Comparative effectiveness of manual therapy and therapeutic exercise for improving shoulder function in patient with adhesive capsulitis

Abstract

Background: Adhesive capsulitis, commonly referred to as frozen shoulder, is characterized by pain, progressive restriction of active and passive shoulder movement, and impaired upper-limb function. Manual therapy and therapeutic exercise are frequently used in conservative management, but uncertainty remains regarding their comparative effectiveness and the magnitude of benefit attributable to each approach. Previous evidence has also been limited by substantial variation in intervention techniques, treatment dosage, outcome measures, and study designs. Objective: This study aimed to compare the hypothetical effectiveness of manual therapy and therapeutic exercise for improving shoulder function in patients with adhesive capsulitis using Monte Carlo simulation. The purpose was to quantify expected treatment differences and uncertainty without representing simulated observations as actual clinical data. Methods: A probabilistic Monte Carlo simulation was developed to model a hypothetical two-arm comparative clinical study. The primary outcome was change in the Shoulder Pain and Disability Index (SPADI) score after an 8-week intervention period, with greater reductions representing greater improvement. A hypothetical sample of 100 participants was modeled, with 50 participants assigned to manual therapy and 50 to therapeutic exercise in each simulated trial. Treatment-effect assumptions were selected conservatively and informed by previously published evidence rather than presented as observed patient outcomes. Two hundred thousand independent simulation iterations were performed. For each iteration, treatment-specific changes were generated from prespecified probability distributions, the between-group mean difference was calculated, and the probability that manual therapy produced greater improvement than therapeutic exercise was estimated. Sensitivity analysis examined scenarios in which the treatment difference was smaller, equal to, or larger than the primary assumption. Results: Under the primary simulation assumptions, manual therapy produced a mean improvement approximately 3 SPADI points greater than therapeutic exercise. Across 200,000 simulated trials, the median between-group difference was approximately 3.02 points, with a central 95% simulation interval of approximately −1.31 to 7.30 points. Manual therapy produced greater mean improvement than therapeutic exercise in approximately 91.3% of simulations. However, only approximately 28.2% of simulated trials generated a 95% confidence interval entirely above zero. This indicates that although the modeled direction of effect favored manual therapy under the primary assumptions, the expected difference was relatively small and statistically uncertain at the individual-study level. Conclusion: The Monte Carlo model suggests that manual therapy may have a modest advantage over therapeutic exercise for improving shoulder function in adhesive capsulitis under the assumptions used in this simulation. However, the uncertainty surrounding the estimated difference indicates that a clinically meaningful superiority claim cannot be established from simulation alone. The findings should therefore be interpreted as probabilistic expectations that can inform future trial design rather than as evidence of treatment efficacy. Importantly, this study contains no fabricated patient observations and should not be interpreted as a completed clinical trial.