JMIR Publications

ABSTRACT

Background:

Digital micro-interventions (brief, evidence-based exercises delivered via mobile devices) show promise for supporting stress self-management in daily life. However, most evaluations focus on feasibility and adherence, leaving the lived, contextual experience of prolonged real-world use underexplored.

Objective:

Using collaborative autoethnography (CAE), this formative study aimed to generate rich, experience-near insights into how a Telegram-based micro-intervention chatbot (MINT: Micro-Interventions for Navigating Tension) was incorporated into, or resisted within, the everyday lives of its own designers and researchers, and to derive design implications useful for other digital health researchers and practitioners developing tools for stress management and mental well-being.

Methods:

Five digital mental health researchers used MINT daily for 4 weeks (June 2025). Data comprised personal reflective diaries, a shared reflexive log, twice-daily ecological momentary assessments, wearable physiological data, and two post-intervention group workshops. Reflexive thematic analysis was performed collaboratively, with workshop discussions serving as the primary site of theme generation.

Results:

Six interconnected themes emerged: (T1) the critical role of timing and emotional congruence in receptivity, (T2) the need for perceived reciprocity and relational warmth, (T3) the role of familiar, self-initiated coping strategies compared to suggested exercises, (T4) how non-engagement during peak stress was indeed meaningful data, (T5) the rapid collapse of use when scheduled reminders were removed, and (T6) the limited current utility of consumer wearables for real-time triggering and the competition from general-purpose large language models.

Conclusions:

This formative CAE highlights subtle but critical mismatches between current micro-intervention design assumptions and the realities of everyday stress management. Successful systems may need to prioritize contextual and emotional receptivity, relational cues, integration with users’ existing coping repertoires, planned fading of reminders, and pathways to genuine autonomy. These exploratory findings offer concrete, experience-driven directions for iterative design of more acceptable and sustainable digital stress-management and mental health tools.