Last update:
21-Sep-2006

The clinical usefulness of a diagnostic test (laboratory test or clinical finding) is determined by the discriminating ability of the test, meaning its ability to correctly classify patients with or without the disease. Likelihood ratios summarize the same kind of information as sensitivity and specificity (likelihood) but provide an alternative, more powerful way of describing the discriminating ability of a diagnostic test. Likelihood ratios express how many times more (or less) likely a test result is to be found in diseased, as compared to nondiseased, people. Their main advantage is that, unlike sensitivity and specificity (which limit the number of test results to just two levels, "positive" and "negative"), they can be generated for multiple levels of the diagnostic test results, taking into account the degree of positivity of the test. When applied to the pretest odds of the target disorder, likelihood ratio generates the posttest odds for that disorder, by means of a simple formulation of Bayes' theorem. Finally, likelihood ratios (positive and negative) provide a more appropriate comparison of the diagnostic abilities of two or more diagnostic tests. Although derivatives of the likelihood ratio provide the same kind of information as likelihoods and their ratios, they are mainly used because they facilitate calculations.

Key words: Discriminating ability, Expected weight of evidence, Likelihood ratio, Potential range of evidence, Weight of evidence.