Can large language models assist with pediatric dosing accuracy?

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Pediatric Patient Safety

Background:

• Medication errors in pediatric care present a persistent challenge.
• Technological advancements have yet to fully mitigate these risks.
• Innovative solutions are crucial for enhancing patient safety.

Purpose of the Study:

• To evaluate the accuracy and efficiency of Large Language Models (LLMs) in pediatric medication dosage calculations.
• To compare LLM performance against experienced nurses in a clinical setting.
• To identify the potential of AI in reducing pediatric medication errors.

Main Methods:

• A cross-sectional study involving 101 nurses and three LLMs (ChatGPT-4o, Claude-3.0, Llama 3 8B).
• Participants completed a nine-question survey on pediatric medication calculations.
• Primary outcomes measured were calculation accuracy and response time.

Main Results:

• LLMs Claude-3.0 and ChatGPT-4o achieved 100% accuracy, surpassing the nurses' average accuracy of 93.14%.
• LLMs demonstrated significantly faster response times (15.7-75.12 seconds) compared to nurses (over 1600 seconds).
• Task performance was influenced by duration and seniority-group interaction, with an overall mean grade of 91.03.

Conclusions:

• Advanced LLMs show perfect accuracy and rapid calculation capabilities for pediatric medication dosages.
• These AI tools hold significant promise for reducing medication errors in pediatric care.
• Further research is warranted to explore the practical integration of LLMs into clinical workflows.