CHAPTER 13 Taking a Closer Look at Fourfold Tables 185
Imagine that you are conducting a study at a primary care clinic. In the study, you
administer a newly developed home pregnancy test to 100 women who come to a
primary care appointment suspecting that they may be pregnant. This is conve-
nience sampling from a population defined as “all women who think they may be
pregnant,” which is the population to whom a home pregnancy test would be
marketed. At the appointment, all the participants would be given the gold stan-
dard pregnancy test, so by the end of the appointment, you would know their true
pregnancy status according to the gold standard, as well as what your new home
pregnancy test result said. Your results would next be cross-tabulated according
to the framework shown in Figure 13-4.
The structure of the table in Figure 13-4 is important because if your results are
arranged in that way, you can easily calculate at least five important characteris-
tics of the experimental test (in our case, the home pregnancy test) from this
table: accuracy, sensitivity, specificity, positive predictive value (PPV), and nega-
tive predictive value (NPV). We explain how in the following sections.
Overall accuracy
Overall accuracy measures how often a test result comes out the same as the gold
standard result (meaning the test is correct). A perfectly accurate test never pro-
duces false positive or false negative results. In Figure 13-4, cells a and d represent
correct test results, so the overall accuracy of the home pregnancy test is a
d
t/ .
Using the data in Figure 13-4, accuracy 33 51 /100, which is 0.84, or
84 percent.
Sensitivity and specificity
A perfectly sensitive test never produces a false negative result for an individual
with the condition. Conversely, a perfectly specific test never produces a false posi-
tive result for an individual negative for the condition. The goal of developing a
screening test is to balance the sensitivity against the specificity of the test based
FIGURE 13-4:
Results from a
study of a new
experimental
home pregnancy
test.
© John Wiley & Sons, Inc.