Predictive values | Brookbush Institute Glossary
See all Glossary Terms

Predictive values

Predictive values in diagnostic testing, are the proportion of true positives and true negatives. The positive predictive value (PPV) and negative predictive value (NPV) describe the accuracy of a diagnostic test; however, unlike sensitivity and specificity, predictive values are largely dependent on the dysfunctions prevalence in the examined population (1). For example, imagine a patient who just tested positive on a special test. If the test was positive, what is the probability that they really have that dysfunction; how worried should they be? Now, imagine the same test performed on a very different individual in a different setting with very different subjective indicators. How do you think that would effect predictive value?
  • Positive predictive value (PPV) - describes the probability of having the dysfunction of interest in a subject with a positive result. Therefore PPV represents the proportion of patients with positive test result that are positive in a given population.
  • Negative predictive value (PPV) - describes the probability of not having the dysfunction of interest in a subject with a negative test result. Therefore NPV represents the proportion of patients without the disease that are negative in a given population.
  1. Šimundić, A. M. (2009). Measures of diagnostic accuracy: basic definitions. Ejifcc19(4), 203