Scholar
Scholar
The Scholar domain tests evidence-based practice: critical appraisal, research design, statistics, teaching, and publication ethics.
Core rules
- Convert uncertainty into a PICO question; choose the design that fits: RCT for efficacy, cohort for prognosis, case-control for rare exposures, diagnostic accuracy study for tests.
- Appraise bias, confounding, blinding, allocation concealment, intention-to-treat, and missing data.
- Interpret effect sizes with confidence intervals; know NNT/NNH, sensitivity, specificity, predictive values, and likelihood ratios.
- Follow reporting standards: CONSORT for trials, PRISMA for reviews, STARD for diagnostic studies, STROBE for observational studies.
- Thesis and publication: ethics approval, sample size, analysis plan, authorship criteria (ICMJE), plagiarism, and predatory journals.
- Teach with objectives, active learning, and feedback; use journal club and case-based learning.
Worked example
A trial reports response in 50% with drug versus 30% with placebo (ARR 20%). Reasoning: NNT = 1/0.20 = 5. Check the CI, trial duration, blinding, and whether response is clinically meaningful. Ask if your patient would have been eligible and whether adverse effects change the decision. A p-value alone does not answer this.
Traps
- Saying significant when you mean clinically important.
- Confusing association with causation.
- Ignoring dropouts and protocol deviations.
- Presenting a thesis without a pre-specified analysis plan or ethics approval.