This workflow is for higher specialty trainees in neurology, usually in their penultimate training year, using BMJ OnExamination as their main SCE Neurology bank. BMJ OnExamination does publish a dedicated SCE Neurology product, so coverage is real here. The principal limitation is not the content but the feedback loop: any personalised or self-selected practice can let a rising overall score sit on top of unattempted domains, so this plan gives you objective rules for when to follow the platform's steer, when to override it and when to move to mixed blocks.
What BMJ OnExamination offers for SCE Neurology right now
BMJ OnExamination lists SCE in Neurology among its Specialty Certificate Examination banks, so unlike some specialties this exam is directly covered.
| Item | Vendor-reported figure (checked 20 July 2026) |
|---|---|
| Question bank | 230+ Neurology SCE questions (vendor-reported), spanning cerebrovascular disease, spinal cord disorders, motor neurone disease, neuro-inflammatory conditions and more |
| Price | From £69.99 for one month up to £189.99 for 12 months, in monthly increments |
| Personalisation | "Select Questions" to customise difficulty and target weak areas; mock tests curated by the BMJ editorial team on recent exam themes; group-learning leaderboards; offline app |
| Free access | 10 free questions per day without subscription |
| Adaptive engine | Performance analytics and customisable selection are offered; no proprietary adaptive algorithm is claimed here, and none is assumed in this workflow |
Figures are vendor-reported as at 20 July 2026 and change; confirm on onexamination.com. Note the bank size: 230+ is a workable set, but it is smaller than some specialist neurology SCE banks, which matters for how quickly you exhaust genuinely unseen items — plan a second, unseen source accordingly.
The exam you are actually sitting
The SCE Neurology is two papers of 100 best-of-five questions each, 200 questions in total, each paper three hours, on one day, computer-based on Surpass at a test centre, one mark per correct answer, no negative marking. Never leave a blank. Pace is roughly 108 seconds per question.
The blueprint is the Federation's Neurology curriculum: cerebrovascular disease, epilepsy and paroxysmal disorders, headache, movement disorders, multiple sclerosis and neuroinflammation, neuromuscular disease (peripheral neuropathy, Guillain-Barré syndrome, myasthenia gravis, motor neurone disease and muscle disease), neuro-infection, neuro-oncology, cognitive disorders, neuro-ophthalmology, spinal cord disease, functional neurological disorders, and the cross-cutting skills of localisation, CSF and neurophysiology interpretation and neuropharmacology. The official blueprint sets the weighting; a vendor's topic list is a convenience, not proof of coverage.
Baseline week: measure before you let anything personalise
Before you lean on "Select Questions" or any curated feed, sit a small blueprint-stratified unseen sample — roughly 40 to 60 questions, mixed and timed — to see which domains are cold before your reading history biases what you practise. The score on a small sample is not a readiness figure, which is why a Q-bank percentage is not your exam score; the value is the domain map. Seal this sample and never revise from it.
First pass: set domain floors the feed cannot hide behind
Set a minimum attempted count per blueprint domain and hold the platform to it. The failure mode with any self-selected or personalised practice is a comfortable overall percentage built on your strong domains — clean cerebrovascular scores masking untouched neuromuscular or neuro-ophthalmology. Domain floors are the fix: your first pass is not "done" until every domain has met its floor, regardless of what the overall figure says.
An error taxonomy that tells you what to do next
Code every miss into one of six types, because each points to a different correction:
- Knowledge gap — you did not know it. One durable note plus a spaced re-test.
- Misread stem — right knowledge, wrong question. Slow-down cue, no content work.
- Premature closure — you settled on a localisation or diagnosis too early. Rehearse the discriminating sign.
- Guideline error — management out of date. Check current NICE, ABN or SmPC/eMC guidance.
- Calculation error — a CSF value, a drug dose or a statistic. Redo it cold.
- Time-pressure error — right with more seconds. A pacing fix, not a reading fix.
A weekly tally tells you whether your real problem is content, comprehension, currency or clock.
Review interval: repeat, transfer, space or read
Match the follow-up to the error. Misread stems need only a note. Knowledge gaps need a short source read and a spaced re-test, not an immediate repeat of the same item. Guideline errors need a primary-source check. The strongest correction for a real gap is a transfer question — a new item on the same principle in a different wrapper — which is where an unseen second bank does the work the seen bank cannot.
When to follow the algorithm, override it, and move to mixed blocks
Because this archetype turns on the follow-or-override decision, make it explicit and rule-based rather than intuitive:
- Follow the platform's steer — its suggested weak-area questions and curated mocks — when a domain is genuinely cold and you need scaffolded exposure to build a base. This is where personalised selection earns its keep.
- Override it when the feed keeps serving domains you are already strong in, when your overall score is rising but a domain floor is still unmet, or when it is drilling you on items you have effectively memorised. Override by manually forcing the neglected domain, not by chasing a higher headline number.
