Abstract
Italy has been one of the most severely affected countries by the COVID-19 pandemic, and the case fatality rate (CFR) estimated based on Italian data is one of the highest worldwide. We analyzed public data from the first 50 days of the epidemic in Italy (from February 24 to April 13, 2020) to evaluate whether evolving testing strategies and capacity could account for trends in the CFR. The CFR increased during the study period, and a significant positive correlation was found between the CFR and the percentage of positive tests among performed real-time PCR tests (positive tests % [POS%]) until March 25, suggesting the surveillance system did not detect a growing number of cases in the initial phase of the epidemic. To avoid distortion due to the delay between the identification of cases and deaths, the expected CFR (expCFR) was calculated, which represents the ratio between the predicted number of cases and deaths at the end of the epidemic based on the best fitting logistic curves of the cumulative numbers of cases and deaths. The expCFR began a downward trend from the 40th day. In the final phase, a decrease in both expCFR and POS% was identified, suggesting an improvement in surveillance. The results of this study suggest data from the first 50 days of the COVID-19 epidemic in Italy were severely affected by ascertainment bias. Insufficient testing and isolation of cases could have facilitated the widespread transmission of COVID-19 in the early stages of the outbreak.