Authors

Samson Cournane

Document Type

Honors Thesis

Major

Computer Science

Advisor(s)

Greg Nelson

Committee Members

Katherine Weatherford Darling, Robert W. Glover, Laura Gurney

Graduation Year

May, 2025

Publication Date

2025

Abstract

Most diabetes-prevention apps prescribe generic guidance without testing what works for each individual. We designed and evaluated From Generalized Advice to Personalized Insights, a mobile platform that embeds brief N-of-1 trials within a CDC-compliant Diabetes Prevention Program. Guided by person-based participatory design (Yardley et al., 2015) and the COM-B model (Michie et al., 2011), we ran Discover-Define-Develop cycles with lifestyle coaches (n = 5). Journey mapping surfaced 29 requirements—two-thirds addressing psychological capability—that shaped a React-Native prototype automating randomized epochs and coach-mediated reflection. From the coaches’ feedback, the mixed-methods evaluation indicated high usability and that patients conducting experiments gained clearer insight into their individual response patterns. Shorter epochs boosted engagement but reduced statistical power, underscoring classic N-of-1 trade-offs. Aligning interface choices with Health-UTAUT constructs preserved workflow fit while minimizing data-entry burden. Through iterative collaboration with diabetes-prevention practitioners, the project developed and embedded self-experimentation protocols into their daily workflows, enabling systematic tracking and analysis of lifestyle interventions.

Share