You have two tests: a screening test, called SCREEN, and a confirmatory test, called CONFIRM. SCREEN has a Sensitivity of 80% and a Specificity of 90%. CONFIRM has a Sensitivity of 98% and a Specificity of 99%. You use SCREEN first on 1000 specimens that you know beforehand have a prevalence of 5%. You then use CONFIRM on the specimens that tested positive by SCREEN.
Fill in the following Table.
SCREEN (on 1000 specimens)
Prevalence
True Positives
False Positives
Positive Predictive Value
CONFIRM (on all the positives from SCREEN)
Prevalence
True Positives
False Positives
Positive Predictive Value
What is the relationship of the PPV on the first test to the Prevalence for the second test?
Now switch the order of testing:
CONFIRM (on 1000 specimens)
Prevalence
True Positives
False Positives
Positive Predictive Value
SCREEN (on all the positives from CONFIRM)
Prevalence
True Positives
False Positives
Positive Predictive Value
What happens when you switch the roles of screening and confirmatory tests (that is, using the better test first)?
(Hint: compare the final PPVs in each case)