13 / Research / May 2026 - Aug 2026
AD-Profile
A patient-profiling web application that captures what clinical records miss for people with Alzheimer's disease, dementia or mild cognitive decline.
- Peer-reviewed papers reviewed
- 11
- Evidence quotes coded
- 196
- Question bank shortlisted
- 80 to 20
- End-to-end API checks
- 29
Overview
New anti-amyloid treatments for early Alzheimer's disease ask a lot of patients: regular infusions, repeated safety MRIs, travel, cost and heavy caregiver involvement. Whether that burden is worth it depends on a person's values, routines, support and constraints, which is exactly the information medical records rarely hold.
AD-Profile is a web questionnaire that patients and care partners complete before an appointment. It gathers preferences, concerns and practical circumstances and hands the clinician a clear, structured profile, so the conversation starts from what matters to the patient. It deliberately does not recommend treatments or judge eligibility.
What I did
Over a four-month Alberta Innovates studentship, I carried the project from groundwork to a deployed product:
- Wrote the research ethics application, including the clinician co-design protocol, consent and data-privacy plan.
- Ran a structured literature review of 11 papers, coding 196 evidence quotes into a 14-concept taxonomy of decision-relevant patient information.
- Turned that taxonomy into an 80-item question bank, each item traceable to published evidence, then shortlisted it to a 20-question instrument.
- Designed all 20 screens as mockups, then built and deployed the full application.
The platform
AD-Profile is built with Next.js, React, TypeScript and Supabase on Vercel. It has separate patient and clinician portals: a one-question-per-screen form with autosave and notes, clinician dashboards, patient summaries, moderated secure messaging and CSV import and export.
Accessibility was designed in rather than added later: screen-reader, keyboard and touch support, text-to-speech narration and a colour-blind-safe mode. On the security side it uses role-based authorization, hashed passwords, signed sessions and rate limiting, verified by a 29-check end-to-end API test suite.
What's next
Clinician co-design sessions continue with the lab in the 2026-27 academic year, using the ethics protocol, question bank and this working prototype as their foundation.