- Project description
- Project partners
- Results
General: What is CorASiV?
Public health departments are facing a great challenge in obtaining available data on the spread of COVID-19, connecting it to other sources of data and analyzing it with regard to ever-pressing questions on how to proceed in combatting the virus. This is why CorASiV actively aids public health departments with visualization and analytics technology. The project addresses various leverage points in decision-making support for public health departments by using a flexible data set to expand both AI methods as well as smart analytical methods and models in combination with visualization. These are then supplied in web-based dashboards for analyzing coronavirus data. CorASiV provides these services and its support on an individual basis to multiple regional public health departments. The primary goal of CorASiV is to provide technology to potentially every public health department as an aid in the second along with, most likely, future waves of coronavirus infections until a vaccine becomes widely available.
Project coordinator Dr. Jörn Kohlhammer presents “CorASiV” (German only)
Partner Institutes
For the CorASiV project, multiple Fraunhofer institutes have come together, bringing with them their specific expertise. The project is coordinated by Fraunhofer IGD, which leverages its years of research in the healthcare sector primarily to develop visual and interactive data analytics methods, also called visual analytics.
Fraunhofer IAIS offers its expertise in healthcare analytics along with its years of experience in the analysis of spatiotemporal data and transmission models as well as in the analysis of documents and voice data.
Fraunhofer IOSB is responsible for designing and developing a digital twin dashboard for integrating consolidated data in the context of the work being done by public health departments on COVID-19, placing special emphasis on data management that complies with data protection laws.
The goal of the Fraunhofer Institute for Digital Medicine MEVIS is to digitally support relevant decision-making processes in public health departments using data-based models, thereby making them more efficient. The project team benefits from the institute’s expertise in clinical decision-making support.
The mission of the Fraunhofer Institute for Industrial Mathematics ITWM is to advance mathematics as a key technology as well as to drive innovation and, along with partners in research and industry, turn that innovation into reality. The Financial Mathematics and Data Science departments focus on decision-making support tools that use statistical methods to generate added value from data and visualize distributions.
The entire project is supported by the medical expertise of Fraunhofer IME in Frankfurt.
Analysis of Groups Exposed to COVID-19
Public health departments often receive inquiries from anxious citizens who have been around an infected person and want to know what their own risk is. Based on current scientific knowledge of incubation time, shedding period and duration of illness, a joint team from ITWM and MEVIS have developed a series of statistical estimates. With a web-based application, it is possible to, e.g., enter when a person began showing symptoms and then analyze the period in which those with whom that person had contact could have been infected and to what probability. Another analytical method assesses the risk for releasing people from group quarantine.
A freely accessible version of the web-based application can be explored here (German only): https://roeger-itwm.shinyapps.io/corona_contact_risk/
Outbreak Detection
Outbreaks are a major driver of the pandemic. An outbreak is when several people are infected by one or more people with the virus, who may be asymptomatic and thus rarely end up in public health department databases. With a combination of network visualizations of the infection process and a temporal representation of the course of the illness in an infected individual, these groupings of exposed people can be made visible. If multiple people become infected at the same time, the probability that they were infected at the same location increases. Coupled with contact tracing information from before and after infection, the result is an indication of an outbreak.
IGD’s visualization can be explored here (German only): https://corasiv.iva.igd.fraunhofer.de/
Geographical Visualization of Infections
One fundamental basis for assessment at public health departments is the geographical distribution of the infections. Infections in a densely populated quarter can spread more easily than in a sparsely populated, rural area. At present, it is still relatively cumbersome to enter the addresses of exposed individuals on a digital map and then to use this representation further. To this end, the team at IOSB lends its expertise in decision-making support systems, particularly those that use geographical data, that emphasize the simplest possible depiction of address information on the proper map section, the choice of temporal focus and the flexible use of sensible colors and icons.
More information can be found here: https://www.iosb.fraunhofer.de/en/kompetenzen/bildauswertung/interactive-analysis-diagnosis/research-topics/augmented-decision-making.html
Data Modeling and Analytics Support
For CorASiV, Fraunhofer IAIS focuses on analytics support, offering its expertise in structured data modeling for healthcare analytics as well as its years of experience with analyzing spatiotemporal data. Problems have been identified whose solving can be facilitated with machine learning/artificial intelligence methods for containment and pandemic management. These scenarios form the basis of cooperation and have unmasked the potential of AI methods in the area of public health. The analysis of the anonymized data follows an agile process in order to keep pace with the dynamic development of the pandemic. The results are regularly evaluated by public health department experts and then presented to the crisis team.
More information can be found here: https://www.iais.fraunhofer.de/en/business-areas/healthcare-analytics.html