Biostatistics & Epidemiology Practice Questions

Biostatistics is only 3–5% of questions but near-guaranteed free points. Absolute risk reduction, cohort design, confidence intervals, and prevalence-sensitivity interactions appear reliably. These four sample questions come from MediLingo's bank of 3,900+ questions across 21 specialties.

Sample questions

Question 1

Stroke risk is 20% on placebo and 10% on treatment. What is the absolute risk reduction?

a) 50%, calculated as 10%/20%
b) 10 percentage points, calculated as 20% − 10% ✓ Correct
c) 2.0, calculated as 20%/10%
d) 5%, calculated as (20% − 10%)/2

Key teaching point: ARR = control risk − treatment risk (the raw difference).

Question 2

Investigators enroll 5,000 factory workers exposed to silica and 5,000 unexposed workers, then follow both groups for 10 years to compare lung disease rates. What design is this?

a) Case-control study
b) Prospective cohort study ✓ Correct
c) Cross-sectional study
d) Randomized controlled trial

Key teaching point: Cohort: start with exposure, follow forward to outcome → relative risk.

Question 3

A study estimates an odds ratio of 2.0 with a 95% CI of 0.5 to 8.0. What does this wide interval mainly reflect?

a) Imprecision, likely from a small sample ✓ Correct
b) Strong evidence of no effect
c) High statistical power
d) Confounding

Key teaching point: Wide CI = imprecise estimate, usually a small sample.

Question 4

A test with 90% sensitivity and 90% specificity is moved from a high-prevalence clinic to a low-prevalence screening population. What happens to its sensitivity and specificity?

a) Both fall
b) Sensitivity falls, specificity rises
c) Sensitivity rises, specificity falls
d) They stay the same ✓ Correct

Key teaching point: Sensitivity and specificity are prevalence-independent; predictive values are not.

Keep going — 3,900+ questions

With streaks, XP and duels. Free for med students in 72 countries; 30-day free trial elsewhere.

More practice questions

© 2026 MediLingo. For educational purposes only. Questions are AI-generated and multi-layer reviewed.