A Tree-Based Scan Statistic for Detecting Signals of Drug-Drug Interactions in Spontaneous Reporting Databases.

Area of Science:

• Pharmacovigilance
• Biostatistics
• Computational Biology

Background:

• Drug-drug interactions (DDIs) pose significant risks for adverse events (AEs).
• Current safety monitoring struggles to identify DDIs due to limitations in clinical trials and existing statistical methods.
• Existing methods often fail to account for the hierarchical nature of AEs and potential reporting biases.

Purpose of the Study:

• To develop an advanced statistical methodology for detecting DDI signals.
• To address the limitations of existing methods by incorporating AE hierarchy and reporting bias.
• To enhance postmarket drug safety surveillance.

Main Methods:

• Developed a novel statistical methodology using tree-based scan statistics.
• Incorporated the hierarchical structure of adverse events (AEs).
• Mitigated reporting bias using a multiplicative interaction model for DDIs.

Main Results:

• The proposed method effectively controlled the type I error rate.
• Demonstrated consistent performance in power, sensitivity, and false discovery rate.
• Showcased robust performance even in the presence of reporting bias.

Conclusions:

• The novel tree-based scan statistic methodology offers a valuable tool for DDI signal detection.
• This approach enhances postmarket drug safety surveillance by accounting for AE hierarchy and reporting bias.
• The method provides a more accurate and reliable way to identify potential drug-drug interaction risks.