Association and Causing Flashcards
If we find an association, does that mean the exposure causes the outcome?
Just because theres an association (even if it is a true association) doesn’t mean there is a causal relationship between the exposure and the outcome
What is the causal pie model?
Helps to think about prevention
Casual pie depicts some exposures, together the exposures are sufficient to cause the outcome
Each exposure is a different component of cause
Eliminating which exposure would prevent the most disease?
To try prevent this disease if we eliminated component cause A which is apart of sufficient cause 1 and 2, these two causes make up the most of disease in the population, hence if we remove exposure A then we do not have as high of a prevalence of a disease in a population
What is a necessary cause?
A component cause that is necessary for a disease to occur is a necessary cause, this must be a part of every sufficient cause, in this example exposure A is found in each cause, if we eliminated exposure A this diesease can’t occur
What are the guidelines for determining if something is causal?
BEST CDS
Biological plausibility
Experimental evidence
Specificity
Temporal sequence
Consistency
Dose-response relationship
Strength of association
What is biological plausibility?
Whether or not the relationship between exposure and outcome is biologically plausible - if it is, then that adds weight to possibility of causal relationship
What is experimental evidence?
Is there any evidence from RCTs or animal experiments - if there is, this adds weight to possible causality
What is specificity?
Consider if the exposure is specifically associated with a particular outcome but not with other outcomes, if it is this adds weight that there may be a causal relationship
What is temporal sequencing?
Know if exposure came before outcome - can’t do in cross sectional study as exposure and outcome measured at same time, can do in cohort study as followed up over time
If exposure does occur before outcome there is a causal relationship
What is consistency?
Look at whether the findings of the study you’re looking at are consistent with other studies, if there is consistency then possibility that it is causal
What is dose-response relationship?
Risk of outcome increases or decreases with increasing or decreasing amounts of exposure, adds weight to possibility of causal relationship, not all relationships are linear so doesn’t always apply
What is strength of association?
The stronger an association is the less likely it is to be due to confounding or bias - not always the case if confounding/bias is really strong
