Anxiety, depression, and fatigue commonly affect adults living with chronic medical conditions and contribute to reduced functioning and quality of life. Digital health interventions can potentially address barriers to access related to geography, mobility, time, and clinician shortages. Prior trials were often single-condition, underpowered, and rarely combined psychological and physical components; evidence on whether human support enhances effects is limited. The eMPower trial aimed to evaluate whether a multicomponent digital intervention reduces anxiety and depression in adults with diverse chronic conditions and whether adding brief human support provides additional benefit.
This was a three-arm, parallel-group, open-label randomized controlled trial conducted fully remotely between February 12, 2023 and December 1, 2024. Recruitment and intervention delivery were online across 13 countries. The trial was registered on ClinicalTrials.gov (NCT05786482) and followed CONSORT reporting standards. The analytic plan preserved prespecified primary and secondary endpoints; all analyses were completed after study end with no interim analyses.
Eligible participants were adults with a self-reported physician-diagnosed chronic physical condition of at least 3 months’ duration, English proficiency, and internet access. At trial start the minimum age was ≥50; this was lowered to ≥18 on May 19, 2023 to improve inclusivity. Exclusion criteria initially included uncontrolled psychiatric conditions, near-daily suicidality, and HADS-Depression >10; the HADS-Depression >10 exclusion was removed on October 21, 2023 to broaden access while retaining exclusions for uncontrolled psychiatric illness and near-daily suicidality. Consent and baseline assessments were completed electronically via REDCap. For a subset (residents of Alberta, Canada providing personal health numbers) diagnoses were verified via provincial records; other reports were reviewed by a study physician using medication lists and available clinical information.
Participants (N = 825) were randomized 1:1:1 using computer-generated permuted blocks within predefined chronic condition strata in REDCap. Allocation was concealed until assignment. Participant blinding was not feasible due to the nature of the interventions. Randomized groups included waitlist control (n = 274), self-directed eMPower (n = 275), and eMPower + human support (n = 276).
The eMPower program was co-designed with patients and informed by behavioral theory (COM-B) and 22 behavior change techniques to support adherence. It combined:
Participants received weekly newsletters highlighting program content and could opt into weekly guided group movement sessions. All content and communications were delivered in English.
In the eMPower + human support arm, participants received weekly telephone check-ins up to 15 minutes in duration from trained nonclinical personnel who completed a 2-day pre-study training in motivational interviewing. Check-ins followed a semi-structured script adapted from prior trials. Both intervention arms received identical digital content and newsletters; human support consisted only of the brief weekly calls.
Control participants were assigned to a waitlist control and continued usual medical care. They received weekly emails with motivational wellness quotes and were offered access to the self-directed program after the 12-week study period. No restrictions were placed on concomitant treatments.
The prespecified primary outcome was change in the Hospital Anxiety and Depression Scale (HADS) total score from baseline to 12 weeks, analyzed via ANCOVA adjusted for baseline HADS total score, chronic condition type, age, and sex. Secondary outcomes included HADS anxiety and depression subscales, fatigue measured by the Modified Fatigue Impact Scale (MFIS), health-related quality of life measured with the 12-Item Short Form Survey (SF-12 mental and physical component summary scores), and the EQ-5D-5L index score (country-specific value sets). The EQ-5D-5L visual analog scale was collected descriptively but was not part of the prespecified confirmatory secondary endpoints. Preplanned exploratory analyses included combined-intervention versus control comparisons and direct comparisons between the two intervention arms. Process measures and additional exploratory outcomes were registered for separate future reports.
The trial sample size was powered for the single prespecified primary comparison: eMPower + human support versus control on the HADS total score at 12 weeks. Analyses followed an intention-to-treat principle. ANCOVA models adjusted for baseline HADS total score, chronic condition type, age, and sex. Details on multiplicity corrections and handling of missing data are reported in the protocol and statistical analysis plan; primary outcome data availability and participant flow are reported below.
Overall, 695 participants (84.2%) completed 12-week assessments. Primary outcome data were available for 222 participants in the eMPower + human support arm, 214 in the self-directed eMPower arm, and 259 in the control arm. In the prespecified primary comparison, the eMPower + human support arm demonstrated a statistically significant improvement in HADS total score versus control with a mean difference of 2.9 points (95% CI 2.0 to 3.8; p < 0.001). In prespecified exploratory analyses, the self-directed eMPower arm also significantly improved HADS total score compared with control (mean difference 2.6 points; 95% CI 1.8 to 3.5; p < 0.001). No significant differences were observed between the two intervention arms for the primary outcome or any secondary outcomes (all p > 0.05).
Both intervention arms were associated with improvements across prespecified secondary outcomes compared with control, including HADS anxiety and depression subscales, MFIS fatigue scores, SF-12 mental and physical component scores, and EQ-5D-5L index scores. The report indicates that secondary outcomes remained statistically significant after multiplicity corrections in the analyses presented.
Participants were encouraged to report intervention-related adverse events. No intervention-related adverse events were reported in any trial arm during the 12-week study period.
The authors identified several limitations: the use of a waitlist control design that does not control for nonspecific intervention effects; reliance on self-reported diagnoses for most participants; a 12-week follow-up window limiting insight into durability of effects; and a predominantly female, highly educated sample that may reduce generalizability. The trial was registered retrospectively ~25 days after the first participant was enrolled because of administrative delays; the authors note no changes to design, outcomes, or analysis plans occurred before registration or before primary outcome data collection. Additional registered process-oriented and exploratory outcomes are planned for separate publications.
A multicomponent digital intervention delivered across diverse chronic conditions significantly reduced symptoms of anxiety and depression at 12 weeks compared with usual care when delivered with brief human support. Exploratory findings showing comparable effects for the self-directed format suggest the potential for scalable, low-resource, “digital first” implementation. The authors recommend longer-term follow-up and cost-effectiveness analyses to inform broader adoption and implementation decisions.
The trial is registered at ClinicalTrials.gov (NCT05786482). Summary statistics and values used to build graphs are reported within the manuscript and supplementary files. Individual participant-level data cannot be publicly shared without approval from the University of Alberta Health Research Ethics Board; de-identified data may be available to qualified researchers subject to ethics approval and a data sharing agreement. Funding sources included the Canadian Institutes of Health Research, Mitacs through partner organizations, and TRIANGLE; funders had no role in study design, data collection, analysis, decision to publish, or manuscript preparation.