ComDiaBeat — Community-Supported Diabetes Management

"You are not alone in the journey."

A behavior-change app for diabetic patients aged 20–70, designed to shift them from inconsistent paper-based glucose logging (1–2×/day) to consistent digital monitoring (2–8×/day) through a deliberate combination of self-monitoring, dialogue support, and community engagement.

Courses: Persuasive Systems Design (811607S-3005) + UX Design & Management (812355A) · University of Oulu

Duration: Jan – Mar 2026 (10 weeks across two courses)

Format: Independent project — sole researcher, designer, and prototyper

Foundational theory: Behavior Change Support Systems (BCSS), Persuasive Systems Design (PSD), Universal Design, Information Architecture

Methods: Semi-structured interviews, competitor benchmarking, card sorting, think-aloud usability testing, NASA-TLX cognitive load assessment, value-sensitive analysis

Live prototype: View on Figma →


1. The Problem

95% of diabetes care is self-management — yet adherence to consistent self-monitoring remains poor. Existing apps treat diabetes as a solo data-tracking problem; in reality, it's a daily behavior, emotional, and social challenge.

Pain points identified through user interviews + competitor analysis

Source Finding
Interviews (n=2 + classroom n=3) Users log on pen-and-paper inconsistently; forget to check; can't identify patterns; feel isolated
Market scan (5 apps from Apple App Store) Advertisements, third-party data sharing, no community features (one had it but was banned), no gamification, none follow a documented behavior-change framework
Literature review Most diabetes apps achieve <25% 90-day retention; behavior change theory rarely applied

The thesis: Diabetes management apps fail not because they can't track data — they fail because tracking alone doesn't change behavior. Sustainable change requires autonomy + feedback + community.


2. Three Core Pillars

🩸 Self-Monitoring

Regular logging of blood sugar, medication, meals, and physical activity — with AI-based pattern detection that tells users what their numbers mean rather than handing them raw charts.

💬 Dialogue Support