This is for international medical graduates using BMJ OnExamination for PLAB 1 who want to read its dashboard honestly before trusting it — or paying for it. It addresses the written PLAB 1 paper, and its principal limitation is the thing you most need to grasp: BMJ OnExamination's "adaptivity" is user-selected difficulty, not a machine-learning engine. Your percentage, your percentile and your place on the daily leaderboard all reflect the questions and difficulty you chose, not your readiness against the GMC blueprint.
What BMJ OnExamination offers for PLAB 1 right now
| Item | Detail (vendor-reported, last checked 19 July 2026) |
|---|---|
| Publisher | BMJ Publishing Group. |
| Bank | PLAB (PLAB 1) — vendor-reported 1,250+ questions, reviewed to reflect current GMC standards; verify the live count on the product page. |
| "Adaptivity" | User-selected difficulty, with recap of challenging questions. There is no ML adaptive feed. |
| Mocks | Editorially curated mock exams and recent-theme tests (vendor-reported). |
| Social features | A daily leaderboard and group ten-question games (vendor-reported). |
| Scale | Vendor-reported 200,000+ registered users. |
| Price | Vendor-reported promotional pricing observed on 19 July 2026 (roughly £4.99 for one month up to around £19.99 for twelve); verify current pricing and access on the product page, as offers change. |
| Components | PLAB 1 written SBA only. It does not reproduce PLAB 2, the OSCE. |
BMJ OnExamination is a credible, editorially governed product from a serious publisher, and difficulty selection plus curated mocks are useful. The risk is not quality; it is misreading the numbers those features generate.
The exam you are actually preparing for
PLAB 1 is a single paper of 180 single-best-answer questions in three hours — about one minute per item — set at the level of a doctor entering the second Foundation year. It is aligned to the GMC's MLA content map, the same framework used for the UK Medical Licensing Assessment; the map was updated in January 2026 and applies from September 2026, with three dimensions: areas of clinical practice, presentations and conditions. PLAB 1 is the applied-knowledge route to UK registration for international graduates; PLAB 2 is a separate clinical-skills exam. Confirm the current format and official sample questions on the GMC website, and treat any third-party "blueprint" as an interpretation of the GMC map rather than the map itself.
Every metric on the dashboard, defined
You cannot read analytics you cannot define:
| Metric | What it is | What it is not |
|---|---|---|
| First-attempt accuracy | Your score the first time you meet a question | Not your exam score; it depends on the difficulty you chose |
| Repeat accuracy | Your score on questions seen before | Largely a memory measure; inflates with exposure |
| Percentile / peer comparison | Where you sit against other users | Not a GMC standard; the peer group is self-selected |
| Leaderboard rank | Position on the daily table | A measure of volume and speed, not readiness |
| Coverage | Share of the bank attempted | Completion, not blueprint coverage |
| Difficulty | The level you chose | A setting you control, not an exam property |
| Time per item | Average seconds per question | Meaningful only in timed, mixed conditions |
The two most misread numbers are first-attempt accuracy, treated as a predicted score when it is conditioned on your difficulty setting, and the leaderboard, treated as a readiness ranking when it mostly rewards fast, high-volume answering.
Why the numbers flatter you: selection and difficulty
Two mechanisms make raw percentages incomparable with a real exam. The first is difficulty selection: set the bank to a comfortable level and your accuracy rises without your readiness changing; set it hard and it falls the same way. Because you chose the difficulty, the percentage is a property of your settings, not your knowledge. The second is recap and repeat exposure: re-attempting challenging questions is good learning, but it lifts repeat accuracy through memory, so a headline average blending first attempts and repeats drifts upward for reasons unrelated to the exam. The honest comparison is always a fresh, timed, mixed block at standard difficulty.
The leaderboard trap
The daily leaderboard is engaging, and engagement helps when studying is a slog. But rank is earned by answering many questions quickly, which pulls behaviour in exactly the wrong direction: toward speed over accuracy, toward easy high-volume topics over hard low-volume ones, and toward questions you can answer fast rather than the ones you get wrong. A candidate can top the leaderboard for a week and be no closer to passing. Use it, if at all, as a nudge to show up daily — never as evidence of readiness, and never to decide which questions you do.
