No one lost to follow‑up.
Somewhere in your data is a patient who was told "we'll see you in three years," and no one ever called. We find them, and we walk them back to care.
Orison reads a health system's or health plan's own clinical data, finds the GI patients whose follow-up was dropped, and drives each one to a completed procedure through the organization's own staff. Built and clinically governed by a board-certified gastroenterologist.
What the engine found in this synthetic panel
Twelve patients, drawn from free-text pathology, endoscopy, radiology, and laboratory reports that no registry was tracking.
How the engine reads a report
Pick a report. The left panel shows the free text exactly as it sits in the record today. The right panel shows what the engine pulls out of it: the finding, the obligation it creates, the guideline behind it, and the due date the calendar never knew about.
The report, as written
What the engine extracted
In the live product, every extraction below a confidence threshold routes to a human review queue, and extraction accuracy is measured against physician-adjudicated ground truth before any feed goes live.
Run the engine live
This one is not precomputed. Edit the report below, or write your own synthetic report from scratch, and the extraction engine reads it in real time using the same guideline rules a gastroenterologist would apply. Synthetic text only, never real patient information. A demo code is required.
Your report (editable)
Live extraction
The result will appear here. Try changing the adenoma size to 9 mm and running it again: the surveillance interval should change, because the guideline changes at 10 mm.
The worklist your team works from
Every open loop, ranked by clinical urgency, with the quality measure and the recoverable value attached. Filter by program lane.
| Priority | Patient | Finding and source | Obligation | Due | Days overdue | Measure | Est. value |
|---|
Priority key: Critical = lost past the safe window. Overdue = due date passed. Clock running = inside the window but unresolved. On track = monitored, no action needed yet. COF-E = proposed HEDIS measure, colonoscopy within 180 days of a positive stool-based test (targeted for measurement year 2027). Values are illustrative single-procedure estimates.
What closing the loop is worth
Move the sliders to your population. Arithmetic only, no magic: positive stool tests times the completion gap, priced per completed colonoscopy. All figures illustrative.
Reference points: real-world completion after positive FIT is 46.7%. NCQA reports plan performance from 24% to 75%. Roughly 1 in 25 completed colonoscopies after a positive stool test finds a cancer.
The patient who never learns their result needed follow-up is the most information-poor person in the system. Orison exists to find them.