Please enable JavaScript.
Coggle requires JavaScript to display documents.
Causal Inference and Public Health - Coggle Diagram
Causal Inference and Public Health
Bradford Hill 1965 criteria
Strength of association - Strong associations are more likely to be causal than weak ones.
Consistency - Repeated observations of an association in different populations and settings increase the likelihood of causality.
Specificity - A specific association (one that is limited to particular diseases and exposures) is more likely to indicate causality.
Temporality - The cause must precede the effect in time.
Dose response relationship - A gradient of risk associated with the degree of exposure supports causality.
Plausibility - A plausible biological mechanism linking the exposure to the outcome supports causality.
Analogy - Similarities between the observed association and other known causal relationships can support causality.
Experimental Evidence - Evidence from experimental studies (such as randomized controlled trials) can strongly support causality.
Coherence - The association should be consistent with existing knowledge and theories.
Association vs Causality
Association and causality are different. Proving causality is difficult.
Causal effect is defined as the comparison of ALL subjects under different conditions
Association is defined as the comparison of different subjects under different conditions
Bias can be a big flaw in studies
Can be limited by
Marginal structural models
Structural nested models
Judea Pearl 1990
Developed DAGs (Directed Acyclic Graphs)
a conceptual representation of a series of activities. The order of the activities is depicted by a graph, which is visually presented as a set of circles, each representing an activity, some of which are connected by lines, representing the flow from one activity to another.
Causal diagrams that help recognize bias and confounding
RCT - gold standard for causation
Related to epidemiological studies
Aim is to receive valid, precise and generalizable estimated of an exposure on a paricipant
Precision opposite of random error.
Validity opposite of bias
terms
Power
Bias
Null hypothesis
Random error
Confounder
Collider
Issues affecting causal inference
Confounding
Colliding
Random error
Systematic error
Selection Bias: Non-random selection into treatment or study participation can affect the results.
Measurement Error: Inaccurate measurement of variables can lead to incorrect causal inferences.
Complex Interactions: Multiple interacting factors can complicate the identification of causal relationships.
Types of Bias
Pre-trial Bias
Selection bias
Flawed design study
Channelling bias
Methods to prove causality
Observational Studies: Including cohort studies, case-control studies, and cross-sectional studies. These studies can identify associations and suggest potential causal relationships.
Randomized Controlled Trials (RCTs): Considered the gold standard for establishing causality because they minimize bias and confounding by randomly assigning participants to exposure or control groups.
Systematic Reviews and Meta-Analyses: Combine results from multiple studies to provide stronger evidence for causality.
Natural Experiments: Situations where external factors create conditions similar to a randomized experiment, allowing for causal inference.
Causal inference is a field of study focused on understanding and establishing causal relationships between variables. It seeks to determine whether a specific factor (the cause) directly affects an outcome (the effect), and to quantify the magnitude of this effect. Causal inference goes beyond mere associations or correlations and aims to uncover true cause-and-effect relationships. This is particularly important in fields like epidemiology, social sciences, economics, and more.
Mendelian Randomization: Uses genetic variants as instruments to determine causal relationships between modifiable risk factors and health outcomes.
Time-Series Analysis: Examines trends over time to infer causality, particularly when interventions are introduced at specific points.
Instrumental Variables (IV): Techniques that use external variables (instruments) that affect the treatment but not the outcome directly, helping to control for unobserved confounding.
Propensity Score Matching: A method that involves matching treated and untreated subjects with similar propensity scores (probability of receiving the treatment) to control for confounding.
Difference-in-Differences (DiD): Compares the changes in outcomes over time between a treatment group and a control group, controlling for time-invariant unobserved differences.
Regression Discontinuity Design (RDD): Exploits a cutoff or threshold in the assignment of treatment, comparing those just above and just below the cutoff to estimate causal effects.
Key Concepts in Causal Inference
Counterfactuals: The idea of comparing what actually happened with what would have happened in the absence of the exposure. This involves considering hypothetical scenarios to understand causality.
Confounding: The presence of a third variable that influences both the cause and the effect, potentially leading to spurious associations.
Causal Diagrams (Directed Acyclic Graphs - DAGs): Visual representations of causal relationships that help identify potential confounders and pathways through which effects are transmitted.
Potential Outcomes Framework: Also known as the Rubin Causal Model, it formalizes causal inference by defining causal effects in terms of potential outcomes under different treatment conditions.