Biostatistics & EBM
for MRCGP AKT & MCPS
The complete free biostatistics revision guide for MRCGP International AKT (evidence domain — 10% of paper) and MCPS Family Medicine Part I. Formula sheets, worked AKT-style examples, study design hierarchy, and GRADE evidence — all in one place.
MRCGP International AKT
Evidence Domain — 10% of paperThe AKT evidence interpretation domain accounts for approximately 20 of 200 questions. This is the domain where most candidates lose easy marks — not because the content is hard, but because they don't revise it systematically. A candidate who scores 90% on clinical questions but only 50% on evidence questions loses 10 marks compared to someone who scores 80% on both.
MCPS Family Medicine Part I
Research Methodology — 5–8% of paperThe CPSP blueprint explicitly lists Research Methodology and Biostatistics as a core component of MCPS Part I. With 200 questions in the paper, this translates to 10–16 questions. Candidates who master this section gain a reliable advantage — these questions are more predictable than clinical scenarios.
18 Formulas You Must Know
Every formula tested in the MRCGP AKT evidence domain and MCPS biostatistics section. Memorise these before your exam.
Diagnostic Test Accuracy
TP ÷ (TP + FN)SnNout — rules OUT if negative
TN ÷ (TN + FP)SpPin — rules IN if positive
TP ÷ (TP + FP)Rises with higher prevalence
TN ÷ (TN + FN)Falls with higher prevalence
Sensitivity ÷ (1 − Specificity)>10 = strong rule-in
(1 − Sensitivity) ÷ Specificity<0.1 = strong rule-out
Risk & Effect Measures
CER − EERAbsolute Risk Reduction
1 ÷ ARRNumber Needed to Treat
ARR ÷ CERRelative Risk Reduction (can mislead)
Risk(exposed) ÷ Risk(unexposed)Cohort studies & RCTs
Odds(cases) ÷ Odds(controls)Case-control studies
EER − CER / 1 ÷ ARIAbsolute Risk Increase / NNH
Epidemiology
New cases ÷ Population at risk × TimeMeasures risk
All cases ÷ Total populationPoint or period
Prevalence ≈ Incidence × DurationKey exam formula
Cases ÷ Exposed population × 100Outbreak investigations
Incidence(exposed) − Incidence(unexposed)Excess risk from exposure
(Total incidence − Unexposed incidence) ÷ Total incidencePopulation Attributable Risk
Study Design Hierarchy
Know which study design answers which research question — and which statistical measure goes with each design.
| Study Design | Research Question | Measure |
|---|---|---|
| Systematic Review / Meta-analysis | What is the overall effect across all studies? | Pooled RR / OR / MD |
| Randomised Controlled Trial (RCT) | Does treatment X cause outcome Y? | RR, ARR, NNT |
| Cohort Study | Does exposure X lead to outcome Y over time? | RR, Incidence Rate |
| Case-Control Study | What exposures are associated with this disease? | Odds Ratio (OR) |
| Cross-Sectional Study | What is the prevalence of X in this population? | Prevalence, Sensitivity/Specificity |
| Case Report / Series | What happened in this patient/group? | Descriptive only |
Worked AKT & MCPS Examples
Step-by-step solutions to the types of biostatistics questions that appear in every MRCGP AKT and MCPS sitting.
AKT-Style: NNT Calculation
Risk MeasuresClinical Scenario
A trial of rosuvastatin in primary prevention shows MI occurs in 3.2% of the treatment group vs 5.8% in the placebo group over 5 years. What is the NNT?
Step-by-Step Solution
ARR = CER − EER = 5.8% − 3.2% = 2.6% = 0.026NNT = 1 ÷ ARR = 1 ÷ 0.026 = 38.5 ≈ 39RRR = ARR ÷ CER = 2.6% ÷ 5.8% = 44.8%Answer & Clinical Interpretation
NNT = 39. You need to treat 39 patients for 5 years to prevent 1 MI. The RRR of 45% sounds impressive — but the NNT of 39 gives the clinically meaningful picture.
AKT-Style: Sensitivity vs Specificity
Diagnostic TestsClinical Scenario
A new point-of-care troponin test is evaluated in 500 patients with chest pain. 100 have confirmed MI. The test is positive in 95 of the MI patients and in 40 of the 400 without MI. What is the sensitivity and specificity?
Step-by-Step Solution
TP = 95, FN = 5, TN = 360, FP = 40Sensitivity = TP ÷ (TP + FN) = 95 ÷ 100 = 95%Specificity = TN ÷ (TN + FP) = 360 ÷ 400 = 90%PPV = TP ÷ (TP + FP) = 95 ÷ 135 = 70.4%Answer & Clinical Interpretation
Sensitivity 95%, Specificity 90%. High sensitivity means few MIs are missed (SnNout). The PPV of 70% means 30% of positives are false — important for clinical decision-making.
MCPS-Style: Study Design Selection
Study DesignsClinical Scenario
Researchers want to investigate whether smoking during pregnancy is associated with low birth weight. They identify 200 babies with low birth weight and 200 normal-weight babies, then ask mothers about smoking history. What study design is this?
Step-by-Step Solution
Outcome (low birth weight) is identified FIRST — cases and controlsExposure (smoking) is measured RETROSPECTIVELYThis is a case-control studyThe appropriate measure of association is the Odds Ratio (OR)Answer & Clinical Interpretation
Case-control study. OR is the correct measure. If OR = 2.5, babies with low birth weight had 2.5 times the odds of having a mother who smoked. Note: OR approximates RR only when the outcome is rare.
GRADE Evidence Framework
Used by NICE, WHO, and all major guidelines. The MRCGP AKT tests GRADE — know the four levels and what downgrades evidence.
Further research very unlikely to change confidence in the estimate. Typically well-conducted RCTs.
RCTs start here
Further research likely to have important impact on confidence. Downgraded RCTs or upgraded observational studies.
Downgraded RCTs
Further research very likely to have important impact. Observational studies typically start here.
Observational studies
Any estimate of effect is very uncertain. Case reports, expert opinion, or heavily downgraded studies.
Case reports / opinion
5 Reasons Evidence Gets Downgraded in GRADE
Risk of Bias
Poor allocation concealment, lack of blinding, high dropout
Inconsistency
Heterogeneous results across studies (high I²)
Indirectness
Different population, intervention, or outcome from question
Imprecision
Wide confidence intervals, small sample size
Publication Bias
Funnel plot asymmetry, selective reporting
Biostatistics FAQ for MRCGP & MCPS
Related Exam Resources
MCPS Family Medicine Exam Guide
8-week study plan, CPSP blueprint, and MCQ bank for MCPS Part I.
MRCGP International AKT Guide
12-week study plan, AKT blueprint breakdown, DVLA tips, and prescribing safety.
MCQ Practice Bank
Past-paper style MCQs with detailed explanations for MCPS and MRCGP AKT.
Interactive Biostatistics Module
Full clinical reference with interactive calculators, 2×2 tables, and worked examples.
Cardiovascular Medicine
Highest-yield specialty for both MCPS and MRCGP AKT — ACS, AF, heart failure.
Endocrinology & Diabetes
Diabetes drug therapy, DKA protocols, thyroid disorders — 12% of MCPS blueprint.
Ready to Master Biostatistics?
Use GPManual's interactive biostatistics module to practice 2×2 table calculations, likelihood ratios, and forest plot interpretation — with instant feedback and worked solutions.
We use Google Analytics to understand how GPs use this tool and improve it. No personal health data is ever collected. You can decline and the site works fully without tracking.
By accepting, you agree to anonymous usage analytics in line with Google's Privacy Policy.