Yes, British clinicians have several tools for clinical evidence search and related work, including iatroX, Praxis Medicine, Medwise and Umbil. They are not interchangeable copies of OpenEvidence. As of 23 September 2026, the useful comparison is their source relevance, available functions and fit for a particular workflow, not an assumption that OpenEvidence remains inaccessible.
This comparison is published by iatroX and includes iatroX among the options. It updates the access question and expands the original guide's clinical-search, calculator, examination and organisational-use scenarios. No country-by-country OpenEvidence registration test or head-to-head clinical accuracy study was performed for this update.
OpenEvidence availability: what changed between April and September 2026?
The original article, published on 3 May 2026, referred to Telecare Aware's 30 April report that OpenEvidence access had been terminated in the UK and EU, citing regulatory uncertainty. That was a dated report about access, not evidence of a permanent prohibition or an independently established explanation of the relevant regulatory position.
Reuters' 22 September reporting uses broader European-access wording. It does not verify UK onboarding. The earlier withdrawal report should therefore not be treated as a settled description of every British clinician's present experience, but the newer wording is also insufficient to announce that UK access has been restored.
A practical availability assessment separates several questions.
| Question | What a useful check establishes | What it does not establish |
|---|---|---|
| Can the site or app be reached? | Technical access from the tested connection | Eligibility to use the service |
| Can the clinician register? | Successful onboarding for that profession and account | Access for every UK professional group |
| Which functions work? | The capabilities available in that account and market | Every feature advertised elsewhere |
| What information may be entered? | The applicable terms and approved information-handling arrangements | Permission inferred from a working text box |
| May the organisation use it? | Approval for a defined use in that organisation | Permission for unrelated settings or automated actions |
These distinctions prevent two opposite errors: dismissing a potentially accessible tool because of an older report, and treating an international announcement as proof of local permission. The main Anthropic and OpenEvidence partnership explainer covers the new initiative; this guide focuses on what British clinicians should actually compare.
Why OpenEvidence became the reference point
The scale figures in the original guide belong to their historical reporting period. OpenEvidence's 21 January 2026 financing announcement reported a US$12 billion valuation, daily use on average by more than 40% of US physicians, a footprint spanning more than 10,000 hospitals and medical centres, and approximately 18 million clinical consultations by verified physicians during December 2025.
Those are company-reported financing and usage measures. They should not be converted into a current UK user count, the number of patients who benefited, a clinical accuracy score or evidence that every hospital had procured an enterprise deployment. The original description of OpenEvidence as a near-default US search tool was an interpretation of its prominence, not a measured designation applying to every doctor.
Its enterprise direction also matters. Mount Sinai's announcement of 31 March 2026 described bringing evidence-based answers into the existing Epic workflow for clinical staff. That establishes an announced integration in that health system, not an equivalent integration for UK organisations.
The historical proposition was free, rapid, cited clinical search. The contemporary buying question is which implementation supports the reader's actual work, rather than which platform has the most impressive headline.
Why the UK Market Is Different
The source required depends on the task. A GP investigating a UK hypertension pathway may need NICE NG136, rather than assuming that an ACC/AHA recommendation answers the same operational question. This is a source-selection example, not a treatment recommendation.
A pharmacist may need the exact product's Summary of Product Characteristics on emc, relevant specialist advice and the local formulary. A useful answer must distinguish general drug information from product-specific information and from local access arrangements.
An MRCP candidate needs the relevant official MRCP(UK) examination content and format, not an assumption that a USMLE-oriented exercise is interchangeable. Some underlying knowledge transfers, but examination conventions and the purpose of practice still matter.
These differences do not prove that a UK-branded product answers every UK question better. An international service can retrieve UK guidance, and a UK-focused service can still omit relevant context. The defensible test is whether the output identifies an appropriate source, represents it accurately and recognises the limits of its applicability.
The Emerging UK Alternatives
iatroX: Clinical Search, Calculators, and Exam Preparation in One Platform
Per iatroX product information, September 2026, Ask-iatroX provides free clinical reference grounded in NICE, CKS, SIGN and SmPC information from emc, with linked sources. There is no trial expiry or professional-verification gate for that free reference access. The methodology describes retrieval, ranking, citation grounding, checking and uncertainty handling; those processes are intended safeguards, not a guarantee of correctness.
