We thank Stamey and colleagues
1 for their comments on our article.
2 We are in complete agreement that appropriate statistical methodology is important in clinical medicine. The alternative statistical approach they provided is robust in its ability to account for potential conditional dependence (although this parameter is difficult to quantify in the clinical setting). The similarity of results between our approach and their alternative technique is striking. It would be interesting to see whether similar results were yielded by applying the authors' alternative analysis to the other three data sets discussed in our original paper. Most important, we would like to reaffirm that both approaches confirm that the true sensitivity of a temporal artery biopsy is substantially lower than 100%. This is critical in the context of clinical practice because this supposed “gold standard” test is imperfect and can frequently yield falsely negative results.