I now have comprehensive content from Park's Textbook. Let me compile the full structured answer.
PSM - Epidemiology & Biostatistics
Based on Park's Textbook of Preventive and Social Medicine
1. DEFINITION & AIMS OF EPIDEMIOLOGY
Epidemiology is the study of the distribution and determinants of health-related states or events in specified populations, and the application of this study to the control of health problems.
Three main aims (IEA):
- Describe the distribution and magnitude of health/disease problems in populations
- Identify aetiological factors (risk factors) in disease pathogenesis
- Provide data for planning, implementing, and evaluating health services
Key distinction from clinical medicine: Epidemiology studies populations; clinical medicine studies individuals. The epidemiologist is interested in the relationship between cases and the population - expressed as a rate.
2. TYPES OF EPIDEMIOLOGICAL STUDIES
A. Descriptive Epidemiology
Describes distribution of disease by:
- When - time distribution
- Where - place distribution
- Who - person distribution
Descriptive studies are usually the first phase of investigation. They formulate hypotheses rather than test them.
B. Analytical Epidemiology
Tests hypotheses. Studies individuals within a population. Two main types:
1. Case-Control Study ("Retrospective")
- Both exposure and outcome have already occurred before study starts
- Proceeds backwards from effect to cause
- Starts with cases (with disease) and controls (without disease), then looks back at exposure
- Measure of association: Odds Ratio (OR) = ad/bc
- OR approximates Relative Risk when the disease is rare
Advantages: Rapid, cheap, few subjects needed, good for rare diseases, can study multiple aetiological factors, no follow-up needed
Disadvantages: Recall bias, selection of control group difficult, cannot calculate incidence directly, one outcome at a time
Biases specific to case-control:
- Recall/Memory bias - cases remember past exposure better than controls
- Selection bias - cases/controls may not represent the general population
- Berkesonian bias - different hospital admission rates for different diseases
- Interviewer bias - knowing who is a case leads to more thorough questioning (eliminated by double-blinding)
- Confounding bias - controlled by matching
2. Cohort Study ("Prospective")
- Starts with exposed and non-exposed people (all disease-free)
- Follows forward in time to see who develops disease
- Directly calculates Relative Risk (RR)
- RR = Incidence among exposed / Incidence among non-exposed
- RR > 1 = positive association; RR = 1 = no association; RR < 1 = protective
Advantages: Direct RR calculation, temporal relationship clear, dose-response can be calculated, can study multiple outcomes
Disadvantages: Expensive, time-consuming, attrition (loss to follow-up), inefficient for rare diseases
C. Experimental Epidemiology
- Investigator controls conditions - applies/withdraws the intervention
- Gold standard: Randomised Controlled Trial (RCT)
- Experimental group vs. control group; outcome compared
- Can be in animals or humans
- Advantages: strongest causal evidence
- Disadvantages: costly, ethical issues, feasibility
3. MEASURES OF DISEASE FREQUENCY
Incidence Rate
"The number of new cases occurring in a defined population during a specified period of time"
Formula:
$$\text{Incidence Rate} = \frac{\text{Number of new cases in a given period}}{\text{Population at risk during that period}} \times 1000$$
- Restricted to acute conditions
- Must always state the unit of time (e.g., per 1000 per year)
Special incidence rates:
| Rate | Formula | Use |
|---|
| Attack Rate | New cases / Population at risk × 100 | During epidemics |
| Secondary Attack Rate | New cases among contacts / Total susceptible contacts × 100 | Household spread |
Prevalence Rate
"All current cases (old and new) existing at a given point in time or over a period"
- More accurately a ratio, not a rate
Types:
- Point prevalence - all cases at a single point in time (most common use)
- Period prevalence - all cases during a defined period
Formula (point prevalence):
