In: Statistics and Probability
Researchers found a crude odds ratio of 1.8 in a case-control study. They stratified their analysis within levels of a binary potential confounder and found an odds ratio of 1.2 in one level of the confounder and an odds ratio of 2.8 in the other level of the confounder.
A. Do you think they should present the adjusted odds ratio in the report of their results?
B. Why or why not?
Odds ratio =
Probability of an event/1- Probability of the event
So odds ratio is defined as the probability of an event happening, divided by the probability of an event not happening. So a higher odds ratio means greater chances of the event happening.
Note: Odds ratio> 1, means increased chances of an event occuring
Odds ratio<1, means decreased chances of the event occuring.
A) Here in the case control study, the odds ratio is already > 1, ie 1.8
But in stratified analysis we find odds ratio of 1.2 in one level and 2.8 in the other level.
So yes we should show this stratified analysis in the report to show greater level of detail.
B) This analysis shows us that the probability of occurrence of one event is relatively higher (with odds ratio 2.8), and then it's lower (closer to 1) for the other variable. So the overall is high because 2.8