Lecture 28: Confounding 2 Flashcards
How do we control confounding in
the study analyses? 3 ways
- Stratification
- Multivariable Analysis
- (Standardisation)
Each can be used to control for confounding and evaluate whether it has occurred
Must have measured potential confounder!
How does stratification work?
- Calculate the measure of association between exposure and outcome (crude, univariate, unadjusted)
- Divide potential confounder into strata (levels)
- For each stratum, calculate the measure of association between the exposure and outcome (i.e. calculate the stratum-specific measures of association)
- Compare stratum-specific measures of association
What are the pros and cons of stratification?
Pros:
- Easy for small number of potential confounders with limited strata
- Can evaluate impact of confounding
- Can identify effect modification
Cons:
- Can leave residual confounding
- Not feasible when dealing with lots of potential confounders with many strata
How can we use stratification to determine if confounding is present? (confounding path) SLIDE 32 for reference
Crude Measure of Association ≈ stratum-specific measure of association?
ie. unadjusted measure is similar to adjusted = No major confounding so Use crude Measure of association
Crude MoA ≠ stratum-specific MoA?
ie. unadjusted is not the same as adjusted = Confounding present so Calculate and use adjusted MoA
What is multivariable analysis?
Statistical method for estimating measure of association whilst controlling for multiple potential confounders
Variety of different techniques recognisable by the term ‘regression’
Can work in situations where stratification won’t
What is standardisation?
When age structures different ANDDDDDDDD disease risk variers with age
What are some limitations of standardisation?
Similar issues as stratification with multiple potential
confounders/number of strata
Multivariable analysis is often more efficient in analytic studies
What are some overall Potential issues: controlling
in study analyses
Residual confounding
Can only control what you’ve measured
How much change indicates confounding?
General guideline is if controlling for confounding changes the measure of association by 10% or more
Use your judgement: has confounding materially changed your interpretation of the measure of association?
Make sure you understand slide 24: evaluating confounding
CHIPPIES
What is confounding vs effect modification?
Confounding involves an additional factor that distorts the true association (or lack thereof) between an exposure and outcome, leading to an incorrect result. It is a nuisance whose influence on the measure of association we wish to remove.
Effect modification also involves an additional factor. Unlike confounding, this factor changes (modifies) the association between exposure and outcome. This is often a very important observation. Effect modification occurs when the association (the ‘effect’) between an exposure and outcome is different for people with and without the additional factor (the ‘effect modifier’) or with different levels of the additional factor.
What is the effect modification pathway? Slide 32 for reference
Effect Modification path
- Calculate crude MoA
- Stratify and calculate stratum-specific MoA
- If Stratum-specific MoA are different = effect modification
- Report stratum-specific MoA
Basically there are differing effects across the strata
What are the Advantages of Controlling for Confounding in Analysis? 3
- Quantify Confounder-Outcome Association:
Allows measurement of how much the confounder (e.g., alcohol) impacts the outcome. - Assess True Confounding:
Compare crude and adjusted associations to determine if a variable truly confounds the relationship. - Identify Effect Modification:
Enables detection of effect modification (e.g., if alcohol changes the effect of coffee on cancer) by including all groups in analysis.
