Featured research (4)
In this work we extend our previously developed compartmental SEIQRD model for SARS-CoV-2 in Belgium. The model is geographically stratified into eleven provinces and a telecommunication dataset provided by Belgium's biggest operator is used to incorporate interprovincial mobility. We introduce variants, seasonality, and vaccines in our model, as their addition has proven critical for the description, forecasting and understanding of the COVID-19 pandemic in Belgium. We then calibrate the model using the daily number of hospitalisations in each province and serological data. We demonstrate how our model can be used to set up hypothetical scenarios to study the combined impacts of new variants, an ongoing nation-wide vaccination campaign and social relaxations. In this way, our model can be used to provide policymakers with relevant insights on the optimal timing of the release of social restrictions. We finally discuss the impact of locally altering social contact and mobility on shielding or containing epidemics and find that lowering social contact is more efficient than lowering mobility to tame a SARS-CoV-2 epidemic.
Given that social interactions drive the spread of infectious diseases amongst humans, one anticipates that human mobility in Belgium affected the spread of COVID-19 during both 2020 "waves". Measures against this spread in turn influenced mobility patterns. In this study, we analyse and mutually compare time series of COVID-19-related data and mobility data across Belgium's 43 arrondissements (NUTS 3 level). First, we confirm that overall mobility did in fact change significantly over the consecutive stages of the pandemic. So doing, we define a quantity that represents the degree of mobility between two arrondissements: the "connectivity index". Second, we analyse spatio-temporal COVID-19-related incidence and hospitalisation data using dynamic time warping and time-lagged cross-correlation. This allows us to quantify time lag and morphological similarities between localised waves. Third, by coupling those analyses, we conclude that mobility between arrondissements is in fact indicative of the spatio-temporal spread of COVID-19 in Belgium, as was previously shown for other EU countries. This conclusion supports the need for mobility analysis and/or control during a pandemic, and by extension motivates the development of our spatially explicit Belgian metapopulation model.
In this report, we bring our epidemiological models in line with the additional measures taken between November 22nd, 2021, and January 28th, 2022, and reassess their impact on the fourth Belgian COVID-19 wave. A larger reduction of leisure contacts results in a faster decrease of COVID-19 hospitalizations and a lower overall burden of disease for the fourth COVID-19 wave. Neglecting the possible impact of the new Omicron variant due to a lack of quantitative data, the threshold of 500 IC patients is forecasted to occur between January 8th, 2022, and January 26th, 2022. It can thus be concluded that pressure on the healthcare system (HCS) will remain high in the near future, which is especially worrisome in light of a potential Influenza epidemic and the emergence of the novel Omicron variant.
In this report, we use two epidemiological models to explore the relative effects of three non-pharmaceutical interventions between November 17th, 2021, and December 24th, 2021 on the COVID-19 related hospitalizations during the fall and winter of 2021 in Belgium. Although the healthcare system will most likely not be overwhelmed by the fourth COVID-19 wave if no measures are taken, it might become severely strained during the following months. Non-pharmaceutical interventions (NPIs) can lower the overall disease burden of the fourth COVID-19 wave. Of the three NPIs considered in this report, school closure and mandatory telework have a smaller effect than reducing leisure contacts on the overall disease burden of the fourth COVID-19 wave. Raising the vaccine coverage is preferred over NPIs to prevent future COVID-19 outbreaks due to the societal impacts associated with the latter.