Sensitivity and Specificity
Jessica Mason, MD, and Mel Herbert, MD
Background
- Sensitivity
- The proportion of patients with the disease who test positive (see formula below)
- How good is the test at picking up anyone who might have the disease?
- Eg, using troponin to test for acute coronary syndrome (ACS)
- Newer troponins are even more sensitive than prior generations
- But this does not mean that everyone with a slightly elevated troponin is having a myocardial infarction
- Specificity
- The proportion of patients without the disease who test negative
- How good is the test at confirming the diagnosis?
- Eg, using troponin to test for ACS
- Many patients who have heart failure, renal disease, or some other cardiac issue such as myocarditis can have an elevated troponin
- Troponin is not very specific to ACS — it still needs to be used in the right clinical setting
- Calculating sensitivity and specificity
- Set up a 2 × 2 table as shown in table 1
- Sensitivity = true positive / (true positive + false negative)
- Specificity = true negative / (true negative + false positive)
- Example with made-up data:
- There are 100 patients
- 10 patients have ACS; of those 10, only 9 have a positive troponin
- 90 patients do not have ACS, but 30 have a positive troponin
- See table 2 for the example 2 × 2 table
- Sensitivity = 9 / (9 + 1) = 90%
- Specificity = 60 / (60 + 30) = 67%
- In this example, troponin is fairly sensitive but not very specific
