A Likelihood Ratio Test Based Method for Signal Detection With Application to FDA’s Drug Safety Data

Three‐Component Mixture Model‐Based Adverse Drug Event Signal Detection for the Adverse Event Reporting System

“…Thus, llr is used to rank signals. They are consistent with their P value‐based rankings . We calculated the BFDR under the EBGM model and set = 2 according to Ahmed et al .…”

“…Under the null, doesn't have a closed form distribution. As a consequence, the authors simulated the null distribution to calculate P values …”

Translational Biomedical Informatics and Pharmacometrics Approaches in the Drug Interactions Research

“…Practically, PRR_025 and ROR_025, the lower bound of 95% confidence intervals for PRR and ROR, are also often used for the signal detection. The likelihood ratio test (LRT) is another frequentist DPA . It assumed that the drug‐induced ADE frequency follows a Poisson distribution.…”

Quantitative Signal Detection and Analysis in Pharmacovigilance

“…While the overall intent of the Bayesian approaches is that overall false positives should be reduced with a limited practical impact on false-negative rate, on an individual combination level there is naturally the possibility that this shrinkage can lead to a false negative; and there has been some research suggesting this to be the case overall for EB05 -the lower 95% confidence limit of the EBGM; for example, see Chen et al (2008) and Hauben et al (2007a) -although there are to our knowledge no studies showing that this is a problem in practice for the IC. Recent variations of frequentist methodology have appeared (Huang et al, 2012;Johnson et al, 2012), some of which have been reported to produce results similar to empirical Bayesian approaches in initial investigations (Johnson et al, 2012). For a summary of different measures of disproportionality see Table 20.3.…”