Subset Analysis for Screening Drug-Drug Interaction Signal Using Pharmacovigilance Database

Area of Science:

• Pharmacovigilance
• Computational Toxicology
• Data Mining

Background:

• Multi-drug combinations are common, necessitating understanding of drug-drug interactions (DDIs) in adverse event profiles.
• Existing DDI signal detection algorithms require simpler models for early adverse event detection.
• Previous subset analysis methods for DDI signal detection may have lacked appropriate criteria, leading to false positives.

Purpose of the Study:

• To present and verify an appropriate criterion for subset analysis in DDI signal detection.
• To improve the accuracy and reduce false positives in identifying DDIs from adverse drug event reports.

Main Methods:

• Utilized the Japanese Adverse Drug Event Report (JADER) database.
• Generated 'hypothetical' true data by combining signals from three established detection algorithms.
• Developed and validated a new subset analysis criterion using machine learning performance indicators.

Main Results:

• The newly proposed subset analysis significantly improved signal detection metrics: Accuracy (0.584 to 0.809), Precision (0.302 to 0.596), Specificity (0.583 to 0.878), Youden's index (0.170 to 0.465), F-measure (0.399 to 0.592), and NPV (0.821 to 0.874).
• The new method substantially reduced the number of false DDI signals compared to previous subset analysis techniques.
• While slightly less accurate than the Ω shrinkage measure model, the proposed method offers a significant reduction in false positives.

Conclusions:

• The newly proposed subset analysis criterion is effective for improving DDI signal detection in pharmacovigilance.
• This approach enhances the reliability of DDI signal detection by minimizing false positives.
• The validated criterion offers a practical tool for early detection of adverse events related to drug combinations.