Improving Assessment of Vaccine Effectiveness by Coupling Test-negative Design Studies with Survival Models

Area of Science:

• Epidemiology
• Biostatistics
• Vaccinology

Background:

• The test-negative design (TND) is a key observational study for vaccine effectiveness (VE) evaluation, particularly for COVID-19 vaccines.
• TND has traditionally been analyzed as a case-control study using logistic regression.
• This limits its application and analytical flexibility.

Purpose of the Study:

• To reframe the TND as a cohort study design.
• To introduce a novel statistical model for TND data analysis.
• To enhance the estimation of vaccine effectiveness against infections and reinfections.

Main Methods:

• Framing the TND as a special case of a cohort study.
• Developing and applying the Prentice, Williams, and Peterson gap-time (PWP-GT) frailty model.
• Conducting extensive simulation studies to compare model performance.
• Analyzing real-world data from the National COVID Cohort Collaborative.

Main Results:

• The TND can be robustly analyzed using cohort study methods.
• The PWP-GT model effectively accounts for recurrent infections and time-varying vaccination status.
• Simulations show the PWP-GT model outperforms traditional methods for TND.
• Real-world analysis estimated Pfizer vaccine effectiveness against Omicron infections and reinfections.

Conclusions:

• The TND offers greater analytical potential when viewed as a cohort study.
• The PWP-GT model provides a superior approach for analyzing TND data.
• This methodology enhances the accuracy of vaccine effectiveness studies, especially for COVID-19.