Digital Physiotherapeutic Shoulder-Specific Training System for Patients with Shoulder Impingement: A Single-Center Noninferiority Randomized Controlled Trial (Preprint)

BACKGROUND Shoulder impingement is a common cause of shoulder pain. Standard management is supervised physiotherapy, but in-person rehabilitation may be limited by travel, cost, and scheduling barriers. Digital health solutions can deliver guided exercises remotely, potentially improving access and adherence. OBJECTIVE To compare a 12-week home-based app-guided shoulder exercise program (DT group) with standard face-to-face physiotherapy (PT group), assessing clinical outcomes and cost-effectiveness. METHODS This prospective, parallel-group RCT enrolled patients with shoulder impingement at a Shanghai orthopaedic center. Participants were randomized to 12 weeks of DT (app-based home exercise with daily sessions and weekly video supervision) or 12 weeks of standard PT (four supervised clinic sessions per week plus home exercises). The primary outcome was shoulder pain at rest on a 0–100 VAS; secondary outcomes included pain on activity (VAS), shoulder function (Constant–Murley score), health-related quality of life (SF-36), patient satisfaction, adherence, and total societal costs. Blinded physiotherapists assessed outcomes at 4, 12, and 24 weeks. We tested non-inferiority with a 1.0-point VAS margin (10 mm) and handled missing data by multiple imputation. Between-group differences over time were evaluated by linear mixed-effects models (with group, time, and interaction), applying Bonferroni correction for multiple comparisons. Cost-effectiveness was evaluated via incremental cost-effectiveness ratios using 1,000 bootstrap replicates. RESULTS A total of 220 patients were randomized. Baseline characteristics were comparable between groups (all P>0.05). At 24 weeks (primary endpoint), mean pain-at-rest VAS had improved by 25.97 in DT vs 23.91 in PT (P=0.812). Pain-on-activity VAS improved by 28.35 vs 24.94 (P=0.558), and Constant–Murley score increased by 25.81 vs 24.38 (P=0.855). SF-36 physical-function scores improved slightly more in DT at 4–12 weeks (P<0.01) but were similar by 24 weeks; other SF-36 domains showed no meaningful differences. Patient satisfaction and adherence were high and similar in both groups (mean sessions/week 5.17 vs 5.15, P=0.848). Mean total cost per patient was lower in DT (¥44,306.69) than PT (¥54,853.65), an average saving of ¥10,546. CONCLUSIONS Home-based digital shoulder rehabilitation produced clinical improvements comparable to conventional physiotherapy. Both groups experienced similar reductions in pain and gains in shoulder function, supporting non-inferiority of the app-based program. Importantly, the digital intervention incurred lower overall costs while achieving similar outcomes, indicating it was more cost-effective than standard care. CLINICALTRIAL The trial was preregistered with the Chinese Clinical Trial Registry (https://www.chictr.org.cn/, registration number: ChiCTR2100042097; registration date: January 13, 2021)