This is the exam-level hub for one question: what should you actually study for ESEGH before you decide you are ready to stop doing new questions? The answer is a checklist, not a study timetable. "I finished the bank" and "I am hitting 75%" are not evidence of coverage; they are evidence of activity. Below is the minimum evidence a candidate should be able to produce, domain by domain, format by format, before saying the ESEGH blueprint is covered. If you cannot produce it, you have found your next block of new questions.
The exam is unforgiving of self-selected practice because it is broad and it is timed. Two 100-question papers cannot be passed by depth in three favourite topics. This checklist is designed to surface the domains and formats your own choices quietly avoid.
The ESEGH exam in brief
According to the Federation of Royal Colleges of Physicians (thefederation.uk, verified 21 July 2026), ESEGH consists of two papers, each three hours, each with 100 best-of-five questions — 200 in total, five options per question, one mark per correct answer, and no negative marking. It is computer-based, administered by Surpass at a test centre. UK candidates are examined against the JRCPTB Specialty Training Curriculum for Gastroenterology; European candidates against the UEMS Blue Book (2017). The 2026 fees were £700 (UK), €800 (ESBGH member countries) and £875 (elsewhere). The blueprint spans luminal and functional GI, hepatology, pancreaticobiliary disease, nutrition, endoscopy, and the general and acute medicine relevant to gastroenterology.
The official curriculum is the authoritative source of scope. Commercial banks approximate it; they do not define it. Every checklist item below is anchored to the official blueprint, not to any single bank's table of contents.
Build your blueprint coverage table
The core instrument is a coverage table with one row per blueprint domain and six columns. It converts a vague sense of readiness into an auditable record. Populate it from your bank analytics and your own coded errors, and update it weekly.
| Domain | Official weight (approx.) | Questions attempted | First-attempt accuracy | Last reviewed | Confidence (1–5) |
|---|---|---|---|---|---|
| Luminal GI (IBD, motility, coeliac) | High | ||||
| Upper GI and GI bleeding | High | ||||
| Hepatology (viral, ALD, MASLD, cirrhosis) | High | ||||
| Pancreaticobiliary | Medium | ||||
| GI malignancy and screening | Medium | ||||
| Nutrition (incl. intestinal failure) | Medium | ||||
| Endoscopy (indications, findings, safety) | Medium | ||||
| Infection and GI emergencies | Medium | ||||
| General/acute medicine relevant to GI | Lower | ||||
| Statistics, guidelines and ethics | Lower |
Two columns do the real work. First-attempt accuracy is the only accuracy that measures learning; anything after a review is memory. Last reviewed exposes the domain you "did" three months ago and have not touched since, which on exam day behaves like a domain you never did. A row with high attempts, high accuracy and a recent review date is covered. Any other pattern is a gap wearing a disguise.
Ten domain-level blind spots self-selected practice hides
These are the domains most likely to stay hidden because candidates gravitate to what they already enjoy and quietly avoid what they find dull or difficult. Self-selected practice is systematically biased toward the topics you are already good at, which is why a candidate can log thousands of questions and still walk into a predictable blind spot. Before publication, this list was reviewed against the gastroenterology curriculum; treat it as the set to actively seek out, not avoid, and require exam-specific clinician review of any domain you are tempted to skip.
- Intestinal failure and home parenteral nutrition — high-stakes, low-frequency, easy to skip.
- Autoimmune and cholestatic liver disease (PBC, PSC, autoimmune hepatitis) — overlapping serology trips people up.
- Vascular liver disorders (Budd–Chiari, portal vein thrombosis) — rare in banks, present in blueprints.
- Neuroendocrine and less common GI tumours — overshadowed by colorectal.
- Motility disorders and functional GI beyond IBS basics — under-drilled, heavily tested.
- Pancreatic cystic lesions and surveillance thresholds — guideline-sensitive and easy to get wrong.
- Nutrition in chronic liver disease and refeeding — crosses two domains, owned by neither.
- Endoscopy complications and sedation safety — procedural knowledge textbooks under-emphasise.
- GI manifestations of systemic disease and drug-induced liver injury — the "general medicine" the exam still assumes.
- Transplant assessment and post-transplant complications — specialised, examinable, routinely neglected.
If any of these has fewer than a couple of dozen attempted questions and no recent review, it is a confirmed content gap and should be your next block — regardless of how good your overall percentage looks.
