Investigating COVID-19 Vaccine Impact on the Risk of Hospitalisation through the Analysis of National Surveillance Data Collected in Belgium

Abstract

The national vaccination campaign against SARS-CoV-2 started in January 2021 in Belgium. In the present study, we aimed to use national hospitalisation surveillance data to investigate the recent evolution of vaccine impact on the risk of COVID-19 hospitalisation. We analysed aggregated data from 27,608 COVID-19 patients hospitalised between October 2021 and February 2022, stratified by age category and vaccination status. For each period, vaccination status, and age group, we estimated risk ratios (RR) corresponding to the ratio between the probability of being hospitalised following SARS-CoV-2 infection if belonging to the vaccinated population and the same probability if belonging to the unvaccinated population. In October 2021, a relatively high RR was estimated for vaccinated people > 75 years old, possibly reflecting waning immunity within this group, which was vaccinated early in 2021 and invited to receive the booster vaccination at that time. In January 2022, a RR increase was observed in all age categories coinciding with the dominance of the Omicron variant. Despite the absence of control for factors like comorbidities, previous infections, or time since the last administered vaccine, we showed that such real-time aggregated data make it possible to approximate trends in vaccine impact over time.

Keywords: COVID-19; SARS-CoV-2; hospitalisation surveillance; risk ratio; vaccination impact.

Figure 1

Approximation of vaccine impact on the risk of hospitalisation through the estimation of risk ratios (RR) between October 2021 and February 2022, based on the analyses of aggregated hospitalisation surveillance data collected in Belgium. (A) We here report RR values estimated for different age categories using a retrospective sliding window of four weeks. RR estimates are associated either with the “at least primary course completed” vaccination scheme (solid curves) or with the “boosted” vaccination scheme (dashed curves). Of note, in Belgium, hospitalised patients having had their booster jab were only registered as “boosted” at their admission from December 8, 2021. Therefore, we are only able to make the distinction between “primary course completed” and “boosted” vaccination schemes for the second half of the study period (15/12/22–28/02/22). Shaded polygons correspond to 95% confidence intervals associated with the “at least primary course completed” vaccination scheme (20). In addition, we also report (B) the progression of the Belgian booster vaccination campaign (shaded polygons *) and the evolution of the number of new hospitalisations in each slide of the retrospective sliding window of four weeks for the different age categories (curves), as well as (C) the temporal evolution of the relative detection frequency of the two main variants of concern (VOCs) circulating during that period in Belgium. (*) Because hidden on the graph by the progression of the booster vaccination campaign for younger age categories, the related progression for the 85+ is highlighted by a dashed curve from early 2022.