Inspiration

We were inspired by how often patients miss appointments or struggle to follow care plans because of everyday barriers like transportation, internet access, or medication costs. Those challenges can be mistaken for a lack of effort. We wanted to help care teams ask a better question: “What’s getting in the way, and how can we help?”

What it does

ALTA lets a patient share a short check-in in their own words. It highlights possible barriers and concerns, then creates a clear summary to help a care team plan a human follow-up. Staff can record what they did and whether the barrier still needs support.

How we built it

We built ALTA as a web app using React, Vite, and Tailwind CSS, with a Node.js and Express API. Clear, explainable rules identify possible barriers and coordination priorities. DeepInfra can draft a staff summary, while the rules—not AI—determine the coordination cues. We use synthetic patient stories in the demo; no real texts are sent.

Challenges we ran into

We had to make the check-in flow understandable and reliable while being clear about what is—and isn’t—saved. We also worked to keep AI in a limited role: it can summarize a message, but it cannot diagnose, prescribe, or make care decisions. Connecting a frontend, backend, and AI service while keeping the API key private was another challenge.

Accomplishments that we're proud of

We’re proud that ALTA follows a patient’s reported barrier beyond the initial check-in. It helps staff prepare a follow-up and record whether support was arranged, the barrier was resolved, or more help is needed. The patient’s words remain visible, and the system doesn’t assume a problem is fixed just because a new message arrives.

What we learned

We learned that useful healthcare technology doesn’t always need to make clinical decisions. Sometimes it can help teams notice practical needs, organize information, and follow through. We also learned that AI is most useful here as a carefully limited assistant, with people staying responsible for decisions.

What's next for ALTA

Next, we’d test ALTA with care coordinators and patients to learn whether the check-ins are clear and the follow-up workflow fits real needs. We’d also improve accessibility, privacy, and secure data storage before any real-world use. ALTA is currently a demo using synthetic data—not a clinical tool.

Built With

Share this project:

Updates

Submission history