To the Editor.
—At the Second Nordic Workshop on How to Practice Evidence-Based HealthCare, the article by Dr Wagner and colleagues entitled "Does This Patient Have Appendicitis?"1 was used as the basis for the discussion groups on the issue of how to use an article about a diagnostic test.The sensitivities and specificities listed in Table 3 of that article do not correspond with the listed likelihood ratios (LRs) if we use the usual method of LR calculation.2,3 For example, the first LRs for right lower quadrant pain in Table 3 are given as 7.31 to 8.46 (positive LR) and 0 to 0.28 (negative LR). When calculated using the sensitivity and specificity data given, the positive LR, calculated as (sensitivity/[1-specifity]), becomes 0.81/0.47, or 1.72; the negative LR ([1-sensitivity]/specificity) is 0.19/0.53, or 0.36. Similar lower LRs were found for the entire Table 3.This means that the power
Witt K, Mäkelä M, Olsen O. Likelihood Ratios to Determine 'Does This Patient Have Appendicitis?': Comment and Clarification. JAMA. 1997;278(10):819. doi:10.1001/jama.1997.03550100045026