Onqi Screening – Solution Architecture
A visual breakdown of how Onqi identifies high-risk patients using structured and unstructured clinical data, NLP pipelines, and eligibility logic.
Architecture Flow
A step-by-step breakdown of how Onqi processes patient data—from input to referral output using modular and scalable components.
1
FHIR / CSV Input
Takes in both structured (FHIR API) and unstructured (CSV uploads) patient data.
2
NLP Extraction
Processes clinical notes using NLP to extract risk indicators and screening-relevant features.
3
Eligibility Rules Engine
Determines patient eligibility using clinical criteria from USPSTF and NCCN guidelines.
4
Backend API + Database
FastAPI backend persists screening results to PostgreSQL and serves structured endpoints.
5
Frontend + Referral Output
Next.js frontend displays results, with downloadable referral letter generation for clinics.
Architecture Diagram
Here's what Onqi generates for eligible patients — a CMS-ready referral letter.
View Sample Referral (PDF)