Key Facts

Epidemiology provides the quantitative framework for understanding disease patterns, evaluating interventions, and informing health policy. Core concepts include disease frequency measures, study designs, measures of association, and assessment of causality.

Disease Frequency Measures

• Incidence rate (person-time): new cases / person-time at risk
• Cumulative incidence (risk): new cases / population at risk over time period
• Point prevalence: existing cases / total population at one time point
• Period prevalence: existing cases over a time period / population
• Relationship: prevalence ≈ incidence × disease duration

Measures of Association

• Relative risk (RR): incidence in exposed / incidence in unexposed (cohort studies)
• Odds ratio (OR): (a/c) / (b/d) in 2×2 table (case-control studies)
• Hazard ratio (HR): instantaneous rate ratio (survival analysis)
• Absolute risk reduction (ARR): risk in control — risk in treatment group
• Relative risk reduction (RRR): ARR / risk in control group
• NNT: 1 / ARR
• NNH: 1 / ARI (absolute risk increase)

Study Design Hierarchy

• Systematic reviews and meta-analyses (highest level)
• Randomised controlled trials
• Cohort studies
• Case-control studies
• Cross-sectional studies
• Case reports and case series (lowest level)

Validity and Bias

• Internal validity: extent to which results are correct for the study population
• External validity (generalisability): extent to which results apply to other populations
• Selection bias: systematic differences between comparison groups
• Information bias: systematic errors in measurement (recall bias, observer bias, detection bias)
• Confounding: extraneous variable associated with both exposure and outcome

Application in Clinical Practice

• Understanding disease prognosis (survival rates, natural history)
• Interpreting diagnostic test performance (sensitivity, specificity, PPV, NPV)
• Evaluating treatment efficacy (NNT, NNH, RRR)
• Screening programme design and evaluation
• Outbreak investigation

Key Concepts for Exams

• Sensitivity: proportion of true positives correctly identified (TP / TP + FN)
• Specificity: proportion of true negatives correctly identified (TN / TN + FP)
• Positive predictive value (PPV): proportion of positive tests that are true positives (depends on prevalence)
• Negative predictive value (NPV): proportion of negative tests that are true negatives
• Likelihood ratio (+): sensitivity / (1 - specificity)
• Likelihood ratio (-): (1 - sensitivity) / specificity

Study Design Selection

• RCT: gold standard for evaluating interventions; randomisation minimises confounding
• Cohort study: follows exposed and unexposed groups over time; measures incidence and RR
• Case-control study: compares exposure in cases vs controls; efficient for rare diseases; measures OR
• Cross-sectional study: measures exposure and outcome simultaneously; measures prevalence; cannot determine causality
• Ecological study: uses population-level data; subject to ecological fallacy

Critical Appraisal

• CONSORT: reporting standard for RCTs
• STROBE: reporting standard for observational studies
• PRISMA: reporting standard for systematic reviews
• GRADE: framework for quality of evidence assessment

Application to Public Health

• Surveillance: monitoring disease trends to detect outbreaks and evaluate interventions
• Screening: applying epidemiological principles to population-level early detection
• Policy: using evidence to inform vaccination programmes, smoking legislation, alcohol pricing
• Outbreak investigation: steps include confirm outbreak, establish case definition, describe epidemiology (time, place, person), generate hypotheses, test hypotheses, implement control measures

Reducing Bias and Confounding

• Randomisation: eliminates confounding in RCTs
• Blinding: reduces observer and participant bias (single, double, triple blind)
• Matching: in case-control studies, match cases and controls on confounders
• Restriction: limit study to subgroup without the confounder
• Stratification: analyse subgroups separately
• Multivariable analysis: statistical adjustment for confounders (regression)

Causation Assessment (Bradford Hill Criteria)

• Strength of association: larger RR/OR more likely causal
• Consistency: replicated across different populations/settings
• Specificity: one exposure leads to one outcome
• Temporality: exposure precedes outcome (only essential criterion)
• Biological gradient: dose-response relationship
• Plausibility: biologically plausible mechanism
• Coherence: consistent with known natural history
• Experiment: intervention reduces outcome
• Analogy: similar exposures cause similar outcomes

• Epidemiological measures directly inform prognosis: survival rates, case-fatality rates, 5-year survival, median survival
• Kaplan-Meier curves: standard method for presenting survival data
• Cox proportional hazards model: identifies prognostic factors (reports hazard ratios)
• Lead-time bias: screening appears to improve survival by detecting disease earlier without actually prolonging life
• Length-time bias: screening preferentially detects slowly progressing disease
• These biases must be accounted for when evaluating screening programme effectiveness

2×2 Table for Diagnostic Tests

• Sensitivity = a / (a+c)
• Specificity = d / (b+d)
• PPV = a / (a+b)
• NPV = d / (c+d)

Levels of Evidence Hierarchy