Core Epidemiology Concepts: Definitions, History, and Measures
Lectures 1–2 – Intro to Epidemiology
Key Definitions:
Epidemiology: study of frequency, distribution, determinants of health/disease in populations + application to control.
Frequency: how often events occur (counts, proportions, rates).
Distribution: patterns by person/place/time.
Determinants: factors influencing occurrence (host/agent/environment).
Historical Figures (name + contribution):
Edward Jenner: cowpox → smallpox vaccine (1st vaccination).
James Lind: scurvy + citrus trial.
Percival Pott: chimney sweeps + scrotal cancer (carcinogens).
John Snow: cholera + Broad Street pump (natural experiment).
Ignaz Semmelweis: handwashing → puerperal fever reduction.
Epi helps: identify patterns, test hypotheses, evaluate interventions, guide prevention/control.
Lecture 3-4: Health, Disease, Prevention
Health: complete physical/mental/social well-being (not just no disease).
Disease: deviation from normal structure/function w/ specific signs/symptoms.
Infected: pathogen present. Asymptomatic: infected, no clinical signs.
Epidemiologic Triad:
Host: gets disease (age, immunity, genetics).
Agent: causes disease (infectious/non-infectious).
Environment: enables transmission (climate, crowding).
Natural History Timeline:
Exposure → Latent period (infection→infectious) → Incubation (infection→symptoms) → Infectious period → Recovery/disability/death.
Prevention Levels:
Primary: prevent occurrence (vaccination, sanitation).
Secondary: early detection (screening)
.Tertiary: reduce complications (rehab).
5 Ws (Descriptive Epi): What/Who/Where/When/Why-How.
Temporal Patterns:
Sporadic: occasional/irregular.
Endemic: constant expected level (ex: 14% osteoarthritis).
Epidemic: > expected. Pandemic: worldwide.
Transmission:
Direct: immediate (bite, contact).
Indirect: airborne/vehicle/vector.
Common vehicle: single contaminated source (salad mix).
Herd immunity: susceptibles protected when enough immune (random mixing assumption).
Lecture 5 – One Health
Interconnected health of humans/animals/environment requiring collaborative, transdisciplinary approach.
Principles:
Systems thinking: interconnected components w/ feedback loops.
Transdisciplinarity: integrate disciplines + stakeholder knowledge.
Benefits: Early zoonosis detection, better resources, complex problem solutions (AMR).
Challenges: Coordination, funding, data sharing, power imbalances.
Lectures 6–8 – Disease Frequency
Why measure? Describe burden, compare groups/time, evaluate interventions.
Proportion = cases/population (0-1 or %). Rate = events/person-time (has time unit).
Prevalence = Existing cases / Population
Point: at single time. Period: during interval.
Tells: probability of HAVING disease (healthcare burden).
Ignores: when disease occurred (not risk of getting).
Incidence Risk = New cases / (Initial NAR – 0.5 × Withdrawals)
Probability of DEVELOPING disease in time t.
NAR = disease-free + susceptible at t=0.
Ex: “67% incidence risk” = 67% initially healthy develop it.
Incidence Rate = New cases / Total person-time at risk
Exact denominator: sum individual time-at-risk.
Approximate: [0.5 × (initial NAR + final NAR)] × ITC
ITC = observation period length.
Interpretation: “X cases per Y person-months” = expect X cases if follow Y people 1 month.
Quick Comparison Table
Measure | What | Numerator | Denominator | Time? |
Prevalence | Existing | All cases | Population | No |
Inc Risk | New prob | New cases | At-risk pop | Implicit |
Inc Rate | New speed | New cases | Person-time | Yes |
Relationship (steady state): Prevalence ≈ Incidence × Duration
Other Measures (all incidence unless noted)
Attack Rate = New cases among exposed / Total exposed
Case Fatality = Deaths from X / Cases of X
Morbidity Risk = New illness cases / At-risk pop
Mortality Risk = New deaths / At-risk pop
Crude: all causes. Cause-specific: one cause.
CFR/attack rate: NOT true rates (no time denominator).
Screening: healthy → detect unrecognized disease. Diagnostic: sick → confirm/treat.
2×2 Table
Disease+ Disease- Total
Test+ a(TP) b(FP) a+b
Test- c(FN) d(TN) c+d
Total a+c b+d n
Formulas (know calc + interpret):
True Prev = (a+c)/n
App Prev = (a+b)/n
Sensitivity = a/(a+c) [Of diseased, % test+]
Specificity = d/(b+d) [Of non-diseased, % test-]
PPV = a/(a+b) [Of test+, % truly diseased]
NPV = d/(c+d) [Of test-, % truly disease-free]
False Results:
1-Sn = False Neg fraction. 1-Sp = False Pos fraction.
Choose test by context:
Minimize FN → High Sn (ex: contagious MRSA).
Minimize FP → High Sp (ex: cull entire flock).
Predictive values depend on: Prevalence ↑ → PPV ↑, NPV ↓.
Series vs Parallel
Series: Both+ to call positive → ↑Sp (↓FP), ↓Sn (↑FN) [Rule-IN]
Parallel: Any+ to call positive → ↑Sn (↓FN), ↓Sp (↑FP) [Rule-OUT]
Validity/Precision:
Validity = distinguishes D vs D- (Sn/Sp).
Precision = repeatability (SD for continuous, Kappa for categorical).
Kappa (agreement beyond chance):
0-0.2 slight, 0.4-0.6 moderate, 0.6-0.8 substantial, 0.8-1 perfect.