OpenEvidence and Tandem are often mentioned in the same breath as "clinical AI," but they represent two genuinely different philosophies. OpenEvidence is a medical evidence search engine built on peer-reviewed literature; Tandem is a regulated, EHR-adjacent decision-support product that lives inside the consultation. They diverge on evidence base, content relevance and, most sharply, on whether a UK clinician can use them at all.
Evidence base
OpenEvidence's defining feature is its grounding. It is trained primarily on peer-reviewed medical literature and built around content partnerships with organisations including the New England Journal of Medicine, JAMA, Cochrane and NCCN. That foundation, plus extraordinary US adoption — reportedly used daily by more than 40% of US physicians across 10,000-plus hospitals — made it the US reference point for source-grounded clinical answers. The demand it proved is real: clinicians want fast, cited answers at the point of care.
It is worth being even-handed about the evidence-quality debate, too. Some peer-reviewed evaluations have raised questions: one study (Low et al.) reported OpenEvidence returned relevant answers to roughly a quarter of open-ended clinical questions, and another evaluation found accuracy below 41% on subspecialty board questions. Those findings sit alongside very large real-world usage and deep literature partnerships, so they sharpen the picture rather than settle it — but they are part of an honest evidence comparison.
Tandem's CDS takes a different tack. Rather than indexing the literature, it answers patient-specific questions from national guidelines and locally uploaded protocols, using the context of the documented visit, with its credibility anchored in Class IIa device assessment.
Content relevance and commercial model
OpenEvidence is US-centric by design. Its content leans on American guidelines and FDA approvals, verification has been built around the US National Provider Identifier, and it has at times surfaced medicines not licensed in the UK or EU. It is also funded by pharmaceutical and medical-device advertising — a structural fact that some clinicians weigh when judging evidence integrity, and one that ad-free competitors have made a point of contrasting themselves against.
Tandem is built for European and UK workflows, integrating with record systems including EMIS and SystmOne, and reaches NHS clinicians through its Accurx partnership.
The decisive factor: UK availability
OpenEvidence withdrew from the UK and EU on 30 April 2026, citing mounting regulatory uncertainty around AI systems, including the EU AI Act, which would treat much clinical AI as "high-risk." As of mid-2026 it remains unavailable to UK and EU clinicians, regardless of credentials, and the withdrawal has become a case study — including a position piece in The Lancet Regional Health – Europe — in how regulatory ambiguity shapes clinical-AI access.
Tandem made the opposite choice: it pursued CE marking as a Class IIa device and stayed. For a UK clinician today, that is the whole difference — one tool is accessible and regulated for the market; the other has left it.
| Dimension | Tandem CDS | OpenEvidence |
|---|---|---|
| Core approach | In-consultation decision support from guidelines + local protocols | Evidence search engine trained on peer-reviewed literature |
| Evidence base | National guidelines, local protocols, visit context | NEJM, JAMA, Cochrane, NCCN and peer-reviewed sources |
| Content relevance | UK/EU workflows; EMIS, SystmOne | US-centric (US guidelines, FDA, NPI verification) |
| Commercial model | Enterprise / EHR-integrated | Free for verified clinicians; ad-funded (pharma/device) |
| Regulatory class | CE marked EU MDR Class IIa | Withdrew rather than operate under UK/EU AI rules |
| UK availability | Available (selected UK customers) | Withdrawn from UK and EU (30 April 2026) |
Where iatroX sits
OpenEvidence's exit left a real gap for UK clinicians who had come to rely on fast, cited clinical answers. iatroX is built squarely for that gap — but for the UK from the outset. It is a free, UK-registered (UKCA-marked Class I) clinical AI grounded in NICE, CKS, SIGN and the SmPC, ad-free, and still available to NHS clinicians.
The honest framing is straightforward. OpenEvidence is a formidable tool with a deep literature base, and Tandem holds a higher regulatory class than iatroX. iatroX's distinct value is not a claim to out-regulate either of them — it is being the UK-native knowledge layer that stayed: source-linked, UK-relevant, free, and designed to inform the clinician's judgement rather than direct it. And its Class I registration, while self-declared rather than notified-body assessed, is not ungoverned — it is backed by a full clinical risk-management file and safety case, independently reviewed.
Frequently asked questions
Is OpenEvidence available in the UK? No. OpenEvidence withdrew from the UK and EU on 30 April 2026, citing regulatory uncertainty around AI systems including the EU AI Act, and remained unavailable to UK and EU clinicians as of mid-2026.
How is Tandem's CDS different from OpenEvidence? Tandem answers patient-specific questions inside the consultation from national guidelines and local protocols, regulated as a Class IIa device. OpenEvidence is a search engine grounded in peer-reviewed literature, US-centric in content, and is no longer available in the UK or EU.
Is OpenEvidence's content suitable for UK practice? Its content is US-centric — leaning on US guidelines and FDA approvals, with US-based clinician verification — and it has at times surfaced medicines not licensed in the UK. UK clinicians need answers grounded in UK guidance such as NICE, CKS, SIGN and the SmPC.
What can UK clinicians use instead? iatroX is a free, UK-registered Class I clinical AI grounded in NICE, CKS, SIGN and the SmPC, ad-free and available to UK NHS clinicians as a knowledge and reference layer.