Assessing ChatGPT Responses to Frequently Asked Questions Regarding Pediatric Supracondylar Humerus Fractures

Area of Science:

• Artificial intelligence in healthcare
• Pediatric orthopedics
• Medical information access

Background:

• Patients increasingly use the internet for medical information.
• AI chatbots like ChatGPT are emerging as potential health information resources.
• Assessing AI's accuracy for specific pediatric conditions is crucial.

Purpose of the Study:

• To evaluate ChatGPT's accuracy in answering frequently asked questions (FAQs) about pediatric supracondylar humerus (SCH) fractures.
• To determine if ChatGPT can serve as a reliable source for pediatric SCH fracture information.

Main Methods:

• Seven FAQs on pediatric SCH fractures were posed to ChatGPT.
• Responses were categorized based on the level of clarification needed (excellent, satisfactory, unsatisfactory).
• Expert opinion (Level V) was used to rate ChatGPT's responses.

Main Results:

• None of ChatGPT's responses were excellent, requiring no clarification.
• Four responses were rated as satisfactory (minimal to moderate clarification needed).
• Three responses were unsatisfactory, requiring substantial clarification or containing inaccuracies.

Conclusions:

• ChatGPT provided partially satisfactory information on pediatric SCH fractures but was unreliable regarding treatment.
• Significant clarification was needed on treatment algorithms, casting, return to sport, and physical therapy.
• ChatGPT is not a dependable source for treating pediatric SCH fractures; physician consultation remains essential.