Core Revision
Revision aid, not clinical advice. This page is educational material written for exam preparation. It is not a substitute for the current guideline or for clinical judgement, and must not be relied on for the care of any individual. Always check up-to-date national guidance and verify independently before acting on anything here.
The Stats Reasoner
Diagnostic-test statistics come from one small table, not a page of formulae. Instead of memorising them, you’ll build the ideas up from a single 2×2 table — and see why a positive result means different things in a common disease and a rare one. Play with the numbers, watch every stat move, then let the tool bring back the ones you find hard until they stick.
Understand mode: build every stat up from one 2×2 table — open a card to see what it means and why.
The one idea · everything lives in one 2×2 table
Core Lens: every diagnostic-test statistic comes from one 2×2 table — read down a column for the test’s own properties (sensitivity, specificity), across a row for what a result means (PPV, NPV), and only the row values change with prevalence.
Deep Dive
Take any diagnostic test and any group of people. Split them two ways at once: does the person actually have the disease (columns), and did the test say positive or negative (rows)? That gives four boxes — the true positives, false positives, false negatives and true negatives — and every diagnostic-test statistic on this page is built from those four boxes.
Read the table down a column and you get the test’s fixed properties (sensitivity, specificity). Read it across a row and you get what a result actually means for this patient (the predictive values) — and those depend on how common the disease is. That single shift, columns vs rows, is the whole topic.
Green = the test got it right; red = the test got it wrong. Hold this picture in your head and the formulae below stop being things to memorise.
Open each concept to see its plain-English meaning and its formula; sensitivity and specificity also carry a memory hook. They’re grouped by the one thing that matters most in the exam: does prevalence change it? After the test statistics come the treatment-effect measures (NNT, NNH, relative and absolute risk) and three charts the AKT can ask you to read.
Rule it OUT
A highly Sensitive test that comes back Negative reliably rules the disease OUT — a sensitive test catches nearly everyone who has it, so a negative is trustworthy.
Rule it IN
A highly Specific test that comes back Positive reliably rules the disease IN — a specific test rarely fires falsely, so a positive is trustworthy.
Why PPV falls when a disease is rare
Core Lens: sensitivity and specificity belong to the test; PPV and NPV depend on who is tested. In a rare disease the false positives outnumber the true positives, so even a good test has a low PPV.
Deep Dive
Sensitivity and specificity are properties of the test — measured in people you already know are sick or well, so they don’t move when the disease gets rarer. The predictive values do move, because they depend on the mix of people being tested.
Picture a good test (99% specific) in a disease only 1 in 1,000 people have. Test 100,000 people: roughly 100 have the disease, but 1% of the other 99,900 — about 999 people — test falsely positive. So a positive result is far more likely to be a false alarm than a real case: the false positives swamp the few true ones, and PPV is low. Make the disease common and that same test looks brilliant. The test never changed — only the prevalence did. That is the single most tested idea in this topic.
Want to express a test’s power in a way that doesn’t depend on prevalence? That’s exactly what likelihood ratios are for — see the card above.
In real studies a test’s specificity can differ between settings (for example primary versus secondary care), because the patients differ; exam questions assume it is fixed.
Type any numbers into the four cells — everything below updates instantly. Green cells = the test got it right, red = it got it wrong. Watch how the predictive values swing when you make the disease rarer.
Loading…
AKT pattern · the traps this calculator exposes
- Sensitivity & specificity are properties of the test — they don’t change with prevalence. PPV and NPV do.
- SnNOUT / SpPIN: a Sensitive test that’s Negative rules a disease OUT; a Specific test that’s Positive rules it IN.
- In a rare disease, even a good test gives a low PPV — lots of false positives swamp the few true ones.
- Likelihood ratios are the prevalence-independent way to express test power: LR+ >10 and LR− <0.1 are strong.
- NNT = 1 ÷ absolute risk reduction (not relative). Watch for questions that quote a big relative reduction to make a tiny absolute one look impressive.
A mix of AKT-style questions and quick cued-recall prompts, all grounded in the definitions and formulae above. You get instant feedback and a running score, and every answer feeds a review schedule — when you run the questions again, the concepts you missed come first. Say the answer out loud before you reveal it; that’s where the memory is made.
Draw the 2×2 every single time
Almost every statistics question in the AKT collapses into a 2×2 table. Don’t try to remember the formulae in the abstract — sketch the box, label disease across the top and test down the side, drop the numbers in, and read the answer off it. Do that on enough questions and the definitions stop being formulae and start being obvious.
Save your progress — and go further
Your work in this tool is saved on this device. Create a free account to get started with Core Revision; the full question banks are part of membership — the tool stays free either way.
Explore other Core Tools
The same calm, reason-first approach runs through every Core Tool.Take this further
This tool is free. If you are preparing for the MRCGP AKT, the full course puts the same reasoning into exam-style questions: a question bank with Core Lens and Deep Dive explanations, a Core Calendar revision plan built around your exam date, and Core Compass to show you which topics need you next.
Questions? Email admin@corerevision.com
A revision aid for the diagnostic-test statistics, treatment-effect measures and chart reading on the AKT. Sources: CEBM glossary; Cochrane Handbook chapters 10 and 13; Bland & Altman, BMJ 1998;317:1572; RCGP AKT data-interpretation topics. Sources last checked October 2026.
Core Revision Ltd, registered in England & Wales, company no. 16956954. Registered office: 71–75 Shelton Street, Covent Garden, London, WC2H 9JQ.
© 2026 Core Revision Ltd. All rights reserved.
Core Revision Ltd is an independent educational platform and is not affiliated with the RCGP.
Written by practising UK GPs for exam revision. Sources are cited on this page and corrections are logged. Education only — always confirm against current guidance. Core Revision is an independent revision resource.