Skip to main content
Medicine 3% exam weight

The 2×2 Table Foundation

Part of the Saudi GP Board study roadmap. Medicine topic medici-006 of Medicine.

→ See in Roadmap Saudi GP Board Study Plans Build 1-Month Plan
By Last updated 3% exam weight

The 2×2 Table Foundation

🟢 Lite — Quick Review (1h–1d)

Rapid summary for last-minute revision before your exam.

Evidence-Based Medicine (EBM) integrates best current research evidence with clinical expertise and patient values to guide primary care decisions. The 2×2 contingency table is the single most-tested framework: it underpins the odds ratio (OR = ad/bc) for case-control studies, the relative risk (RR = incidence in exposed ÷ incidence in unexposed) for cohort and RCT data, and all four diagnostic measures (sensitivity, specificity, PPV, NPV).

Three high-yield numbers to memorise: ARR = risk in control − risk in treatment; NNT = 1/ARR; I² statistic quantifies heterogeneity in a meta-analysis. Sensitivity rules out disease (SnNout); specificity rules in (SpPin). A p-value < 0.05 confirms statistical—but never clinical—significance.

🟡 Standard — Regular Study (2d–2mo)

Standard content for students with a few days to months.

The 2×2 Table Foundation

Every cross-tabulation of an outcome against exposure or disease status produces four cells: a (exposed + outcome), b (exposed, no outcome), c (unexposed + outcome), d (unexposed, no outcome). From these you derive all epidemiological and diagnostic indices. In a case-control study (outcome first), prevalence is fixed by design so only the OR is valid. In a cohort study or RCT (exposure first), you can compute true incidence, making RR and ARR the appropriate measures.

Measures of Therapeutic Effect

  • RR = [a/(a+b)] / [c/(c+d)] — ratio of risk; RR = 1 means no effect.
  • ARR = control event rate − experimental event rate.
  • NNT = 1/ARR — patients you must treat for one extra patient to benefit (rounded up to next whole number).
  • NNH = 1/ARI — counterpart for harm.

A 95% confidence interval (CI) that crosses 1.0 means the result is not statistically significant at α = 0.05.

Diagnostic Test Evaluation

MeasureFormulaClinical Use
SensitivityTP / (TP+FN)High → rule out (SnNout)
SpecificityTN / (TN+FP)High → rule in (SpPin)
PPVTP / (TP+FP)Probability of disease given positive test
NPVTN / (TN+FN)Probability of no disease given negative test
LR+Sensitivity / (1−Specificity)>10 = strong rule-in
LR−(1−Sensitivity) / Specificity<0.1 = strong rule-out

Hierarchy of Evidence and Bias Control

Evidence ranks: systematic reviews/meta-analyses > RCTs > cohort > case-control > case reports. RCTs minimise confounding through randomisation and allocation concealment; intention-to-treat analysis preserves randomisation despite dropouts. Observational studies require adjustment for confounding via stratification, matching, or multivariable regression.

🔴 Extended — Deep Study (3mo+)

Comprehensive coverage for students on a longer study timeline.

Critical Appraisal Tools and Meta-Analysis Pitfalls

Systematic reviews follow PRISMA reporting standards; RCTs use CONSORT; observational studies use STROBE; qualitative studies use SRQR. Critical appraisal questions for therapy studies (CASP) probe randomisation quality, baseline comparability, blinding of assessors, follow-up completeness, and treatment-effect size with precision.

A meta-analysis pools effect estimates, but its validity hinges on heterogeneity assessment. The I² statistic quantifies the percentage of variability due to between-study differences rather than chance: I² < 25% (low), 25–50% (moderate), 50–75% (substantial), >75% (consider not pooling). A fixed-effect model assumes one true effect; a random-effects model is preferred when I² is substantial. Publication bias—where positive trials are preferentially published—is screened with funnel plots and Egger’s test; a missing bottom-left quadrant of small negative studies suggests bias.

Clinical Significance vs Statistical Significance

A large trial may detect a tiny absolute effect (e.g., RR = 0.98, p < 0.001) that is statistically significant but clinically trivial. Always interpret the point estimate, the width of the CI, and the NNT in the context of cost, side effects (NNH), and patient preference.

Connections to GP Practice

Saudi GP Board candidates apply these concepts when writing journal clubs, formulating clinical practice guidelines (graded A–C), interpreting pharma advertisements, counselling patients about screening (e.g., breast, colorectal, diabetes), and rationalising referrals. A screening test with high sensitivity but low prevalence (common in primary care) yields low PPV—producing false positives and over-investigation. Use LR+ and LR− with pre-test probability (Fagan nomogram) to individualise test interpretation.

Common Mistakes

  • Computing RR in a case-control study (impossible denominators).
  • Reporting p-value as the probability the null is true (it is the probability of these data, or more extreme, given the null).
  • Treating NNT from one trial as transferable to a different population without checking baseline risk.
  • Forgetting that PPV/NPV change with prevalence even when sensitivity/specificity are fixed.

Practice Prompts

  1. A cohort study finds 30 events among 500 aspirin users and 50 among 500 non-users. Calculate RR, ARR, NNT, and interpret whether aspirin is protective.
  2. A screening test has sensitivity 90% and specificity 80% applied to a population with disease prevalence 2%. Compare PPV at 2% and 20% prevalence and explain why mass screening can paradoxically harm patients through false positives.

Content adapted based on your selected roadmap duration. Switch tiers using the selector above.

Sources & verification