Exam aid · Calculations and formulas

Biostatistics & Epidemiology

The 2 × 2 table, risk measures and test selection on one page.

Practise Research Methodology MCQs

Tables to memorise

Cover the right-hand column and test yourself before checking.

Diagnostic tests (2 × 2 table)

MeasureFormula
SensitivityTP / (TP + FN)
SpecificityTN / (TN + FP)
Positive predictive valueTP / (TP + FP)
Negative predictive valueTN / (TN + FN)
Accuracy(TP + TN) / total

Risk and effect

MeasureFormula
Relative riskRisk exposed / risk unexposed (cohort, RCT)
Odds ratio(a × d) / (b × c) (case–control)
Absolute risk reductionControl risk − treatment risk
Relative risk reductionARR / control risk
NNT1 / ARR
NNH1 / absolute risk increase
IncidenceNew cases / population at risk over time
PrevalenceAll cases / population at a point in time
Case fatality rateDeaths / cases × 100
Herd immunity threshold1 − 1/R0

Which test?

Data / questionTest
Two means, normal dataStudent's t-test (paired if same subjects)
Three or more meansANOVA
Two groups, non-normalMann–Whitney U (Wilcoxon if paired)
Proportions / categoriesChi-square (Fisher's exact if small)
Linear associationPearson (Spearman if non-normal)
Time to eventKaplan–Meier with log-rank test
Predict binary outcomeLogistic regression

One-liners examiners love

  • Sensitive tests rule OUT (SnNout); specific tests rule IN (SpPin).
  • PPV rises with prevalence; NPV falls with prevalence.
  • A 95% CI for RR or OR that includes 1 means no significant difference.
  • Type I error (α) = false positive; type II (β) = false negative; power = 1 − β.
  • Median and IQR describe skewed data; mean and SD describe normal data.

Memory aids

SnNout and SpPin
Sensitive test Negative rules out; Specific test Positive rules in.

Now turn it into marks

Attempt MCQs on these topics while the sheet is fresh. Anything you get wrong goes to your incorrect list for the next round.

Most useful for

More cheat sheets