Core measures of disease frequency

🟢 Lite — Quick Review (1h–1d)

Rapid summary for last-minute revision before your exam.

Evidence-Based Medicine (EBM) in Saudi GP Board Topic 4 is the discipline of converting a primary-care clinical doubt into a searchable question, finding the best external evidence, critically appraising it, and applying it to an individual patient. The framework is PICO (Patient/Problem, Intervention, Comparison, Outcome), and the evidence hierarchy ranks systematic reviews/meta-analyses above randomised controlled trials (RCTs), which sit above cohort, case-control, and cross-sectional studies.

Sensitivity = TP / (TP + FN); Specificity = TN / (TN + FP) — both prevalence-independent.
PPV / NPV change with prevalence; OR ≈ RR only when the outcome is rare (<10%).
NNT = 1 / ARR; a 95% CI crossing the null means the result is not statistically significant.

Exam weight ≈ 3%, mostly single-best-answer MCQs on interpreting a 2×2 table or an abstract.

🟡 Standard — Regular Study (2d–2mo)

Standard content for students with a few days to months.

Core measures of disease frequency

Prevalence is a snapshot proportion: existing cases divided by the population at that point. Incidence rate requires person-time at risk in the denominator and is the only valid measure for new events. The relationship Prevalence ≈ Incidence × Duration explains why chronic, long-lasting diseases (e.g. diabetes) dominate primary-care workloads even with low incidence.

Measures of association

For cohort studies and RCTs, calculate Relative Risk (RR) = Incidence_exposed / Incidence_unexposed. For case-control studies, only the Odds Ratio (OR) = (a×d)/(b×c) is valid because incidence is unknown. Absolute Risk Reduction (ARR) and the derived NNT = 1/ARR translate efficacy into something a GP can discuss with a patient.

Diagnostic test interpretation

A 2×2 table yields four operating characteristics. Sensitivity and specificity are intrinsic to the test; PPV and NPV shift with prevalence — a useful screening test in a low-prevalence community can yield more false positives than true positives.

Measure	Formula	Clinical use
Sensitivity	TP / (TP + FN)	Rule-out: a negative test rules out disease (SnNout)
Specificity	TN / (TN + FP)	Rule-in: a positive test rules in disease (SpPin)
PPV	TP / (TP + FP)	Probability of disease given a positive test
NNT	1 / ARR	Number of patients to treat to prevent one bad outcome

Hypothesis testing and bias

Set H₀ (no difference) against H₁; α = 0.05 is the Type I error threshold, power = 1 − β is set at 0.80 by convention. Randomisation and blinding counter selection, information, and confounding bias — the three traps most tested in Saudi GP Board MCQs.

🔴 Extended — Deep Study (3mo+)

Comprehensive coverage for students on a longer study timeline.

Sample size and confidence intervals

For comparing two proportions, n = [Z_{α/2}√(2p̄q̄) + Z_β√(p₁q₁ + p₂q₂)]² / (p₁ − p₂)², with p̄ = (p₁ + p₂)/2. For estimating prevalence, n = Z²·P(1−P) / d², where d is the desired margin of error. Underestimating cluster effects or expected dropout (typically 10–20%) is the single most common reason family-medicine audits finish underpowered.

Intention-to-treat vs per-protocol

Intention-to-treat (ITT) analyses every randomised patient in their allocated group, preserving randomisation and reflecting real-world effectiveness. Per-protocol analyses only adherers, inflating efficacy but inviting confounding by adherence behaviour. SCFHS-aligned critical-appraisal tools (CONSORT for trials, STROBE for observational studies) explicitly grade this distinction.

Adjacent topics and exam traps

EBM sits beside preventive-medicine screening criteria (Wilson & Jungner) and clinical-practice guideline appraisal (AGREE II). Candidates regularly lose marks by (1) quoting RR reduction without ARR/NNT, (2) treating a non-significant p-value as equivalence, or (3) using OR where RR is requested in an RCT stem.

Bias type	Where it arises	Mitigation
Selection	Non-random allocation	Randomisation, concealed allocation
Information	Differential measurement	Blinding outcome assessors
Confounding	Extraneous prognostic factor	Restriction, matching, multivariate adjustment
Recall	Case-control interviews	Use objective records

Practice prompts

A screening test has 95% sensitivity, 90% specificity, and is applied where disease prevalence is 1%. Calculate NPV and explain why a positive result needs confirmatory testing.
An RCT reports RR = 0.70 (95% CI 0.42–1.08). State the point estimate, interpret the CI, and compute NNT if control event rate is 20%.

Continue your study

View this topic in your Saudi GP Board roadmap — see where “Topic 4” fits in your personalised plan
Build a quick revision plan — 1-day sprint covering highest-weight topics
Saudi GP Board exam overview — pattern, eligibility, and syllabus
All Medicine notes — browse sibling topics in this subject

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

Core measures of disease frequency