Artificial intelligence fails to outperform orthopaedic surgeons: A systematic review

Area of Science:

• Orthopaedic Surgery
• Artificial Intelligence
• Clinical Decision Support

Background:

• Orthopaedic surgery faces challenges like high patient volumes and long waiting lists.
• Artificial intelligence (AI) is increasingly used for data analysis, patient triage, and imaging interpretation in orthopaedics.
• AI implementation requires demonstrated performance comparable to human surgeons.

Purpose of the Study:

• To systematically review and evaluate the performance of AI relative to surgeons in orthopaedic practice.
• To determine AI's value as a complementary tool in orthopaedic surgery.

Main Methods:

• Systematic review of studies using OVID Medline up to August 13, 2025.
• Categorization of included studies into decision-making, management plans, clinical knowledge, quality control, and patient FAQs.
• Comparative analysis of AI performance against surgeons and residents.

Main Results:

• AI demonstrated high sensitivity (97%) but lower specificity (33%) and accuracy (65%) than surgeons in identifying patient improvements.
• AI performed comparably or superiorly to surgeons in emergency scenarios and answering patient FAQs (scoring higher in empathy, accuracy, completeness, and quality).
• AI showed limited accuracy in knee osteoarthritis radiographic staging (35% vs. >80% for surgeons), while residents outperformed AI in examinations (74.2% vs. 47.2%).

Conclusions:

• AI has the potential to enhance efficiency and patient communication in orthopaedics.
• Concerns regarding AI bias, quality risks, overconfidence, and outdated information limit its ability to replace human expertise.
• Clinician-led design and validation are crucial for the safe and effective integration of AI in clinical practice.