Estimating weekly excess mortality at sub-national level in Italy during the COVID-19 pandemic

Abstract

In this study we present the first comprehensive analysis of the spatio-temporal differences in excess mortality during the COVID-19 pandemic in Italy. We used a population-based design on all-cause mortality data, for the 7,904 Italian municipalities. We estimated sexspecific weekly mortality rates for each municipality, based on the first four months of 2016-2019, while adjusting for age, localised temporal trends and the effect of temperature. Then, we predicted all-cause weekly deaths and mortality rates at municipality level for the same period in 2020, based on the modelled spatio-temporal trends. Lombardia showed higher mortality rates than expected from the end of February, with 23,946 (23,013 to 24,786) total excess deaths. NorthWest and NorthEast regions showed one week lag, with higher mortality from the beginning of March and 6,942 (6,142 to 7,667) and 8,033 (7,061 to 9,044) total excess deaths respectively. We observed marked geographical differences also at municipality level. For males, the city of Bergamo (Lombardia) showed the largest percent excess, 88.9% (81.9% to 95.2%), at the peak of the pandemic. An excess of 84.2% (73.8% to 93.4%) was also estimated at the same time for males in the city of Pesaro (Central Italy), in stark contrast with the rest of the region, which does not show evidence of excess deaths. We provided a fully probabilistic analysis of excess mortality during the COVID-19 pandemic at sub-national level, suggesting a differential direct and indirect effect in space and time. Our model can be used to help policy-makers target measures locally to contain the burden on the health-care system as well as reducing social and economic consequences. Additionally, this framework can be used for real-time mortality surveillance, continuous monitoring of local temporal trends and to flag where and when mortality rates deviate from the expected range, which might suggest a second wave of the pandemic.

Non-COVID-19 patients in times of pandemic: Emergency department visits, hospitalizations and cause-specific mortality in Northern Italy

“…Our findings indicate that the reduction of ED visits and hospitalizations started two weeks before the beginning of the national lockdown, after school closures and the first Italian case hospitalized for COVID-19 in Codogno [1], when no cases of local transmission had been still recorded in BMA. As already reported [2,6,[19][20][21][22][23], a possible explanation is that the population response is likely to be more affected by the national level authority risk message than the real local situation. Several possible reasons have been put forward to explain, at least partially, a reduction in ED visits and hospitalizations [6,7], such as lifestyle changes or fear of the contagion, as well as the increasingly stringent lockdown measures and the postponement of elective procedures, or the sense of civic responsibility of the population [7].…”

Epidemiological characteristics of COVID-19 cases and estimates of the reproductive numbers 1 month into the epidemic, Italy, 28 January to 31 March 2020

“…We found that those diagnosed before 4 March 2020 had an overall CFR of around 20%. Subsequent studies with longer follow-up clarified this aspect better, including studies evaluating overall population excess mortality, which are also more comparable across countries [33][34][35].…”

Non-COVID-19 patients in times of pandemic: decreased emergency department visits and increased out-of-hospital mortality in Northern Italy

“…Our findings indicate that the reduction of ED visits and hospitalizations started two weeks before the beginning of the national lockdown, after school closures and the first Italian case hospitalized for COVID-19 in Codogno [1], when no cases of local transmission had been still recorded in BMA. As already reported [2, 6, 18-20], a possible explanation is that the population response is likely to be more affected by the national level authority risk message than the real local situation [21]. Several possible reasons have been put forward to explain, at least partially, a reduction in ED visits and hospitalizations [6, 7].…”