Blueprint audit: attempted distribution vs the MLA content map
Do not trust the home-screen average. Audit the distribution of questions you have attempted against the GMC MLA content map: one row per area of clinical practice, recording how many questions you have attempted, your first-attempt accuracy and a confidence flag. The method is set out in the completion-is-not-coverage pillar; the point is that a high overall percentage can sit on top of areas you have barely touched. Pay particular attention to what a self-selected feed under-samples: clinical ethics and law, safeguarding, prescribing and medication safety, statistics and evidence-based practice, and image- or data-interpretation items — the low-volume, high-consequence areas that decide borderline passes.
The jurisdiction question every IMG should ask
For an international graduate, the most dangerous gap is not a topic you do not know — it is a topic you "know" from a different health system. PLAB 1 is set in UK practice: UK guidance (NICE, CKS), UK prescribing referenced to the SmPC/eMC rather than any national formulary you trained on, UK ethics and consent, UK safeguarding pathways, and the standards of the GMC's Good Medical Practice. When you review a missed question, ask explicitly whether your original answer reflected UK practice or your home jurisdiction, and correct against a UK source with a date. A bank cannot flag this for you; it is a discipline you impose — and the single highest-yield adjustment most IMGs can make.
What a credible readiness signal requires
A number is only a readiness signal under exam-like conditions. Require all five: unseen (questions you have never met, so you test knowledge not recall); timed (about one minute per item); mixed (all areas interleaved); no assistance (no hints, recap or look-ups); and a sufficient sample (a 20-item set is indicative, but base real decisions on larger, repeated blocks). Any percentage generated without these — including most of what a difficulty-selected, recap-enabled feed produces — is practice data, not readiness data.
Override rules: force what the feed under-samples
Because the feed follows your choices, override it deliberately. Force blocks of: clinical ethics and law; safeguarding of children and adults; prescribing and medication safety; statistics, critical appraisal and evidence-based practice; and image, ECG and data-interpretation items. Force at least some blocks at standard or hard difficulty so your accuracy is not resting on an easy setting. Schedule these overrides rather than hoping to reach them — the feed will always steer you back toward the comfortable middle.
Worked dashboard: turning analytics into next week's quotas
Take a hypothetical dashboard. A candidate shows 71% overall first-attempt accuracy and sits third on the weekly leaderboard, which looks reassuring. The blueprint audit tells another story: cardiology and respiratory each show 200+ attempted at 76%, while ethics and law shows 18 attempted at 55%, prescribing safety 22 attempted at 61%, and data interpretation barely sampled — with difficulty on "easy" for six weeks. The correct reading is that 71% is a real average over an unrepresentative, easy-weighted sample, and the rank reflects volume, not readiness. Next week's quotas write themselves: standard-difficulty blocks in ethics and law, prescribing safety and data interpretation; a switch of default difficulty to standard; and one fresh, timed, mixed, unassisted block for an honest number. Note what we did not do: produce a predicted pass mark. No platform can responsibly convert a self-selected sample into a probability of passing, and you should distrust any that claims to.
Worked example: seven days for an international graduate
A realistic week balancing content review with UK exam conventions, using BMJ OnExamination for one defined job and iatroX for unseen measurement — no proprietary-algorithm claims.
| Day | BMJ OnExamination (one job) | iatroX (unseen measurement) |
|---|---|---|
| Mon | 40 questions, standard difficulty; ethics + prescribing | — |
| Tue | Review; tag every UK-vs-home-jurisdiction correction | 1 fresh 20-item mixed block, timed |
| Wed | 40 questions; safeguarding + data interpretation | — |
| Thu | Blueprint audit: update attempted counts vs MLA map | 1 fresh 20-item mixed block, timed |
| Fri | Curated mock, timed, standard difficulty | — |
| Sat | Review mock by area; force under-sampled blocks | — |
| Sun | Rest, or read UK guidance for two missed topics | Compare unseen accuracy vs baseline |
This keeps BMJ OnExamination as the learning and mock engine and iatroX as the honest measurement layer — an application of the two-Q-bank rule.
