The EPIWORK project proposes a multidisciplinary research effort aimed at developing the appropriate framework of tools needed for the design of epidemic forecast infrastructures to be used in by epidemiologists and public health scientists. The project is a truly interdisciplinary effort, anchored to the research questions and needs of epidemiology research by the participation in the consortium of leading epidemiologists, public health specialists and mathematical biologists (DOWNLOAD THE PROJECT PRESENTATION).

Epidemic researchers along with informatics, computer science, complex systems and physics leading scientists, will tackle most of the much needed development in epidemic forecast of modeling, computational and ICT tools such as i) the foundation and development of the mathematical and computational methods needed to achieve prediction and predictability of disease spreading in complex techno-social systems; ii) the development of large scale, data driven computational models endowed with a high level of realism and aimed at epidemic scenario forecast; iii) the design and implementation of original data-collection schemes motivated by identified modelling needs, such as the collection of real-time disease incidence, through innovative web and ICT applications; v) the set up of a computational platform for epidemic research and data sharing that will generate important synergies between research communities and countries.

Concepts and Objectives

Background

Work Areas

Objectives

Methodology

Progress beyond the state of the art

Modelling and theoretical foundations

Data-driven computational platform

ICT monitoring and reporting system

S/T methodology and associated work plan

Concepts and Objectives

Background

The current threats of pandemic influenza, HIV and XDR tuberculosis, and recent threats such as SARS and the release of bioterrorist agents, raise major urgent concerns with regard to public health preparedness, risk management and decision making. Mathematical models have become important tools in analyzing the spread and control of these threats and assist decision makers in taking proper prevention and containment/mitigation measures. They are used in assessing the impact of infectious disease epidemics and pandemics to human health. Their role in comparing, planning, implementing and evaluating various control programs, is of major importance for public health decision makers.

Recent years have also witnessed a tremendous progress in data gathering, development of new informatics tools, and increase in computational power. The spread of epidemics is indeed inevitably entangled with human behavior, social contacts, and population mobility and mixing. A huge flow of quantitative social, demographic and behavioural data is becoming available, that trace the activities and interactions of individuals, social patterns, transportation infrastructures and travel fluxes. Improved techniques and methodologies support the inter-linkage and integration of datasets with geo-referenced information and economical and transportation databases (BTS, GIS). This has caused a qualitative change in the ways we model epidemic contagion processes. Visualization and analysis tools able to cope with multiple levels of representation are being developed along with computer simulations that provide experiments not feasible in the real world. A quest for innovative technologies comes also from the disease outbreak detection point of view. Innovative technologies can improve the traditional disease-surveillance systems (EISS), providing faster and better localized detection capabilities and resulting in a broad practical impact. For the first time, epidemic processes can be studied in a comprehensive fashion in a manner that addresses the complexity inherent to the biological, social and behavioural aspects of health related problems. In other words we are in the position to envision the development of large data and computational forecast infrastructures aimed at the realistic prediction, and containment, of the diffusion and impact of infectious diseases.

However, it has become clear that any new and significant scientific progress in epidemic research requires a novel interdisciplinary integration of expertise, techniques, and methodologies that interfaces computational and data-intensive science with mathematical epidemiology and an innovative use of Information and Communication Technologies (ICT) to gain knowledge on human, social and economical systems. Indeed, the development of predictive tools based on data driven modelling requires a combined research effort integrating and assimilating a vast range of mathematical, statistical and informatics expertise and ultimately requires a flexible large-scale interface with high quality surveillance systems at the European and global scale. While the EC has supported research networks dedicated to the development of mathematical, computational and statistical models for infectious diseases and databases construction (MODELREL, INFTRANS, POLYMOD, IRIDE, EUPHIN), in most of the cases the modelling and databases projects proceeded in parallel tracks, targeting different needs and communities. However, the vision of a working epidemic forecast infrastructure does require an integrated approach to data, modelling and surveillance that builds on the results of previous projects and systematically redesigns the collaborative interaction among the stakeholders of the various activities.

