Padua, May 7 (beraking latest news Salute) – How effective have the closures been in containing the virus in Italy? A team of researchers from the University of Padua, using new and sophisticated artificial intelligence techniques, has accurately reconstructed the influence that restrictions had on the behavior of Italians. The study, published in ‘Scientific Reports’ (Springer Nature), showed that social distancing and awareness of the risk of infection increased day by day in the first months of 2020, allowing to control the epidemic, but also that “this level awareness tends to decrease quickly as soon as some restrictions are removed “, explains Gianluigi Pillonetto, Department of Information Engineering of the University of Padua and first author of the study.
“This – explains the expert – first occurred in August 2020, a season in which the adoption of social distancing measures and the use of precautions such as masks has greatly decreased due to the relaxation of Italians during the holidays, in particularly in the younger population, and then in mid-October 2020, due to the return to schools and closed workplaces. Unfortunately, these phenomena were the key to triggering the second wave “.
The study shows that the daily data on the number of people hospitalized in intensive care in Italy due to Covid-19 are the most informative to understand how the epidemic really evolves. It also reveals that the actual number of infected subjects tends to be significantly underestimated even by antibody tests. This suggests that in people affected by Covid-19, particularly in asymptomatic people, the antibody level may be poorly detectable and drop rather quickly, even after a few months.
“The population serological study conducted in Italy using the first generation of tests with high sensitivity and specificity showed in the middle of last year that 2.5% of Italians had contracted the virus. Our study, on the other hand, estimates a percentage at least double, around 5 %, not excluding that it could already reach 12% in mid-2020 – continues Pillonetto -. A consequence of this result is that lethality from Covid-19 can be very overestimated. In 2020, using only the number of deaths and the number of positive cases diagnosed with nasopharyngeal swab, lethality was around 14% “.
“Our new approach based on artificial intelligence instead leads to a new estimate of around 1%, a lethality value certainly closer to that reported by studies that have carried out an accurate assessment of the prevalence of SARS infection. -CoV-2. On the other hand it is difficult to test the population of asymptomatic or paucisymptomatic subjects with both molecular and antigenic swabs and even more to trace the contacts of the positives when the incidence of infection exceeds 50 cases per week per 100,000 residents. It is possible that in a medium-high incidence situation, when, for example, swabs reveal 1,000 new infected subjects per day, the true number could be around 14,000 “, concludes Pillonetto.
COPYRIGHT today © breaking latest news