Digital health technologies are reshaping home-based physical therapy. A recent clinical study led by researchers at Yeouido St. Mary’s Hospital demonstrates that utilizing an AI-driven rehabilitation application for adhesive capsulitis (commonly known as a frozen shoulder) yields superior clinical outcomes compared to traditional video-watching routines.

The research was spearheaded by Professor Chang So-yeon of the Department of Rehabilitation Medicine and Professor Yoon Eun-ji of the Department of Orthopedic Surgery. To evaluate the app's efficacy, the team enrolled 18 patients diagnosed with frozen shoulders, dividing them equally into an AI application cohort and a YouTube video exercise cohort.

All participants initially received a single ultrasound-guided steroid injection into the shoulder joint before embarking on a 12-week home rehabilitation regimen. Blinded evaluations were systematically carried out at baseline and at weeks 2, 4, 8, and 12.

At the conclusion of the 12-week study period, data revealed that patients utilizing the AI application experienced a statistically notable enhancement in side external rotation—the movement of rotating the arm outward while keeping the elbow close to the body—outperforming the control group that relied solely on video instruction.

Furthermore, measurements tracking health-related quality of life via the EQ-5D-5L index indicated a more substantial recovery trajectory among app users.

The software central to the trial, branded as "ANAPA ME," was engineered by Orthocare, a company headed by Yeouido St. Mary's orthopedic surgeon Professor Kim Jong-ho. Powered by a CNN-MLP framework, the software leverages standard smartphone camera lenses to track 32 distinct anatomical landmarks in real time. This allows for automated repetition counting and precise posture adjustments without requiring extra wearable hardware.

The platform also bridges the gap between patient and clinician through an integrated web dashboard, enabling medical professionals to track compliance remotely and tailor exercise regimens. Recently, the technology secured regulatory clearance from South Korea's Ministry of Food and Drug Safety to proceed with broader multi-center trials.

"Our findings validate the viability of smartphone-based rehabilitation tools that deliver instant feedback without cumbersome sensors," noted Professor Chang. She added that subsequent largescale, multi-center research will focus on assessing long-term health outcomes and economic efficiency.

The study's comprehensive findings have been featured in the October 2026 edition of the prominent digital health journal npj Digital Medicine.

[Copyright (c) Global Economic Times. All Rights Reserved.]