Dear all,
I am new in this forum. I have an ongoing research and I am dealing with the statistic.
In our study we are trying to compare 2 different diagnostic tools (GCS and FOUR), or prognostic tools if you wish, in their ability to predict the "cerebral death" of our patients.
To be clear, if the GCS is positive we decleare the patients as having an "imminent brain death" (IBD); the same, of course, for the second scale (FOUR). Then we have looked at the outcome: "brain death" diagnosed according to the Italian law. I have just calculated the AUROC for the 2 methods and, then, I have made a Chi2 for the two AUROC (I have used the STATA command "Compare the ROC Curves"). The problem is: " Is this procedure right?" Are the two tests in object prognostic tools instead of diagnostic tools? I mean, at the moment in which patients has the two tests positive, it means that they are going to develop "brain death in the future" but they are not already in the "brain death"state .
Of course the sensibility of the two tests is 100% for both of them (thanks God).
How should I compare this two test? Do you have any suggestion, please?
Moreover, our study is a pilot study in order to plan a bigger one. In the case we would like to calculate the sample size, how should I do it? Thanks to everybody for your suggestions.