Science Model for Will Gruber's Nibble Build

This page documents how evidence is translated into product behavior. Full internal notes are maintained in the repository research document.

ARFID can be persistent over multi-year periods in youth cohorts, which supports building for long-term daily use.

Pediatric ARFID can involve significant medical complexity and hospitalization burden.

CBT-AR youth data shows meaningful food-variety gains but incomplete full remission, reinforcing the need for adherence-friendly daily tools.

Parent-led ARFID models demonstrate feasibility, supporting in-app parent scripts and home exposure coaching.

Gamification helps behavior/knowledge most when tied to structured behavior mechanics, not decoration alone.

Action planning and implementation intentions improve and sustain child dietary behavior.

Decision Rules

• Every core game loop feature must map to a behavior-change mechanism.
• No pressure-to-eat rewards. Process actions only (look, smell, touch, tiny taste attempt).
• If a feature could increase shame, urgency, or family conflict, redesign it.
• Keep child session length short enough to avoid turning into school-like workload.
• Any medical guidance surface must include explicit care-team escalation language.

Engagement

Mission draft choice quality, completion rate, and next-day return.

Clinical Fit

Exposure process adherence, safe-food progression, and distress trends.

Safety

Clear red-flag pathways with immediate care-team escalation prompts.