Sensitivity and specificity

  • The sensitivity of a test is the percentage of results that will be (correctly) positive when HIV is actually present. Lower rates of sensitivity will produce more false negative results.
  • The specificity of a test is the percentage of results that will be (correctly) negative when HIV is not present. Lower rates of specificity will produce more false positive results.

Sensitivity and false negative rates are two different ways of stating the same thing – a test's ability to correctly identify infected individuals. Similarly, specificity and false positive rates are two different ways of measuring a test's ability to identify uninfected individuals.

This is perhaps best illustrated by example. Imagine a population of 100,000 people, of whom 10,000 are actually HIV-infected and 90,000 HIV-uninfected. The goal of a screening procedure would be to correctly identify all 10,000 – and only those 10,000 – HIV-positive individuals. A test with 99% sensitivity would correctly identify 9900 of the 10,000 infected individuals (99%). One hundred infected people would test falsely negative (a false-negative rate of 1%).

If the same test procedure had a specificity of 90% (an unrealistically low value used to simplify this example), it would correctly identify 90% (81,000) of the HIV-uninfected individuals as HIV-negative. However, 9000 HIV-uninfected persons would test falsely positive (a false-positive rate of 9000/90,000 = 10%).

The best strategy, then, is to use a two-part testing procedure: a test with high sensitivity (to detect as many HIV-positive individuals as possible, allowing some false positives but very few false negatives), followed by a confirmatory test with high specificity (to eliminate as many of the false positives as possible). This is the strategy used in practice (as described in detail in the section on Confirmatory tests).1

The calculations for sensitivity and specificity can be expressed as equations in the following way:

  • Sensitivity = (true positive results) / (true positive results + false negatives)
  • Specificity = (true negative results) / (true negative results + false positives)
Related Links

References

  1. Altman DG, Bland JM Statistics notes: Diagnostic tests 1 - sensitivity and specificity. BMJ 308:1552, 1994
This content was checked for accuracy at the time it was written. It may have been superseded by more recent developments. NAM recommends checking whether this is the most current information when making decisions that may affect your health.