Epidemiology Basics and History

Epidemiology Basics

• Epidemiology studies the distribution and determinants of disease frequency in human populations. This involves examining how often diseases occur and the factors that influence their occurrence within groups of people.
• Prevalence is the proportion of a population with a disease at a specific point in time. Example: in a population of 20 people, 4 have depression; Prevalence = 4/20 = 20%.
• Incidence is the rate of new cases of a disease in a population during a specific period; in a group of 16 people at risk, 2 develop depression; Incidence = 2/16 = 12.5%.
• Key aims of epidemiological research include planning healthcare, pinpointing disease causes and preventions, identifying high-risk groups, improving overall health, and reducing suffering.
• Psychiatric epidemiology focuses specifically on mental health outcomes in populations.

History of Epidemiology

• Hippocrates is considered the father of modern medicine.
• John Snow investigated cholera, challenging the miasma theory.
• Emile Durkheim was a pioneer in psychiatric epidemiology, studying suicide rates, noting differences in suicide rates between Catholics and Protestants, potentially linked to differing social controls.

Suspected Causes of Illness

• Methods used to investigate suspected causes include: lab studies, disease surveillance, case reports, theoretical reasoning, and clinical observation (determining outcome and exposure).

Making Inferences from Association

• Association does not equal causation.
• To determine if a relationship is causal, consider:
  • Chance (statistical significance): Random variation in results, not necessarily a causal link; we need to understand the uncertainty involved and use statistics.
  • Bias (selection bias, measurement bias): Systematic errors introduced during a study's design or conduction; the results are not a true picture of the target population.
  • Confounding: An alternative reason for an observed link between exposure and outcome; other factors may be influencing the apparent link.
  • Reverse causality: Whether the outcome is actually causing the exposure.

Chance in Epidemiology

• Random error is inherent in studies.
• P-values and confidence intervals (CIs) quantify uncertainty, helping to make inferences about the broader population.
• Larger samples provide more precise CIs (narrower width) and smaller p-values, reducing the likelihood of chance occurrences. A 95% CI is derived from the difference between groups and the standard error. A smaller standard error (achieved with larger sample sizes) indicates a more precise estimate, leading to a narrower confidence interval and a smaller p-value.

Bias

• Bias is systematic error, distorting inferences.
• Selection bias arises from differences between the study sample and the target population (e.g., low response rates, drop-out). Case-control studies are susceptible to this bias.
• Measurement bias (or information bias) arises from inaccurate/inconsistent measurement of exposures or outcomes. This includes observer bias (observer expectations) and recall bias (illness impacting memory).

Confounding

• Confounding is an alternative explanation for an observed association, potentially creating spurious associations or masking true ones.
• Confounders are third variables linked to both the exposure and outcome.

Dealing with Confounding

• Randomization is the ideal approach.
• Adjusting for confounders in statistical analyses. Multivariate techniques used in analyses will change observed associations.
• Residual confounding represents unknown, unaccounted-for confounders.

Reverse Causality

• Reverse causality suggests the outcome might be causing the exposure.
• Longitudinal studies are useful, measuring exposure before the outcome and following individuals over time.

Causal Inference (Bradford Hill)

• Temporality: Exposure must precede the outcome.
• Strength: Stronger associations suggest a higher likelihood of causality.
• Dose-response: Increased exposure is linked to increased risk.
• Consistency: Similar results in various studies across different populations.
• Specificity: A single exposure is associated with a specific outcome.
• Coherence: Alignment with existing scientific knowledge.
• Plausibility: Biological or other logical reasons exist.
• Analogy: Similar situations likely have similar outcomes.
• Experimental evidence: Supported by animal or experimental research.

Study Designs

• Cross-sectional studies: Examine characteristics and prevalence in a population at a specific time. These are often survey-based, and focus on prevalence and disease distribution (age, sex, social factors).
• Ecological studies: Compare disease frequencies between groups over time within a population, using existing data, aiming to expose correlations and patterns.
• Case-control studies: Identify cases and controls to assess associations with exposures.
• Cohort studies: Follow a population over time to measure risk factors and disease development.
• Randomized controlled trials: Participants are randomly assigned to different exposure groups to evaluate cause-and-effect relationships.