Subgroup and Sensitivity Analyses

Delivery-mode subgroup analysis will compare in-person interventions with remote or digital interventions. Age-group subgroup analysis will compare infants and toddlers, preschool children, and school-age children. Country income level will be analysed as low- and middle-income countries versus high-income countries. Sensitivity analysis will exclude studies rated high risk of bias to test whether observed estimates change. Subgroup analyses will explore heterogeneity in intervention effects when sufficient data are available. Results were consistent across trimester of first dose, vaccine product, and number of doses received. A time-varying Cox model for very preterm birth produced a significant increased hazard, but the authors cautioned against overinterpreting it. Excluding individuals with documented SARS-CoV-2 infection did not substantially change results. The apparent increased risk of preterm birth after only one dose was attributed to reverse causation.