dcyphr | COVID-19 and Italy: what next?


This is a paper published during the exponential rise of COVID-19 cases in Italy, before the country experienced its peak. It uses statistical models and information gathered from Hubei, China to predict the exponential growth in the number of COVID-19 in Italy. The time frame of the predictions is over, but the methods shared remain valuable and they may be used in other regions. The data presented in the paper anticipated a spike in cases which would overwhelm the country’s intensive care capacity. It urged the government to move in over 20,000 doctors and 5000 ventilators in preparation for this spike.


China’s early, relative to the rest of the world, spike and decline in COVID-19 cases made it a source of outbreak statistics for many other nations. This paper makes use of the data available from Hubei, China to anticipate the progression of the virus in Italy. The European nation took many steps, such as internal travel restrictions, to try and contain the virus, but numbers continued to rise. With only 5200 hospital beds available and a survival time of 1-2 weeks for nonsurvivors of COVID-19, this study warns the government of the virus’s likelihood to overwhelm the nation’s healthcare system. 


The number of COVID-19 patients in Italy was published every day starting on Feb 21, 2020. Using this data, the researchers constructed an exponential model with the value of the exponent being r=0.225 (1 per day). This growth pattern was consistent with reported infected patient numbers until 17 days after the start of the model.

The trend following this 17 day period suggested that more than 30,000 patients would be infected by March 15. With infection duration between 15 and 20 days, a basic reproductive number found was 2.76-3.25 (similar to that in Wuhan). The model predicted that the Italian healthcare network would reach full capacity within a few days. Hospitals would reach their capacity by March 14, 2020, assuming half of the 5200 ICU beds in the nation were open to COVID-19 patients. 

The paper also compared the curve of Hubei (population 50 million) to the data gathered in Italy (population 60 million) so far. A divergence in the Chinese region’s predicted vs recorded patient numbers was noted. If Italy experienced a similar divergence, the paper predicts the number of newly infected patients could decrease within 3 or 4 days from March 11.


While most of the modeling in this paper is based on data from Hubei, China, it is noted that a comparison between Italy and this Chinese region is difficult. The draconic measures taken in China to reduce infection are likely impossible in Italy. Additionally, genetic differences and differences in the percentage of the population with diabetes, heart conditions, or other comorbidities are hard to take into account. Lastly, the use of antiviral drugs or other pharmaceutical measures to reduce infection were different in both nations and were not taken into account in the models presented. 

Regardless, the paper clearly states that it is addressing government officials, imploring them to prepare for an inevitable spike in cases.