A1396
Title: Detection of ADR signals in spontaneous reporting systems under potential reporting rate changes
Authors: Tatsuhiko Anzai - Institute of Science Tokyo (Japan) [presenting]
Kunihiko Takahashi - Institute of Science Tokyo (Japan)
Abstract: Detection of Adverse Drug reaction (ADR) signals in spontaneous reporting systems is a central task in pharmacovigilance, and disproportionality methods such as the reporting odds ratio (ROR) are widely used. However, these measures rely on stable reporting rates across Drug-reaction combinations, an assumption that may be violated when reporting rates change non-uniformly, as observed during large-scale disruptions such as the COVID-19 pandemic. The adjusted ROR extends the contingency table with four fluctuation parameters representing deviations in cell-specific reporting ratios. These parameters are estimated by minimizing a divergence between observed and predicted counts, with predictions derived from a regression model assuming stable reporting rates. The adjusted ROR rescales the observed ROR to recover a signal metric closer to the true odds ratio even when reporting conditions are unstable. Application to the Japanese Adverse Drug Event Report (JADER) database demonstrates that the adjusted ROR eliminated spurious signals that emerged under pandemic-related underreporting while preserving signals for Drug-ADR pairs with genuine pharmacological plausibility.
