We present a graphical method for analyzing the cost related to false positive (FP) and false negative (FN) results of diagnostic tests.
We created a function relating cost to the sensitivity (S) and specificity (E) of the diagnostic test. Isocost straight lines were obtained, the gradient of which represents the ratio of false positive cost/false negative cost. The various diagnostic tests are plotted in the ROC space as the point (1-E, S).
Isocost straight lines allowed us to see if the cost of a new test is lower or the same as previous tests.
Graphical analysis of the cost of a diagnostic test is helpful in deciding whether or not to introduce new diagnostic tests.