Confounding II

Question 1

What is stratification?

Answer 1

Stratification is calculating measure of association for each stratum of potential confounder and comparing them. Can be used to control for confounding and evaluate whether it has occurred

Question 2

How do you do stratification?

Answer 2

1. Calculate the measure of association between exposure and outcome (crude, univariate, unadjusted)
2. Divide potential confounder into strata (levels)
3. For each stratum calculate measure of association between the exposure and outcome
4. Compare stratum specific measures of association

Question 3

Pros of stratification

Answer 3

• Easy for small number of potential confounders with limited strata
• Can evaluate impact of confounding
• Can identify effect modification

Question 4

Cons of stratification

Answer 4

• Can leave residual confounding
• Not feasible when dealing with lots of potential confounders with many strata

Question 5

What is multi variable analysis?

Answer 5

Multi variable analysis is a statistical method for estimating MOA whilst controlling for multiple potential confounders - can work where stratification can’t
Similar issues as stratification with multiple potential confounders/number of strata
Multivariable analysis is often more efficient in analytic studies

Question 6

How can we evaluate confounding?

Answer 6

Control for confounding in study analysis and then compare crude and adjusted MOA, look to see how much the MOA changed, general guideline for assessing confounding is if MOA changes by 10% or more there is evidence of confounding

Question 7

What is effect modification?

Answer 7

Association between exposure and outcome differs across strata of the effect modifier (stratum associations are different)

Question 8

How do we identify between effect modification and confounding?

Answer 8

Stratum specific MoA are different we are looking at effect modification
If stratum specific MoA are similar we are looking at confounding