Skip to content

PreOp Navigator

PreOp Navigator is an autonomous pre-operative agent that turns dense perioperative guidelines into an executable workflow, keeping clinicians in control while eliminating the coordination burden that causes surgical del

Code Video Cerebral Valley

Core trick

A deterministic protocol engine decides which pre-operative requirements remain, while an agent handles the outreach to complete each one.

Tags

healthagents

Description

PreOp Navigator is an autonomous pre-operative agent that turns dense perioperative guidelines into an executable workflow, keeping clinicians in control while eliminating the coordination burden that causes surgical delays and preventable complications. It's not a checklist or a chatbot — it's a tireless coordinator that works a case for weeks, so the guideline stops living in a 90-page PDF and starts running itself.
ayushenagpal/abridge-healthcare-hackathon★ 00 forksHTMLCreated Jul 18, 2026 · last push Jul 18, 2026

README excerpt

# PreOp Navigator

An autonomous pre-operative agent powered by a **deterministic clinical protocol
engine**. It owns the pre-operative workflow from surgical referral until a
patient is **Ready To Schedule** — the protocol engine decides *what* is
required; the Navigator agent decides *how* to get it done; the clinician owns
every medical decision.

**Live demo:** https://ayushenagpal.github.io/abridge-healthcare-hackathon/
(auto-deploys from `main` via GitHub Pages).

Built for the Abridge "Future of Agentic AI in Healthcare" hackathon. See
[`docs/`](docs/) for the PRD, architecture, clinical rules, demo case, and the
[implementation plan](docs/implementation-plan.md).

**Demo case:** Frank Delgado, 70 M, elective open right hemicolectomy — a
dual-risk pathway exercising both the **cardiac** spine (RCRI → DASI →
NT-proBNP → cardiology → echo) and the **pulmonary** spine (**ARISCAT** score →
advisory optimization bundle). Two views of the same run: a **Graph**
(dependency DAG) and a **Timeline** (patient journey by touchpoint, split into
"what we know" at intake vs the "required workup" the analysis generates). The
final packet includes a **patient-counseling** section (advisory optimizations
+ medication holds).

## What it demonstrates

1. **The protocol engine — not the LLM — generates the readiness graph.**
   `src/core/protocol/` is a pure, deterministic implementation of a
   clinician-reviewed subset of the 2024 AHA/ACC Perioperative Guideline. Same
   patient state a

Team