The modelling frameworks need to be enlarged in order to deal with the complex features of present techno-social systems. These systems are characterized by large sizes (thousands to millions of elements), irregular and scale-dependent structure, time dependence and emergent phenomena and patterns. The challenge is therefore to determine the long range, long term behaviour of the epidemic which cannot be trivially inferred from the short range interactions included in the models. This implies the use and development of techniques dealing with complex systems features as well as developing new models and mathematical tools able to attain quantitative predictability in large scale techno-social systems. As in weather forecasting, a large scale epidemic forecast infrastructure needs to rely on sophisticated computational tools to integrate present data and huge libraries of previous epidemic patterns into realistic modelling framework. Data gathering has to be informed on the modelling needs as models have to be refined according to the accessible data. Real time data acquisition and sharing should be considered as a key concept as it is in weather predictions; nobody would think to predict weather on the basis of the meteorological situation of two months ago. Data and algorithm sharing to the community at large, still in real time, is an essential requirement. Finally the development of appropriate user interface that would make modelling power available to researchers outside the close circle of mathematical and theoretical epidemiologists is necessary to make an epidemic forecast infrastructure a concrete and useful tool for policy makers and health institutions. This can be ultimately done only by the opportune use of modern ICT technologies.

Work Areas

Building on a multidisciplinary research approach, the projects aims at developing the approriate framework of tools and knowledge needed for the design of epidemic forecast infrastructures. It is based on a highly synergistic, interdisciplinary effort among computational modellers, medical scientists, epidemiologists, and computer scientists. The project aims at exploring the following work areas as the major research themes directly matching the objectives of this proposal and representing the necessary steps toward the future construction of an epidemic forecast infrastructure:

Modelling and theoretical foundations. The research effort ranges from the analysis of stylized models that can provide basic insights in the epidemiological theory to computational approaches for large scale (spatially extended) simulations aimed at realistic scenario analysis. The latter are used as in silico epidemic experiments that, including the detailed complexity of real-world population mobility and host-pathogen interactions, allow researchers to assess the epidemic evolution and to test interventions to mitigate its impact on the population. The integration of detailed data of techno-social systems requires basic theoretical and algorithmic foundations for the understanding of the fundamental principles and mechanisms that govern epidemic behaviour in large scale complex multi-scale network and individual based epidemic models. In this area a crucial point is the understanding of the impact of large scale complex features (scale invariance, extreme heterogeneity, large-scale fluctuations) of interaction and communication networks on epidemic spreading patterns. Progress beyond the state of the art:

Data-driven computational platform. Given the complexity and increasingly interconnectedness of our world, with no closed boundaries between countries, any epidemic forecast effort is crucially dependent on the accessibility to transnational demographic and mobility data, uniform standards to collect detailed spatio-temporal disease incidence, immunization and pathogen evolution data. Such a data integration and assimilation capacity requires the design, implementation, deployment, and maintenance of a computational platform for epidemic research and data sharing. The platform will allow detailed, accurate and reliable simulations for real time and predictive modelling of epidemic events, available to researchers, health-care professionals and policy makers.

ICT monitoring and reporting system. A novel monitoring infrastructure would provide crucial advantage to modellers for real-time data feed to forecast algorithms. This Work Area has the ambition to generalize and evolve the system developed in the Netherlands and in Portugal for the web collaborative monitoring of disease incidence (Influenzanet). This system is conceived to inform and educate the population about the disease and to collect real-time information on population health through web-services. Graphic representation, processing and analysis of data on the progression of the disease, is provided in real time. Population-based real-time monitoring, under development in Sweden, promises further refinement of patient-initiated disease reporting using state-of-the-art telecommunication and Internet techniques, guaranteeing the link to the underlying population that generates the cases. Standardized methods would allow monitoring the evolution of an epidemic across country boundaries in a consistent way, with real time alerts to rapidly identify public health emergencies, and estimate key epidemiological parameters to devise and test methods to contain them.

The work areas targeted in this paper need to be developed in the presence of a mutual feedback and strong collaborative approach among the researchers involved in the project, fostering crucial flows of information and knowledge production in the data-modelling-application process. This level of integration is possible only within a large integrated project providing a framework and shared vision for the overarching goals of the project.

