To path indicator Subpages of “Press and Public Relations“ To navigation by target groups To navigation by topic To quick access To footer with other services
To content
Close-up of a camera and a press card © Harald Richter​/​Shotshop.com
Cooperation with University of Göttingen and Charité

Reliable Predictions of Coronavirus Infections

-
in
  • Research
  • Top News
  • Press Releases
Portrait of Prof. Markus Pauly © Felix Schmale​/​TU Dortmund
Professor Markus Pauly wants to make predictions of pandemic infections more reliable.
Are there ways to reliably predict the progression of COVID-19 or other pandemic infections? This is the topic of a research project in which three statistics professors are working together: Dr. Tim Friede from University Medical Center Göttingen, Dr. Frank Konietschke from the Institute of Biometry and Clinical Epidemiology at Charité – Uni­ver­si­täts­me­di­zin Berlin and the Berlin Institute of Health (BIH), and Dr. Markus Pauly from the Chair of Mathematical Statistics and Applications in Industry at TU Dortmund University. The project is due to start in the spring and will be funded by the Volkswagen Foundation for 18 months.

“Bayesian and Nonparametric Statistics – Integration of Two Opposing Theories to the Advantage of Predictive Studies on COVID-19” is the project’s rather cumbersome title. “In the context of the coronavirus pandemic, we’re currently finding that there is still a lot of uncertainty as far as predicting the progression of COVID-19 infections is concerned,” says Professor Friede – both on an individual level as well as at the level of society as a whole. Determining the risk factors and thus predicting severe disease progression are a vast statistical task, into which various types of data from sometimes small case analyses are fed. For example, predictive models are based on binary data (yes/no) and/or time-to-event data that are collected in clusters, such as in a family or at a single location. The three statisticians now want to develop statistical models that allow better predictions on the basis of existing data. “In the process, we want to minimize the risk of wrong decisions and evaluate the quality of our predictive models with the help of innovative analysis methods,” says Professor Pauly.

Results can support decision-making in intensive care

“The predictive models we develop are then tested on data sets, such as those currently being collected at Charité, for example,” says Professor Konietschke. In their research, the scientists want to unite research approaches used by their guild – Bayesian and nonparametric statistics. The result should be an accurate prediction with a sound assessment of risk and uncertainty, which could serve as guidelines for patient care and for decision-making in politics.

In this sense, the project can help, for example, to base capacity-planning decisions in intensive care units on a sound scientific forecast. Even if this will hopefully no longer be necessary for the current coronavirus pandemic by the time the project ends in late 2022 – the outbreak of the next pandemic is, according to many scientists, only a matter of time. And then the research results produced by the three statisticians, who have been working together for a long time, will be needed again.

Contact for further in­for­mation:

Back

To top of page