JMIR Publications

ABSTRACT

Background:

Digital therapeutics for back pain may improve access to structured exercise-based care, but large-scale real-world data on longitudinal outcomes, engagement effects, and non-completion patterns remain limited. ViViRA is a prescription digital therapeutic for non-specific back pain listed as a Digitale Gesundheitsanwendung (DiGA) in Germany.

Objective:

This post-marketing clinical follow-up (PMCF) study investigated pain trajectories, dose–response relationships, baseline predictors of clinically meaningful improvement, and patterns of programme non-completion

Methods:

We performed a retrospective analysis of de-identified real-world data from 21,803 patients who activated ViViRA between 2020 and 2025 and provided at least two pain ratings. Pain intensity was self-reported using the 11-point Numerical Rating Scale (NRS, 0–10). Longitudinal pain trajectories were examined using linear and quadratic mixed-effects models with individual random slopes, operationalised both by weeks since activation and by completed session number. Responders were defined as patients achieving an absolute pain reduction of ≥2 NRS points from onboarding to their best observed score within 90 days. Baseline predictors of clinically meaningful improvement were assessed using binary logistic regression models.

Results:

Pain followed a curvilinear trajectory in both analyses: initial improvement was followed by attenuation after approximately Week 8 or Session 20-34. Clear dose-response relationships were observed for both usage frequency and session adherence. Recommended users (≥ 3 sessions per week) achieved a predicted pain reduction of -1.32 NRS points (- 25.6%) at Week 12; non-users showed near-zero sustained improvement (-0.13 NRS, -2.5%). Patients with 75% session completion achieved -1.91 NRS (-35.2%) by Session 39, compared with -0.34 NRS (-6.3%) at 25% completion. Overall, 49.6% of patients with a baseline pain score ≥ 2 NRS met the MCID criterion. Results were consistent when a relative criterion (≥30% reduction) was applied (46.8%; κ = 0.843). The strongest independent predictors of response were higher initial pain score (OR = 1.33 per NRS point), acute pain duration, female sex, and absence of baseline pain medication and physiotherapy. Only 22.1% of patients completed all 39 sessions. However, more than one third of non-completers (36.6%) had already improved by ≥ 2 NRS points at their last report (18.6% reached a final pain score ≤ 2 NRS), suggesting that a meaningful share of apparent dropout reflects early success rather than disengagement.

Conclusions:

In this large real-world PMCF cohort, ViViRA use was associated with engagement-dependent reductions in back pain, with the greatest improvements occurring during the early treatment phase and approximately half of patients achieving clinically meaningful pain relief. Higher use frequency and greater session completion were consistently associated with larger pain reductions. Although overall completion was low, non-completion did not uniformly indicate treatment failure, as a substantial subgroup of patients discontinued after achieving a clinically meaningful pain reduction. These findings highlight the importance of patient engagement, baseline symptom severity, and differentiated interpretation of dropout when evaluating digital therapeutics for non-specific back pain in routine care.