Question

1. A psychiatrist devised a test for depression based on a short number of questions. To evaluate an independent investigator used the test on a population of 400 psychiatric outpatients for whom a more extensive test that can be viewed as a gold standard had been used. Among the 50 outpatients found to be depressed according to the gold standard, 40 patients were positive for depression based on the short-test. Among 350 patients found not to be depressed according to the gold standard, 18 patients were found to be positive for depression for the short-test.

Construct a table showing these results according to the test and gold standard then calculate the following :

(A) The sensitivity of the short-test (2)

(B) The specificity of the short-test (2)

(C) The positive predictive value in the population studied (2)

(D) The negative predictive value in the population studied (2)

(E) The true prevalence of depression in the population studied (1)

(F) If the researcher asked 40 patients to repeat the short-questionnaire a few days later (in which time you can assume depression status should not change) and found they all were still classified the same way what feature of the test does this demonstrate is good? (1)

Answer 1

The 2x2 contingency table of frequencies is as follows.

	Positive result	Negative result	Marginal frequency
Depressed	40 (True positive)	50-40= 10 (False negative)	50
Not depressed	18 (False positive)	350-18= 332 (True negative)	350
Marginal frequency	58	342	400

(A)

Sensitivity of short test

(B)

Specificity of short test

(C)

Positive predictive value of short test

(D)

Negative predictive value of short test

(E)

True prevalence of depression in the population studied

(F)

Same result over different times proves a good consistency of the short test procedure.