The model learned to accurately predict the actual number of cases of COVID-19. During its development, scientists from the Georgia College of Medicine used well-known factors, such as population density and age distribution in a particular area. The work was published in the journal Infection Control and Hospital Epidemiology.

When compiling the model, mathematicians took data only from open sources (for example, from the website of the World Health Organization), and then added to the information about the population density and the proportion of the population of the district that lives in cities. After that, the researchers divided the entire population into three age groups: from 0 to 14, from 15 to 64, and from 65 or more.

The model takes into account the number of cases of infection before the first recorded peak, as well as the date ranges for these peaks as an indicator of the trend towards an increase in the number of registered cases.

Researchers then tested the model on data collected in Italy. The analysis showed that for each confirmed case there are four unconfirmed ones – due to the extremely high population density in the country. At the time of the first recorded peak on March 9, thus, the actual number of cases in the country was an additional 30 thousand cases.

A calculation based on data from China showed that at the time of the peak in the country, there were actually between 12 million and 89 million infected with the new type of coronavirus. This is from about 149 to 1.49 thousand unacknowledged cases of infection for one confirmed.

“The actual preparedness for a pandemic depends on how accurate we know the actual number of people infected. At higher rates, we will be able to better assess how long the virus will remain active and how negative it can lead to infection. Without these indicators, health systems and doctors cannot prepare for an epidemic”.

Shrinivas Rao, lead author of the study