---

Epidemiological Study Designs

---

Overview: The Hierarchy of Study Designs

Observational                    Experimental
├── Descriptive                  └── Randomized Controlled Trial (RCT)
│   ├── Cross-sectional               └── Field Trial
│   └── Longitudinal                  └── Community Trial
└── Analytical
    ├── Case-Control
    └── Cohort

---

1. Cross-Sectional Study (Prevalence Study)

• Simplest form of observational study
• Also called a "prevalence study" - measures disease prevalence, not incidence
• Often likened to a "photograph" of the population at a single moment

• Measuring the magnitude (burden) of disease in a community
• Studying distribution of disease by person, place, and time
• Generating causal hypotheses to be tested by analytical studies
• More useful for chronic diseases than for short-lived (acute) ones

• Cannot establish time sequence - you cannot determine whether exposure preceded disease (no causal inference)
• Does not yield incidence rates
• Provides little information on natural history of disease

---

2. Case-Control Study (Retrospective Study)

1. Both exposure and disease have already occurred before the study begins
2. Study proceeds backward - from effect to cause
3. Uses a control/comparison group to support or refute an inference

• Must be comparable with cases for known confounding factors (age, sex, occupation, social class)
• Controls represent the population that gave rise to the cases

• Relatively quick and inexpensive
• Requires fewer subjects
• Useful for rare diseases and diseases with long latency
• Can study multiple exposures simultaneously

• Susceptible to recall bias (cases may remember past exposures differently than controls)
• Susceptible to selection bias in choosing controls
• Cannot calculate incidence rates directly
• Cannot study rare exposures

---

3. Cohort Study (Prospective / Longitudinal Study)

1. Cohorts are identified prior to the appearance of disease
2. Study groups are observed over time for frequency of disease
3. Study proceeds forward - from cause to effect

• Framingham Heart Study (US Public Health Service) - prospective, cardiovascular risk factors
• Doll and Hills - prospective study on smoking and lung cancer
• Royal College of GPs - oral contraceptives and health outcomes

• Good prior evidence of exposure-disease association (from case-control studies)
• Exposure is rare, but incidence is high among the exposed (e.g., X-ray workers)
• Follow-up is feasible (stable, accessible cohort)
• Ample funds available

• Can calculate true incidence rates and relative risk
• Time sequence is clear (exposure before disease)
• Minimizes recall bias
• A well-designed cohort study is the most reliable observational method for showing association between a risk factor and disease

• Expensive and time-consuming (may take decades for chronic diseases)
• Subject to losses to follow-up
• Not suitable for rare diseases (need very large numbers)

---

4. Randomized Controlled Trial (RCT)

• Randomization ensures both groups are comparable at baseline, controlling for known AND unknown confounders
• Blinding reduces observer and subject bias:
  • Single-blind: Subject unaware of allocation
  • Double-blind: Both subject and investigator unaware
  • Triple-blind: Subject, investigator, and data analyst are unaware
• Direction: Cause → Effect (prospective, experimental)
• Measure: Can calculate absolute risk reduction, NNT, relative risk
• Gold standard for evaluating therapeutic interventions

• Field trial - conducted in the community on healthy individuals (e.g., vaccine trials)
• Community trial - entire communities randomized (e.g., fluoridation of water)
• Clinical trial - patients with disease are randomized

---

Comparison at a Glance

---

Biases to Know

---

Evidence Hierarchy (Exam Favourite)

Systematic Reviews & Meta-analyses (highest)
         ↓
Randomized Controlled Trials
         ↓
Cohort Studies
         ↓
Case-Control Studies
         ↓
Cross-Sectional Studies
         ↓
Case Reports / Expert Opinion (lowest)

---