The COVID-19 crisis caused by the SARS-CoV-2 virus that has spread quickly and affected large world population has highlighted in clear and compelling terms the importance of crisis response planning and management of epidemics. The effectiveness and success of the mitigation measures depend significantly on decision-makers relying on timely, evidence-based and informed decision support. The key pieces of information for managing an epidemic are the number of people infected and the spread of the disease in a community. However, due to the limited availability of reliable tests to confirm individual diagnosis during the latest COVID-19 pandemic and the substantial fraction of people showing no or only mild symptoms, it has been demonstrated by serological surveys in blood donors that the numbers of infected people have been greatly underestimated.
Scientists have shown that this virus has also excreted in an infected person’s stool. If the toilet is connected to a sewer system, such as the case in most high-income counties, the excrement ends up in the sewerage network and ultimately at the sewage treatment works. The increased circulation of a virus in communities will raise the amount of viral RNA flowing into sewers. Clearly, there is the potential for observing trends in virus circulation in the population by analysing sewage.
In low to middle-income countries or areas affected by natural or anthropogenic disasters, however, sewer systems are not available to all or they are present in some communities only. The management of epidemics in non-sewered areas can also benefit from a better understanding of the links between onsite non-sewered sanitation and infectious diseases.
To develop a framework for a digital epidemic observatory and management system (DEMOS) based on combining non-invasive viral and infectious disease detection methods in large populations. The system will provide effective evidence-based and informed decision-support throughout the entire epidemic management cycle, i.e., preparedness, early warning, response, mitigation and recovery (Fig. 1).
Data and information gathered by the DEMOS non-invasive epidemic risk management system are obtained directly or indirectly through data-mining the network of monitoring devices that cover both sewered and non-sewered parts of a community. Devices include: (1) automatic or semi-automatic samplers for collecting sewage samples from sewer mains, open drains, faecal sludge transfer stations and sewage and faecal sludge treatment plants, (2) MEDiLOO® – smart toilets and toilet retrofit modules for precision health monitoring, (3) clinical and epidemiological data collected about the virus and disease in the community, (4) socio-economic data for the community, and (5) any other relevant data (e.g., social media and contact tracing apps).
Sewer Surveillance Data
Collected sewage and faecal sludge samples are transferred to and analysed at a laboratory where parameters of interest are determined (e.g., SARS-CoV-2) using quantitative Real-Time Polymerase Chain Reaction (qPCR). The qPCR methodology is well established, accurate and capable of detecting the genetic material of the virus from collected samples (Medema et al., 2020). However, estimating the number of infected people is still a scientific challenge, due to uncertainties linked to inter-individual variability in shedding through faeces and other variables linked to sewer systems. This will be feasible as more sewer surveillance data becomes available and when combined with clinical and epidemiological information (e.g., the number of diagnosed or hospitalised people due to the infection). Nevertheless, the current system is suitable to rapidly determine trends in virus circulation in communities, which is the key to monitoring all stages of the epidemic management cycle (Fig. 1)
MEDiLOO® Smart Toilet Data
Another source of information is obtained in a non-invasive way from a novel system of MEDiLOO® toilets (Brdjanovic et al., 2015; Brdjanovic 2020). These toilets can be equipped with sensing equipment for monitoring of human excreta (stool, urine, and sweat), internal body fluids (e.g.. blood), organs (eyes, heart and skin) and breath. The concept entails the application of (a combination of) various technologies, including specific sensors, optical identification cameras, hyperspectral cameras, test strips, and lab-on-a-chip. A collection of symptoms related to one or more of infectious diseases, for example, fever, coughing, diarrhoea, blood in the stool, dehydration, hemoglobinuria, slow or fast pulse, red or yellow eyes, jaundice, brown urine and light stools, can be sensed to support a diagnosis. Further research is needed to investigate data analytics requirements for validated and reliable ways to confirm the pre-diagnosis.
Clinical and Epidemiological Data
Clinical and epidemiological information about the community ranging from a small group of individuals to large cities include, for example, transmission and incubation periods, shedding rates and virus sequences, will also be included as a data source. However, this data can be sporadic and rely on testing of individuals or treating those with symptoms. The key challenge is in the linking of non-invasive monitoring data at a large scale (i.e., sewered and non-sewered areas) with clinical and epidemiological information to better estimate the total number of infected people in a community.
A multitude of social and economic criteria can be considered as potential determinants for the spread and severity of an infection. These may include, for example, healthcare infrastructure (e.g., medical resources and health coverage), societal characteristics (e.g., social connectedness and household size), economic performance (e.g., GDP and employment levels), natural environment (e.g., air and water pollution) and demographic structure (e.g., population size, structure and density). Linking this type of data with non-invasive and clinical/epidemiological data can provide important clues for developing socio-economic mitigation measures and policies directed at reducing the impact of ongoing or potential future epidemic crises.
Digital data from mobile phones and footprints from web searches and social media can be accessed to support community surveillance, contact tracing and evaluation of health intervention policies. The key challenges are in accessing this type of data rapidly in a legal, proportionate and ethical manner while observing individual’s right to privacy and confidentiality.
Digital Epidemic Observatory (DEMOS)
The key objective for DEMOS is to combine various data types with data analytics, simulation and optimisation models to enable the observatory to support effective evidence-based and informed decision-support throughout the entire epidemic management cycle. Such an observatory requires a tailored cloud-based infrastructure and associated web-based tools to enable users to access data and model outputs that can support decision making at various stages of the epidemic cycle. Machine Learning (ML) and statistical tools, together with complexity science simulation (e.g., System Dynamic or Agent-Based Modelling) and optimisation tools (e.g., Gondwana®, Van Thienen and Vertommen, 2015) will be used to model the severity and spread of infection, assess the level of risk as well as forecast the potential effects of mitigation measures and policy interventions on the epidemic evolution and impacts.
One of the key research challenges will be to identify what determinants/variables to include in models such that they can explain the extent of the epidemic and provide a sound basis for making robust decisions. Furthermore, the above objective is tightly coupled to the need to better understand how different data sources and digital components can be incorporated into an integrated digital environment. This involves research challenges related to the standards required for interoperability and the infrastructure needed to transfer, store, filter and quality assess these large volumes of data.
The project is highly interdisciplinary and will bring together health, epidemiology, water, environmental, computer, instrument and social scientists with statisticians, modellers and stakeholders.