Characterizing vaccine-associated risks using cubic smoothing splines.

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• Estimating short-term risks associated with childhood vaccines presents significant challenges.
• Existing methods may not adequately capture transient increases in risk following vaccination.

Purpose of the Study:

• To propose and evaluate a novel statistical approach for assessing short-term vaccine-related risks.
• To apply the method to analyze vaccine safety data for common childhood immunizations.

Main Methods:

• Utilized cubic smoothing splines to estimate daily event risk post-vaccination.
• Compared spline-predicted incidence rates with a log-linear trend excluding vaccination days.
• Employed a model-based bootstrap for confidence intervals.
• Applied to measles, mumps, rubella (MMR) and MMR-varicella (MMRV) vaccines in 1-year-old children.

Main Results:

• Identified significant spikes in fever, rash, and adenopathy diagnoses 9-11 days post-vaccination.
• Negative control outcomes showed predicted incidence consistent with modeled risks.
• Observed a potential increased risk for congenital malformation diagnoses, possibly due to healthcare-seeking behavior ('provider visit effect').

Conclusions:

• The proposed spline-based method effectively detects short-term vaccine-associated risks.
• The approach shows promise for improving vaccine safety surveillance systems.
• Further investigation is needed for unexpected findings like congenital malformations.