Ripple

Inspiration 💡

1 in 8 Canadians are unaware they are living with pre-diabetes. Prediabetes is increasingly shifting younger. Recent data notes that nearly 2.6% of Canadian adolescents are currently in the prediabetic range. We want to catch it early to protect Canadian youth. Ripple provides clinicians and patients an easy-to-use, all-in-one solution to chronic health risk prevention.

What it does 🩺

For the care team

Patients are ranked by risk trajectory, with a "stopped logging" badge.
Each flag has a points ledger: every signal, its points, its reading and a guideline citation.
Lab charts show projected threshold dates, and BP charts mark where a shift began.
Metabolic syndrome appears as "3 of 5 criteria, 2 unknown".
A clinician validates or dismisses every flag. The AI never messages a patient on its own.

For the patient

Snap a meal, a BP monitor screen or a lab report, or type how you feel.
Do a weekly sleep check-in and see a timeline and messages from the care team.

How it works ⚙️

Risk engine: TypeScript rules built on Canadian guidelines. It spots slow trends and decides who gets flagged.
Gemini: extracts data out of photos and text and writes the explanation, citing only sources from the engine's list.
Tiger Data: stores the readings and runs the trend math, 3x faster than our app code on demo data.
Vultr: patient data stays in Canada. A nightly job re-runs the engine for every patient.
All data is synthetic and labeled as such.

What we learned 🌱

Clinicians act on flags they can check in 30 seconds, so the explanation matters more than the score.
Four to six irregular lab points needs robust stats, not plain regression.
Keeping the AI out of the decision made the system easier to trust and to test.