By Siân Jenkins & Sacha Darwin
Epidemiology is an established pillar of medicine that deals with the incidence, distribution and control of disease. Riskaware’s BioAware platform provides a way to visualise and interact with a suite of disease modelling tools, that aid with epidemiological investigation of current or potential biological threats and associated mitigation strategies. This includes modelling the impact of non-infectious and infectious diseases on a population and, in turn, the expected cost and strain placed on healthcare provision.
The need for technology such as that provided by BioAware, has been demonstrated by the COVID-19 pandemic. Scientists are able to analyse the ongoing pandemic through disease modelling and associated tools within the platform. With an interactive and accessible interface, it provides visual results which both scientists and decision makers can understand in order to plan their response and justify it to the public.
What is BioAware?
BioAware is a sophisticated suite of epidemiological tools for analysing and investigating the impact of biological incidents.
BioAware empowers users with the ability to thoroughly explore the impact of diseases on populations through disease modelling and forecasting. Powerful inference techniques allow parameter estimation based on real-world data, allowing users to create realistic models that can be used to forecast the potential threats from diseases. By modelling scenarios with and without mitigation techniques, the need for and type of mitigation strategy can be explored.
This is useful for advanced planning for the possible occurrence of a biological event or when monitoring an ongoing event. In addition, novel biosurveillance techniques allow state inference, providing situational awareness and improved forecasting, to aid in the monitoring and planning for ongoing biological events.
How BioAware Works
Our platform aids current epidemiological modelling by providing a framework that unifies multiple forms of disease analysis and inference, with at- a-glance interpretation of outputs that facilitate quick and easy understanding. The flexible and extensible nature of the system means that bespoke disease models and tools can easily be incorporated into the suite, along with tailored output graphics.
The platform allows bespoke epidemiological models to be analysed using the sophisticated analysis tools held within BioAware. This includes algorithms that utilise user uploaded data to infer parameters and states for the models, allowing BioAware to provide more accurate results.
Bespoke disease models can be incorporated into the platform through a plug and play methodology. The results from forecasting these models over time are displayed using a combination of different statistical views, such as:
- Spaghetti plots
- Box plots
- Frequency trees
Mitigation strategies can be included within these models, to explore their impact on the spread of disease, allowing their impact to be assessed and for different implementation intensities to be explored.
The suite of tools that can be applied to the disease models include algorithms for:
- Parameter estimation
- Parameter optimisation for goal attainment
- State inference for situational awareness
- Early warning and detection of disease outbreaks ahead of traditional diagnostic tests
Empowered by Unique Capabilities
The state inference and early warning and detection of disease outbreaks is provided through a novel biosurveillance algorithm, unique to the BioAware platform. It uses particle filters and Bayesian networks to fuse data from multiple sources, with the purpose of inferring information on the state of health of the population.
This is used to provide situational awareness to decision makers. It includes critical information such as the presence of a new infection in the population and the number of people at each stage of that infection. This data can then be used to forecast future potential demand on health services, allowing decision makers to determine and explore mitigation techniques through the platform.
By fusing multiple sources of data, the impact of erroneous data is reduced, resulting in more accurate results, including reduced false alarms during outbreak detection compared to other similar algorithms.
BioAware can be tailored to use any form of data source such as:
- Hospital Accident and Emergency (A&E) and admissions data
- General Practitioners (GPs) data
- Healthcare product sales data
- Social media monitoring
- Weather data
What does a disease outbreak look like in BioAware?
BioAware can be used to model outbreaks that are infectious or non-infectious, including those naturally occurring and those arising from a deliberate biological attack. The platform contains multiple disease models, allowing scientists to choose the correct model that corresponds to the disease of interest.
The disease models can be used to generate forecasts for the spread of disease, demonstrating the rate at which it will spread through the population. Plots for the number of people at each stage of the infection over time are created, allowing scientists and decision makers to determine if intervention from authorities is required, as some outbreaks will die out naturally.
Monte Carlo outputs allow a range of scenarios to be explored simultaneously, providing an overview of the possible worst and best case scenarios. These are complemented with diagrams that interpret the plots in a friendly and diagrammatic way, demonstrating the number of people affected at each stage of infection.
The properties of the disease model determine the granularity of the outputs. For example, by including healthcare centres in the models, the impact of outbreaks on healthcare facilities can be determined by including them in the modelling process. By also including mitigation strategies, their impact to slow the spread of infection can also be determined and the level of implementation required can be determined.
Spatial models can be used in conjunction with disease models to determine the source of outbreaks such as legionnaires, a non-infectious disease, spread through contaminated water sources. Dispersion models can be utilised to determine those affected by airborne agents. These capabilities allow public health bodies to organise and provide timely and targeted relief to the affected populous.
Who is BioAware suitable for?
The information provided by BioAware benefits entire populations. The end goal of using this kind of tool is to inform public health decision making and facilitate the creation of informed mitigation strategies, to limit the spread of the disease and protect the health of the population. National agencies and health organisations can use BioAware to prepare and plan for biological incidents, both active and those seen as potential threats.
Inference techniques can be used to determine the status of active biological incidents within the population based on available observations. Health agencies can use this to determine trends of the disease and respond accordingly. They can also forecast the potential future impact of the event and determine if mitigation strategies are required. In some cases, such strategies can potentially be incorporated into a disease model within the platform and the level of mitigation required can be explored. Similar approaches can be used to explore the impact of potential threats and the impact of mitigation strategies.
During COVID-19, Riskaware have used BioAware capabilities to combine the outputs from a wide range of epidemiological models about the virus to create a clear, comprehensive image of the possible outcomes.
YOU MAY LIKE: Coronavirus: Can we slow the spread?
Large commercial corporations can also make use of BioAware capabilities and disease modelling. Using store card data, purchasing trends, and product sales they can identify and predict an increase in the demand for health-related products, such as flu remedies.
This information is extremely helpful for drug companies, pharmacies, and general stores – allowing them to more effectively plan their logistics regarding distribution and production of such products.
What is unique about BioAware?
BioAware is unique in providing a platform that brings together multiple tools and techniques, for more reliable results, and using enhanced graphics that enable all types of users to work together. Providing a more accessible way to gain actionable insights, users are more empowered to create targeted mitigation strategies in response to current and future biological threats.
The platform also incorporates novel inference algorithms that provide parameter and state estimates. Ultimately this facilitates more accurate situational awareness and forecasting from our disease modelling, as well as early warning for disease outbreaks. By combining multiple data sources through Baeysian data fusion, our epidemiological models see through the noise in data and provide greater situational awareness about the progress of an outbreak and the health of a population.