Research Methodology, Epidemiology and Evidence-Based Practice
Study designs
Randomised controlled trials, cohort, case-control, cross-sectional, ecological, and qualitative designs. Know strengths and weaknesses, selection bias, confounding, and the meaning of internal and external validity.
Measures of effect
Incidence, prevalence, relative risk, odds ratio, absolute risk reduction, number needed to treat, sensitivity, specificity, positive and negative predictive values. Understand that predictive values depend on prevalence.
Appraisal and synthesis
Levels of evidence, systematic reviews and meta-analysis, heterogeneity, publication bias, confidence intervals, p-values, and the difference between statistical and clinical significance. Understand screening principles (Wilson-Jungner criteria).
Traps
Confusing odds ratio with relative risk; interpreting a p-value as the probability the hypothesis is true; ignoring prevalence when interpreting a screening test; confusing correlation with causation.
Worked rule of thumb
If a test has 90% sensitivity and 90% specificity in a population with 1% prevalence, most positives will be false positives. Always calculate or estimate positive predictive value using prevalence.