key: cord-0824138-nuq650bt
authors: Di Bari, M.; Balzi, D.; Carreras, G.; Onder, G.
title: Extensive testing may reduce COVID-19 mortality: a lesson from northern Italy
date: 2020-05-01
journal: nan
DOI: 10.1101/2020.04.24.20078709
sha: 66b30862ff642368aa6df2eefa1ec53fdfa3e88a
doc_id: 824138
cord_uid: nuq650bt

We examined data on the progression of COVID-19 epidemics in four regions in northern Italy. Lombardy, Emilia-Romagna, and Piedmont had an extremely steeper increase in mortality with increasing number of tests performed than Veneto, which applied a policy of broader swab testing. This suggests that the strategy adopted in Veneto, similar to that in South Korea, is effective in containing COVID-19 epidemics and should be applied in other regions of Italy and countries in Europe.

Initially, epidemiological surveillance and strategies for swab testing were under control of regional healthcare authorities. On February 25, the Italian Ministry of Health issued more stringent testing policies for application of swabs to identify COVID-19 cases, prioritizing patients with respiratory symptoms and possible COVID-19 contacts who required hospitalization. Most regions promptly complied with these recommendations, whereas Veneto maintained its policy, implemented after the occurrence of the first cases, of extensive testing and isolation of positive cases (3). Surprisingly, the debate stemming from these different regional policies valued international more than Italian evidences (4). We aimed at assessing, using data from the first All rights reserved. No reuse allowed without permission.

was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 1, 2020. . https://doi.org/10.1101/2020.04.24.20078709 doi: medRxiv preprint month of the Italian experience, how different policies for swab testing may impact on the initial progression of COVID-19 epidemics.

Data were obtained from the reports of the Italian Department of Civil Protection, issued since February 24, which include daily number of swabs performed and deaths from COVID-19 in each region (5). We compared Lombardy, Emilia-Romagna, and Piedmont, three regions in northern Italy that closely followed the recommendations for restrictive COVID-19 testing, and Veneto, which applied a policy of broader testing (3).

The cumulative number of tests performed in the first month, from February 24 through March 27, and COVID-19 cumulative mortality from March 2 through April 3, were indexed by population in each region (6) . The 7-day lag time between COVID-19 testing and mortality was allowed because death occurs usually 7+ days after clinical onset and diagnosis (2) .

Piecewise linear regression was applied, separately for the four regions, to identify the breakpoints in the slope of the number of tests performed over time, and to examine whether the relationship between the cumulative number of tests performed through each date (independent) and mortality (dependent variable) followed a different progression over time in the four regions.

The effectiveness of the two testing strategies was estimated by regressing the number of tests (dependent) and the cumulative lagged mortality (independent variable), separately for the two most distant scenarios of Lombardy and Veneto: the slope of these regressions represents the number of tests associated with one death.

The proportion of positive cases was calculated as the percent ratio of the 3-day moving average of positive cases over the 3-day moving average of tests performed, to compensate for delays and imprecisions in the daily reporting of data. One-way ANOVA was used to evaluate differences in the proportion of positive cases across the four regions, applying the Games Howell test for unbalanced variances for post-hoc comparisons. Pearson's r correlation coefficient was used All rights reserved. No reuse allowed without permission.

was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 1, 2020. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 1, 2020. 

We observed that extensive swab testing, applied since the beginning of the epidemics, may reduce the spread of COVID-19, by identification of a high number of positive cases that can be eventually isolated (3,7). Four regions in the same area of Italy, with similar demographics, economics, and healthcare system, which were almost simultaneously hit by the virus, adopted different strategies for COVID-19 outbreak containment. In Veneto, where extensive testing was applied (3), the increase in COVID-19 mortality was milder than in the other regions, which initially clustered in a steeper relation between the number of tests and mortality. Accordingly, the proportion of positive tests was lower in Veneto than elsewhere, whereas the rate of daily increase in mortality in Emilia-Romagna, initially similar to that in Lombardy, declined when the rate of daily increase in the number of tests performed became steeper. Thus, whereas with its policy for extensive testing Veneto was efficaciously containing the spreading of the disease, Lombardy, Piedmont, and initially also Emilia-Romagna, were somehow chasing the virus, using tests more to confirm clinically plausible diagnoses than to contain the epidemics.

To estimate the spread of the disease, we used COVID-19 mortality instead of the number of positive tests, which depends heavily on the policy for testing: other factors being the same, the All rights reserved. No reuse allowed without permission.

was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 1, 2020. . https://doi.org/10.1101/2020.04.24.20078709 doi: medRxiv preprint 6 broader the criteria for testing, the wider the denominator, the lower the proportion of positive tests, and vice versa. A 7-day lag was allowed to identify deaths, as this is the minimal interval to attribute death to COVID-19 (2) .

Being the first western country facing COVID-19 outbreak, Italy represents a living-lab to evaluate the effectiveness of practices to contrast it (8). In agreement with data from South Korea (9), a broader policy for swab testing, such as that applied in Veneto, appears to contribute successfully to contain COVID-19 threat.

Critical care utilization for the COVID-19 outbreak in Lombardy, Italy: early experience and forecast during an emergency response

Case-fatality rate and characteristics of patients dying in relation to COVID-19 in Italy

Popolazione residente al 1° gennaio

Suppression of COVID-19 outbreak in the municipality of

All rights reserved. No reuse allowed without permission.was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which this version posted May 1, 2020. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which this version posted May 1, 2020. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which this version posted May 1, 2020. . https://doi.org/10.1101/2020.04.24.20078709 doi: medRxiv preprint