High engagement and differential completion of emotion-focused exercises in a digital wellbeing application for bipolar disorder patients: real-world observational study

Introduction

Bipolar disorder affects approximately 1–2% of the European population and is characterized by recurrent mood episodes despite available pharmacological treatments. Digital health interventions may complement standard care by supporting daily self-monitoring and delivering structured therapeutic exercises. However, determinants of user engagement and exercise completion in real-world digital tools remain poorly understood.

RESEARCH QUESTION

Which usage patterns and intervention components are associated with high engagement and completion rates in a digital wellbeing application for individuals with bipolar disorder?

Aims

Primary aim: To assess engagement and usage patterns of the digital wellbeing application Leopa in a real-world cohort of individuals with bipolar disorder using a 30-day rolling retention metric.

Secondary aim: To characterize engagement with and completion rates of in-app therapeutic exercises across a wide range of selected protocols.

Methods

Design

Single-arm observational pilot study.

Participants

Adults with self-reported bipolar disorder recruited via patient associations (N=57), with ≥30 days of follow-up available.

Data collection: Passive behavioral telemetry, including number of sessions, daily activity, feature-level usage, and completion rates across nine therapeutic exercises.

Intervention content: Exercises derived from established therapeutic frameworks (CBT/Beck, Barlow protocols, and Acceptance and Commitment Therapy), adapted for mobile delivery.

Statistical analysis: Descriptive statistics with Wilson-derived 95% confidence intervals for key proportions; sensitivity analyses applied to retention definitions.

Results

Primary outcome — Engagement (rolling retention)

The application demonstrated sustained engagement over 30 days:

Day 1: 94.7% (54/57)
Day 7: 86.0% (49/57)
Day 14: 77.2% (44/57)
Day 21: 71.9% (41/57)
Day 30: 70.2% (40/57)

(95% confidence intervals available)

Secondary outcome — Therapeutic exercise use

A total of 340 exercise protocols were initiated across nine exercise types, with 77 completed (overall completion rate: 22.6%).

Completion rates varied by exercise type:

Emotion acceptance : 41.9% (18/43)
Emotion exploration : 34.4% (11/32)
Thought analysis : 30.6% (15/49)
Beck-based cognitive exercises were most frequently initiated (n=104) but showed lower completion (14.4%)

Overall, emotion-focused exercises demonstrated higher completion rates than cognitive exercises.

Limitations

Limitations include small sample size, self-reported diagnosis, single-arm design, and recruitment through a self-selected volunteer cohort, limiting generalizability.

Conclusions

Real-world data suggest that digital interventions incorporating mood monitoring and emotion-focused exercises can achieve high engagement in individuals with bipolar disorder. A 30-day retention rate of 70.2% compares favorably with published benchmarks for self-guided mental health applications.

Differences in completion rates across exercise types highlight the importance of intervention design and user experience. Emotion-focused approaches may be particularly suited for mobile delivery in this population.

Collecting qualitative and quantitative feedback from users will bring insights to further develop and customize the solution by providing the right features to users.

These findings support further development and evaluation of digital therapeutics, including integration of adaptive and AI-driven personalization strategies to enhance engagement and potentially improve clinical outcomes for the patients.

KEYWORDS

Bipolar disorder
Digital therapeutics
mHealth
Patient engagement
Retention
Feasibility study

References

No reference