- Move to mixed blocks once every domain floor is met and first-attempt accuracy on cold domains has stabilised. Mixed, timed, random blocks remove the topic cue and rehearse pacing and no-blank discipline together. You do not need to have finished the 230+ bank to make this switch.
None of this assumes inside knowledge of a proprietary selection algorithm — it is a discipline you impose on any personalised feed using your own domain data.
Exit criteria that are not "I finished the bank"
You are ready on signals, not on completion: every domain floor met against the official curriculum; stable first-attempt performance on unseen, mixed, timed blocks; pacing near 108 seconds with nothing blank; retention holding on spaced re-tests; and calibration against the Federation's published SCE Neurology sample questions.
A seven-day plan around clinical work
BMJ OnExamination does one defined job — targeted weak-area practice and curated mocks — while iatroX supplies unseen transfer questions and timed measurement. No proprietary-algorithm claims are made; the override decisions are yours.
| Day | Main job | Time |
|---|---|---|
| Mon | BMJ OnExamination: "Select Questions" on the weakest domain (follow the steer); code misses | 40 min |
| Tue | iatroX: unseen mixed transfer block; check the concept generalised | 30 min |
| Wed | BMJ OnExamination: override to force a neglected domain the feed keeps skipping | 40 min |
| Thu | Spaced re-tests of Monday and Wednesday knowledge-gap items | 30 min |
| Fri | BMJ OnExamination curated mock, timed; review pacing errors only | 45 min |
| Sat | Federation sample questions, timed; calibrate against bank scores | 45 min |
| Sun | Rest or one short iatroX block to keep retrieval warm | 20 min |
Decision checklist: continue, supplement, switch or stop
| Signal | Action |
|---|---|
| Domains cold, curated mocks still useful | Continue and follow the platform's steer to floor each domain |
| Strong on seen items, shaky on unseen transfer | Supplement with unseen iatroX blocks now |
| Genuinely unseen items running low in a 230+ bank | Switch more volume to a second bank for fresh coverage |
| Floors met, unseen scores stable, sample calibrated | Stop adding volume and rehearse under exam conditions |
Decide on measurable gaps, not on novelty or sunk cost.
Bottom line
BMJ OnExamination genuinely covers SCE Neurology and its curated mocks are a real asset, but a smaller bank plus any personalised feed makes the follow-or-override discipline essential. Set domain floors, follow the steer only where a domain is cold, override when the feed protects your strengths, move to mixed blocks on coverage and stable performance, and keep an unseen source for a clean readiness signal.
Frequently asked questions
Is BMJ OnExamination enough for SCE Neurology on its own? It can carry a large share of blueprint coverage, since it publishes a dedicated Neurology SCE bank of 230+ questions (vendor-reported, 20 July 2026). The gap it leaves is a clean readiness measure: once you have worked its items, your score is on seen questions, so pairing it with a small unseen source for timed measurement is the sensible way to finish, especially given the modest bank size.
Which SCE Neurology component does BMJ OnExamination not reproduce well? Genuinely unseen, timed practice under mixed conditions is the component that thins out fastest, because a 230+ bank is exhausted sooner than a larger specialist set. Its curated mocks help, but once items are familiar they measure recall, not readiness — which is exactly the job a second unseen bank should take.
How many BMJ OnExamination questions should I complete per day for SCE Neurology? A reviewable 30 to 50 questions a day, coded and corrected, is more useful than large binges you cannot review, and it also stretches a modest bank across your revision window rather than exhausting it early. Keep some days lighter for spaced re-tests of earlier weak areas.
When should I stop using BMJ OnExamination and move to mixed mocks? Move once every domain floor is met and first-attempt accuracy on cold domains has stabilised, not when you have seen every item. Draw mixed mocks from unseen questions so they measure readiness; if the bank's unseen items are running low, shift mock volume to a second source.
How should I combine BMJ OnExamination with iatroX without duplicating practice? Give each a single job — BMJ OnExamination for targeted weak-area practice and curated mocks, iatroX for unseen transfer questions and timed measurement — and never revise the same item in both. Under the two-Q-bank rule, keeping iatroX blocks unseen preserves them as a clean readiness signal.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 20 July 2026. Question counts, prices and features are vendor-reported and change without notice; verify each figure on onexamination.com before relying on it. Disclosure: iatroX operates a cross-specialty UK question bank and clinical-knowledge tool that competes for revision time, so its role here is confined to jobs BMJ OnExamination does not claim — unseen-question measurement and spaced retrieval, not specialty-specific SCE content. Corrections are welcome via the feedback route on iatrox.com.
References: Federation of Royal Colleges of Physicians — Specialty Certificate Examination in Neurology; BMJ OnExamination SCE in Neurology; iatroX — Your Q-Bank Percentage Is Not Your Exam Score; iatroX — question-bank comparison hub.