To the Editor: Dr Kline and colleagues1 concluded that D-dimer assay and ADSM should be used together to exclude the presence of a PE. They reported a sensitivity of 98.4% and a specificity of 51.6% for the 2 tests combined. In their study population, the prevalence of PE was 16.8%, and if both tests were normal, they estimated a 0.75% posterior probability of having a PE.
The authors considered the test results jointly and dichotomized the test results to at least 1 result being abnormal or to both results being normal. They report sensitivity, specificity, positive likelihood ratio (LR), and negative LR for this dichotomized joint test. However, since 4 possible combinations of test results were possible, information was lost. Reporting categorical LRs would have been more useful.