Modeling infectious disease dynamics towards informed public health interventions, with applications on COVID-19 and cholera

Emerging and existing infectious diseases pose a constant threat to individuals and communities across the world. In many cases, the burden of these diseases is preventable through public health interventions. However, taking the right decisions and designing effective policies is an intricate task: epidemics are complex phenomena resulting from the interaction between the environment, pathogens, individuals, and societies. Modeling offers a principled way to reason about infectious disease dynamics from scarce and biased information and to guide decision-makers towards effective policies. This thesis tackles selected topics in cholera and COVID-19 modeling towards informed public-health decisions. These two contrasting diseases were associated by a twist of fate, but also through the lens of a common modeling approach: compartmental, SIR-based, models are conditioned on the available evidence using computer-age statistical inference frameworks. A set of five models is developed, each tackling a different facet of the spread and control of these two infectious diseases. Each model aims at answering questions related to either the understanding of the mechanisms behind disease transmission, the projection of the future dynamics under different scenarios, or the assessment of the effectiveness of past interventions. Moreover, a novel application of epidemiological models to the formal design of control policies is proposed. Optimal control provides a rigorous framework to identify the most effective control measures under a set of operational constraints, providing a benchmark on what it is possible to achieve with the available resources.The results presented in this thesis range from scientific insight on the relationship between cholera and rainfall in Juba, South Sudan to the COVID Scenario Pipeline which produces reports used to inform the response to the COVID-19 pandemic of different governmental entities. Furthermore, the effectiveness of the non-pharmaceutical interventions against COVID-19 in Switzerland is evaluated; and so is the probability of eliminating cholera from Haiti under different scenarios of mass vaccination campaigns. Finally, the development of an optimal control framework towards the effective spatial allocation of vaccines against SARS-CoV-2 in Italy closes this conversation of models.The present thesis demonstrates how infectious disease modeling enables informed decision-making by projecting the uncertainties under the light of the available evidence. It also highlights the effort needed to tailor the models and inference methods to the specificities of the transmission setting and the research question considered. From insights on transmission pathways to weekly reports aimed at decision-makers, it explores different applications of infectious disease modeling. Methods developed along the way enrich the toolbox available to modelers, to guide policy decisions further towards a reduction of the burden of infectious diseases on communities.

Infectious Disease Containment Based on a Wireless Sensor System

Marcel Salathé, Xiao Sun

Infectious diseases pose a serious threat to public health due to its high infectivity and potentially high mortality. One of the most effective ways to protect people from being infected by these diseases is through vaccination. However, due to various resource constraints, vaccinating all the people in a community is not practical. Therefore, targeted vaccination, which vaccinates a small group of people, is an alternative approach to contain infectious diseases. Since many infectious diseases spread among people by droplet transmission within a certain range, we deploy a wireless sensor system in a high school to collect contacts happened within the disease transmission distance. Based on the collected traces, a graph is constructed to model the disease propagation, and a new metric (called connectivity centrality) is presented to find the important nodes in the constructed graph for disease containment. Connectivity centrality considers both a node's local and global effect to measure its importance in disease propagation. Centrality based algorithms are presented and further enhanced by exploiting the information of the known infected nodes, which can be detected during targeted vaccination. Simulation results show that our algorithms can effectively contain infectious diseases and outperform other schemes under various conditions.

Ieee-Inst Electrical Electronics Engineers Inc2016

Hydrological, Anthropogenic and Ecological Processes in Cholera Dynamics

Lorenzo Righetto

The present Thesis deals with understanding, measuring and modelling epidemic cholera. The relevance of the endeavour stems from the fact that mathematical epidemiology, properly guided by model-guided field validation, is a reliable and powerful tool to monitor and predict ongoing epidemics in time for action, and to save lives through evaluation of the effectiveness of mitigation policies or the deployment of medical staff and supplies. In recent years, waterborne diseases – and cholera in particular – have consolidated their role as a major threat for developing countries, where sanitation conditions are poor and the vulnerability to extreme events is highest. Several important features controlling the dynamics of occurrence and spreading of the disease in a region are studied in the present Thesis, from both theoretical and experimental perspective. Understanding the basic processes that regulate cholera infection is key to build reliable prediction tools. In this Thesis several driving mechanisms of cholera are investigated, with the objective of connecting together hydro-climatology, ecology and epidemiology in a comprehensive framework. First, the role of water volume fluctuations is analyzed partly through a bifurcation study in a mostly theoretical assessment. Such work is then particularized in a field campaign carried out in rural Bangladesh, where hydro-climatological variables and Vibrio cholerae concentrations have been monitored for more than a year in one of the ponds constituting the local water reservoir. Concomitantly, a procedure for the detection of Vibrio cholerae, based on flow cytometry, is tested in the field. Further emphasis is also given to the role of human mobility in disseminating the disease among different communities. In particular, the contribution of human mobility in the dispersal of vibrios along the hydrologic network is specifically analyzed. All the knowledge collected in these studies is then used to add essential details to a modeling framework that is applied to the dramatic case of the Haiti epidemic. A spatially explicit model, taking into account both hydrological transport and human mobility, is developed to simulate the spreading of the disease since its onset. It is also shown that the resurgence of the disease, coinciding with the rainy season of June-July 2011, can only be reproduced if hydrological forcings are considered. The framework is tested by forcing it with synthetic rainfall scenarios and projecting epidemiological outputs. It is shown that the model can quantify correctly the number of cases in a given time span, even when calibrated with limited information. This result allows then to use it as a tool to assess a priori the effectiveness of intervention policies, such as vaccination and sanitation. The effect of these two is tested both in the short- and in the long-term, with different results. Such endeavour represents the ultimate goal of the work presented in this Thesis – albeit further effort is needed to link together public health management and mathematical epidemiology in this field.

EPFL2013

Seasonality in cholera dynamics : A rainfall-driven model explains the wide range of patterns of an infectious disease in endemic areas

Theo Baracchini

An explanation for the spatial variability of seasonal cholera patterns has remained an unresolved problem in tropical medicine. No simple and uniﬁed theory based on local climate variables has been formulated, leaving our understanding of seasonal variations of cholera outbreaks in diﬀerent regions of the world incomplete. A mechanistic model for the Bengal region, which encompasses the variety of seasonal patterns worldwide, may provide a unique opportunity to gain insights on the conditions and factors responsible for endemicity around the globe, and therefore, to also revise our understanding of the ecology of Vibrio cholerae. Through the analysis of a unique historical dataset, we propose the ﬁrst mechanistic, rainfall-driven, SIR-based stochastic model we are aware of for the population dynamics of cholera, capable of capturing the full range of seasonal patterns in this large estuarine region. Parameter inference was implemented via new statistical methods that allow the computation of maximum-likelihood estimates for partially observed Markov processes through sequential Monte-Carlo. The results indicate that the hydrological regime is a decisive driver determining the seasonal dynamics of cholera. It was found that rainfall and longer water residence times tend to buﬀer the propagation of the disease in wet regions due to a dilution eﬀect, while also enhancing cholera incidence in dry regions. This indicates that overall water levels matter and appear to determine whether the seasonality is unimodal or bimodal, as well as whether it is pre-, post-, or in-phase with the monsoon. We present evidence that the environmental reservoir is responsible for the persistence of the disease, and therefore its endemicity. Given the undeniable interplay between the seasonality of cholera and the environment, a deeper understanding of the underlying mechanisms could allow for the better management and planning of public health policies with respect to climate. In terms of disease prevention and mitigation strategies this is of paramount importance today, as changes in the population dynamics of infectious diseases are expected in response to fast anthropogenic climate change.