Preserve Context at Handoff
The patient flow requests a summary and short issue title from the conversation before creating the doctor request. This gives the doctor an entry point into the discussion rather than an empty chat.
YHack 2024 · TEAM PROJECT
Telemedicine App
Carry an AI-assisted intake conversation into a doctor chat.
The Problem
A patient describing symptoms in an initial chat may need to repeat that information when a clinician joins. My teammates and I wanted to carry that context into the doctor chat so patients would not have to start over.
The Approach
I worked with three teammates to build the prototype in 24 hours at YHack 2024. The prototype combines an OpenAI-assisted intake chat with a request to speak to a doctor. That request generates a short conversation summary and issue title. A doctor dashboard exposes the context and supports real-time patient–doctor messaging through Firebase.
Separate AI and Doctor Responses
The chat route checks whether a doctor is active and stops generating AI responses for that interaction. Separate patient and doctor views support the change in who is responding.
Real-Time State in a Hackathon Prototype
Firestore listeners update chat state and messages in both views. Next.js and Material UI provide the interface, allowing the team to focus on the intake-to-handoff flow within the time limit.
Prototype Outcome
We built a working intake, summarization, and doctor-messaging flow within the 24-hour hackathon. It demonstrates how a patient’s context can carry into a doctor chat; it remains a prototype and has not been clinically tested.