Which self-reported bio-psychosocial variables predict recovery in rotator cuff tendinopathy? An international prospective cohort study

• Published in: Physiotherapy theory and practice pp. 1 - 14
• Main Authors: Delen, Mehmet, Tayfur, Abdulhamit, Gulle, Halime, Birn-Jeffery, Aleksandra, Morrissey, Dylan
• Format: Journal Article
• Language: English
• Published: England 16.05.2025
• Subjects: recovery; rotator cuff; Outcome predictor; physical activity; tendinopathy
• ISSN: 0959-3985; 1532-5040; 1532-5040
• DOI: 10.1080/09593985.2025.2502509

It is unclear what baseline factors predict recovery from rotator cuff (RC) tendinopathy, making evidence-based prognostic estimates difficult. The study aimed to identify predictors of RC tendinopathy recovery to guide clinical decision-making. People with RC tendinopathy were recruited to an international prospective cohort study. Participants completed a detailed baseline survey in which 41 plausible bio-psychosocial and demographic factors were measured. Recovery was established via the Global Rating of Change scale at monthly follow-ups for a year. Univariate and multivariate Cox proportional-hazards regression analyses were conducted to identify predictors and to build models. Model performance (internal validity) was evaluated using bootstrapping. Seventy-three people with RC tendinopathy (43.9 ± 14.0 years; Shoulder Pain and Disability Index = 37.7 ± 24.4; 45 females) were recruited, providing 15,284 days at risk (208 ± 129 days). The recovery rate was 47%. Participants typically recovered around month seven. Being moderately active (hazard ratio (HR) = 2.23) and having higher health status (HR = 1.03) were associated with recovery. Demographic and shoulder severity variables did not improve the model performance nor individually predict recovery. The final model partially predicted recovery with near acceptable performance (optimism-corrected Harrell's C discrimination = 0.66 and Calibration Slope = 0.99). The inadequate recovery rate in RC tendinopathy (under 50%) indicates a necessity for enhancements in treatment strategies for this condition. Demographic variables and shoulder pain severity are not useful factors to predict RC tendinopathy recovery. However, clinicians may consider assessing baseline health status and activity level to guide decision-making as these variables partially predict recovery. Strategies to include these factors in intervention development should be considered.