The original May 2026 guide described more than 80 clinical calculators and more than 15 adaptive examination question banks, naming PLAB, UKMLA, MRCP, MRCGP AKT, MRCEM, PSA, free MSRA access and specialist diplomas. Those are historical product descriptions, not the current number of examinations covered. The September 2026 product specification supplied by iatroX covers more than forty examinations and includes adaptive sequencing and spaced repetition. Readers should use the current examinations directory for their particular route rather than infer coverage from an older count.
The wider proposition is a connected learning process. Socratic Tutor opens on an attempted question and uses targeted follow-ups to identify a misconception. The study planner takes the examination date and daily study time into account and adapts to quiz performance. Simulations and CPD tools are part of the same paid subscription, not separate add-ons, according to iatroX's September 2026 product information.
The published September 2026 price is £99 paid upfront for a year, equivalent to £8.25 per month billed annually, or £29 paid monthly. Free reference and free question access remain separate from that subscription. The value is several useful learning methods for one relevant goal, not access to unrelated examinations a clinician will never sit.
The earlier guide also described iatroX's UK clinical-reference component as UKCA-marked and MHRA-registered as a Class I medical device. That historical regulatory statement is not independently verified by this article and should not be extended to every educational feature. Readers assessing regulated use need the current intended purpose and supporting documentation; a regulatory description is not an accuracy result or an examining body's endorsement.
For an individual who needs a UK source check, related question practice and remediation of a recurring knowledge gap, iatroX is a relevant option. It should be assessed on those tasks rather than on an unsupported claim that no other platform combines reference and education.
Praxis Medicine: New Entrant with UK-Focused Positioning
The original guide identified Praxis Medicine as founded by Douglas Stark, previously a Voi co-founder, with Balderton Capital and Creandum backing. Breakit's 14 April 2026 report described 70 million SEK in funding. That amount remains in its reported currency and historical period, rather than being converted into a comparison with another company's financing.
The May 2026 product description listed NICE Guidelines, NICE CKS, NHS Digital and Europe PMC among its stated sources. This was UK-focused positioning, not an independent evaluation of which sources supported each answer.
There is a specific primary record for its early integration intentions. A 10 April 2026 NHS England Developer Community post described Praxis as a product for UK-licensed healthcare professionals, operating exclusively in the UK, and seeking NHS Website Content API access as a first step towards broader integrations. An application and an integration ambition do not establish that access was subsequently granted or that a particular NHS workflow is supported today.
Praxis therefore belongs in the comparison as a clinical-search option whose current eligibility, product status and source behaviour should be checked directly. Its funding and founder history are useful company context, but neither substitutes for examining the current service.
Medwise: NHS Enterprise Guideline Search and Individual Access
Medwise should no longer be described here as unavailable to individual clinicians. Its login page, checked on 23 September 2026, explicitly advertises free access for healthcare professionals and students. That is a current published access proposition, not a report that every possible account configuration was tested.
The organisational use case remains relevant. Medwise's work has been described as integrating global, national and local clinical guidance, including institution-specific material. Innovate UK Business Connect's 4 March 2025 account discusses that customisable clinical-search approach. An individual account should not be assumed to include a particular Trust's private policies or enterprise configuration.
The May version of this guide also referred to an HRA-listed prospective pilot comparing Medwise search with manual hospital intranet search and called it the first study of its kind in the UK. The priority claim and comparative results are not established by the material reviewed for this update. A study listing would describe planned research, not itself demonstrate clinical benefit.
Similarly, the historical statement that Medwise's website referred to thousands of clinicians was a vendor reach claim, not a count of paying subscribers, retained users or improved patient outcomes. Neither the earlier scale description nor an enterprise label is a reason to ignore the current individual-access route.
Medwise is worth assessing for someone seeking another clinical-search interface, and separately for an organisation investigating local-document retrieval. Those are distinct deployment questions, even when they concern the same supplier.
Umbil: Ward-Focused Clinical Workflow Assistant
Umbil's public website, checked on 23 September 2026, presents a UK clinical-workflow assistant with clinical question-answering and document-generation functions, including referral and handover-related work. That gives it a different emphasis from a platform chosen mainly for examination practice.
The original guide's source-checking caveat remains important. Readers should confirm the exact current source collection, document dates and whether product-specific SmPC information from emc is included for the task they intend to perform. A general claim of UK guidance does not establish the provenance of every medicines statement.