Objectives

A fully operational, accurate and reliable epidemic forecast infrastructure nowadays faces problems related to the lack of appropriate models to understand how an infectious disease spreads in the real world, lack of extensive and accurate epidemiologically relevant data (from societal data to epidemic surveillance data), lack of understanding of the interplay among the various scales of the problem (from the host-pathogen interaction, to human-to-human transmission, to the interaction with the environment) and most importantly lack of communication among the different areas of research which proceed almost independently, crucially hampering a significant progress in a highly interdisciplinary field of research. The present project intends to fill this gap. Through computational thinking, complex systems concepts and data integration tools relevant for epidemiological understanding at all levels, it will provide a set of radical, paradigm-changing results enabling a novel approach to the modelling, forecast and policy making approach to infectious diseases. The projects overarching goals are:

Methodology

We address the above issues with the intention to initiate a community approach to the problem and develop an armoury of tools and knowledge that will produce a quantum leap in the state of the art of epidemic research. The project involves a number of different groups that represents the whole spectrum of disciplines needed for a fruitful and truly interdisciplinary approach to the problem. The team includes computer scientists and physicists experts in data analysis and integration, and computational modelling. Two groups of mathematical epidemiologists lead the basic modelling definition and constructions. Three groups of epidemiologists representing centres for disease control in the respective countries supervise the data analysis and acquisition and work on validating the results and tools developed in the project. One consultancy and project management firm supervises the deployment of the proposed Internet-based monitoring system on a global European scale.

While our project might sound very ambitious, it revolves around a clear set of tasks that will lead to substantial improvements in the data gathering and integration processes and will allow the development of new classes of realistic epidemic models able to cope with the complex realities of techno-social systems. The research considers most of the much needed development of modelling, computational and ICT tools in epidemic research such as the foundation and development of the mathematical and computational methods needed to achieve prediction and predictability of disease spreading in complex social systems. Furthermore, the development of large scale, data driven computational models endowed with a high level of realism, the design and implementation of original data-collection schemes including real-time disease incidence data and the set up of a computational platform for epidemic research and data sharing will generate crucial synergies between research communities and countries, fostering a collective approach to the prevention, detection, and timely response to any public health emergency of national and international concern.

Progress beyond the state of the art

Infectious diseases remain a serious medical burden all around the world with 15 million deaths per year estimated to be directly related to infectious diseases. The emergence of new diseases such as HIV/AIDS, the severe acute respiratory syndrome (SARS) and the eventual rise of a new influenza pandemic represent a few examples of the serious problems that the public health and medical science research need to address. While for centuries, mankind seemed helpless against these sudden epidemics, in recent time our ability to control future epidemic outbreaks is to a high degree facilitated by the advances in modern science. The cures for a number of dangerous pathogens are available and thanks to the genetic revolution new drugs to prevent and reduce the health consequences of new epidemics can be developed and manufactured faster with respect to the past. On the global scale, the World Health Organization and other agencies have put in place cooperative infrastructures for disease control and surveillance that allow a timely identification and track of emergent disease evolution and spread, enabling the alert of national and international health services and the deployment of containment measures (WHO-EPR). On the other hand, however, the increasing population of urban areas, the massive interconnectivity among world regions and the human mobility are factors which accelerate the spread and diffusion of old and new diseases. In this context, diseases and epidemics can have far reaching effects on a very short time scale (WHO-SARS). As a result, we demand ever-increasing predictive power to anticipate future outbreaks and evaluate associate risks. The successful containment of the next pandemic event is not just linked to our medical infrastructure but also on our capacity to predict its diffusive pattern and optimize medical and mitigation policies. In such containment processes, the ability to forecast how a disease might spread on the local and global level (as much accurately as we can now do for the weather) is essential for the identification and development of appropriate control strategies. In this perspective, computational power and ICT advances allow us for the first time to ambitiously imagine the creation of computational epidemic forecast infrastructures able to provide reliable, detailed and quantitatively accurate predictions of global epidemic spread.