Format checklist: endoscopy images, liver data and guideline differences
Coverage is not only about topics; it is about the forms a question can take. Verify that you have practised each format deliberately, not just stumbled across it:
- Endoscopy and imaging. Can you interpret an endoscopic image, a CT or MRCP, and a liver ultrasound well enough to answer a management question built on the finding, not just the label?
- Liver data. Can you read a liver-function panel, a synthetic-function trend, non-invasive fibrosis scores and ascitic-fluid analysis, and act on them under time pressure?
- European versus UK guideline differences. Where European (UEMS/ESGE/EASL-informed) and UK (NICE/BSG) practice diverge — surveillance intervals, treatment thresholds, screening — do you know which framework each candidate cohort is examined against?
These formats are covered in depth in the companion piece, what MCQ banks cannot prepare you for in ESEGH. For this checklist, the test is simple: if you cannot point to deliberate, timed practice in each of the three formats, you have a format gap even if your topic coverage is complete.
Interpretation checklist
Beyond images and liver data, ESEGH quietly tests several interpretation skills. Tick each only if you have practised it as a discrete skill:
- Laboratory trends — iron studies, autoimmune serology, tumour markers read as a pattern, not a single value.
- Radiograph and cross-sectional images — recognising the finding and translating it into a next step.
- Calculations — MELD, Child-Pugh, Glasgow-Blatchford, Maddrey and similar scores computed correctly under time pressure.
- Statistics and critical appraisal — sensitivity, specificity, likelihood ratios and study-design flaws, as they appear in evidence-based-practice items.
- Ethics and consent — capacity, endoscopy consent and end-of-life decisions in advanced liver disease.
Recency checklist
Gastroenterology guidance moves. A correct answer learned two years ago can be the wrong answer today. For every guidance-sensitive topic, record the date and jurisdiction of the source you learned it from:
- Hepatitis B and C treatment thresholds and regimens.
- MASLD/MASH terminology and fibrosis-assessment pathways (the shift away from "NAFLD").
- IBD biologic and small-molecule sequencing.
- Colorectal and Barrett's surveillance intervals.
- Variceal bleeding prophylaxis and management.
- Pancreatic cyst surveillance thresholds.
If a topic on this list was last reviewed against a source you cannot date, treat it as unverified and re-anchor it to current guidance. Note which framework applies to you — European or UK — because divergence here is a deliberate examiner trap.
Performance checklist
Coverage of content is necessary but not sufficient. Before you stop doing new questions, verify performance under exam-like conditions:
- Unseen, timed, mixed blocks. Your accuracy on fresh, mixed, timed questions — not your all-time bank average — is the readiness signal that matters.
- Speed. At roughly 1.8 minutes per question, can you finish a paper without a late-stage scramble?
- High-confidence errors. Track questions you were sure of and got wrong; these are the dangerous ones, because you will not flag them to revise.
- Retention. Re-test domains three to four weeks after you "finished" them; if accuracy has decayed, they are not covered.
- Official-material calibration. At least once, sit official-format material under exam conditions as a clean check, as set out in the ESEGH official-material workflow.
The stop-or-continue decision tree
Use the measured gap, not a feeling, to choose your next activity:
- Continue new questions in any domain where attempts are low or first-attempt accuracy is below your target band.
- Consolidate (space and re-test, stop adding new items) in domains that are high-accuracy but not recently reviewed.
- Simulate (full timed mocks) once every domain is at least adequately covered and the binding constraint is stamina and pacing.
- Seek teaching for a domain that stays low after repeated new questions — a persistent gap is usually a conceptual misunderstanding a person can fix faster than more MCQs.
- Rest when accuracy is stable and high across the board and further churn is producing fatigue, not learning.
The one-page checklist and a worked example
Copy the one-page version below and keep it beside your revision. You have covered ESEGH when you can tick every line honestly:
- Every blueprint domain has adequate attempts and a review date within the last month.
- All ten common blind-spot domains have deliberate, recent practice.
- Endoscopy-image, liver-data and guideline-difference formats each have timed practice.
- Every guidance-sensitive topic is dated and jurisdiction-labelled to current sources.
- Unseen, timed, mixed-block accuracy is stable in your target band.
- High-confidence errors are logged and re-tested.
- One clean official-format calibration has been sat under exam conditions.
