Objective:To assess the potential of ChatGPT and Baidu as tools for online health information consultation related to orthopedic conditions by simulating patient searches.Methods:Patients diagnosed with lumbar disc herniation,knee osteoarthritis,congenital scoliosis,and Klippel-Feil syndrome(n=50 per disease)at Peking Union Medical College Hospital and Beijing Puren Hospital were selected to complete an online questionnaire.The responses were categorized by etiology,symptoms,disease progression,treatment,and prognosis.Finally,two questions were selected for each category,resulting in 10 questions per disease and a total of 40 questions.These questions were used for searches on ChatGPT and Baidu,and the results were compared.Results:For the 40 orthopedic-related disease questions,the overall responses from ChatGPT were superior to those from Baidu,with a statistically significant difference(P<0.001).For common orthopedic diseases,ChatGPT performed better than Baidu did(P=0.042),particularly in questions about prognosis(P=0.040).For rare orthopedic diseases,ChatGPT's responses were better than those of Baidu(P<0.001),especially for questions related to etiology(P=0.017),symptoms(P=0.013),and prognosis(P=0.036).The accuracy of Baidu's responses for common diseases was significantly greater than that for rare diseases(P<0.001),whereas there was no statistically significant difference in ChatGPT responses between common and rare diseases(P=0.445).Conclusions:Compared to Baidu,ChatGPT offers more credible answers for orthopedic-related disease questions and can meet the general medical consultation needs of the majority of patients.
关键词
人工智能/大语言模型/腰椎间盘突出症/膝骨关节炎/先天性脊柱侧凸/Klippel-Feil综合征
Key words
Artificial Intelligence/Large Language Model/Lumbar Disc Herniation/Knee Osteoarthritis/Congenital Scoliosis/Klippel-Feil Syndrome
维普
万方数据