Hünkar Cagdas Bayrak, Bekir Karagöz, Özlem Bayrak
Acta Orthopaedica et Traumatologica Turcica - 2025;59(6):415-420
Septic arthritis is a serious joint infection requiring urgent intervention. This study comparatively evaluated the performance of three large language model-based chatbots (ChatGPT, Claude, and Perplexity) in responding to septic arthritis scenarios. The evaluation was based on 24 scenario-based clinical questions, and each chatbot's response was independently assessed by two senior experts. The results showed that all three chatbots demonstrated high performance but had different strengths and weaknesses. ChatGPT and Claude provided more comprehensive and detailed responses, while Perplexity offered more concise and reference-supported answers. These findings suggest that the selection of chatbots should be based on the intended use, prioritizing clarity and practicality in clinical settings and evidence-backed detail in research contexts.
