Dear Statalist,

I am working on a clinical trial data set. I have 169 observations, a golden standard binary variable with missing values (25 missing values) and a continuous variables with no missing data.

I have applied multiple imputation with mi chained equation.



I found the same cutoff (threshold = 47) for my m=20 imputations. My questions is, do you recommend any particular method for calculation Sensitivity and Specificity for this threshold? Using Rubins rules ?


Then I used the following program to calculate AUC for my ROC curve.
* https://www.stata.com/support/faqs/s...-imputed-data/
Is there any program that allows for calculating Sensibily/specificity ?

Thank you !

Kristina