Epidemiology Terms: Sensitivity, Specificity & Risk Measures

Key Epidemiology and Biostatistics Terms

This document defines core epidemiology and biostatistics concepts used in clinical research and public health.

Diagnostic Test Measures

• Sensitivity: The probability that an individual with the disease tests positive. High sensitivity is useful to rule out disease when the test is negative.
• Specificity: The probability that a healthy individual tests negative. High specificity is useful to confirm disease when the test is positive.

Population Prevalence and Incidence

• Prevalence: The total number of existing cases in the population divided by the total population at a given time.
• Incidence: The occurrence of new cases of a disease in a population over a specified period.
• Cumulative incidence (individual risk): The probability or risk that an individual will develop the disease over a specified time period.

Agreement and Reliability

• Kappa coefficient (Cohen's kappa): Measures the degree of nonrandom agreement between observers. Values range from -1 to +1, where higher positive values indicate better agreement beyond chance.

Measures of Association and Effect

• Relative Risk (RR): Used in cohort studies; measures the strength of association between exposure (risk factor) and disease. RR < 1 suggests a protective factor, RR > 1 suggests increased risk, RR = 1 indicates no association.
• Odds Ratio (OR): Commonly used in case-control studies; expresses the odds of exposure among cases relative to controls. OR is especially useful when the disease is rare.
• Absolute Risk Reduction (ARR): The absolute difference in event rates between control and treatment groups; used to evaluate the direct reduction in the number of new cases due to a preventive or therapeutic measure.
• Relative Risk Reduction (RRR): The proportional reduction in event rates between groups; typically expressed as a percentage of cases prevented by the intervention.
• Number Needed to Treat (NNT): The number of subjects who must receive the intervention to prevent one additional adverse outcome (or to achieve one additional cure). NNT = 1 / ARR.

Study Designs

Descriptive and Ecological Studies

• Descriptive studies: Describe the characteristics, patterns, and frequency of a health problem in a population (person, place, time).
• Ecological studies: Analyze data at the group or cluster level (not individual-level data) and can identify associations that require further individual-level study.

Cross-Sectional (Prevalence) Studies

• Cross-sectional studies (prevalence studies): Examine the relationship between a disease (often chronic) and variables at a single point in time in a population.

Analytical Studies: Observational and Experimental

• Analytical studies: Aim to establish associations between risk factors and disease and can be observational or experimental.
• Observational analytical studies:
  • Cohort studies: Prospective follow-up studies that move from exposure (cause) to outcome (effect); they allow measurement of incidence and RR.
  • Case-control studies: Compare individuals with the disease (cases) to similar individuals without the disease (controls) to assess prior exposures; they estimate OR.
• Experimental studies (clinical trials): The investigator assigns the exposure or intervention and typically randomizes the sample. Randomized controlled trials provide strong evidence for causality when well designed.

Types of Errors and Bias

• Random error: Variability due to chance, reduced by larger sample sizes and precision.
• Systematic error (bias): Nonrandom error that can distort results. Common types include:
  • Selection bias: Systematic differences in characteristics between those selected and those not selected for the study.
  • Information bias: Systematic errors in measurement or classification of exposure or outcome.
  • Confounding: Extraneous variables that are associated with both the exposure and outcome and can distort the observed association if not controlled.

Clinical Trial Phases

• Phase I: First administration to humans (often healthy volunteers, n ≈ 20–80). Focuses on safety, tolerability, and initial pharmacokinetics; may include single-dose toxicity assessment. Typically no control group.
• Phase II: Provides information on dose–response, efficacy signals, and further safety in a larger group of patients.
• Phase III: Large randomized trials to establish efficacy and monitor adverse effects compared with standard treatment or placebo; data used for regulatory approval.
• Phase IV (post-marketing surveillance): Pharmacovigilance after market authorization to detect uncommon or long-term adverse reactions not identified in earlier phases.

Note: The definitions above are concise explanations of fundamental concepts used in clinical epidemiology and research methodology. Proper study design, careful measurement, and appropriate analysis are essential to obtain valid and clinically useful results.