A modified self-controlled case series method to examine association between multidose vaccinations and death

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• The self-controlled case series (SCCS) method is used to study time-varying exposures and health outcomes.
• Analyzing the association between infant vaccination schedules and unexplained sudden unexpected death (uSUD) presents unique challenges due to multiple exposures and terminal events.

Purpose of the Study:

• To propose and evaluate an adapted SCCS method for analyzing vaccination risks, specifically focusing on uSUD.
• To compare a new SCCS adaptation, truncating observation periods by vaccination schedule, with a recently introduced method for complex exposure scenarios.

Main Methods:

• Development of a novel SCCS adaptation that truncates observation periods based on vaccination schedules, utilizing known minimum intervals between doses.
• Simulation studies using realistic age distributions and vaccination schedules to compare the new method with existing SCCS adaptations.
• Application of the SCCS methods to a real dataset of unexplained sudden unexpected death (uSUD) cases from the German case-control study on sudden infant death (GeSID).

Main Results:

• The adapted SCCS method with truncated observation periods demonstrated high power (≥90%) to detect a 16-fold risk increase after the 4th vaccine dose and 80% power to detect a 2-fold risk increase.
• Despite excluding 50% of cases, the new method showed only minor differences compared to the more complex SCCS adaptation.
• Reanalysis of GeSID data yielded slightly higher point estimates for vaccination risk using SCCS methods compared to the original case-control analysis.

Conclusions:

• The proposed SCCS adaptation offers a simpler approach to analyzing vaccination risks, particularly for events like uSUD, by truncating observation periods.
• The method is capable of detecting significant risk increases associated with infant vaccinations.
• While effective, the truncation approach may lead to the exclusion of a substantial proportion of potential cases.