|
||||||||||||||||
|
||||||||||||||||
The following HTML text is provided to enhance online readability. Many aspects of typography translate only awkwardly to HTML. Please use the page image as the authoritative form to ensure accuracy. 
FIGURE 9.2 Probability distributions of diseases and test results. Figure 9.2 displays test results on the horizontal axis, and the probability of a given test result on the vertical axis for patients with disease (going up) and patients with no disease (going down). The three vertical lines display three different potential cutoffs or criteria of positivity, A, B, and C. The heavy line B is a middle ground that represents a compromise between sensitivity and specificity. The line to the left (A) represents a more lenient—“lax”—definition of positivity; this definition increases sensitivity (including a greater proportion of patients with disease to the right of the criterion) while decreasing specificity (including a greater proportion of patients without disease to the right of the criterion). The line to the right (C) represents a more strict—“rigid”—definition of positivity; this definition decreases sensitivity (including a smaller proportion of patients with disease to the right of the criterion) while increasing specificity (including a smaller proportion of patients without disease to the right of the criterion). Those performance characteristics can be transformed into a variety of interpretation aids, of which the receiver operator characteristic (ROC) curve is particularly useful. The Receiver Operator Characteristic Curve for a TestA critical question in interpreting screening tests is how and where to establish a criterion of positivity—that is, the cutoff (cutpoint) for that particular test as
|
|
|||||||||||||||
