How to calculate
Divide true positives by all reference positives for sensitivity. Divide true negatives by all reference negatives for specificity. The table also shows predictive values and observed accuracy.
Worked example
Use True positives 90 · False positives 20 · False negatives 10 · True negatives 80. The result is 90%.
Input and result notes
Use reference-standard classifications or validated performance estimates. Results describe the study or statistical model for the entered inputs. Enter zero for any cell with no observations.
FDA guidance on evaluating diagnostic tests