🤒Intro to Epidemiology Unit 5 – Association and Causation

Key Concepts

• Association refers to the relationship between two variables, such as an exposure and an outcome, in which a change in one variable is linked to a change in the other
• Causation implies that an exposure directly leads to an outcome, establishing a cause-and-effect relationship
• Correlation measures the strength and direction of the relationship between two variables (positive, negative, or no correlation)
• Confounding occurs when a third variable influences both the exposure and the outcome, leading to a spurious association
• Bias is a systematic error in the design, conduct, or analysis of a study that can lead to incorrect conclusions about the association between variables
  • Selection bias arises when the study participants are not representative of the target population
  • Information bias occurs when the data collected about the exposure or outcome is inaccurate or incomplete
• Temporality is a key criterion for causation, requiring that the exposure precedes the outcome in time
• Dose-response relationship strengthens the evidence for causation, showing that increasing levels of exposure lead to increasing risk of the outcome

Types of Association

• Positive association indicates that as the exposure increases, the outcome also increases (smoking and lung cancer)
• Negative association suggests that as the exposure increases, the outcome decreases (physical activity and cardiovascular disease risk)
• No association means that there is no relationship between the exposure and the outcome
• Linear association describes a straight-line relationship between the exposure and the outcome
• Non-linear association refers to a relationship that follows a curved pattern (U-shaped or J-shaped)
  • U-shaped association shows that both low and high levels of exposure are associated with the outcome (alcohol consumption and mortality)
  • J-shaped association indicates that low levels of exposure may be protective, while high levels increase the risk (body mass index and mortality)
• Monotonic association means that the relationship between the exposure and the outcome is consistently increasing or decreasing

Measures of Association

• Relative risk (RR) compares the risk of the outcome in the exposed group to the risk in the unexposed group
  • RR = (Risk in exposed) / (Risk in unexposed)
  • RR > 1 indicates a positive association, RR < 1 suggests a negative association, and RR = 1 implies no association
• Odds ratio (OR) compares the odds of the outcome in the exposed group to the odds in the unexposed group
  • OR = (Odds in exposed) / (Odds in unexposed)
  • OR interpretation is similar to RR, but OR tends to overestimate the strength of association when the outcome is common
• Attributable risk (AR) measures the absolute difference in risk between the exposed and unexposed groups
  • AR = (Risk in exposed) - (Risk in unexposed)
• Population attributable risk (PAR) estimates the proportion of cases in the population that can be attributed to the exposure
  • PAR = (Prevalence of exposure) × (RR - 1) / [1 + (Prevalence of exposure) × (RR - 1)]
• Incidence rate ratio (IRR) compares the incidence rates of the outcome between the exposed and unexposed groups
  • IRR = (Incidence rate in exposed) / (Incidence rate in unexposed)
• Hazard ratio (HR) is used in survival analysis to compare the hazard rates between the exposed and unexposed groups
  • HR = (Hazard rate in exposed) / (Hazard rate in unexposed)

Causation vs. Correlation

• Causation implies that the exposure directly leads to the outcome, while correlation only suggests a relationship between the two variables
• Correlation does not necessarily imply causation, as the association may be due to confounding, reverse causation, or chance
• Reverse causation occurs when the outcome influences the exposure, rather than the exposure causing the outcome (weight loss and cancer)
• Establishing causation requires meeting specific criteria, such as the Bradford Hill criteria
• Randomized controlled trials (RCTs) are considered the gold standard for assessing causation, as they minimize confounding and bias
  • In RCTs, participants are randomly assigned to the exposure or control group, ensuring that confounding factors are evenly distributed between the groups
• Observational studies, such as cohort and case-control studies, can provide evidence for association but cannot definitively establish causation due to the potential for confounding and bias

Bradford Hill Criteria

• Strength of association: A strong association between the exposure and outcome supports causation
• Consistency: The association is observed across different studies, populations, and settings
• Specificity: The exposure is associated with a specific outcome, rather than multiple outcomes
• Temporality: The exposure precedes the outcome in time
• Biological gradient (dose-response relationship): Increasing levels of exposure lead to increasing risk of the outcome
• Plausibility: The association is biologically plausible based on existing knowledge
• Coherence: The association is consistent with existing theories and knowledge
• Experiment: Experimental evidence, such as from RCTs, supports the causal relationship
• Analogy: Similar exposures have been shown to cause similar outcomes

Confounding and Bias

• Confounding occurs when a third variable is associated with both the exposure and the outcome, leading to a spurious association
  • Age is a common confounder, as it is often associated with both exposures and health outcomes
  • Socioeconomic status can confound the relationship between education and health outcomes
• Confounding can be addressed through study design (randomization, matching) or statistical analysis (stratification, regression)
• Selection bias arises when the study participants are not representative of the target population
  • Healthy volunteer bias occurs when healthier individuals are more likely to participate in a study
  • Loss to follow-up bias happens when participants who drop out of a study differ from those who remain
• Information bias occurs when the data collected about the exposure or outcome is inaccurate or incomplete
  • Recall bias arises when participants' ability to remember past exposures or outcomes differs between groups
  • Measurement bias occurs when the tools or methods used to assess the exposure or outcome are inaccurate or inconsistent
• Bias can lead to an overestimation or underestimation of the true association between the exposure and outcome

Study Designs for Assessing Association

• Cohort studies follow a group of individuals over time to assess the relationship between an exposure and an outcome
  • Prospective cohort studies enroll participants before the outcome occurs and follow them forward in time
  • Retrospective cohort studies identify a cohort based on past records and follow them up to the present
• Case-control studies compare the exposure history of individuals with the outcome (cases) to those without the outcome (controls)
  • Cases and controls are matched on key characteristics to minimize confounding
  • Case-control studies are useful for rare outcomes or long latency periods
• Cross-sectional studies assess the exposure and outcome at a single point in time
  • Cross-sectional studies can identify associations but cannot establish temporality or causation
• Ecological studies compare exposure and outcome data at the population level, rather than the individual level
  • Ecological fallacy occurs when associations observed at the population level do not hold true at the individual level
• Randomized controlled trials (RCTs) randomly assign participants to the exposure or control group to minimize confounding and assess causation

Real-World Applications

• Epidemiological studies have established the causal link between smoking and lung cancer, leading to public health interventions and policies
• The association between air pollution and respiratory health outcomes has informed environmental regulations and guidelines
• Studies on the relationship between diet, physical activity, and chronic diseases have shaped public health recommendations and interventions
• Investigations into the association between occupational exposures (asbestos) and health outcomes (mesothelioma) have led to workplace safety regulations
• Research on the link between infectious agents (human papillomavirus) and cancer (cervical cancer) has informed vaccination programs and screening guidelines
• Studies assessing the association between social determinants of health (education, income) and health outcomes have highlighted the importance of addressing health inequities
• Pharmacoepidemiological studies evaluate the association between medication use and adverse events, informing drug safety and regulation
• Investigations into the relationship between environmental exposures (lead) and child development have led to policies to reduce exposure and mitigate harm