Worked example (invented data). A candidate reports 78% overall and "the bank is finished". Their coverage table tells a different story: hepatology attempts 640, first-attempt accuracy 82%, reviewed last week — covered. Nutrition attempts 41, accuracy 54%, last reviewed nine weeks ago — a gap. Endoscopy-image items: 60 attempted, but almost all reviewed, so first-attempt data is thin — a format gap. On unseen, timed, mixed blocks they score 66%, well below their 78% all-time average — the tell-tale sign that the 78% is memory, not readiness. The decision tree is unambiguous: continue new questions in nutrition and vascular liver disease, drill fresh endoscopy-image items, and keep taking unseen timed blocks until that 66% closes toward the all-time figure. "Finishing the bank" was irrelevant; the gaps were specific and measurable. Notice what the exercise did not produce: a predicted exam score. Coverage work tells you where to spend your next fortnight, not what you will score on the day — a prediction no honest reading of a question bank can give you, and one you should distrust wherever a product offers it. This is exactly the discipline argued in question-bank completion is not coverage.
Frequently asked questions
How do I know whether I have covered the full ESEGH blueprint? You know it when your coverage table shows adequate attempts, target-band first-attempt accuracy and a review date within about a month for every domain, and when your unseen, timed, mixed-block accuracy is stable. Coverage is an auditable state, not a feeling or a completion percentage. If any blueprint domain — especially one of the ten common blind spots — has thin attempts or a stale review date, you have not covered it, however high your overall average looks.
Can one question bank be enough for ESEGH? Occasionally, but rarely, and only if that single bank genuinely maps the whole blueprint, is current against guidance, and includes endoscopy-image and liver-data practice. Most candidates need a second source to cover a domain the first neglects or to provide unseen items once the first bank is memorised. The safe test is not the number of banks but the coverage table: if one bank fills every row honestly, one is enough; if it leaves gaps, add a second using the two-Q-bank rule so you do not simply duplicate practice.
What should I measure instead of my overall Q-bank percentage for ESEGH? Measure first-attempt accuracy on unseen, timed, mixed blocks, broken down by blueprint domain and by format, plus your rate of high-confidence errors and your retention after three to four weeks. Your overall percentage blends fresh and reviewed questions, easy and hard domains, and rewards you for repeating items you already know. The domain-and-format breakdown on unseen questions is the number that predicts exam-day performance; the headline percentage mostly measures how much of the bank you have seen before.
When should I stop doing new ESEGH questions? Stop adding new questions in a specific domain once its first-attempt accuracy is stable and high and it has been recently reviewed — pouring more new items into a covered domain is low-value. Stop doing new questions altogether only when every domain is covered, your unseen timed accuracy is stable, and your remaining constraint is stamina and pacing rather than knowledge. At that point, switch from new questions to full timed simulation and targeted review of logged errors.
Which ESEGH resource should I use for my weakest component? Match the resource to the failure type. For a knowledge gap, use a specialty bank such as StudyPRN or RevisionPro SCE for gastroenterology- and hepatology-specific depth, and re-anchor concepts to current European or UK guidance. For an image or data-interpretation gap, use image-rich material and atlases plus deliberate timed practice. For a transfer or retention gap, use a general unseen-MCQ layer with spaced retrieval — the job iatroX is built for — to confirm that fixed concepts hold up on new stems. Choose by the coded weakness, not by which resource is most convenient.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 21 July 2026. Exam-format facts are from the Federation of Royal Colleges of Physicians; any figures attributed to StudyPRN, RevisionPro SCE or other providers are vendor-reported and were correct to the best of our checking on that date — verify current details on the provider's own page. Domain weightings are described as approximate because the exam body publishes a blueprint rather than a fixed per-topic count; confirm the current blueprint on thefederation.uk. Disclosure: iatroX operates a competing question bank; in this hub its role is confined to unseen-MCQ measurement and spaced retrieval, jobs a specialty bank does not claim, and it is not presented as a specialty-specific ESEGH bank. Corrections are welcome via the feedback route on iatrox.com.
References: the Federation ESEGH page (thefederation.uk); UEMS "Blue Book" gastroenterology curriculum (2017); JRCPTB Gastroenterology curriculum; the ESEGH modality-gap article; Your Q-Bank Percentage Is Not Your Exam Score; the iatroX comparison hub.