Sensitivity and Specificity

In describing how useful a test or symptom is for making a diagnosis, doctors often throw around terms like ‘sensitivity’ and ‘specificity’. Here is an attempt at clarification.

The sensitivity of a test is the proportion of those people who actually have the disease who have a positive test result. When a test has a high sensitivity, a negative result can help rule out the diagnosis. For example, the sensitivity of serum iron measurement for diagnosing iron-deficiency anemia is 90%; therefore if a person does not have a low serum iron, it is highly unlikely that the person has anemia. To help remember this, think of ‘SNOUT’ – SeNsitivity negative rules OUT.

The specificity of a test is the proportion of people without the disease who have a negative test result. When a test has a high specificity, a positive result ‘rules in’ the diagnosis. For example, the specificity of a low serum iron level for diagnosing anemia is quite high, at 85%; therefore if a person has a low serum iron, it confirms the diagnosis of iron deficiency anemia. Think ‘SPIN’ – SPecificity high rules IN.

These two estimates have limited usefulness; for instance, they are influenced by the overall frequency of a particular disease. They are therefore often replaced by other calculations, such as the ‘predictive value’ and the ‘likelihood ratio’. More on these later, if there’s sufficient interest.