Smartphone Video-Based AI Gait Analysis Reports to Support Remote Follow-up After Total Knee Arthroplasty

This study evaluates whether an artificial intelligence (AI)-assisted gait report generated from smartphone walking videos can support remote follow-up after total knee arthroplasty (TKA). Patients complete standardized walking video recordings and symptom assessments before surgery and at approximately 6 and 12 weeks after surgery. The system analyzes the videos and provides gait information, including walking speed, step characteristics, gait symmetry, gait phases, knee movement, and changes over time. A separate clinician evaluation examines whether these AI-generated gait reports improve remote clinical decision-making. Clinicians review postoperative cases using either usual remote information, including patient-reported symptoms and the original walking video, or the same information supplemented with the AI-generated gait report. The study evaluates the clinical acceptability of management recommendations, agreement with in-person clinical assessments, clinician confidence, need for additional information, and decision time. The study also evaluates the feasibility, usability, and trust associated with home-based smartphone gait assessment from both patient and clinician perspectives. The AI gait report is intended to support postoperative monitoring and communication and does not replace clinical diagnosis or scheduled in-person follow-up.