Estimates of the added benefit from COVID-19 vaccine booster doses can inform seasonal vaccine recommendations. We set out to estimate COVID-19 vaccine effectiveness (VE), avoiding common sources of bias. We chose a modified screening method, using only vaccinated cases of symptomatic COVID-19 recorded in the mandatory infectious disease reporting system, with data from a vaccination registry. Effect estimates were obtained by Bayesian logistic regression, comparing outcomes within dose strata between immunized (15-21d after vaccination) and not yet immunized (up to 7d). VE estimates were calculated recursively and relative VE estimates were obtained to quantify the added benefit of booster doses. We found VE for clinical illness to be around 90% during the predominance of the SARS-CoV-2 Alpha variant. During the Delta period, VE was distinctly lower, but did not decrease much further. The first booster dose added substantial protection during the Delta period, but that benefit became marginal during the Omicron period in the 18+. The effect of second booster doses, which became widespread only during the Omicron period, was modest. Using a novel, vaccinated-only study design we found that two doses of a COVID-19 vaccine offered substantial protection from symptomatic COVID-19, even during the Omicron period. One booster dose was highly effective during the Delta period, but the effect became modest during the Omicron period, when second booster doses offered, but a marginal benefit.
Foppa et al. (Thu,) studied this question.