$$\text{Point Prevalence} = \frac{\text{All current cases at a given point in time}}{\text{Estimated population at same point}} \times 100$$
Prevalence-Incidence Relationship
$$\boxed{P = I \times D}$$
- P = Prevalence, I = Incidence, D = Mean duration of disease
- Long duration disease → high prevalence (e.g., tuberculosis)
- Short/rapidly fatal disease → prevalence ≈ incidence (e.g., food poisoning)
- Analogy: Prevalence = photograph (snapshot); Incidence = film (continuous record)
4. MEASURES OF ASSOCIATION
| Measure | Used In | Formula | Interpretation |
|---|
| Relative Risk (RR) | Cohort study | Incidence (exposed) / Incidence (unexposed) | Strength of association |
| Odds Ratio (OR) | Case-control study | ad/bc | Approximates RR for rare diseases |
| Attributable Risk (AR) | Cohort | Incidence (exposed) - Incidence (unexposed) | Excess risk due to exposure |
| Population Attributable Risk | Cohort | AR × Prevalence of exposure | Impact at population level |
5. SCREENING & DIAGNOSTIC TESTS
2x2 Contingency Table
| Disease + | Disease - |
|---|
| Test + | a (True Positive) | b (False Positive) |
| Test - | c (False Negative) | d (True Negative) |
Key Formulae
| Parameter | Formula | Meaning |
|---|
| Sensitivity | a/(a+c) × 100 | Ability to detect true positives - "does the test pick up disease?" |
| Specificity | d/(b+d) × 100 | Ability to detect true negatives - "does the test exclude disease?" |
| PPV (Positive Predictive Value) | a/(a+b) × 100 | If test is +ve, probability truly has disease |
| NPV (Negative Predictive Value) | d/(c+d) × 100 | If test is -ve, probability truly disease-free |
| False-negative rate | c/(a+c) × 100 | = 1 - Sensitivity |
| False-positive rate | b/(b+d) × 100 | = 1 - Specificity |
Example from Park's (hypothetical figures):
- a=40, b=20, c=100, d=9840, Total=10,000
- Sensitivity = 40/140 × 100 = 28.57%
- Specificity = 9840/9860 × 100 = 99.79%
- PPV = 40/60 × 100 = 66.66%
- NPV = 9840/9940 × 100 = 98.9%
Key Points About Screening
- Sensitivity and specificity are properties of the test itself - they do not change with disease prevalence
- PPV increases with increasing prevalence of disease in the population - the same test has higher PPV in a high-risk group
- High sensitivity = good for ruling OUT disease (SNOUT: Sensitive test, Negative result, rules OUT)
- High specificity = good for ruling IN disease (SPIN: Specific test, Positive result, rules IN)
Difference: Screening vs. Case-finding vs. Diagnosis
- Screening - testing asymptomatic people NOT seeking care (e.g., neonatal screening)
- Case-finding - testing people seeking care for OTHER reasons (e.g., VDRL in antenatal clinic)
- Diagnosis - confirming disease in people with signs/symptoms
6. BIAS & CONFOUNDING
Bias = any systematic error in determining the association between exposure and disease.
Types of Bias
| Type | Description | Control |
|---|
| Recall/Memory bias | Cases recall past exposures better than controls | Objective records |
| Selection bias | Cases/controls not representative of general population | Careful selection criteria |
| Berkesonian bias | Different hospital admission rates for different diseases | Population-based studies |
| Interviewer bias | Interviewer probes cases more thoroughly knowing the hypothesis | Double-blinding |
| Confounding | Third variable associated with both exposure and outcome | Matching, stratification, multivariate analysis |
Confounding is a key concept: A confounder is a variable that is associated with both the exposure and the disease but is not on the causal pathway. For example, in studying alcohol and lung cancer, smoking is a confounder.
7. QUICK REVISION TABLE - Study Designs at a Glance
| Feature | Cross-sectional | Case-Control | Cohort | RCT |
|---|
| Direction | Snapshot | Backward | Forward | Forward |
| Measure | Prevalence | OR | RR, AR | RR |
| Cost | Cheap | Moderate | Expensive | Very expensive |
| Time | Short | Short | Long | Varies |
| Rare disease | Poor | Good | Poor | Poor |
| Rare exposure | Poor | Poor | Good | Good |
| Causality proof | Weakest | Moderate | Strong | Strongest |
| Bias risk | Moderate | High | Moderate | Lowest |
Source: Park's Textbook of Preventive and Social Medicine, 9789382219163