Toward a more resilient society

Prevent

Dr. Alexander Stolz, Head of Safety Technology and Protective Structures at the Fraunhofer Institute for High-Speed Dynamics, Ernst-Mach-Institut, EMI.
© Fraunhofer / Philipp Horak
Dr. Alexander Stolz, Head of Safety Technology and Protective Structures at the Fraunhofer Institute for High-Speed Dynamics, Ernst-Mach-Institut, EMI.

How can the worst case be prevented?

The curse of preventive action is that while it may save lives, it wins little recognition. Yet recent months have shown just how important prevention is. “If we had rigorously monitored the spread of this novel coronavirus at an early stage worldwide, we could have taken a lot of preventive measures, including an early tightening of border controls, an increase in flu vaccinations so as to avoid parallel waves of influenza and COVID-19, and the stockpiling of protective masks and medication,” Stolz explains. In other words, proper monitoring procedures are vital in this phase. However, even a good early-warning system is of little use when information is ignored. News of the emergence of an unknown respiratory disease in China was first issued via the international early-warning system ProMED to recipients including the Robert Koch Institute on December 31, 2019. Yet it was 78 days before the first measures were implemented in Germany.  

To identify relevant information in the mass of data, it takes not only efficient software and AI methods but also useful graphics. The Fraunhofer Institute for Computer Graphics Research IGD in Darmstadt develops such visualization tools for doctors, epidemiologists and health authorities. These include data-analysis graphics for population studies and tools for creating and comparing patient cohorts.

To identify relevant information in the mass of data, it takes not only efficient software and AI methods but also useful graphics. The Fraunhofer Institute for Computer Graphics Research IGD in Darmstadt develops such visualization tools for doctors, epidemiologists and health authorities. These include data-analysis graphics for population studies and tools for creating and comparing patient cohorts. Wherever data is patchy and there are no existing sources to make up for this lack, mathematics can help answer key questions. In the case of a pandemic, such questions include: how will the infection spread, and which measures are best suited and most effective? 

A major factor in the coronavirus equation is the number of unknown cases. This helps not only to explain differing mortality rates but also to predict future developments and specific measures. In order to determine the number of unknown cases in different regions of Germany, researchers at the Fraunhofer Institute for Industrial Mathematics ITWM modified a statistical model originally developed for use in vehicle development. Their calculations suggested there were 298,000 unidentified infections in Germany at the end of April, of which 40 percent were asymptomatic.

Mathematicians from Fraunhofer ITWM also developed a special simulation model to help the authorities with decisionmaking. This model uses time-variable parameters that are estimated on the basis of the recorded number of cases and then matched with further statistical data. This enables an assessment of the impact of specific measures on the rate of infection. In April, working with the Max Planck Society, the Helmholtz Association and the Leibniz Association, Fraunhofer researchers then used various mathematical models in order to determine the most promising strategy for the pandemic: a rigorous containment of new infections until effective contacttracing becomes possible, followed by an adaptive approach in which new cases are traced back to their origin and measures to limit contact can be introduced.