Diagnostic testing, pre- and post-test probabilities, and their use in clinical practice

Abstract

In the process of establishing a diagnosis, the clinician calculates the disease probabilities, which keep changing as data is gathered from clinical history, physical examination and laboratory and imaging data. This paper sets out to explain, in a simple and practical way, how the characteristics of a test can be applied in order to determine post-test probability, that is, the probability of a specific disease being present. Before applying the test, we start with a pretest probability that corresponds to the prevalence of the disease. With each additional test this probability changes, until a threshold is reached that includes or excludes the clinical diagnosis. To calculate the post-test probabilities, the sensitivity and specificity of the test are used to calculate the positive and negative predictive values for a determined pretest probability, as well as likelihood ratios (through graphics or nomograms, approximations or equations). Among the test characteristics, the most useful for calculating post-test probabilities are likelihood ratios, which have several advantages over sensitivity and specificity.