Healthcare evidence collection with Jev
Inboxes mix scheduling, benefits, records, and language that your protocol treats as urgent. Jev can route administrative text you already collected. It is not a medical device, not a diagnosis engine, and not a substitute for a licensed clinician.
This unofficial page is the evidence collection slice of the healthcare operations routing pack. Intent: apply the Jev (TypeSafe System One) decision model to healthcare operations routing evidence collection. Primary search language: Healthcare Jev evidence collection. Confirm patterns on docs.typesafe.ai. This site does not sell, issue, or proxy TypeSafe keys. Use a credential you already have from the console or a documented gateway.
Independent angle (cover ≠ clone): Admin-queue routing against your protocol excerpt — explicit non-device / non-diagnosis limits. Not a clone of a clinical-protocol product page.
Healthcare use-case context
Evidence collection for healthcare operations routing happens before POST /v1/systemone. Jev does not browse your warehouse, retriever, or ESP. You gather the patient or staff message (admin text) facts, filter them, then ask snap questions. This slice is where fan-out cost math belongs: batch questions, do not re-send state.
Hub: Use cases. Compare, when the other tool is the real job: healthcare protocols.
Evidence Collection inputs
Collect:
- The message text as received (after your privacy filter)
- The current protocol excerpt the Noul names
- Channel / site if the question binds them
Never send:
- Asking Jev what the diagnosis is or which drug to prescribe
- Unfiltered EHR novels
- Images of rashes or IDs
Shape the payload like this once the gather step finishes:
{
"message": { "id": "inb-19", "text": "Need to move Thursday labs; also dizzy since last night." },
"protocol": { "urgent_admin": "Chest pain, trouble breathing, or sudden neuro language → nurse_oncall queue." },
"context": { "site": "clinic-a", "channel": "patient_portal" }
}Decision signals and actions
Each evidence field should change a named answer:
| Id | Type | Job |
|---|---|---|
queue |
Choice | scheduling / benefits / records / nurse_oncall / other |
completeness |
Score | missing facts → enough for that queue |
urgent_match |
Noul | Matches protocol.urgent_admin phrasing? |
Queue + completeness + urgent_match in one call. Do not add a second “what is the disease?” question — that is out of scope for this unofficial pack.
Do not treat a Noul of 0.5 as a “medium” healthcare operations routing score — it means yes and no are equally likely. Conjunctions stay in your code.
Guardrails and escalation
If the gather step fails (empty patient or staff message (admin text), redaction stripped everything, retriever empty), fail closed on closing a patient message or giving medical reassurance. Do not invent evidence so Jev has something to say. TypeSafe’s confidence-gated examples use a lower bar for recoverable reads than for irreversible actions. Those numbers are illustrations. For healthcare operations routing, treat auto_close_message as the high bar (closing a patient message or giving medical reassurance). Tune on labels — see offline evaluation.
Evaluation and rollout notes
Your eval set should include thin-evidence cases, not only happy patient or staff message (admin text)s. Label queue gold from ops, urgent-match gold from protocol authors, and whether a nurse would have wanted the ping. Pin jev-1.13.0 (the versioned id) after you fit thresholds. jev-latest and the marketing line jev-1.13 can move. Log the response model. TypeSafe’s published list price for jev-1.13 is $0.042 per million input tokens (vendor claim — confirm on the models page); output tokens are free on that same page. Unused distractors still bill as input.
Official Python and JavaScript SDKs read TYPESAFE_API_KEY and retry documented 429/529. This site does not sell, issue, or proxy TypeSafe keys. Use a credential you already have from the console or a documented gateway.
Pack map
| Slice | Page |
|---|---|
| Graph and primitives | decision workflow |
What may enter state |
input contracts |
| What to gather first | you are here |
| Atomic rules | policy checks |
| Act / review / abstain | confidence thresholds |
| Reviewer payload | human handoff |
| What to persist | audit trail |
| How it breaks | failure modes |
| Labeled replay | evaluation |
| Shadow → canary | production rollout |
FAQ
Should evidence live in the question text?
Put facts in state and point instructions at message.text, protocol.urgent_admin, context.channel. Criteria stay stable so you can replay.
When do I split calls? Queue + completeness + urgent_match in one call. Do not add a second “what is the disease?” question — that is out of scope for this unofficial pack.
Where is the rest of the Healthcare pack? Start with Healthcare input contracts and Healthcare decision workflow. Cluster hub: Use cases.
Is this a medical device or diagnostic product? No. This unofficial page is operations routing. Confirm any regulated use with your own counsel and docs.typesafe.ai.
Can we send imaging? No. State is text. Vision stacks stay elsewhere; pass labels only if a question names them.
What this page does not claim
- Not a medical device, diagnosis, or treatment recommendation.
- No sensitivity/specificity numbers.
- Not official TypeSafe.
- Official TypeSafe status, or that jev.pro issues API keys.
- That a schema-constrained answer is automatically factually correct.
Disclaimer
This is an independent unofficial site and is not affiliated with TypeSafe AI; official documentation is available at https://docs.typesafe.ai.
Primary documentation: https://docs.typesafe.ai. Hub: Use cases.
Sources
Public TypeSafe or adjacent documentation only. No private claims.