blog

A single letter grade cannot carry everything a UK clinician actually needs to know. A more honest interface would separate hierarchy, certainty, guideline status and applicability...

Starting from the literature and starting from the guideline are both defensible approaches to clinical AI, and each carries risks the other does not. The safest answer at the poin...

OpenEvidence asks how certain the published evidence is. iatroX asks what the best-supported and UK-relevant answer actually is. These are related but genuinely different questions...

EvidenceGrade tells a UK clinician how strong the cited evidence is. It does not tell them whether NICE recommends it, whether it is licensed here, or whether the NHS pathway even ...

Whether a citation is real, how strong the underlying evidence is, and whether it applies to this patient in this healthcare system are three separate questions. Clinical AI needs ...

High certainty and clinical recommendation are two different things entirely. A high grade can sit next to a treatment that should not be used, and confusing the two is a genuinely...

A single clinical answer often bundles diagnosis, efficacy, harms, and monitoring together. A single evidence grade risks flattening genuinely different levels of certainty into on...

iatroX's evidence hierarchy for UK point-of-care questions runs from authoritative national guidance through systematic reviews to individual trials, with clinical judgement determ...

A meta-analysis sits near the top of the evidence hierarchy for a reason, but the hierarchy is a route towards the likely best evidence, not an automatic quality verdict on any spe...

EvidenceGrade shares a name and a lineage with the GRADE framework, but a real-time system built for point-of-care speed cannot fully replicate what a deliberative, weeks-long expe...

Different clinical studies are not interchangeable data points. Averaging them into a single grade, however tempting for a clean interface, risks concealing exactly the disagreemen...

OpenEvidence already grounds its answers in peer-reviewed literature. EvidenceGrade exists because peer review, on its own, tells a clinician far less than it seems to.