Decision checklist: continue, supplement, switch or stop
- Continue BMJ OnExamination where your blueprint audit is improving and standard-difficulty accuracy is rising on fresh questions.
- Supplement with unseen measurement whenever your headline percentage stops telling you anything new.
- Switch or add a bank only for a measurable coverage gap the current one under-samples, not for a better-looking dashboard.
- Stop expanding when coverage across the MLA areas is even, unseen timed scores are stable, and remaining errors are careless rather than conceptual.
Bottom line
BMJ OnExamination is a well-governed PLAB 1 resource, and its difficulty settings, curated mocks and daily prompts are useful for coverage and momentum. But its numbers are practice data, not readiness data: the percentage is conditioned on the difficulty you chose, the leaderboard rewards volume, and neither is a GMC standard. Read the dashboard as a coverage tool, audit your attempted distribution against the MLA content map, correct for UK jurisdiction, and generate your readiness signal separately on fresh, timed, mixed blocks.
Frequently asked questions
Is BMJ OnExamination enough for PLAB 1 on its own? For learning content and mock practice it can carry a large share of PLAB 1 revision, with a vendor-reported bank of 1,250+ questions and curated mocks. For measuring readiness it is not enough alone, because its headline numbers are conditioned on the difficulty you select and lifted by recapping seen questions. Pair it with a source of fresh, unseen questions.
Which PLAB 1 component does BMJ OnExamination not reproduce well? It does not reproduce PLAB 2, the OSCE, at all — it is a written-SBA resource. Within PLAB 1 it does not reproduce exam-standard difficulty unless you set it there, and left on an easy setting it will not reproduce the real paper's challenge; it also under-reproduces low-volume areas — ethics and law, prescribing safety, data interpretation — unless you force them.
How many BMJ OnExamination questions should I complete per day for PLAB 1? Roughly 30–50 first-attempt questions a day is sustainable, but the conditions matter more than the count: standard difficulty, at least as much review as answering, and volume weighted toward the areas your blueprint audit shows you have under-sampled. Volume for its own sake — the behaviour the leaderboard rewards — is the least useful way to spend the time.
When should I stop using BMJ OnExamination and move to mixed mocks? Interleave mixed, timed mocks throughout rather than saving them for the end, and shift the balance decisively toward full, timed, mixed, standard-difficulty mocks once your blueprint audit is even across the MLA areas, your pacing sits near one minute per item, and your accuracy no longer depends on an easy setting. That transition, not bank completion, is the signal.
How should I combine BMJ OnExamination with iatroX without duplicating practice? Use BMJ OnExamination as your learning and mock engine and iatroX only for fresh, unseen measurement, so you never test yourself on questions you have already recapped. Learn and review on BMJ OnExamination, then run blueprint-sampled, timed, unassisted blocks in iatroX for the honest readiness number the leaderboard and difficulty-selected percentage cannot give you.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026. Exam format reflects the published GMC PLAB 1 structure and the MLA content map (updated January 2026, applying from September 2026) at that date; verify the current format and official sample questions on the GMC website before relying on any detail. BMJ OnExamination question counts, prices, mock and leaderboard features and user numbers are vendor-reported and should be confirmed on the product page, as promotional pricing and features change.
Disclosure: iatroX operates its own question bank and clinical-knowledge tools, so it competes with the resources discussed here. This is written as an exam-level method rather than a product pitch, and iatroX is positioned only for the job it suits — fresh, unseen questions for readiness measurement — not as a replacement for BMJ OnExamination's learning content and mocks, or for PLAB 2. Corrections are welcome through the feedback route on iatrox.com. Compare resources on the iatroX comparison hub.
References: General Medical Council — PLAB 1 format, the MLA content map and official sample questions (gmc-uk.org); BMJ OnExamination PLAB product page (onexamination.com); NICE, CKS, SmPC/eMC and NHS content for UK jurisdiction and guideline currency; iatroX, why your Q-bank percentage is not your exam score and the completion-is-not-coverage blueprint-matrix method.