phases 1 and 2 of covid19 epidemic in the three geographical areas of italy: an estimation of italian government measures based on a bayesian changepoint detection method

Background: based on data from the ministry of health, which highlighted the earlier onset of covid-19 epidemic in italy, compared with the europe, we would like to present a statistical elaboration on the impact of measures taken by the government, during the phase 1 and the start of phase 2. methods: after the implementation of a bayesian changepoint detection method, we looked for a best fit model, based on the first part of time series data, in order to observe the progress of the data in the presence and absence of the restriction measures introduced. results: both the implementation of changepoint detection method and the analysis of the curves showed that the decree that marked the start of lockdown has had the effect of slowing down the epidemic by allowing the start of a plateau between 21 and 25 march. moreover, the decree that decided the beginning of phase 2 on 4 may did not have a negative impact. conclusion: this statistical analysis supports the hypothesis that stringent measures decreased hospitalization, thanks to a slowing down in the evolution of the epidemic compared with what was expected.