skip to main content
iatroX JournalUK Guidelines

emc vs NICE CKS vs SPS vs MHRA vs Peer-Reviewed Research: Which Prescribing Source Should You Use?

Featured image for emc vs NICE CKS vs SPS vs MHRA vs Peer-Reviewed Research: Which Prescribing Source Should You Use?

The characteristic prescribing-information error is not consulting a bad source; it is consulting a good source that was designed to answer a different question, and then trusting the answer because the source was good. emc will not tell you where a medicine sits in a pathway. NICE will not give you the excipient list of a specific formulation. A trial will not tell you your trust's first-line choice. Each source is excellent at its purpose and silently unhelpful outside it, so the skill worth building is the routing matrix, question to source, and this article is that matrix with three worked conflicts attached.

The question-to-source matrix

Licensed indication, contraindications and product particulars: the relevant SmPC on emc, the regulatory record for the exact product, strength and formulation, updated by companies within ten days of approved changes. Condition management and treatment sequencing: NICE guidance, CKS for primary-care presentation, SIGN where applicable, the comparative layer that decides where a medicine belongs, not just what it is. New safety warnings and recalls: MHRA communications, the live regulatory channel that no static summary, AI or otherwise, should be assumed to have absorbed, which is the recency problem in one sentence. Practical medicines optimisation and unusual populations: SPS and specialist medicines-information services, built for the questions that fall between the SmPC's authorised text and the guideline's typical patient. Local first-line choices, formulary restrictions and referral pathways: local policy, which operationally outranks the national default because it encodes local resistance data, commissioning and agreed shared care. Emerging or contested interventions: peer-reviewed research, read as evidence with a certainty level, not as permission. And rapid cross-source orientation: an evidence-navigation layer such as Ask-iatroX, which is best understood as the intelligent front door that helps you find the relevant corridor, not the owner of every room in the building.

Three worked conflicts

Licensed but not first-line: a medicine holds a marketing authorisation for a condition, and the guideline recommends something else first; both sources are right, the SmPC describes what is authorised, the guideline judges what is preferred, and the prescriber's job is knowing which document answers "can I?" and which answers "should I, here, now?". Guideline-supported but off-label: national guidance recommends a use outside the marketing authorisation; the SmPC has not erred and neither has the guideline, the states are simply different, and the answer must carry both labels plus the documentation and consent behaviour off-label use requires, treated fully in its own article. National recommendation versus local antimicrobial policy: the national default and the trust's policy diverge because local resistance patterns and stewardship decisions exist; operationally, local policy governs, and an answer that quotes the national line without flagging the local layer is incomplete in the way that causes real errors. In each conflict, notice the shape: no source was wrong, and the error would have come from asking one source to do another's job.

Building the reflex

Two habits install the matrix. Name the question type before opening anything, licensing, pathway, safety, optimisation, local, evidence, orientation, ten seconds that route the next two minutes correctly. And let synthesis point, sources decide: an Ask-iatroX answer that surfaces the guideline position, the relevant SmPC link and the research context has done the orientation job, and the click-through to the exact document is where the decision's weight transfers to the source built to carry it, which is the division of labour the whole matrix exists to teach. The prescriber who works this way satisfies the competency framework's evaluation requirement as a byproduct of speed, not in tension with it.

Frequently asked questions

Which single source causes the most misrouted questions?

Guidelines, because their authority tempts product-level reliance: sequencing questions they answer superbly, formulation and licensing particulars they were never built to hold, and the SmPC habit is the corrective.

Where do interaction resources fit in the matrix?

As the optimisation layer's specialist tools alongside the SmPC's interaction sections and pharmacist advice; for consequential combinations, more than one layer should agree, and silence from one layer is never clearance.

Does the matrix change for supplementary prescribers?

The routing is identical and one layer gains force: the clinical management plan joins local policy in the operational tier, and answers must fit inside it regardless of what national sources would permit.

How do MHRA alerts reach me if I only check static sources?

They largely do not, which is the recency problem in operation: safety communications are a live channel, worth subscribing to for the medicine classes you prescribe most, because no summary, human or AI, should be assumed to have absorbed this morning's alert.

What if the local policy itself looks out of date?

Escalate rather than freelance: flag it to the medicines-optimisation or governance owner, document the query, and follow the current version meanwhile unless it poses immediate safety concern, in which case specialist advice is the route.

Start at the medicines front door →

Back to Journal