Agreement Between Video-Based and In-Person Assessment in Patients with Knee Pain—A Prospective Repeated-Measures Pragmatic Study

Background: Digital health has accelerated telehealth uptake, yet evidence comparing video-based musculoskeletal assessment with traditional in-person examination is limited. This study evaluated the concurrent validity and interrater reliability of video-based physiotherapy assessment versus face-to-face assessment in patients with knee pain. Methods: Patients with knee pain underwent randomized consecutive in-person and video-based assessments by experienced musculoskeletal physiotherapists. Clinical diagnoses were categorized into seven groups (red flag, yellow flag, arthrogenic, tendinopathy, patellofemoral pain, muscle sprain, neurogenic). Primary outcomes were intermethod agreement and Cohen’s kappa; sensitivity, specificity, PPV, NPV, and interrater reliability for video assessments were also reported. Results: Forty-five participants (mean age 38 ± 6.5 years; 55.6% female) completed the study. In-person and video-based assessments produced identical diagnoses in 43/45 cases (Cohen’s κ = 0.92, p < 0.001). Telehealth accuracy was high across all diagnostic categories (90–100%). Interrater agreement between video-based assessors was 93.3% (κ = 0.89, p < 0.001). Agreement between assessments was moderately associated with KOOS (r = 0.312, p = 0.037). Conclusions: In this selected pragmatic sample, video-based physiotherapy assessment demonstrated high concurrent agreement and excellent interrater reliability with face-to-face assessment. Given the study’s sample size, repeated-measures design, and lack of an independent reference standard, these results indicate feasibility and intermethod agreement rather than diagnostic equivalence. Video assessment may be a feasible option for triage and management in selected settings, but further research in larger, more diverse populations and evaluation against independent reference standards is required.