confounding vs effect modification - dreamz4usmle

[ArchivalUser]

can anyone explain.

[ArchivalUser]

A confounder is related to both exposure and outcome. It can be eliminated when you stratify groups according to confounder
Example: alcohol consumption increases lung cancer, the confounder her is smoking
Alcoholics are most likely to be smokers and smoking is associated with lung cancer (smoking is associated with both alcohol consumption and lung cancer)
To understand it, keep it simple, alcohol is a false risk factor and smoking is the real (confounder) risk factor
When you divide this study subjects into smokers and nonsmokers, you will find that alcohol doesn't increase lung cancer in either group.
You should not report that alcohol increase lung cancer.

Effect modification:
effect modifier: can modify "increase or decrease" the effect exposure on outcome "it comes between them". You can't eliminate it by group stratification according to effect modifier,
Example: Women on OCP are at risk of thrombosis, right? Smoker women on OCP have higher risk of DVT than nonsmoker
Smoking increases "modifies" effect of OCP on DVT
When you stratify study subjects into smokers and nonsmokers, you will find OCP increases DVT among smokers more than nonsmokers "both have high risk in comparison to general population"
You should report that smoking increases the risk of DVT in smoker women on OCP more than nonsmoker women on OCP.

Hope I was able to explain it!

[ArchivalUser]

thanks a lot lamonti, you surely described it very well.

[ArchivalUser]

👍

[ArchivalUser]

Thank you