Methods
We retrieved daily data on new cases of SARS-CoV-2 (i.e., individuals with positive test results), number of tests (reverse transcription polymerase chain reaction; RT-PCR) performed, deaths, and admissions to Intensive Care Units (ICU) in each Region, from February 24th to March 18th, 2020, obtained from the Health Ministry website2. Demographic, socioeconomic and healthcare organization data were retrieved from the National Institute of Statistics (ISTAT)3.
As an index of different screening strategies, the number of tests/positive test results (T/P) ratio as of March 7th, 2020, was considered. The subsequent evolution of the epidemic was assessed through the cumulative number of deaths and of new severe cases, between March 23th and 25th, inclusive; the latter were defined as a composite of death and admission to ICU. This work is based on publicly available data, needing no ethical approval.
The association of those two outcomes with the number of T/Pratio at March 7th was assessed using a linear regression model. For each confounder significantly associated with outcomes, multivariate linear regression models were applied to assess the independent contribution of T/P, assuming two-sided p<.05 as significant. Analyses were performed on SPSS (SPSS-Inc., Chicago, IL, USA) 25.0.