AI-supported rehabilitation showed an early benefit in a shoulder-pain trial

A Seville trial involving 46 adults found modest early benefit from RehBody for rotator-cuff-related shoulder pain. First published on 27 July, the paper appears in the October issue.[1]

The rehabilitation task is already well defined. A 2025 international clinical guideline recommends active exercise, including motor-control or resistance work, as an initial treatment for rotator-cuff tendinopathy. It also calls for education adapted to the patient’s circumstances and understanding. The guideline identifies continuing uncertainty over exercise dosage and the value of different supervision arrangements. These are practical questions about delivering treatment: what help reaches someone between appointments, and how can the programme respond when a movement is difficult?[2]

The treatment people actually received

Both groups received four clinic sessions and home exercises. RehBody added camera feedback, different exercises and remote physiotherapist support. Assessors were blinded.[1]

Following the programme, 3 of 23 app-group patients and 10 of 23 controls received further treatment. Two authors worked for the developer. The result concerns the combined programme.[1]

From a progress graph to a decision

For a rehabilitation service, the practical question is how to reproduce the support. A useful specification would identify who reviews exercise records, how often review happens, and how the prescription changes in response. Staff time and patient outcomes could then be evaluated together, making the complete service the unit of comparison.

Report an error or suggest a correction