Moving beyond sensitivity and specificity: Using likelihood ratios to help interpret diagnostic tests

Abstract

Properties of diagnostic tests have traditionally been described using sensitivity, specificity, and positive and negative predictive values. These measures, however, reflect population characteristics and do not easily translate to individual patients. Likelihood ratios are a more practical way of making sense of diagnostic test results and have immediate clinical relevance. In general a useful test provides a high positive likelihood ratio and a small negative likelihood ratio.