Evaluating the Performance of ChatGPT-5 in Answering Common Patient Questions on Developmental Dysplasia of the Hip and Periacetabular Osteotomy
This study demonstrates that ChatGPT-5 provides highly relevant, accurate, readable, and clinically safe responses to common patient questions regarding developmental dysplasia of the hip and periacetabular osteotomy, positioning it as a valuable adjunctive tool for orthopedic patient education despite some variability in completeness.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
When a person's hip joint does not form correctly, a condition known as developmental dysplasia of the hip, the ball of the thigh bone does not sit securely in its socket. This misalignment can cause pain, instability, and early wear and tear on the joint. To fix this, surgeons often perform a procedure called a periacetabular osteotomy, where they carefully cut and reposition the bone around the socket to create a better fit. For patients facing these complex medical situations, understanding what to expect is crucial. In the past, people turned to the internet for answers, but online medical information has often been inconsistent, difficult to read, or simply wrong. Recently, a new type of computer program capable of holding conversations and generating text has emerged as a potential source of medical information. These programs, known as large language models, can answer questions in plain language, but it has remained unclear whether they can provide safe, accurate, and complete advice for specific, serious orthopedic conditions.
A team of researchers set out to test the latest version of one of these conversation programs, specifically asking it to answer twenty-five common questions that patients might have about hip dysplasia and the corrective surgery. The researchers did not ask the computer to answer pre-selected questions; instead, they first asked the program to generate the list of twenty-five most common questions itself, covering symptoms, diagnosis, the surgery, risks, and recovery. Once the list was created, they asked the program to provide clear, patient-friendly answers to each one. Two expert hip surgeons then read every single question and answer pair. They judged the responses based on several factors: whether the answer actually addressed the question, if the medical facts were correct, how easy the language was to understand, whether the answer included enough detail to be useful, and if the advice was safe to follow. They also measured how easy the text was to read using standard tools that analyze sentence length and word complexity.
The results showed that the computer program performed remarkably well in almost every category. The answers were consistently relevant, meaning they directly addressed what the patient was asking. The language used was exceptionally clear and easy to understand, scoring at the highest possible level for readability. The surgeons found the information to be highly accurate and clinically safe, with no responses that were misleading or dangerous. The overall quality of the answers was very high, with an average score near the top of the scale. However, the researchers did notice one area where the computer struggled slightly: completeness. While the answers were correct, they sometimes lacked specific numbers, detailed explanations, or the kind of personalized context that a doctor would provide during a real consultation. For instance, the computer might explain that recovery takes several months but might not specify the exact timeline for returning to different types of work or sports without further prompting.
This study suggests that the newest generation of these conversation tools is a significant step forward in making medical information accessible. Unlike earlier versions of similar programs, which sometimes produced text that was too complex for the average person to read, this latest model successfully translated complex surgical concepts into language that a patient could easily grasp. The experts concluded that while the computer should never replace a conversation with a real doctor, it can serve as a valuable helper. It can provide a solid, safe, and easy-to-read foundation of information that helps patients prepare for their appointments and understand their condition better. The findings indicate that as these tools continue to improve, they may become a standard part of helping patients navigate difficult medical decisions, provided they are used to support, rather than replace, the judgment of a medical professional.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.