To reach this goal, knowledge as well as resources need to be accessed, shared and integrated among researchers working in this area – epidemiologists, computer scientists, mathematical biologists, information scientists, medical scientists. Strong collaboration, shared information and integration of different expertise are the crucial points of this proposal and the key to address a complex problem that involves elements non-trivially interacting with one another at different time and length scales, from virus to host to population to environment. Feedback loops connecting the different areas of research are fundamental in allowing further improvement in a multidisciplinary field of research that otherwise would be hampered by transdisciplinary boundaries. While the EC has supported research targeting the mathematical, computational and statistical models for infectious diseases and databases construction (MODELREL, INFTRANS, POLYMOD, IRIDE, EUPHIN), a true progress in knowledge production and understanding the spread of human infectious diseases requires a systematic redesign of the approach in terms of integration, multidisciplinary effort and collaborative interactions among the stakeholders of the various activities, building on the results and experience reached in previous projects. A realistic and reliable large-scale epidemic forecast infrastructure will involve such a high level of complexity and realism that its progress in parallel non-communicating tracks, targeting different needs and communities is bound to fail. Real data needs to be analyzed to refine and design appropriate modelling approaches, and models require data as input ingredient. Data availability across physical and non-physical boundaries (such as national borders, linguistic barriers, and disciplines boundaries) as well as uniform and standardized approaches are an essential requirement. Availability to the community at large of data, information and computational approaches in epidemic research through an appropriate user interface would make modelling power a concrete and useful tool for assessing scenarios, predicting epidemic evolution, managing health emergencies, benefiting a large audience of users, including researchers, non-experts, policy makers, and health institutions. On the other hand, the above research plan needs the constant inputs and contribution of epidemiologists, mathematical biologists and public health experts. The basic research questions and the practical use and implementation of the predictive power derived by modelling and computational results must be informed and defined with epidemic and health researchers.

The aim of the project is to produce non-incremental advances in the capability of forecasting the spreading of infectious diseases, relying on modelling and computational tools able to provide scenario forecast, risk assessment and containment measure testing contextually to the onset of health crisis as well as for recurrent and seasonal events. In other words, the capability of statistically predicting incidence, morbidity, mortality spatio-temporal patterns for epidemics and pandemic events and to evaluate medical and non-medical interventions such as drugs deployment, vaccination, social distancing. The present state of the art in the field has serious limitations hindering our forecast capabilities, and the present proposal intends to tackle those limitations both in the modelling, data and experimental areas. For this reason the consortium gather experts in a wide range of disciplines ranging from physics and computer science to theoretical epidemiology and public health. The top rate expertise of the consortium in mathematical epidemiology, epidemiology and public health anchors the project to the actual questions and challenges of the field and provides a truly multidisciplinary approach. In the following we provide a discussion of the advancement beyond the state of the art offered by the present proposal in each of the main Work Areas targeted by the project.

Modelling and theoretical foundations

A first limitation of the state of the art in epidemic modelling and forecast is the large gap between the classical epidemiology theory, dominated by stylized models, and the theoretical understanding of the many complex facets encountered in the realistic description of epidemics. Epidemic models are largely based on random populations or very simplified contact patterns. The multiscale complexity of the population-disease-environment system and the dynamical aspects of contacts and mobility of individuals, as well as other time dependent features such as seasonal effects and other external environmental drivers are generally not included in the model formulation. Analogously, the complexity added by the interaction between pathogen and host, vaccination procedures, short and long-term immunity, and reinfection mechanisms has been investigated only at the level of very simple compartmental models. The mathematical epidemiologists and modellers in the consortium intend to finally lay the foundation of new models stylized enough as to remain analytically tractable, yet nevertheless accurate to simulate realistic heterogeneous epidemic processes. In the case of complex systems, this task amount to distinguish different classes of parameters and to identify which ones are really relevant in the description of the behaviour of the system, obtaining models which stay at a reasonable level of precision but still captures enough realism in order to be useful in practical situations.

Complexity of the population-disease-environment system
The work carried out during the project will define models considering that the contact patterns among individuals over which the disease spreads are highly heterogeneous, leading to the emergence of high -risk individuals and potential super-spreaders, and thus making the variations in people’s interacting behaviour a crucial determinant of individual-level risk. This will determine how heterogeneities due to population aggregations and spatial structuring control the epidemic size and the extinction threshold as well as competition and evolution between disease strains. The project will also focus on the interplay between seasonality and pathogen competition and evolution and their complex effects in epidemic evolution. These results allow a new take on the design of vaccination intervention and the effect of reinfection thresholds and pathogen dynamics. The proposed work is therefore providing major theoretical and algorithmic advances in crucial areas such as

• Characterization of heterogeneities due to population aggregations and spatial structuring.
• The impact of seasonality and other external environmental drivers in techno-social systems of epidemic relevance.
• New vaccination and containment strategies taking advantage of the systems’ complexity.
• The effect of temporal correlations and evolution (reinfections, immunity, antigenic variations) at the population level.

