DrVirgo
Forum Hero
Topics: 1083
Posts: 3,460
|
Please explain the difference between Confounding Bias and EFFECT MODIFICATION
If anyone can give another example of "Effect Modification" that would be great!
Thanks!
Attached Files:
DSC05384 (Small).JPG (49 KB, 82 downloads)
___________________
Our greatest glory is not in never falling, but in rising every time we fall.
|
| sprint123
Forum Guru
Topics: 129
Posts: 870
|
Dear DRVirgo,
Here is how I think of these two!
When you think of confounding---Think of the specific variable should have relationship with both the effect and the outcome. Example---When you are studying the effects of smoking and esophageal cancer ,ALCOHOLISM is a definite confounding factor since patients who smoke can also have high probability of drinking alcohol and alcohol per say is a risk factor of esophageal cancer.
When you think of Effect modification----It is a part of natural history of the disease FOR WHICH YOU COULD NOT DO ANYTHING. In this example you could not CONTROL for family history since it is a natural phenomenon accompanying the disease but in the previous case you could CONTROL for ALCOHOL by stratification!
Hope this helps!
|
| tongsaa
Forum Junior
Topics: 15
Posts: 52
|
|
| DrVirgo
Forum Hero
Topics: 1083
Posts: 3,460
|
Kaplan gives this example for Confounding:
Comparing relationship between exercise and heart disease in two populations when one population is younger and the other is older. Are differences in heart disease due to exercise or age?
-In this case age CAN be a confounder. correct?
___________________
Our greatest glory is not in never falling, but in rising every time we fall.
|
| melody1997
Forum Newbie
Topics: 0
Posts: 1
|
Smoking is not related to estrogen but both are related to DVT = effect modification
Smoking is related to drinking and both are related to esophageal cancer = confounding
|
| SmokyWaters
Forum Elite
Topics: 6
Posts: 458
|
IT DEPENDS UPON THE DESIGN OF STUDY
IF YOU ARE TAKIN A FACTOR AS CONFOUNDING OR EFFECT MODIFIER
YES DR VIRGO, AGE CAN BE CONFOUNDER HERE
|
| zubr
Forum Junior
Topics: 14
Posts: 61
|
How do you know if they related or not?
Maybe some relationship exists but has not been proven yet?
|
| Vietnamese
Forum Elite
Topics: 12
Posts: 285
|
Confounding factor= Not able to be modified (ex: Smoke + alcohol or smoke alone---> altered risk of esophagus cancer but alcohol alone is unclear for cancer risk; then alcohol is confounder)
Effect Modification= Able to be modified (ex: family hx of breast cancer alone and OCP alone are risks for breast cancer, then each of them is effect modification).
Edited by Vietnamese on 06/04/08 - 02:05 AM
___________________
Nothing is impossible.
|
|
| |
| | | | | | | | |
