IN THIS ISSUE

Extract

Assessing the Probability That a Positive Report is False

The genomic revolution presents exciting opportunities to learn about the etiology of cancer and other complex diseases. In a commentary, Wacholder et al. (p. 434) argue that the practice of using the P value alone to declare a finding to be statistically significant is no longer appropriate for deciding which of the many reports of associations between genetic variants and common cancer sites are truly significant. They propose, instead, that investigators use the false-positive report probability (FPRP)—the probability of no true association between a genetic variant and a disease, given a statistically significant finding—to evaluate whether a finding is noteworthy. The authors show how to calculate FPRP and demonstrate how FPRP can be used in the design, analysis, and interpretation of molecular epidemiology studies. The authors conclude that the FPRP approach helps formalize what investigators have always been performing informally, that is, tempering enthusiasm for surprising findings with consideration of plausibility.