--- module: 065-02 language: en chapter: 65 title: "Epidemiology, Screening, Vaccination, Prevention, and Population Health" module_title: "Population measurement, screening trade-offs, vaccine strategy, and equitable prevention" source_sha256: 553b91ab51b1246d8e2411422ec8efc104a7b1480ff4a4e5017079c3fe1ba413 --- # Population measurement, screening, vaccines, and equity ## The population frame ### Attention moves to distributions ### Air, work, housing, income decide patient numbers ### Who reaches it, who bears the burden ## Measuring occurrence ### Cumulative incidence: probability over interval ### Incidence rate: events per person-time ### Prevalence rises with incidence or duration #### Treatment extends survival, raises prevalence ### Case fatality changes with milder diagnosis ### Crude rates reflect age and sex structure ### Direct standardisation: common reference population ### Indirect: observed versus expected events #### Not the actual local burden ## Causation and effect size ### Association is not causation ### Confounder links exposure and outcome ### Randomisation averages known and unknown confounding #### Loss to follow-up and crossover remain ### Observational adjustment leaves unmeasured confounding ### Relative risk looks impressive for rare events ### Number needed to treat inherits time horizon #### Changes with baseline risk ## Screening justification and biases ### Offered without symptoms for patient-important outcomes ### Earlier finding alone is not justification ### Assay with poor follow-up is not a programme ### Lead-time bias: death unchanged ### Length bias finds slower lesions ### Overdiagnosis: real but harmless pathology #### Better confirmation cannot correct it #### Overtreatment follows ### Randomised harm reduction beats stage or survival ## Screening thresholds and decisions ### Low prevalence magnifies false positives #### Imaging, biopsy, anxiety, labels ### Higher threshold: specificity up, disease missed #### Delayed diagnosis versus unnecessary follow-up ### Same period and denominator for all outcomes ### Symptomatic testing is not screening ### Declining is legitimate and revisable ### Uptake easiest among the well served #### Outreach, interpreters, navigation needed ### Track confirmation and outcomes by group ## Vaccine platforms and effectiveness ### Live attenuated: broad response #### Unsafe in cellular immunodeficiency, pregnancy ### Non-live platforms cannot replicate infection ### Conjugation enables memory in young children ### Efficacy in trials, effectiveness in practice ### Severe disease better protected than infection ### Herd threshold rarely exact ## Vaccine delivery and safety ### Right product, person, and interval ### Cold chain deviation not judged by appearance ### Temporal association is not causation #### Stimulated reporting creates clusters #### Compare observed with expected rates ### Contraindications narrower than myths ### Non-live vaccines before immunosuppression ### Catch-up continues, not restarts ## Cardiovascular prevention ### Based on absolute risk #### Same lipid fall prevents more at high risk ### Confirm blood pressure before long-term labels ### Risk calculators are not individual probability #### Miss structural disadvantage ### Lifestyle is not a precondition for drugs ### Aim: fewer infarcts, strokes, kidney failure ### Overtreatment if benefit lies beyond life expectancy #### Frailty, falls, polypharmacy, bleeding ### Neither stop by age nor continue automatically ## Cancer, infection, and dependence ### Screening differs by organ and risk #### Ovarian markers, thyroid ultrasound overdiagnose ### Interrupt the actual transmission route ### Resistance prevention starts with fewer infections ### Nicotine replacement separates nicotine from combustion ### Relapse prompts another treatment attempt ### Opioid harm reduction keeps abstinence access ## Equity and implementation ### Social categories are not fixed biology ### Equity modifies resources for comparable benefit ### Averages conceal worsening minorities ### Separate invited, participating, lost, harmed #### Absolute and relative gaps may diverge ### Examine system mechanisms, not culture ### Process measures matter only with outcomes ### Transition plan preserves care and data