Recommendations for estimating and reporting vaccine effectiveness by time since vaccination: a COVID-19 case study

Category	Detail
Condition	COVID-19
Key Mechanisms	Vaccine effectiveness (VE) changes over time since vaccination (TSV), influenced by waning immunity and viral variant/sublineage evolution
Target Population	Individuals vaccinated against COVID-19, across age groups and vaccine types/brands
Care Setting	Observational case–control studies including test-negative design studies in healthcare settings

Estimating VE by TSV is essential to understand protection dynamics and enable meaningful comparisons across variants, age groups, and vaccine brands.
Case–control studies require precise definitions of study population, index dates, vaccination exposure, and outcomes for accurate VE by TSV estimation.
Two main modeling approaches for VE by TSV are recommended: categorical TSV analysis and continuous TSV modeling, depending on study objectives.

Define study population, period, index date (e.g., symptom onset), vaccination exposure (≥14 days before index date), and outcome (e.g., lab-confirmed infection).
Restrict cases by variant/sublineage using sequencing or study period to control for immune escape effects.
Calculate TSV by subtracting vaccination date from index date; use symptom onset for cases and recruitment date for controls.

Use high-quality, complete vaccination and index date data to avoid imprecise or biased VE estimates.
Code unvaccinated or reference group participants as 0 days since vaccination; note they may have other vaccine doses.
Plot vaccinated cases and controls by TSV day to guide modeling and interpret VE trends.

Monitor VE changes over time to detect waning immunity and variant-specific differences.
Apply appropriate modeling approaches (categorical or continuous TSV) based on research questions.
Report VE estimates by TSV to allow valid comparisons across vaccine brands and variants.

Bias from incomplete or inaccurate vaccination and index date data can lead to over- or underestimation of VE.
Failure to account for TSV can confound comparisons between vaccine brands or variants due to differing vaccination or epidemic timing.
Ignoring variant/sublineage circulation timing may misattribute VE decline to waning rather than immune escape.

Vaccinated individuals in observational case–control studies, including test-negative designs

VE estimates must be stratified or modeled by TSV to accurately reflect protection levels and inform vaccine brand or booster recommendations

Define and document all key study parameters including index dates, vaccination status, and outcome definitions clearly in protocols and statistical analysis plans.
Ensure high completeness and quality of vaccination and index date data; use missing data methods cautiously.
Use graphical descriptive analyses of TSV distributions by case status to inform modeling strategy.
Select modeling approach (categorical vs continuous TSV) aligned with study objectives and audience needs.
Report VE estimates by TSV with transparent methods to facilitate comparability across studies and contexts.

Clinical Scorecard: Guidelines for Assessing and Presenting Vaccine Effectiveness Over Time Since Vaccination: Insights from a COVID-19 Case Analysis