For a clinician considering assistance with a referral or handover, the evaluation should inspect omissions, unsupported additions and the effort needed to check the final document. Current tier entitlements and prices should be taken from the supplier's live terms; no unverified price is assigned here.
ChatGPT and General AI Tools: With Caveats
General-purpose AI should not be dismissed as incapable of UK-source retrieval or guided learning. Capabilities depend on the product, configuration, supplied sources and task. OpenAI's Study Mode announcement of 29 July 2025 explicitly describes guided questioning and structured support, so "general AI only gives instant answers" is not an accurate contrast.
The earlier guide grouped ChatGPT, Claude and Gemini as useful for writing, broad research and administrative work while warning about unsupported medical content. The warning remains a reason to check outputs, not a reason to assume that every specialist product outperforms every general-purpose model.
For evidence synthesis, distinguish finding papers from critically appraising them. For UK clinical questions, specify the jurisdiction and inspect the actual source. For patient-specific use, establish the approved product and information-handling arrangement rather than assume that consumer and organisational versions are equivalent.
A general tool may be preferable for a broad writing or research task. A specialist interface may be more convenient when it reliably reduces the work of source selection, checking or structured practice. Those are scenario-based judgements requiring a relevant comparison, not permanent rankings.
Which Tool Should UK Doctors Use?
Start with the job, then compare the current implementations.
For a daily "what does NICE say?" question, assess whether the answer retrieves the relevant recommendation, shows the source and avoids importing an inappropriate pathway. iatroX's UK grounding is relevant to that task, but the source and output still need checking. A different tool that performs the task well should not be rejected because it is internationally positioned.
For MRCP revision alongside clinical reference, examine the learning sequence. Can the learner attempt a question, understand the error, practise a related distinction and revisit it later? iatroX's September 2026 combination of banks, tutoring, planning and simulations addresses that wider objective. The claim should be about the workflow available, not exclusivity.
For clinical calculations, the examples in the original guide remain NEWS2, QRISK3, Glasgow-Blatchford and CHA2DS2-VASc. Check the intended population, input units, underlying tool version and interpretation in the calculator resource. Validation of an original clinical score should not be confused with an independent validation of every software implementation or every use outside its intended population.
For organisational retrieval of local Trust policies and formularies, assess the specific Medwise or other enterprise configuration. Ask how documents are approved, updated and removed, and how local guidance is distinguished from national guidance. Do not assume that an individual account gives access to the organisation's private collection.
For ward documentation, examine the actual outputs and review steps in a workflow-oriented product such as Umbil. For broad literature exploration or writing, a general-purpose AI tool may fit the task, provided its references and interpretation are verified.
A practical comparison using one fictional working day
Imagine a trainee who encounters an unfamiliar guideline question during a supervised clinic, later revises the same topic for MRCP, and needs to discuss a local referral process with a colleague.
The evidence question calls for a source-linked answer and inspection of the relevant guidance. The examination task calls for unaided practice and feedback, not simply rereading the answer. The referral question may depend on a local document that no public search tool has permission to retrieve.
Trying to force all three jobs into one conversational answer can conceal the differences. A more useful toolkit may combine a reference application, a structured learning service and the organisation's approved local source. Equally, a single platform can be convenient where it genuinely performs several of those tasks well.
The verdict is therefore by scenario. Choose for UK relevance, citation visibility, speed after checking, practical integration, educational depth, calculator suitability and mobile usability. Confirm current access separately. No clinician needs to settle the whole medical AI market before choosing a tool that helps with one clearly defined part of the day.
Frequently asked questions
Is OpenEvidence currently available to UK clinicians?
The April 2026 withdrawal report and September 2026 European-access wording do not, together, establish a verified UK onboarding result. Check the current registration and eligibility information for the intended account, and keep organisational permission separate from technical access.
Is Medwise only available through an NHS organisation?
No: Medwise's public login page, checked on 23 September 2026, advertises free access for healthcare professionals and students. Access to a particular organisation's local policies or enterprise configuration is a separate question.
Is iatroX free or does it require a subscription?
Per iatroX product information, September 2026, Ask-iatroX and free question access have no trial expiry or verification gate. The combined learning subscription costs £99 upfront annually or £29 monthly and includes question banks, Socratic Tutor, the study planner, simulations and CPD tools.