9.2: Confounding II ✅

Question 1

Dealing with Confouding

Answer 1

1) Identify potential confounders

2) Adjust for potential confounders

3) Compare crude and adjusted estimates

Question 2

Statistical Adjustment

Answer 2

Aims to eliminate the confounding effect for potential confounders in any exposure-outcome association

–> remove any effects of potential confounders, therefore providing a more clean estimate for the exposure-outcome association

Adjusted for that confounder e.g. age-adjusted estimates

Question 3

Stratification

Answer 3

Involves performing the analysis for the exposure-outcome association separately in the categories of the confounder -e.g. for smokers and non smokers

This will give us 2 different estimates which need to be combined

Question 4

What happens if the exposure-outcome association is substantially different for the categories of the connfounder?

Answer 4

Then the estimates can’t be combined

We present them separately

This is known as Effect Modification

Question 5

Crude estimates

Answer 5

Simply estimates for exposure-outcome associations (Odds Ratio, mean difference, etc.) before applying any adjustment.

Question 6

Adjusted estimates

Answer 6

Estimates for exposure-outcome associations after applying statistical adjustment for any potential confounder

Question 7

Judging the effect of adjustment

Answer 7

The different effects are:

1) Association becomes weaker
2) Association disappears 
3) Association appears
4) Association becomes stronger 
5) Association is reversed 
6) Association remains unaffected

Adjusting the mediators will have the same effect on the exposure-association as the confounder
->only the interpretation changes.

Question 8

Over-adjustment

Answer 8

A factor which lies in a casual pathway is treated as a confounder which can lead to a underestimate

Question 9

Residual Confounding

Answer 9

When some are adjusted for confounders but not all

->which can lead to an overestimate of the true exposure-outcome association.