Evidence engine

Text with provenance

Most clinical chatbots return text. MyRadAssistant returns text with provenance — the evidence engine is the layer that makes the output verifiable.

Three-tier evidence model

Tier 1
Strongest evidence
Clinical guidelines (RCR, ESR, ACR, NICE), systematic reviews and randomised trials. Statements supported here are rendered assertively.
Tier 2
Supporting evidence
Cohort studies, case-control studies, large case series and society position statements.
Tier 3
Weakest evidence
Small case series, case reports and narrative reviews. Statements supported only here are hedged — "some evidence suggests", "in a small case series".

Tier assignment is rules-based, mapping source type to tier deterministically. Expert consensus validation of the tier rules is pending.

Tier-aware retrieval

When multiple sources address a question, retrieval preferentially surfaces higher-tier evidence. You can see the tier of every cited source in the response.

Citations

Every claim in a response is anchored to a specific source passage. Click the citation in the response and you land on the exact text in the source document. We are progressively migrating from post-hoc similarity matching to native generation-time citations (Anthropic API) for more reliable anchoring.

On the roadmap

  • Retraction Sentinel (Phase 2): automatic flagging of cited papers that have been retracted or expressed concern about.
  • Disagreement detection (Phase 2): when sources disagree, surface the disagreement explicitly rather than hiding it inside a single synthesis.
  • UMLS terminology grounding (Phase 2): standardise terminology in queries and retrieval.
  • GraphRAG over RadLex (Phase 3): differential diagnosis support driven by RadLex's 30,000+ terms and 55,000+ disease-finding relationships.

Credentials and regulatory status

We publish only the certifications, registrations, and regulatory statuses we currently hold and that can be independently verified. Anything in progress, applied for, or planned is deliberately excluded from this page.

CredentialStatusVerify
MHRA Class I medical deviceSelf-certified — Function 1 only (see scope below)MHRA public device register
NHS DSPTStandards Met — ODS U9L1B, 2025-26 v8, published 13 April 2026, valid to 30 June 2027NHS DSPT register
ICO registrationZB621582, valid to 3 November 2026ICO public register
ODS codeU9L1BNHS Organisation Data Service
RCR iRefer content licenceFull licence signed — RCR iRefer referral guidelinesLicence on file
Radiopaedia content licenceFull licence signed — Radiopaedia reference contentLicence on file

MHRA scope statement

Our MHRA Class I self-certification covers Function 1 only — the grounded, citation-anchored question-answering tool — registered as a Class I medical device with the MHRA (manufacturer account 38102, GMDN 61087).

The report-critique tool is not a medical device: it is a professional medical-education tool and sits explicitly outside the scope of the device registration.

MyRadAssistant is a clinical decision-support reference tool, not a diagnostic device. It does not process patient-identifiable data and is not for use at the point of care on an individual patient.

What the evidence engine cannot do

The evidence engine is only as good as the corpus. New high-quality evidence published yesterday might not be in today's index until the next ingest cycle. Always treat a citation as the starting point for verification, not the end of inquiry.

Provenance on every answer.
Start your free trial.

Cancel anytime