Pilot Home Rehabilitation Trial Finds No Clear Clinical Edge in Adhesive Capsulitis
Adhesive capsulitis is a slow-moving disorder, but the pain and stiffness it creates can dominate daily function for months. In this pilot randomized study, a low-supervision home rehabilitation system paired with asynchronous smartphone monitoring failed to produce a detectable advantage over standard home exercise instructions when added to conventional outpatient rehabilitation. The intervention was designed to make home therapy more visible and more engaging, yet the clinical signal never separated cleanly from control.
All 30 adults in the study completed the protocol, and no adverse events were reported. That makes the safety picture straightforward, but the efficacy picture remains far less generous. The trial’s exploratory structure also matters: the outcome hierarchy was assigned only after data collection and analysis had already been completed, so every result has to be read as hypothesis-generating rather than confirmatory.
Baseline Imbalance Narrowed the Interpretability of the Signal
Participants were randomized 1:1 at a university-affiliated outpatient rehabilitation department. Both groups received conventional outpatient rehabilitation twice weekly for 4 weeks, but the intervention arm also used a gamified smartphone-based home program called HappyGoGo with asynchronous system monitoring. The control arm received standard home exercise instructions.
The two groups did not begin from the same clinical starting point. Baseline imbalance favored neither interpretation nor confidence: the intervention group was more restricted in range of motion and reported higher pain before treatment began. In a rehabilitation trial built around change from baseline, that kind of asymmetry can obscure a true treatment effect or make an apparently favorable one look larger than it is. It also raises the possibility that the intervention arm entered the study with more severe disease, which would make equal end-of-study outcomes harder to achieve even if the digital program had some benefit.
Fourteen clinical outcomes were analyzed with analysis of covariance, adjusting for each corresponding baseline value and using Benjamini-Hochberg correction for multiplicity. A sensitivity analysis also examined unadjusted change scores. Across the adjusted analyses, no outcome reached nominal significance, and every 95% confidence interval crossed the null. The statistical frame was therefore consistent: there was no reliable between-group separation on the measured endpoints.
Pain Improved Numerically, but the Estimate Stayed Inconclusive
Pain provided the strongest numerical hint of benefit. The adjusted mean difference favored the intervention by 1.30 points, with a 95% confidence interval from -2.64 to 0.04 and a p value of 0.057. The estimate moved in the expected direction, but it stopped just short of conventional significance and remained compatible with no effect at all.
In the unadjusted analysis, pain and internal rotation reached nominal significance, but neither survived correction for multiple comparisons. That sequence is important in a small pilot trial with 14 outcomes. When many endpoints are examined in parallel, scattered nominal positives can appear by chance, especially when the sample is small and baseline imbalance is present. In practical terms, the data do not support a claim that the asynchronous home system improved pain, motion, or function beyond standard care.
The clinical rationale for the intervention still makes sense on paper. Adhesive capsulitis often requires repeated stretching, mobility work, and sustained participation at home, and digital monitoring can theoretically improve adherence by reducing the invisibility of unsupervised exercise. Gamification may reinforce repetition and provide feedback loops that keep patients engaged long enough for soft tissue remodeling and capsular stretching to accumulate. In this study, however, the biologic plausibility did not translate into a measurable outcome advantage over 4 weeks.
Engagement Was Trackable, but It Varied Widely
The system did succeed at one of its most practical tasks: documenting participation. Engagement data were retrieved for 14 of 15 intervention participants, with a mean engagement rate of 61.0% and a range from 7.1% to 96.4%. That spread is clinically revealing. Some patients used the platform frequently, while others barely interacted with it, and such variation can blunt any average treatment effect in a small cohort.
Wide engagement dispersion is common in home-based rehabilitation, where symptom burden, motivation, technological comfort, and daily routine all shape adherence. A digital program can only influence recovery if it is actually used. Here, the asynchronous system proved feasible from a tracking standpoint, but feasibility does not guarantee therapeutic potency. A platform that captures behavior accurately can still fail to move pain or mobility if participation remains inconsistent or if the exercise dose is insufficient for a condition as stubborn as adhesive capsulitis.
What the Pilot Can and Cannot Support
This study does not establish that low-supervision asynchronous home rehabilitation should replace or outperform standard instructions. It does show that the system could be integrated alongside routine outpatient care without reported safety problems and with automated documentation of use. Those are useful operational findings, especially for clinics trying to measure home adherence rather than infer it indirectly.
Still, the confidence intervals were too wide to determine whether a clinically meaningful effect exists. Small sample size, short follow-up, and the post hoc assignment of outcome hierarchy all limit certainty. The baseline imbalance further clouds interpretation because the intervention arm started worse off on pain and range of motion. When the starting lines differ, the finish line becomes harder to read.
For clinicians, the most defensible reading is restrained. A gamified, asynchronously monitored home program for adhesive capsulitis can be delivered safely and can record participation, but this pilot did not demonstrate a measurable clinical advantage over conventional home exercise advice during a 4-week rehabilitation window. The next question is not whether digital monitoring can exist in this setting; it already can. The more important issue is whether a larger, prospectively planned trial with balanced baseline severity and a longer treatment horizon can prove that better monitoring translates into better shoulder recovery.