(C) the point at the end of the normal values
(C) would be the correct answer if we were trying to maximize specificity, which is defined as the true negatives (TN) divided by the sum of people who don't have the disease, or the true negatives plus the false positives (FP+TN), or specificity = TN/(FP+TN). The easiest way to maximize this is to maximize the numerator or to maximize the number of TNs. To maximize the number of true negatives in order to maximize specificity, we must be able to say that for every healthy person tested, they must be negative. This implies that the threshold for marking someone as not diseased must be the highest normal person's value, which ends up being (C) on the chart. The further right you go on the chart, specificity stays at 100%, but the more sensitivity decreases. At (E), everyone tests negative and so sensitivity is 0%.
to maximize sensitivity, we need a setpoint at the lowest diseased value (every diseased person tests positive)
to minimize sensitivity, we need a setpoint at the highest diseased value (every diseased person tests negative)
to maximize specificity, we need a setpoint at the highest normal value (every normal person tests negative)
to minimize specificity, we need a setpoint at the lowest normal value which is usually 0 (every normal person tests positive)
submitted by โshak360(19)
(A) the leftmost value (correct)
Sensitivity is defined as the fraction of true positives (TP) to the number of people who have the disease, which is true positives plus false negatives (TP+FN), or sensitivity = TP/(TP+FN). The easiest way to maximize this value is to maximize the numerator or to maximize the number of TPs. To maximize the number of true positives in order to maximize sensitivity, we must be able to say that for every diseased person tested, they must be positive. This implies that the threshold for marking someone as diseased must be the lowest diseased person's value, which is the leftmost value.