Vaccination schedule and efficacy

Since December 2020, China has given conditional approval or emergency use approval for seven COVID-19 vaccines. The National Health Commission recommends that inactivated vaccines are administered on a two-dose schedule with an interval of ≥21 days, recombinant subunit vaccines administered on a three-dose schedule with an interval of ≥28 days, and recombinant adenovirus type-5-vectored vaccines administered one dose. For simplicity, in the main analysis, we modelled the administration of an inactivated vaccine developed by the Beijing Institute of Biological Products⁶⁴, which entails a two-dose schedule across all age groups with an interval of 21 days. In separate sensitivity analyses, we explored an interval of 14 and 28 days (SA16 and SA17).

China approved its first local COVID-19 vaccine (developed by Sinopharm) for general public use on 31 December 2020, with an estimated VE of 79.3%²¹. In the main analysis, we used a VE of 80% against infection in individuals aged 20–59 years. In the developed model, vaccination confers partial protection, that is, vaccinated individuals are 80% less likely to develop infection upon an infectious contact. Sensitivity analyses using a VE of 60% (SA9) and 90% (SA10) were performed separately. The alternative values of VE were selected on the basis of published upper efficacy of vaccines of 94–95% and in such a way to cover a plausible efficacy range of forthcoming vaccines^65–67.

Phase 2 clinical trials demonstrated that vaccine immunogenicity was lower among older individuals than in younger adults⁶⁴. And for other inactivated vaccines such as influenza vaccine, a lower VE is observed in children compared with young adults⁶⁸. Accordingly, we assumed an age-dependent VE. In particular, given a baseline efficacy VE among individuals aged 20–59 years (80% in the main analysis), we assumed a 50% lower VE in individuals <20 and ≥60 years of age (namely 40%). A scenario without age-specific variations in VE was explored as sensitivity analysis SA24.

Individuals vaccinated with the first dose could still develop infections without any immune protection, while the second dose vaccination could produce the expected VE after an average of 14 days. In the main analysis we assume both natural infection-induced and vaccine-induced immunity to SARS-CoV-2 infection does not wane within the considered time horizon (2 years). In additional sensitivity analyses, we considered an average duration of vaccine-induced protection of 6 months (SA12) and 1 year (SA25). We also consider a sensitivity analysis assuming that vaccination is effective in preventing symptomatic illness but not infection (SA11), and another one assuming an all-or-nothing vaccine, that is, the vaccine confers full protection to VE percent of vaccinated individuals (SA18).