COVID-19 and the associated restriction of lung function that can develop into a life-threatening illness presents a new challenge to society, government, commerce and, above all, research. The number of verified SARS-CoV-2 infections around the world is in the millions and, in Germany, hundreds of thousands of people have already tested positive for the new virus. According to virologists, we all will be dealing with this disease for some time still -- until vaccines and medications become widely available or enough of the population becomes immune.
What can help
Until then, it is important to extract as much information as possible from the available data and develop solutions for health professionals, epidemiologists, public health authorities, etc. to help manage the current crisis and be prepared for the later waves of infections that are expected. This includes software solutions that compile available data, help analyze it and thus improve our knowledge of the COVID-19 disease. At the same time, the goal of such efforts must always be to put tools in the hands of those fighting this disease that help them in their work -- treating the infected as well as protecting the population from SARS-CoV-2 infection.
What we are doing
For many years now, Fraunhofer IGD has developed software solutions that help health professionals and others in the healthcare sector. These include data analysis tools for population studies as well as for forming and comparing patient cohorts. The institute also focuses on the analysis of medical imaging data for extracting biomarkers. Lastly, the visual presentation of data and analytical results also play an important role in our work.
In the context of COVID-19, Fraunhofer IGD’s expertise can be used for various issues and the institute is already working on providing appropriate solutions:
Analyzing and visualizing COVID-19 data from German state public health authorities
Developing methods for the exploratory visual analysis of COVID-19 patient cohorts
Classifying and quantifying pulmonary changes in the ultrasound images of COVID-19 patients
Correlating image-based and clinical biomarkers for the early detection of COVID-19 patients with a potentially critical disease progression
Visual analytics for COVID-19: details
As stated in the federal government of Germany’s 4/15 resolution on additional COVID-19 measures, one primary focus of the coming weeks will be on better understanding how the virus spreads and the probable links between infections. Visualization and the visual analysis of data in particular can help immensely in this endeavor.
The researchers at Fraunhofer IGD have been working with medical data and the visual analysis of this medical data for many years. COVID-19 is a prime example of a problem that cannot be analyzed with automated methods alone. Medicine and the entire world are in uncharted waters with COVID-19, so AI methods have no teaching data from which they can learn. Only through the capabilities of medical experts will we come up with helpful and successful ways to stem the tide and optimize healthcare processes as quickly as possible -- until a vaccine is available.
This is where technologies for visual correlation analysis (finding relationships in complex data volumes and visualizing them in a helpful manner), similarity-based cohort analysis (what age group, with what preexisting conditions, with what symptoms and with what treatment lives longer than five days in the ICU) and data visualization (what visualization helps the general population, what one is more suitable for medical experts in assessing the situation) come in. Together with health professionals from various hospitals around Germany, Fraunhofer IGD is currently working on these very solutions.