These results and the new model structures will change parameter estimation schemes and the project will develop new tools for estimating key parameters on real data and evaluate the prediction quality of the models. By combining the different theoretical developments and parameter estimation methods the Consortium epidemiologists expect to make significant advances to the understanding of what components must be included in early warning systems for infectious diseases.

The structure of interacting populations on many magnitude and spatial scales
Another shortcoming of the present state of the art in epidemic modelling is represented by the lack of understanding and characterization of the complex structure of metapopulation models. Large scale models of epidemic evolution in fact crucially dependent on the ability to quantify the topological and dynamical features of human mobility network and spatially structured models (the basic brick of epidemic forecast computational approaches). This in turn requires the development of theoretical frameworks providing order parameters, the identification and extraction of universal features and the quantification of differences and similarities on an international, national and regional level. Despite considerable effort and advances in the study of human transportation networks, e.g. investigations of statistical properties of specific human transportation networks, no systematic investigation of traffic networks incorporating all spatial scales exists today. As such our understanding of human mobility within heterogeneous and spatially structured populations is limited. The project envisions overcoming two of the most challenging aspects of complexity in this context, namely, the structure of interacting populations on many magnitude and spatial scales beyond and within international, interregional, intercultural and linguistic boundaries and their effective interaction by means of multi-length scale transportation and mobility networks. The project will combine the analysis of extensive transnational transportation dataset with new sources of pervasive online data such as i) Eurobilltracker.com that investigates the dispersal of individual bank notes registered at the online bill tracking system; ii) Geocaching.com, a modern, international GPS treasure hunt in which trackable items are transported by travelling humans from place to place and iii) bookcrossing.com an internet administered game of similar nature. These data sets provide an unprecedented amount of information on particular aspects of human mobility with very high spatio-temporal accuracy and across international boundaries and will define a class of proxy mobility networks that complementing regular transportation data will to provide a global multiscale picture of human dynamics. The proposed research will provide a first principles construction of the meta-population description that goes beyond the usual identifiable structured contexts such as cities, town and villages, or – on a smaller scale – schools, workplaces and homes. The complexity of the techno-social systems makes these intuitive modules opaque or arbitrary and the project will work on defining novel community structure identification algorithms to identify the effective community structure in multi-scale mobility networks.

Network-network duality framework
The combination of the results on contact patterns among individuals and the metapopulation structure represents another major conceptual and theoretical advance proposed by the project: the understanding of human interactions in a network-network duality ansatz. The statistical properties of contact networks are strongly influenced by the way individuals behave as agents in a spatially structured meta-population. Contemporary research on disease dynamics has so far not succeeded in merging these two aspects. We anticipate bridging this gap in a network-network duality framework. The key result is the understanding of the reciprocal impact of social networks and mobility networks. To what extent are contact networks determined by mobility networks and vice-versa? In this approach we plan to develop mathematical tools and theoretical frameworks that allow the development of models in which social contact networks and mobility networks are modelled simultaneously.

Data-driven computational platform

Modelling platform
The computational approach to the realistic modelling of infectious diseases is at the moment a research area witnessing a few competing groups developing computational models tailored for specific world regions or systems, which require specific expertise and are not accessible to researchers outside the circle of experts in the field. This often implies re-inventing the wheel and the duplication of computational efforts. In addition, these modelling approaches are not critically examined and their actual predictive power is rarely assessed. The research agenda of Epiwork has a twofold approach to overcome these problems:

• The collaboration of epidemiologists, computational modellers and physicists will allow tackling foundational issues related to complexity and predictability in computational epidemiology.
• The development and implementation of a modelling platform available through the web to a wide range of users to simulate epidemics by explicitly including the complexity of the real world.

