Research, evidence-based practice and quality improvement
This domain tests whether you can find, appraise and apply evidence, and whether you can run a small improvement cycle. Questions ask you to rank evidence, interpret a statistic, or distinguish audit from research.
Asking and answering a question
- Frame the question with PICO: population, intervention, comparison, outcome. A focused question determines the search strategy.
- Levels of evidence, highest first: systematic reviews and meta-analyses of randomised trials, randomised controlled trials, cohort studies, case-control studies, cross-sectional studies, case series, expert opinion.
- Appraise validity (was the design sound, were groups comparable, was allocation concealed and blinded, was analysis by intention to treat) before applying results to your patient.
Statistics you must read
- Sensitivity is the proportion of true cases correctly identified; specificity is the proportion of non-cases correctly excluded. Positive predictive value depends heavily on prevalence, so a good test in a low-prevalence population still yields many false positives.
- Absolute risk reduction is the difference in event rates; number needed to treat is 1 divided by the absolute risk reduction.
- Statistical significance is not clinical importance: a small difference in a huge trial can be significant and irrelevant.
Quality improvement
Use plan-do-study-act cycles, measure before and after, and keep changes small and testable. Audit compares current practice against an agreed standard and closes the loop with re-audit; research generates new generalisable knowledge and needs ethics approval. Incident reporting, root-cause analysis and mortality review are quality tools, not research.
Worked example
A trial reports death from 4% to 2% with a new drug. Reasoning: absolute risk reduction is 2 percentage points, so the number needed to treat is 1 divided by 0.02, which equals 50. Fifty patients are treated to prevent one death; the nurse weighs this against cost, harms and patient preference.
Traps that cost marks
Confusing incidence with prevalence; calling an association causal; treating expert opinion as strong evidence.