sens and spec for 2 by 2 table

sens and spec for 2 by 2 table

by Laura Koth -
Number of replies: 1

help for diagt, diagti in  stata use the definitions of specificity and sensitivity using a 2x2 table :

Sensitivity (sens.) as the proportion of diseased patients correctly identified = a/(a+b).  Specificity (spec.) is the proportion of healthy patients correctly identified = d/(c+d).

so, when using the episensi command, I am not understanding how we now are supposed to come up with separate sensitivities and separate specificities needed for the command:

dseca(pdf_option)       define the sensitivity among the cases

dspca(pdf_option)       define the specificity among the cases

dsenc(pdf_option)       define the sensitivity among the noncases

dspnc(pdf_option)       define the specificity among the noncases

can someone help me? perhaps I'm missing a really basic point here.

In reply to Laura Koth

Re: sens and spec for 2 by 2 table

by Francois Rerolle -

This is to allow sensitivity and specificity to vary by case/control status. If it doesn't vary, then the sensitivity in cases is the same as it is in the controls and you will provide the same input for dseca and dsenc. Similarly for specificity.


Does that make sense?