Clinician-controlled evidence
Research and validation
Current evidence in this release is engineering verification and software checks, including checks against synthetic cases. The public materials reviewed here do not provide a dated, version-specific clinical-validation statement or report clinical study results; software checks do not establish diagnostic accuracy or patient benefit.
Which claim is supported by a test, which by a simulation and which still needs a real clinical study?
Software tests can check that implemented functions behave as specified. Synthetic cases can exercise software paths and expose technical failure modes. Neither is evidence that a clinician makes more accurate diagnoses or that patients benefit.
Measure whether clinicians find sources, notice conflict, resist unsupported suggestions, correct drafts and understand what the system has not established.
Known failure modes, excluded populations, unavailable features, data constraints and regulatory scope should change as the product changes—not remain a static reassurance page.
The public materials reviewed here do not provide a dated, version-specific clinical-validation statement or clinical study results. Engineering tests and synthetic-case checks do not establish clinical validity, diagnostic accuracy or patient benefit.
Contact SheldonDx to discuss a proposed evaluation. Any participation, data use and publication would need to be agreed in advance.
Discuss the current evaluation state, known limitations and an appropriately governed pilot.
Email to request the research brief