Evaluation of CHAT GPT-4 Artificial Intelligence Responses to Common Problems in Patients after Hip Replacement with Expert Opinion

Abstract

Aim: In this study, it is aimed to evaluate the CHAT GPT-4 artificial intelligence responses to common problems in patients after hip replacement with expert opinion.
 
 Materials and Methods: In the study, an 11-question form created by the researcher by scanning the literature for the information needs of patients after hip replacement surgery was used. Ten orthopedic specialists evaluated the accuracy of the answers generated by Chat GPT-4 with a 5-point Likert-type score scale (5: Strongly agree, 1: Strongly disagree).The maximum score that can be obtained from the score scale is 55, and the minimum score is 11. Scores are summarized with descriptive statistics.
 
 Results: All of the experts stated that artificial intelligence correctly answered the signs of wound infection (4.80±0.42), practices aimed at preventing home accidents (4.50±0.52), and in which cases a doctor should be consulted (4.50±0.52). 50% of the experts, the antithrombotic treatment time of artificial intelligence (3.80±1.39), 60% of the time, will my prosthesis come off? He stated that he answered the question (3.20±0.91) partially true/false. The total mean score obtained from the score scale was found to be 44.40±3.83.
 
 Conclusion: When Chat GPT was evaluated by experts, it was determined that it gave mostly accurate information. However, it is thought that it may be more limited in individualized treatment and care practices. More studies are needed to confirm the accuracy of the results.