LRs are defined by the ratio of the probability of the test result in the population carrying the disease versus the probability in the non-diseased population. The answer to that question lies in applying Bayes’ theorem : pretest odds multiplied by the likelihood ratio (LR) of the laboratory test result give the posttest odds ( S1 Appendix). For example, how much does a D-dimer result of 1000 μg/L, double the recommended cut-off, increase or decrease the clinical suspicion of thrombosis. Many modern laboratory techniques provide quantitative test results and it would be important to know how much a particular test result would increase or decrease the odds for a specific diagnosis. Laboratory tests can significantly contribute to this effort by modifying the probabilities of a diagnosis. The production of patient-specific information is the main objective of the clinical laboratory. Diagnostics is about connecting these two types of information. Two types of medical information can be distinguished, patient-specific and knowledge-based. Medical diagnostics is an information processing endeavor based on probabilities. The use of such a method would help in standardizing quantitative test results, which are not always comparable between different test providers, by reporting them as LRs for a specific diagnosis, in addition to, or instead of, quantities such as mg/L or nmol/L, or even indices or units. Here, a simple distribution-independent method is described to reach this goal using Bézier curves that are defined by tangents to a curve. Since the exact distribution of test results in diseased and non-diseased people is often not known, the calculation of such LRs for quantitative test results is not straightforward. These LRs are equal to the slope of the tangent to the ROC curve at the corresponding point. However, since many diagnostic tests are of quantitative nature, it would be desirable to estimate LRs for each quantitative result. Similarly, LRs can be calculated for certain ranges of test results. In this way, positive and negative likelihood ratios (LR+ and LR-) can be calculated to be used in Bayes’ way of estimating disease probabilities. Receiver operating characteristic (ROC) analysis is widely used to describe the discriminatory power of a diagnostic test to differentiate between populations having or not having a specific disease, using a dichotomous threshold.

Curva ROC de Microsoft Excel