AI Based Telerehabilitation for Patellofemoral Pain Syndrome

Overview

Patellofemoral pain syndrome (PFPS) is a common cause of anterior knee pain in young, active adults and is often aggravated by prolonged sitting, stair use, squatting, and running. Exercise-based physiotherapy is the standard first-line treatment, but access to regular in-person supervision can be limited. This study compares two ways of delivering the same evidence-based exercise programme for PFPS: AI-based telerehabilitation, delivered remotely through the Physitrack platform with AI-assisted movement screening (MediaPipe), and conventional face-to-face physiotherapy. The study aimed to determine whether AI-based telerehabilitation is as effective as, or more effective than, conventional physiotherapy in reducing pain and improving physical function, and to evaluate how usable participants found the telerehabilitation platform

This randomized controlled trial enrolled 48 individuals with PFPS (mean age 29.48 ± 6.40 years, range 19-40; 28 female, 20 male), screened using a structured demographic/SNAPPS questionnaire and an AI based movement screen built on Google's MediaPipe pose-estimation model, which tracked angle of knee. Eligible participants were randomly allocated using the gold-fish-bowl method to the Telerehabilitation Group (n=24; mean age 29.33 ± 6.75, 16 female/8 male) or the Conventional Physiotherapy group (n=24; mean age 29.63 ± 6.16, 12 female/12 male). Both groups followed an identical exercise protocol delivered three times weekly for four weeks (12 sessions of approximately 40-45 minutes). The Telerehabilitation Group received the programme remotely via the Physitrack platform and PhysiApp, including video-guided exercises, an AI-enabled assistant, remote monitoring, and messaging with the treating physiotherapist. The Conventional Physiotherapy group received the identical exercise content face to face in a physiotherapy clinic under direct therapist supervision. Outcomes were assessed at baseline and after 4 weeks: pain (Numeric Pain Rating Scale, NPRS), knee function (Anterior Knee Pain/Kujala Scale, AKPS), knee-related quality of life (KOOS Patellofemoral Subscale, KOOS-PF), knee flexion/extension range of motion and quadriceps (Q) angle via goniometer, and for the telerehabilitation group only platform usability (Telehealth Usability Questionnaire, TUQ). Data were analyzed in SPSS version 22 using paired and independent-samples t-tests or their non-parametric equivalents as appropriate. All 48 randomized participants completed the trial with no dropout.