A 4-week smartphone app using cognitive behavioral therapy and mood tracking reduced depression, anxiety, and stress scores by 14.5 points in working adults, compared to a 3.3-point increase in controls. The effect was robust , though engagement didn't predict individual outcomes.
Work stress carries documented mental health costs. Smartphone-based interventions offer theoretical scalability, but real-world evidence in occupational settings remains limited. This randomized controlled trial tested Neurum, an app combining ecological momentary assessment (real-time mood tracking) with brief CBT exercises and mindfulness content, across 4 weeks in 201 working adults.
The results showed a clear directional effect. Participants in the intervention group saw DASS-21 scores (a standard 21-item scale measuring depression, anxiety, and stress) drop by an average of 14.5 points over 4 weeks. The control group moved in the opposite direction, increasing 3.3 points on the same scale. The between-group difference was statistically significant (P<0.001, η² = 0.252), meaning the app accounted for roughly 25% of variance in outcome. Stress reduction drove most of the effect: the intervention group improved 6.7 points on the stress subscale while controls worsened by 2 points.
However, engagement data introduced nuance. Participants completed an average of only 6 exercises and logged 9.7 mood entries over the full 4 weeks, spending about 5 minutes per day in the app. Critically, the analysis found no significant relationship between frequency of app use and magnitude of symptom improvement. This suggests either that brief, consistent engagement suffices, or that unmeasured factors (expectancy, app design, timing of use relative to stress events) matter more than volume.
Attrition posed a notable limitation. Twenty-one participants dropped from the intervention group versus only 2 from control, yielding a final sample of 178 (21 fewer than enrolled). The study did not report reasons for dropout or analyze whether people who continued differed systematically from those who left. This 21% loss in the active arm is higher than typical for short digital interventions and suggests the app may not suit all users despite its stated ease of access.
The study establishes proof-of-concept that structured digital mental health tools can reduce workplace stress symptoms in a non-clinical population over a short timeframe. For individuals already working with stress symptoms, or seeking supplementary support alongside other practices like journaling, breathwork, or social connection, app-based CBT represents a scalable option.
The lack of dose-response relationship is worth noting: more app engagement did not predict better outcomes. This cuts both ways. It suggests people don't need to use the tool exhaustively for benefit, lowering the barrier to adoption. It also means your engagement patterns might not track neatly with symptom improvement, so consistency and timing may matter more than raw frequency.
The high dropout rate warrants caution about generalizability. If you're someone who finds app-based interventions engaging, this study offers encouragement. If you've abandoned similar tools before, that's a valid signal. The 4-week timeframe also limits inference about durability: we don't know whether symptom improvement persisted after the app was removed or whether ongoing use was necessary.
| Parameter | Details |
|---|---|
| Study Design | Randomized controlled trial |
| Sample Size | 201 enrolled; 178 analyzed (85 male, 93 female; mean age 34.6 years) |
| Intervention | Neurum smartphone app: mood journaling, cognitive behavioral therapy exercises, mindfulness, self-regulation activities; 4 weeks |
| Control | No intervention; waitlist |
| Primary Outcome | DASS-21 (Depression, Anxiety, Stress Scale-21); Cronbach α = 0.87 |
| Key Findings | Intervention group: -14.5 points (95% CI inferred from F-statistic); Control group: +3.3 points; F(1,176) = 59.36, P<0.001, η² = 0.252 |
| Secondary Findings | Stress subscale: Intervention -6.7 vs Control +2.0; F(1,176) = 64.68, P<0.001. Mean engagement: 6.27 exercises, 9.74 mood logs, 4.95 min/day. No significant association between engagement and outcome (P value not reported). |
| Attrition | 21 dropouts (intervention), 2 dropouts (control); χ² = 18.26, P<0.001 |
| Setting | Working adults, non-clinical sample |
| Journal | JMIR mHealth and uHealth |
| PubMed ID | 42710031 |
ProtocolEngine provides general health information based on published research. This is not medical advice. Consult a healthcare professional before starting any supplement or health protocol.