Automated lab accessioning
Nexession reads scanned and faxed requisitions, checks them, and delivers clean electronic orders into the laboratory information system you already run. A trained reviewer confirms every order before it is submitted.
Keyed by handEvery field, every form
The problem
Many referring practices still send requisitions by fax or mail. Your team keys each one by hand, and the mistakes land in the fields that cost the most to fix.
From scan to submitted order
The engine handles the routine majority. Anything it is not certain about goes to a trained reviewer, never a guess.
Nexession identifies the form, aligns it, and reads each field. Checkboxes are measured with classical computer vision so every result can be checked by hand. Written fields are read with optical character recognition and scored for confidence.
Patient identity and every machine read order detail are confirmed by a trained reviewer. The system enforces this on the server, so an order cannot be submitted until the confirmation is complete. Anything unclear is flagged, never guessed.
See the safety rules
Identity confirmedEnforced on the server
The order is sent to your laboratory information system as a standard HL7 message with acknowledgement tracking. If delivery fails, the order is held for retry rather than lost, and each message is linked back to the source image.
The order record
Nexession produces one structured order record for each requisition. Every field records where the value came from, how sure the system was, and who confirmed it. That is what answers an auditor who asks why a value was accepted.
Why Nexession
A wrong test code or a mis-keyed identifier is a clinical and billing event, not a typo. Three rules are fixed in the software and cannot be turned off.
Marks are measured with classical computer vision and stored with the exact numbers behind each decision, so a reviewer or an auditor can recheck any result by hand. No language model ever reads a checkbox.
Diagnosis codes, test selections, and provider identifiers are copied from the form. They are never inferred, defaulted, or normalized into a billable code. A blank field stays blank; an unreadable one goes to a person who reads the image.
A trained reviewer confirms patient identity and order content before any order is sent, enforced on the server rather than in the interface. Every access and edit is written to a permanent, append-only log with the before and after value.
Fits the lab you already run
Orders arrive in your LIS over your interface, so the referral relationship stays intact and the keying tax goes away.
Delivered as OML^O21 today, FHIR and other targets ready
Order record that decouples reading from delivery
Changes required from your referring providers
Security and compliance
Nexession runs on HIPAA eligible infrastructure under a Business Associate Agreement. The controls are part of the design, not a later addition.
Where the product stands
We say this plainly because trust is the point. The engine reads, validates, reviews, and delivers on a set of degraded synthetic requisitions. Live work with protected health information happens only inside a covered environment.
A focused pilot
Agree on the measures first, then run a short pilot on your own paper. Compare review time against keying time and measure the drop in holds and callbacks.
