Continuous Post-Market Sequential Safety Surveillance with Minimum Events to Signal

Area of Science:

• Pharmacovigilance
• Biostatistics
• Public Health

Background:

• The Vaccine Safety Datalink project uses near real-time surveillance for rapid adverse event detection.
• Continuous sequential methods allow near-continuous data evaluation while maintaining statistical integrity.
• Current methods may signal adverse events with very few observed cases.

Purpose of the Study:

• To explore modifications of continuous sequential monitoring for vaccine safety surveillance.
• To evaluate methods that require a minimum number of observed events before rejecting the null hypothesis.
• To assess continuous sequential analysis with a delayed start.

Main Methods:

• Investigated modified continuous sequential monitoring designs.
• Analyzed scenarios with a minimum event threshold for hypothesis rejection.
• Evaluated delayed-start continuous sequential analysis.
• Provided tables for critical values, statistical power, and average signal times.

Main Results:

• Requiring a minimum number of observed events can increase statistical power and decrease the expected time to signal.
• Continuous sequential analysis with a delayed start is primarily beneficial for logistical reasons when initial data is available.
• Maintained the overall alpha level across all evaluated methods.

Conclusions:

• Modified continuous sequential monitoring offers improved efficiency and power for vaccine safety surveillance.
• These enhanced methods can lead to faster and more reliable detection of adverse events.
• The findings support the optimization of post-market vaccine safety monitoring systems.