Model created to estimate number of COVID-19 cases in each autonomous community
A team of mathematicians and statisticians from the UAB, Humboldt University in Berlin and the UPC are developing a model to estimate the daily number of new cases of people with COVID-19, given the impossibility to gather this data directly from the population due to so many light and asymptomatic cases.
31/03/2020
Spain has become the fifth country with the largest amount of COVID-19 cases. Although several measures have been taken to lower the impact of the outbreak and contribute to flattening the curve of cases, official data is based on numbers reflecting the most severe cases. Therefore, it is unknown how many other light or asymptomatic cases may exist and are also contributing to the expansion of the pandemic.
The protocols used to determine cases in Spain mainly includes people with severe symptoms, but now the authorities have announced a new protocol with rapid tests to detect the virus in populations with light symptoms who currently are not included in the statistics. In any case, the size of the population infected, but with no symptoms, will be higher than that registered until now.
In 2016, researchers from the Department of Mathematics at the UAB developed a method by which to analyse underrepresented data in statistics (Under‐reported data analysis with INAR‐hidden Markov chains, Amanda Fernández‐Fontelo, Alejandra Cabaña, Pedro Puig, David Moriña, Statistics in Medicine, July 2016). The method allows for an accurate estimation of the number of cases not registered officially, with a series of applications on public health, such as the monitoring of the number of real cases of infections caused by the human papillomavirus (HPV), botulism and also real cases of victims of domestic violence (Untangling serially dependent underreported count data for gender‐based violence, Amanda Fernández‐Fontelo, Alejandra Cabaña, Harry Joe, Pedro Puig, David Moriña, Statistics in Medicine, July 2019).
The UAB researchers, in collaboration with the Department of Computing Science at the UPC, the Humboldt-Universität zu Berlin and the Centre for Mathematic Research, are using this method to update daily the situation of COVID-19, and particularly to quantify the unreported cases within the official registry of people infected in Spain. The results allow for a more realistic picture of the pandemic in real time, as well as more precise fundamental data such as real mortality rates and basic reproduction number, necessary for professionals and politicians to take decisions. The analysis is designed so that it can be easily reproduced with data from other countries.
More information:
Link to follow daily updates on the new, unregistered cases and representation of the outbreak.
https://underreported.github.io/
In the photo, the team of researchers: David Moriña (CRM), Amanda Fernández (HU-Berlin), Pere Puig (UAB), Alejandra Cabaña (UAB) and Argimir Arratia (UPC).