Photo: © 2020 EPFL
With the EPFL Blue Brain Project’s determination to make our computing resources and expertise available for the fight against COVID-19, we brought our experience in software development to team up with the Foundation for Innovative New Diagnostics (FIND). FIND is a global non-profit organization focused on diagnostics, currently co-convening the Access to COVID-19 Tools (ACT) Accelerator Diagnostics Partnership alongside The Global Fund, as a key part of the global response to the pandemic.
According to the World Health Organization (WHO), diagnostic testing for COVID-19 is critical to tracking SARS-CoV-2 (the virus responsible for COVID-19), understanding epidemiology, informing case management, and suppressing transmission. With diagnostics emerging as one of the most pressing issues in the COVID-19 crisis, Blue Brain has collaborated with FIND to develop a Diagnostic Implementation Simulator for SARS-CoV-2 diagnostics.
The Diagnostic Implementation Simulator (Dx Implementation Sim) uses modeling data to rank the potential impacts of deploying different testing strategies for COVID-19 on key outcomes. The Dx Implementation Sim model is based on the well-accepted SIR (susceptible/exposed, infected and recovered) theoretical model of epidemics. To model the spread of a specific infection, such as COVID-19, in a specific country, modelers add large volumes of information about the disease and about the country into the theoretical SIR model.
Richard Walker, who was with the Blue Brain from inception until 2018, came out of retirement to bring his considerable skills and experience to the project, by leading the implementation of a new version of the SIR model, which was adapted to meet the special needs of COVID-19. In the implementation, Richard worked in close collaboration with FIND and with a team from Ateneo de Manila University, Manila, the Philippines.
The validity of the Dx Implementation Sim model was tested by fine-tuning the values for the timing and severity of government interventions to give the most accurate possible prediction of the number of deaths on a day-by-day basis in Switzerland, Italy, Hubei Province in China (where the pandemic started), and South Korea. The results showed that the model is capable of accurately predicting the course of the epidemic. However, all the countries used for validation were middle- to high-income countries. As high-quality data are often lacking for lower income countries, the model may over- or under-estimate the absolute number of deaths and exact duration of the epidemic in these countries. Therefore, while the Dx Implementation Sim cannot replace expert epidemiological data, it does provide indicative data for more general use, with predefined parameters set to ensure that realistic outputs can be obtained for a range of different country settings.
Several Blue Brain employees were redeployed with great enthusiasm to work full time on the tool, which is both desktop and mobile compatible. The Blue Brain team comprised of front and back end engineers, project managers and visual designers who worked tirelessly to translate the requirements for a web interface to the model into a working platform in a matter of days.
“Our work in digitally reconstructing and simulating the mouse brain has led us to create a powerful computing architecture, and accumulate extensive software development and design expertise,” explains Blue Brain Founder and Director, Prof. Henry Markram. “Accordingly, we made these and our other resources available to help tackle the immense global threat of COVID-19. Our team science approach meant when we were contacted by FIND, we were able to immediately put together a cross-functional team who made a seamless transition to quickly develop the Dx Implementation Sim”.
To access the Diagnostic Implementation Simulator, click here.
About the Diagnostic Implementation Simulator, clic here.
Questions on using the Simulator – FIND - email@example.com
Questions about the modeling – Richard Walker
Media enquiries - Kate Mullins, Blue Brain Communications Manager
This tool is not intended to replace detailed epidemiological models for country decision making or the estimates of deaths and of epidemic duration coming from such models.