---

Epidemiological Study Methods — K Park (PSM)

What is an Epidemiological Study?

---

CLASSIFICATION AT A GLANCE

EPIDEMIOLOGICAL STUDIES
│
├── A. OBSERVATIONAL STUDIES (no intervention)
│   ├── 1. Descriptive Studies
│   │   ├── Case Reports / Case Series
│   │   ├── Cross-Sectional (Prevalence) Study
│   │   └── Ecological (Correlational) Study
│   │
│   └── 2. Analytical Studies
│       ├── Case-Control Study (retrospective)
│       └── Cohort Study (prospective / retrospective)
│
└── B. EXPERIMENTAL STUDIES (intervention done)
    ├── Randomized Controlled Trial (RCT)
    ├── Field Trial
    └── Community Trial

---

A. OBSERVATIONAL STUDIES

1. DESCRIPTIVE STUDIES

---

(a) Case Report / Case Series

• What it is: Detailed description of 1 patient (case report) or a small group (case series).
• Purpose: Raises an alarm; generates hypothesis.
• Example: First AIDS cases described in homosexual men (1981) → led to hypothesis.
• Limitation: No comparison group; cannot prove causation.

---

(b) Cross-Sectional Study (Prevalence Study / Survey)

• What it is: A snapshot — disease and exposure measured at the same point in time in a defined population.
• Purpose: Measures prevalence; useful for planning health services.
• Example: A survey measuring prevalence of hypertension in a district on a single day.

• Advantage: Quick, cheap, gives prevalence data.
• Limitation: Cannot determine cause-effect; unsuitable for rare or acute diseases.

---

(c) Ecological (Correlational) Study

• What it is: Compares disease rates across whole populations (not individuals) with exposure data.
• Example: Countries with high fat intake vs. countries with high breast cancer rates → correlation found.
• Limitation: Ecological fallacy — what is true for a population may not be true for an individual.

---

2. ANALYTICAL STUDIES

---

(a) Case-Control Study (Retrospective)

• Cases = people who already have the disease.
• Controls = people without the disease (similar in other ways).
• Both groups are asked: "Were you exposed to the risk factor in the past?"
• Measure of association: Odds Ratio (OR)

         Cases   Controls
Exposed    a        b
Not Exp    c        d

Odds Ratio = (a×d) / (b×c)

• Good for rare diseases
• Quick, inexpensive
• Can study multiple exposures for one disease

• Subject to recall bias (patients may remember differently)
• Cannot calculate incidence or relative risk directly
• Selection of controls is difficult
• Temporal relationship hard to establish

---

(b) Cohort Study (Prospective / Longitudinal / Follow-up Study)

• Take a group of disease-free people.
• Divide them into exposed and non-exposed.
• Follow up over time → see who develops the disease.
• Measure of association: Relative Risk (RR) = Incidence Rate Ratio

              Disease   No Disease
Exposed         a           b
Not Exposed     c           d

Incidence in exposed    = a/(a+b)
Incidence in unexposed  = c/(c+d)
Relative Risk (RR)      = [a/(a+b)] / [c/(c+d)]

• Can calculate incidence and relative risk
• Temporal relationship clearly established (cause before effect)
• Can study multiple outcomes from one exposure
• Less recall bias

• Very expensive and time-consuming
• Large numbers needed
• Loss to follow-up (attrition bias)
• Not suitable for rare diseases

Historical Cohort (Retrospective Cohort): Uses old records. Exposure was in the past, disease identified now. Faster and cheaper than prospective.

---

Quick Comparison: Case-Control vs Cohort

---

B. EXPERIMENTAL STUDIES

---

(1) Randomized Controlled Trial (RCT) — Clinical Trial

• Individuals randomly assigned to intervention (treatment) group or control (placebo/standard) group.
• Randomization eliminates selection bias and balances confounders.
• Blinding: Single-blind (patient unaware), Double-blind (patient + investigator unaware), Triple-blind (+ analyst unaware).
• Gold standard for testing therapeutic interventions.
• Example: Testing a new drug for TB — half get the drug, half get placebo.

---

(2) Field Trial

• Conducted in the field (community), not a hospital.
• Participants are healthy people at risk (not yet diseased).
• Purpose: Test preventive measures (vaccines, prophylactic drugs).
• Example: Salk polio vaccine trial — vaccine given to healthy children in field.
• Very expensive; requires huge numbers.

---

(3) Community Trial (Community Intervention Study)

• The community (not the individual) is the unit of randomization/intervention.
• Example: Fluoridation of water supply in some towns vs. not in others → compare dental caries rates.
• Used when individual randomization is not feasible.

---

SPECIAL CONCEPTS TO REMEMBER

Attributable Risk (AR)

• How much of the disease in exposed people is due to the exposure?
• AR = Incidence in exposed − Incidence in unexposed

Population Attributable Risk (PAR)

• How much disease in the whole population is due to the exposure?
• Important for public health policy.

Bias Types (often asked)

---

ONE-LINE SUMMARY FOR QUICK REVISION

---

EXAM TIPS (K Park Specific)

1. Definition of epidemiology (John Last): "Study of the distribution and determinants of health-related states or events in specified populations and the application of this study to control health problems."
2. Case-control is best for rare diseases; cohort is best for rare exposures.
3. Only experimental studies can establish causation.
4. Cross-sectional gives prevalence; cohort gives incidence.
5. Odds ratio approximates relative risk when disease is rare (rare disease assumption).
6. The strongest evidence hierarchy: Systematic Review/Meta-analysis > RCT > Cohort > Case-Control > Cross-Sectional > Case Report.

---