# Standard-of-Care Deviation Software (Beta) > Standard-of-care deviation detector software (beta): flag departures from the expected care pathway, provisional behind a mandatory clinician sign-off gate. Canonical page: https://medrecords.ai/product/standard-of-care-deviation-detector/ --- NewWe now read the actual MRI and CT imaging — not just the radiologist's 1-page report. [See how →](https://medrecords.ai/product/imaging/) [Home](https://medrecords.ai/) ›[Product](https://medrecords.ai/product/) › — Standard-of-Care Deviation Detector STANDARD-OF-CARE DEVIATION DETECTOR BETA ## Where the care timeline departs from the expected pathway, cited and provisional. Standard-of-care deviation detector software compares the documented care timeline against the expected diagnostic and treatment pathway for the condition, flagging departures as they occur. Every flag ships labeled provisional until a qualified clinician reviews, approves, and signs it off — the AI's judgment is never final on its own. Every flag ships labeled "AI-assisted, provisional" behind a mandatory clinician Approve / Revise / Reject sign-off gate, exactly like Case Merit & Standard-of-Care Assessment. The AI's judgment is never final without a qualified clinician's sign-off. [Join the beta ](https://medrecords.ai/demo/?src=beta-standard-of-care-deviation-detector) [Book a demo](https://medrecords.ai/demo/) Adams, Timothy — right knee · Case #IME-4812 AI-ASSISTED · PROVISIONAL Care pathway comparison — every flag cited On pathway — conservative management, weeks 1-6 cited Departure flagged — treatment gap, Jun 12 gap Matched guideline — expected pathway UNSIGNED REF Sign-off gate — required Approve Revise Reject BETA · in active development with early customers Sits on the same cited engine as **Case Merit & Standard-of-Care Assessment** — matched literature in, provisional flag out. ### Compared against the expected pathway, not a guess. The detector reads the documented care timeline against the diagnostic and treatment pathway matched to the condition, the same literature-matching layer that already powers Medical Literature & Standard-of-Care Matching. Where the timeline departs, it's flagged with the guideline it departs from, cited both ways. Built on live Literature & Standard-of-Care Matching Every departure flagged with its matched guideline Pathway inputs Matched guideline — expected course — live Documented care timeline — cited Departure candidates — flagged Review workflow Departure flagged — provisional — awaiting review Clinician reviews full reasoning trail — cited Approve, revise, or reject — signed Only a signed flag ever moves downstream. ### Same engine, same sign-off gate as Case Merit Assessment. Standard-of-Care Deviation Detector runs the same provisional-and-signed workflow already live on Case Merit & Standard-of-Care Assessment: every flag is labeled AI-assisted and provisional, and nothing moves downstream until a qualified clinician reviews it and signs off. AI-assisted, provisional label on every flag Mandatory Approve / Revise / Reject gate What the AI provides · Departure flags, cited to the matched pathway — · Full reasoning trail behind every flag — · Timeline compared against the expected course What only the clinician provides · The judgment call — approve, revise, or reject — · The signature that makes it final The boundary ### AI flags the departure. The clinician signs the finding. Standard-of-care deviation is a clinical judgment with professional consequences, and this feature is built to respect that line, exactly like Case Merit & Standard-of-Care Assessment. The platform compares the documented care against the matched pathway and flags where it departs; the qualified clinician decides whether to adopt that finding. Audit-grade, source-linked, and legally defensible, because the human signature is structural, not cosmetic. Every flag ships labeled AI-assisted and provisional behind a mandatory Approve / Revise / Reject sign-off gate, and the AI's judgment is never final without a qualified clinician's sign-off. No automated finding quietly becomes a determination. That is the design, and it does not soften. ### How it works. 3 steps, with the clinician's decision as the only exit. 01 The pathway is matched Medical Literature & Standard-of-Care Matching supplies the guidelines and expected diagnostic and treatment course for the condition. 02 A provisional flag is drafted Departures from the expected pathway are flagged, labeled AI-assisted and provisional, with the reasoning trail attached. 03 The clinician signs — or doesn't Approve, revise, or reject at the mandatory gate. Only a signed flag ever moves downstream. ### Who signs the deviation flags. The clinicians and counsel whose judgment the gate protects. [ Med-mal teams Standard-of-care departures flagged with their literature basis, before the expert budget is spent. For malpractice work ](https://medrecords.ai/solutions/medical-malpractice/) [ Medical evaluators A structured, cited comparison to test their own read of the care against. For evaluators ](https://medrecords.ai/solutions/medical-evaluators/) [ Law firms Early, cited signal on whether the documented care departs from the expected pathway. For law firms ](https://medrecords.ai/solutions/law-firms/) FAQ ### Standard-of-care deviation detector, answered. No, by design, never. Every flag ships labeled "AI-assisted, provisional" behind a mandatory, visible sign-off gate, exactly like Case Merit & Standard-of-Care Assessment: the qualified clinician reviews the flag and its reasoning, then approves, revises, or rejects it. The AI's judgment is never final without the clinician's sign-off. Medical Literature & Standard-of-Care Matching supplies the guidelines and the expected diagnostic and treatment pathway for the condition, already live today. The detector compares the documented care timeline against that matched pathway and flags where it departs; it does not introduce new, unsourced facts. Same underlying engine and the same sign-off standard, narrower question. Case Merit & Standard-of-Care Assessment renders a broader scored judgment reviewed by counsel or a clinician. Standard-of-Care Deviation Detector asks one focused clinical question, whether the documented care departs from the expected pathway, reviewed by a qualified clinician. Yes, in beta. It's live and testable now on real files, and we're refining it hands-on with the clinicians and firms using it early. If your case mix is a good fit, we'll work with you directly. Yes. Every flag carries its full reasoning trail: which part of the timeline departs, which guideline it's measured against, and where each input traces in the record. If you can't audit it, you can't sign it, so the reasoning is always on the table. ### Related capabilities. The cited layer this detector is built on. [ Case Merit & Standard-of-Care Assessment The same provisional, sign-off-gated engine, applied to the broader case-merit question. Explore ](https://medrecords.ai/product/case-merit-assessment/) [ Literature & Standard-of-Care Matching Guidelines and peer-reviewed sources matched to the record's diagnoses, the pathway this detector compares against. Explore ](https://medrecords.ai/product/medical-literature-standard-of-care-matching/) [ Case Outcome Benchmarks Success and value ranges from comparable resolved cases, cited to your record. Explore ](https://medrecords.ai/product/benchmarks/) ### Run a care timeline through it yourself. Standard-of-Care Deviation Detector compares the documented care against the matched pathway and flags departures, provisional until a qualified clinician signs off. Join the beta to run it on a file, or book a demo to see the full reasoning trail. [Join the beta ](https://medrecords.ai/demo/?src=beta-standard-of-care-deviation-detector) [Book a demo](https://medrecords.ai/demo/) [All capabilities →](https://medrecords.ai/product/)