Cancer Epidemiology, Prevention, and Screening
Core concepts
Master incidence vs prevalence vs mortality, crude vs age-standardised rates, 5-year survival, and lifetime risk. Understand study designs used in epidemiology - cohort vs case-control vs ecological - and their bias profiles (selection, recall, lead-time, length-time). Know screening principles (Wilson–Jungner): important disease, accepted treatment, detectable latent stage, suitable test, test acceptable, natural history understood, cost-effective. Apply sensitivity, specificity, PPV, NPV, and understand how prevalence drives PPV. Recognise effective screening programmes: mammography (breast), low-dose CT (high-risk smokers), FIT/colonoscopy (colorectal), HPV testing (cervical), and the limits of PSA screening. Understand primary prevention (tobacco control, HPV/HBV vaccination, sun protection, weight/alcohol) and chemoprevention (tamoxifen in high-risk breast; aspirin in Lynch syndrome). Know attributable risk and population attributable fraction.
Traps
Questions confuse lead-time bias (earlier diagnosis without survival benefit) with length-time bias (slow-growing tumours over-represented) and overdiagnosis. PPV is often mis-calculated by ignoring prevalence. Distinguish screening from early diagnosis. Do not assume a test with high sensitivity has high PPV in a low-prevalence population.
Worked example / rule of thumb
Sensitivity 90%, specificity 95%, prevalence 1% in 10,000 people: TP=90, FN=10, FP=495, TN=9,405. PPV = 90/(90+495) ≈ 15%. Rule of thumb: when prevalence is low, even good tests generate many false positives - PPV falls.