The SARS-CoV-2 effective reproduction rate has a high correlation with a contact index derived from large-scale individual location data using GPS-enabled mobile phones in Germany

Ruediger, S.; Konigorski, S.; Edelman, J.; Zernick, D.; Lippert, C.; Thieme, A. H.

The novel coronavirus (SARS-CoV-2), which started in Hubei, China in December 2019, has caused an ongoing pandemic. Due to pauci-symptomatic cases, the virus may spread invisibly in a community. Healthcare systems have repeatedly been challenged with a rapid onset of patients. In the absence of vaccination, non-pharmaceutical interventions (NPIs) like interpersonal distancing were implemented in several countries and have been key to effectively reduce viral spreading. In Germany after an exponential growth of cases numbers in March 2020, NPIs were able to effectively control the pandemic and sufficiently reduced the daily reported new infections allowing for partial release of NPIs. Since these measures were unable to remove the virus entirely from the population, responsible behavior and interpersonal distancing are still required. Methods are needed which can rapidly identify SARS-CoV-2 outbreaks. To this end, we developed a novel statistical method evaluating the earliest possible events of infections, the contacts between individuals, which is essential for virus transmission. We derived the contact index, an index for the contact intensity of the population from spatial proximity between individuals as proxy for physical interaction based on complex network science. In an application, we estimated the contact index from GPS mobile phone data of about 1 million users in Germany, and investigated its association with infection rates in Germany. The results show that the contact index is able to model the time evolution of new infections of SARS-CoV-2. It shows a strong association with the effective reproduction number of the virus about seven days later in all observed phases in Germany (Pearson correlation r=0.88): 1) the early phase of the first wave with the highest reproduction rate, 2) phase of strict NPIs (lockdown) with the lowest reproduction, 3) release of NPIs accompanied with an increase of reproduction. This work presents the foundation to create a geographic information system that can display daily updated human contact data, which we plan to further extend to a full early warning system for SARS-CoV-2.