Epidemiology
Core concepts
Epidemiology questions test whether you can move fluently between measures of frequency (incidence, prevalence), measures of association (risk ratio, rate ratio, odds ratio), and measures of impact (attributable risk, population attributable fraction). Know that incidence requires a population at risk and a defined time period, whereas prevalence is a snapshot and rises with longer duration of disease. Distinguish cumulative incidence from incidence rate (person-time denominator).
Study designs and bias
Be able to rank designs by strength of evidence and to name the bias each invites: selection bias in case-control sampling, recall bias in retrospective interviews, loss to follow-up and healthy worker effect in cohorts, and confounding whenever a third variable is associated with both exposure and outcome without lying on the causal pathway. Understand effect modification as a real biological interaction (report stratum-specific effects) versus confounding (adjust it away).
Causality and screening
Memorise the Bradford Hill viewpoints but remember they are considerations, not a checklist. For screening, apply Wilson-Jungner criteria and know lead-time bias, length bias and overdiagnosis - the classic trap is a screening programme that improves survival time without reducing mortality.
Worked rule of thumb
In a cohort, if 20 of 100 exposed and 10 of 100 unexposed develop disease, risk ratio = 0.20/0.10 = 2.0; attributable risk = 0.10 (10 per 100 exposed); population attributable fraction ≈ attributable risk × prevalence of exposure in the population divided by overall risk. A common exam trap is confusing odds ratio with risk ratio when the outcome is common - they diverge, so state that OR overstates RR when disease is frequent.