Research methods, evidence-based practice, and biostatistics
Domain 8 - Research methods, evidence-based practice, and biostatistics
This domain tests critical appraisal: study design, bias, effect measures, and applying evidence to an individual patient.
Core rules
- Hierarchy: systematic reviews and RCTs for therapy; cohort studies for prognosis and rare exposures; diagnostic accuracy studies for tests.
- Intention-to-treat preserves randomization; per-protocol analysis can exaggerate effect.
- Report effect size with confidence intervals, not just p-values.
- NNT = 1/ARR, where ARR = control event rate minus treatment event rate.
- Check validity: randomization, blinding, follow-up, confounding, and applicability.
Worked example
An RCT reports mortality of 10% with control and 8% with treatment, RR 0.80 (95% CI 0.65–0.98), p = 0.03. ARR = 2%, so NNT = 50. The confidence interval excludes 1, so the result is statistically significant, but the upper bound is close to no effect; consider cost, adverse effects, and whether your patient resembles the trial population.
Traps
- Treating p < 0.05 as proof of clinical importance.
- Confusing odds ratio with risk ratio.
- Ignoring confidence intervals and loss to follow-up.
- Applying trial results to patients excluded from the trial.