Let's look at an example.

A clinical trial is conducted to evaluate a diagnostic screening test designed to detect chromosomal foetal abnormalities (CFA). CFA are confirmed using amniocentesis (considered here as the gold standard).

Image source

The screening test is performed on a random sample of 200 pregnant women, who later undergo an amniocentesis. The following 2 x 2 cross-tabulation table summarizes the screening test results:

	CFA +	CFA -	Total
Test +	a 14	b 64	78
Test -	c 6	d 116	122
Total	20	180	200

 

 

 

 

 

 

 

The test characteristics are as follows:

SE = P (Test+ | CFA+) = 14/20 = 70.0 %
In words: 70% of all chromosomal foetal abnormalities (classified through amniocentesis) are detected by the new CFA screening test, and 30% are missed (not detected).

SP = P (Test- | CFA-) = 116/180 = 64.4 %
In words: 64.4% of CFA-negative (healthy) women have a negative test result, and 35.6% have a false positive test result.

At this point one has to consider the consequences especially of a false negative, true positive and false positive test result in order to decide whether the new test is helpful in the given situation.