Evidence-Based Medicine, Biostatistics, and Drug Information
What the exam tests
Evidence-based medicine questions test critical appraisal, biostatistics and drug information retrieval. You must distinguish statistical significance from clinical importance and apply results to a patient.
Core rules
- Frame questions with PICO: population, intervention, comparator, outcome.
- Randomized controlled trials best answer therapy questions; cohort studies for prognosis; diagnostic accuracy studies for tests.
- Intention-to-treat analysis preserves randomization; per-protocol analysis can exaggerate benefit.
- Absolute risk reduction (ARR) = control event rate − treatment event rate; NNT = 1/ARR. Relative risk reduction can look large when ARR is tiny.
- Confidence intervals show precision; a p-value does not measure effect size.
- Sensitivity and specificity depend on the threshold; likelihood ratios update pre-test probability.
- Hierarchy of drug information: primary literature, secondary databases, tertiary compendia. Verify with more than one source.
Worked example
A trial reports myocardial infarction in 2.0% of controls and 1.4% of treated patients. ARR = 0.6%, so NNT = 167. The relative reduction is 30%, which sounds impressive, but 167 patients must be treated for one year to prevent one event. Compare that with cost, bleeding risk and adherence.
Traps
Quoting relative risk alone; confusing statistical significance with clinical relevance; ignoring loss to follow-up; using surrogate outcomes as if they were patient-important; accepting a single website as drug information.