Research, Evidence-Based Practice and Biostatistics: Critical Appraisal, Study Design and Interpretation of Clinical Evidence
What the exam tests
Critical appraisal, study design, and interpretation of clinical evidence.
Core rules and concepts
- Levels of evidence: systematic reviews and meta-analyses, randomised controlled trials, cohort studies, case-control studies, case series.
- Validity: randomisation, allocation concealment, blinding, intention-to-treat analysis, and confounding.
- Diagnostic statistics: sensitivity, specificity, positive and negative predictive values; predictive values depend on prevalence.
- Statistics: a p-value is not clinical significance; use 95% confidence intervals, MCID, and number needed to treat.
- Applicability: consider the Kuwait context, patient values, and resource availability.
Worked example
A study reports 90% sensitivity and 80% specificity for a diagnostic test; prevalence is 10%. Reasoning: positive predictive value is low despite high sensitivity because prevalence is low. Interpret the test with prevalence in mind and do not overestimate the post-test probability.
Traps that cost marks
Equating statistical significance with clinical importance; ignoring prevalence; overlooking confounding; ignoring publication bias.