| Summary: | There is a current demand for “safety signal” screening, not only for single drugs but also for drug-drug interactions. The detection of drug-drug interaction signals using the proportional reporting ratio (<i>PRR</i>) has been reported, such as through using the combination risk ratio (<i>CRR</i>). However, the <i>CRR</i> does not consider the overlap between the lower limit of the 95% confidence interval of the <i>PRR</i> of concomitant-use drugs and the upper limit of the 95% confidence interval of the <i>PRR</i> of single drugs. In this study, we proposed the concomitant signal score (<i>CSS</i>), with the improved detection criteria, to overcome the issues associated with the <i>CRR</i>. “Hypothetical” true data were generated through a combination of signals detected using three detection algorithms. The signal detection accuracy of the analytical model under investigation was verified using machine learning indicators. The <i>CSS</i> presented improved signal detection when the number of reports was ≥3, with respect to the following metrics: accuracy (<i>CRR</i>: 0.752 → <i>CSS</i>: 0.817), Youden’s index (<i>CRR</i>: 0.555 → <i>CSS</i>: 0.661), and <i>F</i>-measure (<i>CRR</i>: 0.780 → <i>CSS</i>: 0.820). The proposed model significantly improved the accuracy of signal detection for drug-drug interactions using the <i>PRR</i>.
|
|---|