Causality Flashcards
Who created the postulates of determining that a disease was caused by a bacteria?
Robert Koch 1890
What are Koch’s postulates?
Consistency
Specificity
Biological coherence
Predictive or experimental performance
Explain consistency re Koch
Agent must be shown to be present in every case of the disease by isolation in pure culture
What is Koch’s specificity?
Agent must not be found in the case of another disease
What is Koch’s biological coherence?
Once isolated, the agent must be capable of reproducing the disease in experimental animals
What is Koch’s predictive performance?
The agent must be recovered from the experimental disease produced
What criteria of causality can be used?
Koch Bradford Hill Susser Rothman's Sufficient Web of causation
What is Bradford Hill’s criteria of causality?
Temporal association Dose-response association Specificity Consistency Plausible biological association High strength of oassociation Absence of reverse causality
Susser’s categories of causation
Strength Specificity - in cause and effect Consistency Predictive Performance Coherence/plausability
How can one measure Susser’s consistency?
Replicability
Subcategories of Susser’s Coherence category?
Theoretical
Factual
Biological
Statistical
Essential categories of Susser’s causality
Association
Direction of prediction
Time order
How is Susser’s association category judged?
By strength of probabilities based on preset expectations of chance occurrences and consistency upon replication.
Explain Susser’s category of Direction of Prediction
Change in one state assumed to be the effect must be consequent on change in independent antecedent states presumed to be the cause.
Which essential criteria of Sussers is more critical in defining causality?
Direction of Prediction
Explain Susser’s time order category
Reversal of time order assures elimination of putative cause
What is sufficient cause?
Minimal set of conditions that produce an outcome rather than just one
In which studies does one see sufficient cause
Epidemiology
What makes up sufficient cause?
Component cause
Importance of component cause
Outcome will not occur by a causal pathway if any of the components are missing
Why is the idea of component cause relevant to public health?
Suggests that it is not necessary to identify all of the component causes to prevent a disease outcome
What is a necessary cause?
A component cause that must be present in every sufficient cause for the outcome
What terms are associated with Rothman’s Sufficient Model?
Sufficient cause
Component cause
Necessary cause
Who created the term web of causation?
MacMahon and Pugh in 1960
What was web of causation used to describe?
Complexity of the relationship among antecedents when investigating cause of disease
What is the concept in web of causation?
Mutual interactions among risk factors which have both direct and indirect effects on outcomes.
Who expanded the web of causation?
Murray and Lopez
What did Murray and Lopez do to the web of causation?
Separated proximal risk factors from distal
What are proximal risk factors?
Established pathophysiological mechanisms
What are distal risk factors?
Based on empirical evidence
What is another name of the model Murray and Lopez created from the web of causation?
Hierarchical causal model
How can positive result from a study be explained?
Selection differences between two groups
Measurement differences between two groups
Confounding factors causing indirect association
Chance causing spurious association (Type 1 error)
Presence of true causal assocation
What is the independent variable?
The one which is having the effect
What is the dependent variable?
Outcome of interest
What criteria are needed for a characteristic to be a confounder?
Must be related to exposure in some way
Must be related to outcome in terms of prognosis or susceptibility
Will not be on causal pathway between exposure and outcome
Distribution of characteristic must be different in groups compared (mean or standard deviation)
What is needed to withstand confounding effects in analytical studies?
Systematic effort to identify and measure potential confounders
Data on distribution of potential confounders across groups must be available
Methods to control confounders in design of study
Restriction
Matching
Randomisation
Methods t control confounders in analysis of study
Stratification
Multivariate methods
What is restriction?
Avoid groups with significant confounding influence in study
Disadvantages of restriction
Will limit sample size
What is matching?
Make sure confounders are equally distributed
Impact of randomisation on confounders
Will ensure two groups are more likely to b similar in terms of confounder distribution
What must one do if randomisation does not lead to equal distribution of confounders?
Consider stratification or multivariates
Which design method helps control both known and unknown confounders
Randomisation only
What is stratification?
Tabulate data for various categories of exposure to confounder separately and analyse variable influence
What are multivariate methods?
Regression methods to analyse effect of various confounders.
What do multivariate methods produce?
Adjusted rates and crude rates
What is an effect modifier?
A third factor existing as a sub-group but nor related to outcome or exposure
How can effect modifiers be eliminated?
Stratified analysis
How do confounders result in bias?
Affect internal validity of a study
