I am doing a research on the probability of diagnosing breast cancer given different independent variables. I define the presence of the variable as +ve-test. I am wondering if I can calculate the probability of the disease in the +ve test-group in a case-control retrospective study. I know that this can be done in cohort study. 

I have a solution and I need you valuable comments.

1) First, from the case control study I will calculate the odd of the +ve-test in the disease-group as well as the normal-group. This can be then converted to the probability of the +v-test given the disease and the normal status ( which is equivalent to the sensitivities of the test).  

2) Second, I will use Bayes` formula for posterior probability and the known prevalence of the disease in the community to derive the probability of the disease given the+ve-test.

Regards.

More Adel Bishara's questions See All
Similar questions and discussions