Research Methods, Statistics, and Evidence-Based Practice
What this domain tests
This domain tests whether you can read a study critically, judge whether it applies to your patient, and translate findings into practice without overstating them.
Core rules
- Know the threats to internal validity: selection, attrition, maturation, regression to the mean, instrumentation, and history.
- In trials, look for randomisation, allocation concealment, blinding, and intention-to-treat analysis. High attrition weakens any result.
- Effect sizes travel better than p-values: Cohen's d of roughly 0.2, 0.5, and 0.8 for small, medium, and large. Number needed to treat equals 1 divided by the absolute risk reduction.
- A p-value is the probability of the data (or more extreme) given the null hypothesis, not the probability that the hypothesis is true.
- Evidence-based practice combines best available evidence, clinical expertise, and the patient's values and context. Guidelines inform, they do not dictate.
Worked example
In a trial, 40% of the treatment group respond versus 25% of controls. Absolute risk reduction is 15 percentage points, so NNT = 1/0.15 ≈ 6.7, meaning about seven patients must be treated for one extra response. Reasoning: report the absolute figure, because relative reductions of 37.5% sound larger and hide the base rate.
Traps that cost marks
"Trend toward significance"; equating correlation with causation; treating a small p-value as a large effect; confusing odds ratios with risk ratios; ignoring publication bias and small sample sizes; and applying trial results to patients excluded from the trial.