Diagnostic test performance: Relationship between the cutoff value, sensitivity, and specificity

This figure illustrates how changing the cutoff value for a diagnostic test impacts the sensitivity and specificity. The curves represent the range of test results for individuals without the disease (solid orange curve; TN) and with the disease (solid blue curve; TP). Since the curves overlap, it is not possible to choose a cutoff value that will correctly identify patients as having or not having the disease 100% of the time. The cutoff value shown in the figure (dashed line) will result in a small percentage of false negatives (striped blue area; FN) and false positives (striped orange area; FP). Lowering the cutoff value will reduce the number of false negatives (ie, better sensitivity) but will do so at the cost of increasing the false positive rate (lower specificity). Raising the cutoff value will reduce the number of false positives (better specificity) but will do so at the cost of increasing the false negative rate (lower sensitivity).