eHignite

Built for the ONC eHignite Challenge. Phase 1 was a strategic brief plus static wireframes; if it advances, Phase 2 requires a functional prototype. Prize money rides on it. Currently sitting with the judges.

Phase 1 was solo: strategy, problem framing, product design, and the AI-orchestration architecture all on me. Phase 2 I'm partnered with one backend engineer to build the functional prototype. Foundation is on the latest Next.js (16) and React 19 so Phase 2 can ship without rewriting it.

Under the 21st Century Cures Act, providers have to give patients their Electronic Health Information on request. In practice that means: a PDF dump from one provider, a different format from another, a third login at a third portal, and a fourth set of paper bills arriving in the mail.

Three failure modes stack on top of each other:

• fragmentation · every provider exports separately. No single view of a patient's own health history.
• illegibility · clinical reports are written for clinicians. Most patients can't parse the labs, the codes, the shorthand.
• paper trail debt · bills arrive late, by mail, from each provider independently. No coordination, no tracking, no signal on what's urgent.

The patient is left to be the integration layer, and most of them can't be.

eHignite is a patient-owned portal that connects to every provider a patient uses. EHI exports flow into a single unified record. From there the system does the work providers don't:

• AI export reader · configured against each export format, returns factual, digestible summaries. Not advice. Translation.
• appointment tracking · pulls next-step suggestions from the last visit notes and surfaces them when they matter.
• sourced AI chat · patient can ask “what did Dr. K say about my blood pressure last visit?” and the system reads back the actual line from the export. Sources every answer. Never gives medical advice.
• live dashboard · vitals from the most recent visit, medications to pick up, prescriptions expiring soon, bills in one place.
• one-click portability · patient can send their unified record to a new provider without re-pulling from every old one.

Primary: patients tired of being the integration layer between their own providers. They want control over their health information and the ability to actually understand it. Especially patients with multiple specialists, chronic-condition management, or recent care transitions. Anyone whose record lives in three or more places.

Secondary: providers receiving information from a new patient. The portability flow gives them a clean, structured handoff instead of fax or PDF chaos.

The product takes a strict position on the role of AI in healthcare: it translates, it surfaces, it summarizes, it sources. It does not diagnose, prescribe, or advise. That line is the entire product. The user trust model collapses the moment it blurs.

The AI layer is configured per export format. Each provider's EHI export carries its own schema, its own clinical-shorthand conventions, its own document structure. The system maps those into a normalized record so downstream features (chat, summaries, dashboard) read from a single shape.

Sourcing as the safety contract. Every AI answer in the chat sources the exact line of the exact export it came from. The user can click through to the original. This is the same discipline as in the PK/PD engine, every claim has to be traceable to a source the user can inspect. In a clinical context, that's not a nice-to-have. It's the difference between “the model said” and “your doctor said.”

Hard boundaries on advice. The AI can read back what a doctor wrote. It cannot interpret what the doctor meant, suggest a different course, or answer “should I…” questions. Those queries hit an explicit refusal surface that points the user back to their provider.

Why this matters for Phase 2. A functional prototype has to demonstrate the orchestration boundary holds under real export data and real chat patterns. The Phase 1 mockup proved the design discipline; Phase 2 proves the engineering discipline.

Both Phase 1 deliverables are linked below. The brief walks through strategy, problem framing, and solution architecture. The wireframes show the screens at static resolution.