Abstract Reaching population immunity against COVID‐19 is proving difficult even in countries with high vaccination levels. Thus, it is critical to identify limits of control and effective measures against future outbreaks. The effects of nonpharmaceutical interventions (NPIs) and vaccination strategies are analyzed with a detailed community‐specific agent‐based model (ABM). The authors demonstrate that the threshold for population immunity is not a unique number, but depends on the vaccination strategy. Prioritizing highly interactive people diminishes the risk for an infection wave, while prioritizing the elderly minimizes fatalities when vaccinations are low. Control over COVID‐19 outbreaks requires adaptive combination of NPIs and targeted vaccination, exemplified for Germany for January–September 2021. Bimodality emerges from the heterogeneity and stochasticity of community‐specific human–human interactions and infection networks, which can render the effects of limited NPIs uncertain. The authors' simulation platform can process and analyze dynamic COVID‐19 epidemiological situations in diverse communities worldwide to predict pathways to population immunity even with limited vaccination.

Reaching population immunity against COVID-19 is proving difficult even in countries with high vaccination levels. We demonstrate that this in part is due to heterogeneity and stochasticity resulting from community-specific human-human interaction and infection networks. We address this challenge by community-specific simulation of adaptive strategies. Analyzing the predicted effect of vaccination into an ongoing COVID-19 outbreak, we find that adaptive combinations of targeted vaccination and non-pharmaceutical interventions (NPIs) are required to reach population immunity. Importantly, the threshold for population immunity is not a unique number but strategy and community dependent. Furthermore, the dynamics of COVID-19 outbreaks is highly community-specific: in some communities vaccinating highly interactive people diminishes the risk for an infection wave, while vaccinating the elderly reduces fatalities when vaccinations are low due to supply or hesitancy. Similarly, while risk groups should be vaccinated first to minimize fatalities, optimality branching is observed with increasing population immunity. Bimodality emerges as the infection network gains complexity over time, which entails that NPIs generally need to be longer and stricter. Thus, we analyze and quantify the requirement for NPIs dependent on the chosen vaccination strategy. Our simulation platform can process and analyze dynamic COVID-19 epidemiological situations in diverse communities world-wide to predict pathways to population immunity even with limited vaccination.

SummaryInfectious disease outbreaks challenge societies by creating dynamic stochastic infection networks between human individuals in geospatial and demographical contexts. Minimizing human and socioeconomic costs of SARS-CoV-2 and future global pandemics requires data-driven and context-specific integrative modeling of detection-tracing, healthcare, and non-pharmaceutical interventions for decision-processes and reopening strategies. Traditional population-based epidemiological models cannot simulate temporal infection dynamics for individual human behavior in specific geolocations. We present an integrated geolocalized and demographically referenced spatio-temporal stochastic network- and agent-based model of COVID-19 dynamics for human encounters in real-world communities. Simulating intervention scenarios, we quantify effects of protection and identify the importance of early introduction of test-trace measures. Critically, we observe bimodality in SARS-CoV-2 infection dynamics so that the outcome of reopening can flip between good and poor outcomes stochastically. Furthermore, intervention effectiveness depends on strict execution and temporal control i.e. leaks can prevent successful outcomes. Schools are in many scenarios hubs for transmission, reopening scenarios are impacted by infection chain stochasticity and subsequent outbreaks do not always occur. This generalizable geospatial and individualized methodology is unique in precision and specificity compared to prior COVID-19 models [6, 16, 17, 19] and is applicable to scientifically guided decision processes for communities worldwide.

Interactions between humans cause transmission of SARS-CoV-2. We demonstrate that heterogeneity in human-human interactions give rise to non-linear infection networks that gain complexity with time. Consequently, targeted vaccination strategies are challenged as such effects are not accurately captured by epidemiological models assuming homogeneous mixing. With vaccines being prepared for global deployment determining optimality for swiftly reaching population level immunity in heterogeneous local communities world-wide is critical. We introduce a model that predicts the effect of vaccination into an ongoing COVID-19 outbreak using precision simulation of human-human interaction and infection networks. We show that simulations incorporating non-linear network complexity and local heterogeneity can enable governance with performance-quantified vaccination strategies. Vaccinating highly interactive people diminishes the risk for an infection wave, while vaccinating the elderly reduces fatalities at low population level immunity. Interestingly, a combined strategy is not better due to non-linear effects. While risk groups should be vaccinated first to minimize fatalities, significant optimality branching is observed with increasing population level immunity. Importantly, we demonstrate that regardless of immunization strategy non-pharmaceutical interventions are required to prevent ICU overload and breakdown of healthcare systems. The approach, adaptable in real-time and applicable to other viruses, provides a highly valuable platform for the current and future pandemics.

Re-opening societies and economies across the globe following the initial wave of the severe acute respiratory syndrome-coronavirus 2 (SARS-CoV-2) pandemic requires scientifically-guided decision processes and policy development. Public health authorities now consider it highly likely that transmission of SARS-CoV-2 and COVID-19 will follow a pattern of seasonal circulation globally. To guide mitigation strategies and tactics in a location-specific manner, accurate simulation of prolonged or intermittent patterns of social/physical distancing is required in order to prevent healthcare systems and communities from collapsing. It is equally important to capture the stochastic appearance of individual transmission events. Traditional epidemiological/statistical models cannot make predictions in a geospatial temporal manner based on human individuals in a community. Thus, the challenge is to conduct spatio-temporal simulations of transmission chains with real-world geospatial and georeferenced information of the dynamics of the disease and the effect of different mitigation strategies such as isolation of infected individuals or location closures. Here, we present a stochastic, geospatially referenced and demography-specific agent-based model with agents representing human beings and include information on age, household composition, daily occupation and schedule, risk factors, and other relevant properties. Physical encounters between humans are modeled in a time-dependent georeferenced network of the population. The model (GERDA-1) can predict infection dynamics under normal conditions and test the effect of different mitigation scenarios such as school closures, reduced social contacts as well as closure or reopening of public/work spaces. Specifically, it also includes the fate and influence of health care workers and their access to protective gear. Key predictions so far entail: (i) the effect of specific groups on the spreading, specifically that children in school contribute substantially to distribution. (ii) the result of reopening society depends crucially on how strict the measures have been during lock-down. (iii) the outcome of reopening is a stochastic process - in the majority of cases, we must expect a second wave, in some cases not. To the best of our best knowledge, the GERDA-1 model is the first model able to predict a bimodal behavior of SARS-Cov-2 infection dynamics. Given the criticality of the global situation, informing the scientific community, decision makers and the general public seems prudent. Therefore, we here provide a pre-print of the GERDA-1 model together with a first set of predictions and analyses as work in progress.