MedElevator Войти в исследования
← Поиск публикаций

DOI 10.5435/jaaos-d-25-00260

Медицинская библиотека · The Journal of the American Academy of Orthopaedic Surgeons · 2025

Comparison of Responses from ChatGPT-4, Google Gemini, and Google Search to Common Patient Questions About Ankle Sprains: A Readability Analysis.

Chavda Hetsinhji, Sontam Tarun R, Skinner William C, Ingall Eitan M, Zide Jacob R

СтатьяПеревод в очереди

Аннотация

INTRODUCTION

Ankle sprains are among the most common injuries treated by foot and ankle specialists. Despite increasing online health information-seeking behavior, existing information on ankle sprains has been criticized by physicians for poor quality and readability. ChatGPT-4 and Google Gemini are artificial intelligence chatbots known for their accessible interfaces and versatility, making them popular for patient self-education. This study evaluated the accuracy and readability of answers to frequently asked questions (FAQs) on ankle sprains from ChatGPT-4, Google Gemini, and Google Search.

METHODS

ChatGPT-4, Gemini, and Google Search each generated 10 FAQs and answers about ankle sprains, with questions classified based on the Rothwell Classification System and answers graded using the Flesch-Kincaid Grade Level, Flesch-Kincaid Reading Ease Score, and Gunning Fog Score. ChatGPT-4 and Gemini were then asked to answer the 10 questions generated by Google Search. Two fellowship-trained foot and ankle surgeons and a current orthopaedic foot and ankle fellow independently graded responses for accuracy and comprehensiveness using a 4-point scale.

RESULTS

Only 10% of FAQs generated by ChatGPT-4, Gemini, and Google Search overlapped. When answering the same 10 FAQs from Google Search, ChatGPT-4 and Gemini provided significantly longer responses than Google Search, with mean differences of 353.1 ± 42.8 words ( P < 0.0001) and 87 ± 35 words ( P < 0.0001), respectively. ChatGPT-4's answers were more difficult to read based on the Flesch-Kincaid Grade Level (3.2 ± 1.7, P = 0.0012) and Flesch-Kincaid Reading Ease Score (22.2 ± 9.3, P < 0.0001). Surgeons ranked ChatGPT-4's answers as the most comprehensive and accurate and Google Search's answers as the least.

CONCLUSION

This study suggests that ChatGPT-4 and Google Gemini provide longer, more complex, yet more accurate and comprehensive answers than Google Search, highlighting their potential as patient education tools for orthopaedic conditions such as ankle sprains.

Для цитирования: Chavda Hetsinhji, Sontam Tarun R, Skinner William C, Ingall Eitan M, Zide Jacob R. Comparison of Responses from ChatGPT-4, Google Gemini, and Google Search to Common Patient Questions About Ankle Sprains: A Readability Analysis.. The Journal of the American Academy of Orthopaedic Surgeons. 2025. DOI: 10.5435/jaaos-d-25-00260.

Предметные рубрики MeSH

Подготовка и перевод публикации
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…

История · было → стало

Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.

Тип публикации указан по данным источника. Уровень доказательности автоматически не присваивается.