Statistical Evaluation of Diagnostic Tests - Part 2 [Pre-test and post-test probability and odds, Likelihood ratios, Receiver Operating Characteristic Curve, Youden's Index and Diagnostic test biases].

I t h e p r e v i o u s a r t i c l e o n t h e s t a t i s t i c a l e v a l u a t i o n of diagnostic tests –Part 1, we u n d e r s t o o d t h e m e a s u r e s o f sensitivity, specificity, positive and negative predictive values. The use of these metrices stems from the fact that no diagnostic test is ever perfect and every time we carry out a test, it will yield one of four possible outcomes– true positive, false positive, true negative or false negative. The 2 x 2 table [Table 1] gives each of these four possibilities along with their mathematical calculations when a new test is compared with a gold standard test.1 In this article, the second in the diagnostic test series, we will discuss single summary statistics that help us understand and use these tests appropriately both in the clinical context and when these summary statistics appear in literature. Before we discuss these, we need to recapitulate a few concepts presented in earlier articles [odds and probability] and also some novel concepts [Bayesian statistics, pre-test and post-test probabilities and odds]. From the example, it follows that Odds = p/1-p, where p is the probability of the event occurring. Probability, on the other hand, is given by the formula p = Odds/1+Odds