Change Point Analysis for Detecting Vaccine Safety Signals

Area of Science:

• Pharmacovigilance and Drug Safety
• Biostatistics and Data Analysis
• Public Health Surveillance

Background:

• Timely detection of adverse event reporting changes is crucial for vaccine safety.
• National immunization programs require robust surveillance systems to identify public safety concerns.
• Existing methods may not be optimal for detecting abrupt shifts in event reporting rates.

Purpose of the Study:

• To evaluate the effectiveness of change point analysis (CPA) methods for vaccine safety signal detection.
• To compare the performance of Bayesian CPA, Taylor-CPA, and EnvCpt using simulated and real-world data.
• To assess the utility of CPA in identifying safety signals within vaccine adverse event databases.

Main Methods:

• Assessed three CPA methods: Bayesian CPA, Taylor-CPA, and EnvCpt.
• Utilized simulated data with varying baseline event rates and change magnitudes.
• Validated performance using the Korea Adverse Event Reporting System (KAERS) database, focusing on human papillomavirus (HPV) immunization adverse events.

Main Results:

• Taylor-CPA showed high accuracy (96-100%) and specificity (98-100%) in simulation studies.
• Taylor-CPA and EnvCpt identified a significant change point in HPV immunization adverse events in Q2/2013 in the KAERS database.
• The detected change point correlated with known public safety concerns.

Conclusions:

• Change point analysis is a viable and efficient tool for early vaccine safety signal detection.
• Taylor-CPA offers superior performance compared to other assessed CPA methods for this application.
• CPA implementation can enhance the proactive monitoring of vaccine safety in national immunization programs.