计算机科学
冲程(发动机)
人工智能
虚拟现实
物理医学与康复
人机交互
钥匙(锁)
组分(热力学)
医学
集合(抽象数据类型)
作者
Fatma Özcan,Merve ÖRÜCÜ ATAR,Yasin Demir,Eda Gürçay
标识
DOI:10.31609/jpmrs.2025-109562
摘要
ABS TRACT Objective: Nowadays, artificial intelligence chatbots are increasingly used by patients to obtain information about the diagnosis, of their diseases.This study aimed to determine the reliability and usability of Chat Generative Pre-trained Transformer (ChatGPT) in stroke-related questions.Material and Methods: A total of 39 questions were prepared according to 3 keywords ("general information", "complications" and "rehabilitation"), which were identified as the most frequently searched keywords in Google Trends, and were evaluated simultaneously by 2 raters according on a 7-point Likert scale for reliability and usability.Results: Inter-rater Cronbach α scores indicated almost perfect to substantial agreement for both reliability and usability scores (α between 0.813-0.949,and α between 0.303-0.857,respectively).The highest mean reliability score was for "general information" (mean 5.2).The lowest average was for the "rehabilitation" section (mean 4.1).The "complications" for rater 1 and "general information" for rater 2 had the highest mean scores for the usability (mean 5.4), and the lowest mean value was recorded in the "rehabilitation" section (mean 4.6, 4.2, respectively).Conclusion: ChatGPT's responses to the stroke-related questions were reliable and useful.However, it should be kept in mind that ChatGPT may provide incorrect and incomplete information, especially in the "rehabilitation" section, which may lead to significant deficiencies in disease management.
科研通智能强力驱动
Strongly Powered by AbleSci AI