JMIR Publications

ABSTRACT

Background:

Premenstrual syndrome (PMS) is highly prevalent and can significantly affect daily functioning and quality of life. Current mHealth applications for menstrual and reproductive health management primarily focus on predicting upcoming menstruation, with PMS often treated as a secondary feature. Furthermore, most of these applications lack tools to support users in interpreting recurring patterns or contextualizing their experiences. Recent advances in artificial intelligence (AI) can offer promising opportunities to address these limitations.

Objective:

We developed and evaluated a mobile health application, PMSync, specifically designed to support PMS self-management. This study aimed to explore user perceptions, needs, and design implications from both female and their male partners for digital menstrual health technologies.

Methods:

PMSync was designed as a smartphone application. The app integrates five main components, including PMS self-screening using the J-DRSP (SF) questionnaire, symptom visualization, cycle phase awareness, journaling, and AI-generated insights, as well as an optional partner-view feature. A qualitative exploratory design was employed, using semi-structured interviews and think-aloud interaction sessions with 18 participants (13 female, 5 male). Audio recordings were transcribed and analyzed using thematic analysis via the Dovetail web-based qualitative analysis platform.

Results:

Our findings suggest that PMS technologies may extend beyond traditional menstrual tracking tools by supporting interpretation, action, and relational communication. PMS self-management may enhance menstrual health literacy by enabling users to recognize and understand stress or unexplained emotional instability related to the menstrual cycle. The partner-sharing feature was found to be useful in highlighting partners’ emotional states and facilitating timely support. Finally, most participants welcomed AI-generated suggestions for both personal use and partner understanding, pointing to a broader human–AI interactional relationship emerging in the use of the system.

Conclusions:

The findings suggest several design implications for future PMS self-management tools. First, PMS technologies should support interpretive sense-making rather than simple symptom logging by integrating educational content with personalized symptom analysis and clear visual explanations. Second, AI-generated guidance should balance clinical and emotional tones, and support user-control, remain transparent, optional, and user-controlled, allowing users to critically evaluate or ignore recommendations while maintaining a supportive and non-authoritative tone. Third, data-sharing features should prioritize selective and context-dependent disclosure, enabling users to share specific information with trusted individuals while retaining full control over access and permissions. Finally, long-term engagement will depend on minimizing interaction burden, supporting irregular cycles and diverse lifestyles, and implementing adaptive reminders that align with users' routines.