sensitivity and specificity
Last reviewed 10/2021
These are statistical terms that are often confused. These terms can be illustrated by way of an example:
At a particular hospital a MIBG scanner is used to detect phaeochromocytomas. A positive scan was reported in 38 cases. An audit showed:
- true positive (TP) 10, true negative (TN) 25,
- false positive (FP) 0, false negative (FN) 3
where:
- patients with phaeochromocytoma TP (person with a phaeochromocytoma and positive test)
- patients with phaeochromocytoma FN (person with a phaeochromocytoma and negative test)
- patients without phaeochromocytoma FP (person without a phaeochromocytoma with a positive test)
- patients without a phaeochromocytoma TN (person without a phaeochromoytoma with a negative test)
| phaeochromocytoma present | phaeochromyctoma absent |
test result positive TP = 10 | test result positive FP = 0 |
test result negative FN = 3 | test result negative TN = 25 |
- sensitivity = 100xTP /(TP+FN) = 100 x 10(10 + 3) = 77%
- i.e. proportion of people with a phaeochromocytoma who are corectly identified by the screening test = 77%
- specificity = 100xTN /(TN+FP) = 100 x 25 /(25+0) = 100%
- i.e. proportion of people without phaeochromocytoma who have had a negative test result = 100%
- positive predictive value = 100xTP/(TP+FP) = 100 x 10/(10+0) = 100%
i.e. the proportion of screening tests that are correct.