**- Receiver-Operator Characteristic
(ROC) Curve**“
. . . is a graphic representation of the relationship between sensitivity and specificity
for a diagnostic test. It provides a simple tool for applying the predictive value method
to the choice of a positivity criterion” (Knapp and Miller, 1992). **
****- Steps in Constructing
ROC Curves **(Knapp and
Miller, 1992):
1. *Draw the curve* by plotting the true positive
rate (sensitivity) against the false positive rate (1-specificity) for several choices of
the positivity criterion.
2. *Using the curve to locate the positivity criterion*.
The point marked by the yellow circle in the upper left corner represents a perfect
diagnostic test. At this point, both sensitivity and specificity are 100%, that is, all
diseased individuals are identified, all healthy individuals are labeled disease-free, and
no disease-free individuals are labeled diseased.
- *When the costs of a false positive and false negative
test result are equal*, set the positivity criterion equal to the point on the ROC
curve closest to the upper left corner (See point #3). At this point, the discriminate
ability of the test is maximized and the number of erroneous diagnoses is minimized.
- *When a false positive (mis-diagnosis) result is
especially undesirable*, set the positivity criterion equal to the point farthest to
the left on the ROC curve (See point #1 or 2).
- *When a false negative (missed diagnosis) result is
especially undesirable*, set the positivity criterion equal to a value toward the right
on the ROC curve (See point #5 or 6). At the point on the ROC curve farthest to the right,
all patients with the disease are detected by the diagnostic test. |