An novel epidemiological model for COVID-19

Abstract

AbstractCOVID-19 is characterized by a large number of asymptomatic and mild cases that are difficult to detect; most of them remain unknown, still having an important role in the transmission of the disease, this make the pandemic difficult to control. The purpose of this research is to develop an epidemiological model that allow to estimate the number of unknown/asymptomatic cases in a given area.The SEIAMPR system, a novel simulation based model for COVID-19 is designed and implemented in Python. The intuition of the model is simple: about 80% of COVID-19 infected people evolve as asymptomatic or with a mild clinical course, many of them remain unknown to the authorities, some of them including those in critical conditions are eventually detected and classified as positive cases. The simulator reproduces this process using an adaptive method integrated with official data.The simulator has been used for modelling the outbreak in 21 regions in Italy. The positive effects of lockdown policies are demonstrated: unknown active cases 12 days after the lockdown (March the 21th) ranged from 284101 to 374038, e.g. many more than all the official cases in Italy, reducing to 10213/20949 the reopening day. The number of unknown active cases at the beginning of June in the Lombardia region ranged from 6813 to 13390 demanding particular attention.SEIAMPR is simple to tune and integrate with official data, it emerges as an up-and-coming tool for reporting the effect of lockdown measures, the impact of the disease on the population, and the remaining unknown active cases for evaluating the timing of exit strategies.