On the foundational issues, the envisioned work will provide a basic understanding of the predictive time scale as a function of information available on the system and its initial conditions. While the analysis of chaotic properties in basic epidemic models has a long tradition, computational epidemiology faces issues related to the large scale of the extended dynamical systems and its complex properties for the first time. The project will leverage on the collaboration of epidemiologists, physicists and modellers. This will allow the assessment of issues concerning the complexity and predictability of epidemic spreading patterns resulting from multi-scale and agent based computational models by integrating techniques used in information theory and statistical physics while at the same time considering the relevant epidemiological parameters, the plausible uncertainties and data bias. The modelling platform will integrate models, real-data and visualization techniques to perform simulations and provide access to the state-of-the-art computational modelling to a wide audience of both experts and non-experts. The aim is to provide a flexible and user-friendly tool for the simulation of a case study, test and validation of specific assumption on the spread of a disease, understanding of observed epidemic patterns, study of the effectiveness and results of different intervention strategies, analysis of risk through model scenarios, forecasts of newly emerging infectious diseases. The platform will be informed and tested by the consortium epidemiologists and public health professionals and is also envisioned as scenario and training tool for public health workers and policy makers, fostering the use of computational and informatics approach beyond the circle of simulations experts. This platform will work on the open source agreement and is supposed to seed a European effort in the computational modelling of infectious diseases at the moment limited to a very few research groups working independently. This will allow the exploitation of research synergies and the eventual nucleation of a large scale computational infrastructure for epidemic modelling in realistic complex techno-social systems. This is the basic component along with the data integration/sharing platform toward the construction of a large scale European epidemic forecast infrastructure.

Information platform
Another limiting factor toward realistic epidemic forecast is the difficulty in assimilating and integrating the ever increasing wealth of datasets needed to support the modelling approach and to extract knowledge and pattern from multiple data source. As of today, data are generally gathered according to different standards and usually collected by national census bureau, health institutes and centres for disease control without coordination. The situation is even more cumbersome for scientific experiment or targeted epidemiological data acquisition. The project envisions a unified and integrated approach for the management of these resources, with the design and implementation of an Epidemic Marketplace Platform, publicly available on the web. The project will define a simple reference format which will facilitate the navigation and use of the datasets. This information platform will provide the community with an unprecedented tool in the epidemic research field with the following enabling features:

• It will support the sharing and management of epidemic datasets and resources as well as their rating, annotation, and selection.
• It will be used as an on-line social networking site that will serve researchers, practitioners, and educators all over the world to foster a virtual community for epidemic research.
• It will support the exchange of resources as well as user interactions. Based on some of the Web2.0 characteristics, users will become active participants, generating information and providing data for sharing, and collaborating online, rather than being satisfied with a passive information consumer/viewer role.
• It will create an exchange point connecting modellers who search for data for deriving their models and those who have data and search for the help of modellers on interpreting their data.
• As collaborations evolve, through direct trustful sharing of data, the platform will also serve as a forum for discussion and the organisation of meetings that will guide the community into uncovering the necessities of sharing data between providers and modellers.

The project has chosen to implement the system on a grid platform based on standards defined by the Open Grid Forum , the Globus Alliance and the OASIS consortium. We intend also to integrate the infrastructure deployed for this project with European grid initiatives such as EGEE . Grid computing aims at the provision of a global ICT infrastructure that enables a coordinated, flexible, and secure sharing of diverse resources, including computers, applications, data, storage, and networks across dynamic and geographically dispersed organizations and communities (sometimes known as Virtual Organizations). Grid technologies promise to change the way organizations tackle complex problems by offering unprecedented opportunities for resource sharing and collaboration. Grid technology can either significantly reduce the cost or time to produce results, or provide resources that are able to deliver services that cannot be economically delivered using conventional networked information systems. The implementation of a system on the computational grid platform for epidemiologic studies opens new perspectives for gathering data on large populations, and – as a consequence – would allow stratification of large scale metropolitan epidemiology studies.

ICT monitoring and reporting system

Data represents an obstacle to progress not only because it is scattered across different repositories in different formats, standard and classification, but also because of its rate of acquisition and experimental design is limited. Real time surveillance data are crucial to rapidly identify public health emergencies, understand global trends and driving factors, feed realistic data-driven models to assess the impact on the population, optimize the allocation of resources to fight against them, and devise mitigation and containment measures to reduce economic, communication, transportation, and – more in general – social disruption. At the moment, however, acquisition of large scale data is not timely and, more importantly, not thought to inform models in real time. In general existing disease surveillance systems (GP sentinel) have several important limitations, in particular in their inability to provide information in real time, and on patterns of household transmission; in their lack of uniform standards for clinical definitions, that vary considerably between countries and even between reporters (EISS). Moreover, age-stratified rates of physician consultation may vary widely with different health care and health insurance systems. Especially for diseases as influenza-like-illness, only a minor (and unknown) fraction of all infected individuals sees a doctor, and frequently after a considerable delay, when a complication has occurred or in case a doctor’s certificate is required (e.g. in Sweden such a certificate is not required until after 1 week). Reporting rate may change unpredictably during an epidemic, making extrapolations of those statistics to the general population uncertain. Moreover, in many European countries the recruitment of sentinel GPs does not ensure proper sampling so that the geographic distribution of reporters does not faithfully represent the distribution of inhabitants, further complicating the attribution of observed cases to the population at risk. Existing surveillance systems also lack the flexibility to cope with new variants of existing pathogens that may result in atypical symptoms. These systems are also very vulnerable in the case of high prevalence and socially disruptive epidemics due to the limited capacity of hospitals and health care centres which induce biases in the number of visits per sentinel physician. In all cases, these limitations are crucially undermining the development of real-time data-driven modelling and forecast capabilities.

