Can ChatGPT provide personalised responses for frequently asked questions regarding anterior cruciate ligament...

Chat Generative Pre-trained Transformer (ChatGPT) has the potential to enhance healthcare delivery by providing accessible, high-quality medical information. However, no prior studies have evaluated its ability to provide personalised and patient-specific information. Our study aims to evaluate ChatGPT's ability to provide personalised responses to frequently asked questions (FAQs) regarding anterior cruciate ligament (ACL) reconstruction.

We curated a list of 13 clinically relevant FAQs regarding ACL reconstruction and created four distinct patient profiles with varying demographics, comorbidities and activity demands. Each patient profile was entered separately as prompts into the latest version of ChatGPT (GPT-4o). ChatGPT's responses to the FAQs were recorded and assessed independently by two orthopaedic consultants who graded the responses on a scale of 0-2 (0, 1, 2) based on their accuracy and degree that it was personalised to each patient profile.

ChatGPT performed well in providing accurate and personalised responses, achieving a mean overall accuracy score of 1.85 and mean overall personalisation score of 1.90. It was able to provide personalised and patient-specific responses by customizing its recommendations to cater to the situation of each distinct patient - achieving mean personalisation scores of 1.96, 1.88, 1.92 and 1.81 for the four respective patient profiles.

In our study, ChatGPT demonstrated the ability to adapt to distinct patient profiles and provide highly personalised responses regarding ACL reconstruction. Despite potential ethical and regulatory challenges, the ability of AI chatbots to deliver individualized information and counselling is an exciting new frontier in personalised medicine which warrants further research.