Now-casting the COVID-19 epidemic: The use case of Japan, March 2020

Stephan Gloeckner; Gerard Krause; Michael Hoehle

The added value of this now-casting approach increases with the length of the reporting delaythe larger the delays the more important is the adjustment. As illustrated per 95% prediction intervals in figure 3 , uncertainty of now-casts is obviously highest for the days are closest to the actual day of the now-cast itself. We used publicly available data for our analysis, which tends to be less complete than the data available to national surveillance agencies. This slightly limits the precision of our now-cast, but also indicates the even higher value of now-casting when used within a nation's surveillance system. For this reason, we are currently building this tool as a permanent functionality into the dashboard of the Surveillance Outbreak Response Management System (www.sormas.org) to further enhance its role in COVID-19 response. [20] [21] [22] [23] Via the open-source R package surveillance applied within the user-friendly shiny application, public health surveillance institutions can easily apply this now-cast approach in their country for COVID-19 or any other disease. No system or country specific configurations are needed. As a specific use case this tool may be of particular value for the challenging risk assessment and risk communication in the context of the Summer Olympic Games in Tokyo July to August 2020 and similar situations elsewhere.