Gemini 1.5 Pro with ZS has the highest consistency.

Prompting techniques significantly influence LLM performance in nutrition tasks.

Abstract

Large language models (LLMs) are fundamentally transforming human-facing applications in the health and well-being domains: boosting patient engagement, accelerating clinical decision-making, and facilitating medical education. Although state-of-the-art LLMs have shown superior performance in several conversational applications, evaluations within nutrition and diet applications are still insufficient. In this paper, we propose to employ the Registered Dietitian (RD) exam to conduct a standard and comprehensive evaluation of state-of-the-art LLMs, GPT-4o, Claude 3.5 Sonnet, and Gemini 1.5 Pro, assessing both accuracy and consistency in nutrition queries. Our evaluation includes 1050 RD exam questions encompassing several nutrition topics and proficiency levels. In addition, for the first time, we examine the impact of Zero-Shot (ZS), Chain of Thought (CoT), Chain of Thought with Self Consistency (CoT-SC), and Retrieval Augmented Prompting (RAP) on both accuracy and consistency of the responses. Our findings revealed that while these LLMs obtained acceptable overall performance, their results varied considerably with different prompts and question domains. GPT-4o with CoT-SC prompting outperformed the other approaches, whereas Gemini 1.5 Pro with ZS recorded the highest consistency. For GPT-4o and Claude 3.5, CoT improved the accuracy, and CoT-SC improved both accuracy and consistency. RAP was particularly effective for GPT-4o to answer Expert level questions. Consequently, choosing the appropriate LLM and prompting technique, tailored to the proficiency level and specific domain, can mitigate errors and potential risks in diet and nutrition chatbots.