Associations and Causality

Environmental Influences and Disease

Focus from supernatural to impact of environmental factors as causes of disease

Theory of Contagion and Modern Understanding

Infection stems from transferrable entities

Miasma Theory's Prevailing Influence

Diseases stemmed from noxious vapors originating from decaying organic material

Spontaneous Generation and its Rejection

Simple life forms could instantly arise from nonliving matter (maggots)

Advancements of Germ Theory

Pasteur: Microbial growth in sterile environments, Koch: systematic approach to linking specific pathogens to distinct diseases

Limitations of Koch

If disease has multiple factors, causative organisms may not be present

Deterministic Models

Specific causes will consistently result in specific effects, like physical laws

Necessary and Sufficient

Cause and effect always present together (rare)

Sufficient but not Necessary

Cause can lead to effect alongside other potential causes, indicating multiple pathways to disease

Necessary but not Sufficient

Cuase must be present for effect to occur, yet not everyone exposed will manifest disease due to other influences

Niether Necessary nor Sufficient

Cause may contribute to effect but alone it does not directly result in the disease (chronic illness)

Sufficient Component Cause Model

Visualizes causality as a collection of component causes in pie chart (tuberculosis: tubercle bacillus)

Probabilistic Models

Incorporate elements of randomness, establishing likelihood of an effect occurring given certain exposure levels

Probability Causality

Quantify likelihood of health effects based on exposure levels. Models incorporate randomness, for chronic diseases these are preferred since they account for complexities and diversity of factors

Null Hypothesis

A hypothesis of no difference in the outcome in a population between the groups being compared. Most commonly used hypothesis in research

Method of Difference

Examines situations where all factors except one are constant

Method of Concomitant Variation

Looks at correlation between the increasing frequency of exposure and an outcome

Operationalization

Determines how these variables will be measured such as using questionnaires to assess habits and medical records to track diagnoses

No Association

Dietary sugar intake, incidence of diabetes is statistcally independent of sugar consumption

Association

Higher sugar consumption leads to increased occurrence (positive), higher sugar consumption leads to fewere rates (negative)

Noncausal

Association could occur due to randomness or involve a confounding factor

Indicrect Causation

Excesss sugar intake might contribute to obesity which increase diabetes risk

Direct Causal Relationship

High sugar consumption directly causes an increase in diabetes occurrences without intermediary factors

Criteria for Causality

Strenght of association, consistency of findings, specificity of relationships (unique outcome), temporality criterion (exposure before outcome), biological gradient (dose-response relationship), plausibility, coherence, experimental evidence, analogy in causality. Comprehensive of all 9 criteria must be conducted to determine causality

Multivariate Causality

Chronic diseases now stemming from multiple causal factors. Epidemiologic triangles visualize causation of each, web of causation emphasizes intricate relationships

Power

Ability of a study to demonstrate an association or effect if one exists

Multi-factorial Causation

Multiple factors

Attributable Fraction

Quantifies potential impact of eliminating specific causal factors on disease incidence

Multi-Causality

Assess impact of risk factors on disease occurrence and potential reductions through elimination

Necessary Causes in Disease Outbreaks

Every sufficient cause infludes a necessary cause (specific foods may be sufficient causes, while presence of harmful bacteria is necessary cause)

Factors in Causation

Predisposing, enabling, precipitating (direct exposure), reinforcing factors (environmental conditions and repeated exposures)

Risk Factors

Correlated with increased likelihood of disease

Interaction of Causes

Combined effect can exceed sum of individual effects of factors

Hierarchy of Causes

Immediate causes distinguished from indirect causes

DPSEEA

Driving forces, pressure, state, exposure, effect, action

Types of Exposure

Chemicals, biological organisms, physical (radiation), societal agents

True Risk Factors vs. Surrogates

When risk factors are unidentified surrogate measures become critical

Dose vs. Exposure

Exposure: external agents, dose: internal biomarkers

Comparative Advantages

External environmental measurements can provide a more stable estimate of true exposure levels due to lower variability in concentration compared to intra-individual variability. Allow for analysis of more agents due to wider array of available analytical methods

Intragroup

Within exposure group

Intergroup

Across multiple grups

Challenges of Retrospective Exposure Assessment

Historical measurement limitations

Challenges in Occupational Studies

Incomplete work histories

Limitations of Ecological Evaluations

Misclassify exposure levels

Advantages of Ecological Evaluations

Most useful when there is significant exposure contrast between groups compared to variability within each group

Dendrogram

Shows hierarchical relationship between clusters

Precision

Measures random error

Validity

Measures systematic error

Misclassification Error

Study subjects inaccurately assigned to exposure categories

Differential

Significant bias as a result. Diseased subjects can provide more accurate exposure data htan non-diseased subjects. Interviewers can introduce biases

Nondifferential

Results towards no association

Internal Validation Studies

Exposure is measured, and more accurate/expensive measurement is collected simultaneously in small subset of cases and controls selected randomly

Uncertainty Analysis

Quantifies variability around exposure point estimates and improves understanding of these estimates' implications on disease risk