Research Methodology, Evidence Appraisal, Ethics and Patient Safety
Domain 12 tests critical appraisal, study design and professional conduct rather than dermatology facts alone.
Evidence and study design
- Randomized controlled trial for therapy, cohort for prognosis, case-control for rare exposures, cross-sectional for prevalence, diagnostic accuracy studies for tests.
- Sensitivity is true positives divided by all diseased; specificity is true negatives divided by all healthy. Positive predictive value depends on prevalence. Likelihood ratios summarise test performance.
- Bias: selection, recall, confounding and measurement. Intention-to-treat analysis and blinding protect validity.
- Levels of evidence (Oxford CEBM) and GRADE inform guideline strength.
Ethics and safety
- Informed consent requires capacity, information and voluntariness; document the discussion.
- Confidentiality, conflicts of interest, vulnerable populations, ethics committee approval, Declaration of Helsinki.
- Patient safety: allergy documentation, wrong-site surgery prevention, medication reconciliation, laser and smoke safety, tracking biopsies and critical results, incident reporting and root-cause analysis.
Worked example
A test has 90% sensitivity and 95% specificity; prevalence is 1%. In 10,000 people there are 90 true positives among 100 diseased and 495 false positives among 9,900 healthy. Positive predictive value is 90 divided by 585, about 15%. Reasoning: low prevalence makes positive results mostly false, so confirmatory testing is needed.
Traps
Do not present sensitivity as the chance of disease. Do not ignore prevalence when interpreting a screening test. Do not obtain consent retrospectively for a procedure already performed.