The project intends to overcome the limitation of the state of the art surveillance systems by proposing an innovative ICT approach based on Web2.0 tools. Starting from the successful experiences of internet-based monitoring systems (IMS) in the Netherlands and in Portugal (Influenzanet) which displayed high values in surveillance, epidemiological analysis and participant recruitment, the project plans to deploy an innovative real-time surveillance system across European countries. The IMS works with the Internet participation of the population to collect real-time information on the distribution of diseases through web-services. The collaborative participation of users is achieved through targeted communication and recruitment. Graphic representation, processing and analysis of data on the progression of the disease, is provided in real time. In addition the project considers mobile telephone as enabling access technology so that larger fractions of the population can be involved in the real time data acquisition. The proposed IMS foresees for the first time the collaboration of epidemiologist and public health practitioners with modellers. The epidemiology teams will take care of defining “golden standard” for different diseases in order to have unified data across European countries. While influenza-like illnesses (ILI) are used in the early deployment of the system, the final IMS will consider other diseases and infections. Population-based real-time monitoring, under development in Sweden, promises further refinement of patient-initiated disease reporting using state-of-the-art telecommunication and Internet techniques, guaranteeing the link to the underlying population that generates the cases. The project will provide also the first comparative assessment of various surveillance methods and will test and calibrate the IMS by the epidemiology teams.

S/T methodology and associated work plan

The project revolves around six distinct scientific work packages (WP1-WP6) aimed at providing a virtuous feedback cycle between tool development, data collection, analysis and modelling. The research plan is structured so as to foster a fruitful interplay between the various components of the project. WP1 and WP2 are aimed at exploring theoretical issues in the area of epidemic modelling in complex, multiscale systems, structured populations and in the presence of the dynamical interplay between social and technological factors, seasonality and climate, health policies implementations. WP3 and WP4 are devoted to the collection and sharing of data on a computational platform and have a two way continuous exchange with WP1 and WP2 of data and algorithms. WP5 and WP6 is aimed at the developing, set-up and deployment of innovative web monitoring and data gathering tools that provide a continuous stream of data to WP3-WP4 and is informed by a constant feedback on the modelling needs in terms of data gathering by WP1 and WP2. The project revolves around a relatively small number of WPs sign of the common research agenda of the consortium teams that work in a coordinated way on the various tasks. This favours a closer interchange of ideas and knowledge among the groups and the various components of the project. The methodology is clearly truly interdisciplinary. Each WP includes several core disciplines expertise and it is anchored to the epidemiology area by the presence of mathematical biologists, epidemiologist and public health experts. These groups will provide the main research questions, the basic disease and parameters choices and the relevant complex features of epidemiological systems as well as their contribution in the development of cross-fertilized and novel approaches targeted in the WPs.
WP 7 is management. This project is initiated by a group of senior scientists, working at the best research institutions in Europe. The Institute of Scientific Interchange provides the management of the project and the coordination of the consortium. Finally WP8 is devoted to outreach and dissemination activities. In summary, the project is structured in the following work packages:

1. Work package 1 – Populations models and contact networks
2. Work package 2 – Spatially structured models and human mobility
3. Work package 3 – Information platform
4. Work package 4 – Computational Modelling Platform
5. Work package 5 – ICT monitoring and reporting system
6. Work package 6 – Reporting systems comparative analysis and validation
7. Work package 7 – Management
8. Work package 8 – Dissemination and exploitation

Each WP has a WP-leader that supervises the work progress and assesses and corrects deviations from project goals. The WP-leader is also responsible for the coordination with the activities of other WPs.

Contact Email:

info [at] epiwork.eu

Coordinator:

Prof. Alessandro Vespignani