Related Experiment Video
Updated: Jun 5, 2025
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
The utility of Chat Generative Pre-trained Transformer as a patient resource in paediatric otolaryngology
Walter M Jongbloed1, Nancy Grover2
1Department of Otolaryngology - Head & Neck Surgery, UConn Health, Farmington, CT, USA.
Objectives:
This study aimed to determine the feasibility of using Chat Generative Pre-trained Transformer (ChatGPT) (https://chatgpt.com) as a patient resource for paediatric otolaryngology conditions and assess the quality of responses generated by ChatGPT when compared with information available on the internet.
Method:
ChatGPT responses to common paediatric otolaryngology conditions were compared with top internet pages for readability (Flesch Reading Ease score, word count), expediency (time taken to generate response), validity (comparison of recommendations to the American Academy of Otolaryngology Head and Neck Surgery guidelines) and consistency (changes in recommendations based on question variation).
Results:
ChatGPT was more expeditious in generating responses with fewer words, albeit with higher reading scores. When compared with accredited guidelines, there was no difference in validity between these sources (internet sources and ChatGPT). Consistent responses were obtained with question variation.
Conclusion:
ChatGPT may be a valuable source for patients and families in providing valid information comparable to internet materials.
