Sensitivity and Specificity should be De-emphasized in Diagnostic Accuracy Studies

摘要: In the era of evidence-based medicine, not only therapies but also diagnostic procedures must undergo critical evaluations. Despite recent efforts, however, methodological framework for evaluations is still markedly incomplete, in contrast to many guidelines randomized trials and observational etiologic studies (1–14). A large proportion accuracy have serious flaws or provide results with limited practical applicability (15,16). Apparently, there a gap between research practice. Improvement remains challenge near future. These conclusions can be drawn from debate among various researchers (1,9–14) about future research, initiated by an article Feinstein (2). The practice was rationale Standards Reporting Diagnostic Accuracy (STARD) initiative. objective STARD initiative improve completeness reporting accuracy, allow readers assess potential bias study evaluate its generalizability (17,18). result checklist 25 items, recently published journals. This encourages investigators report necessary elements helps better understand reports their quality. are impressive. We fully subscribe selection which indeed major importance ensure proper understanding research. reservations first item, recommends use sensitivity specificity as medical subject headings (MeSH) characterizing studies. do agree such we believe that parameters our view, these relevance practice, estimation should necessarily pursued this article, briefly explain thinking.