Clinical Sciences (pathophysiology, biochemistry, genetics, immunology, statistics)
Diagnostic statistics
Sensitivity = TP/(TP+FN); specificity = TN/(TN+FP). Positive predictive value = TP/(TP+FP) and is heavily dependent on prevalence, whereas sensitivity and specificity are not. The positive likelihood ratio = sensitivity/(1 − specificity); a ratio above 10 markedly raises probability. Number needed to treat = 1/absolute risk reduction.
Study design and evidence
Randomised controlled trials best support causality; cohort studies yield relative risk; case-control studies yield odds ratios and are prone to recall bias. Watch for selection bias, lead-time bias in screening, publication bias and confounding. A p-value is the probability of the observed data (or more extreme) if the null hypothesis were true - not the probability that the null hypothesis is true. Confidence intervals convey precision and clinical relevance.
Genetics
Know autosomal dominant (vertical transmission, e.g. familial hypercholesterolaemia, Huntington disease), autosomal recessive (consanguinity, carrier parents), X-linked (affected males, carrier females) and mitochondrial inheritance (maternal). Penetrance and variable expressivity explain why family histories may be misleading.
Immunology and biochemistry
Complement deficiency predisposes to Neisseria infection; IgA deficiency is the commonest primary immunodeficiency; anaphylaxis is IgE-mediated and treated with intramuscular adrenaline.
Worked example: sensitivity 95%, specificity 90%, prevalence 1% in 10,000 people gives 95 true positives and 990 false positives, so PPV is 95/1085 ≈ 9%. Even an excellent test cannot rule in a rare disease - stressing the value of pretest probability.