Hostname: page-component-7c8c6479df-94d59 Total loading time: 0 Render date: 2024-03-28T20:09:53.165Z Has data issue: false hasContentIssue false

Understanding the Factors Affecting COVID-19 Mortality in Italy: Does a Relationship Exist With a Sharp Increase in Intensive Care Unit Admissions?

Published online by Cambridge University Press:  15 October 2021

Giulia Lorenzoni
Affiliation:
Unit of Biostatistics, Epidemiology and Public Health, Department of Cardiac, Thoracic, Vascular Sciences, and Public Health, University of Padova, Italy
Danila Azzolina
Affiliation:
Unit of Biostatistics, Epidemiology and Public Health, Department of Cardiac, Thoracic, Vascular Sciences, and Public Health, University of Padova, Italy Department of Medical Sciences, University of Ferrara, Ferrara, Italy
Aslihan Şentürk Acar
Affiliation:
Department of Actuarial Sciences, Hacettepe University, Ankara, Turkey
Luciano Silvestri
Affiliation:
Unit of Biostatistics, Epidemiology and Public Health, Department of Cardiac, Thoracic, Vascular Sciences, and Public Health, University of Padova, Italy
Paola Berchialla
Affiliation:
Department of Clinical and Biological Sciences, University of Turin, Turin, Italy
Dario Gregori*
Affiliation:
Unit of Biostatistics, Epidemiology and Public Health, Department of Cardiac, Thoracic, Vascular Sciences, and Public Health, University of Padova, Italy
*
Corresponding author: Dario Gregori, Email: dario.gregori@unipd.it.
Rights & Permissions [Opens in a new window]

Abstract

Objective:

The present study aims to explore whether a relationship exists between the immediate sharp increase in intensive care unit (ICU) admissions and the mortality rates in Italy.

Methods:

Official epidemiological data on coronavirus disease (COVID-19) were employed. The forward lagged (0, 3, 7, 14 days) daily variations in the number of deaths according to the number of days after the outbreak started and the daily increases in ICU admissions were estimated.

Results:

A direct relationship between the sharp increase of ICU admissions and mortality rates has been shown. Furthermore, the analysis of the forward lagged daily variations in the number of deaths showed that an increase in the daily number of ICU admissions resulted in significantly higher mortality after 3, 7, and 14 days. The most pronounced effect was detected after 7 days, with 250 deaths (95% CI: 108.1-392.8) for the highest increase in the ICU admissions, from 100 to 200.

Conclusions:

These results would serve as a warning for the scientific community and the health care decision-makers to prevent a quick and out-of-control saturation of the ICU beds in case of a relapse of the COVID-19 outbreak.

Type
Original Research
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
© The Author(s), 2021. Published by Cambridge University Press on behalf of Society for Disaster Medicine and Public Health, Inc.

The coronavirus disease (COVID-19) outbreak posed severe challenges to the health care systems of the countries worldwide. The dramatic and unexpected sharp increase of the demand for health care assistance led to a shortage of the health care resources in a short time, especially for what concerns intensive care units (ICUs). Reference Grasselli, Zangrillo and Zanella1,Reference Zhuang, Cao and Zhao2

Italy is the first European country where the COVID-19 started spreading, Reference Harari and Vitacca3 initially from 2 Northern regions, Lombardia and Veneto, and then to the rest of Italy. Reference Sebastiani, Massa and Riboli4 Since the beginning of the epidemic, the regional governments had to face the potential local health care systems crisis related to the outbreak. Previous experience of such a health crisis was only in Wuhan city, China, whose health care system is not comparable with the Italian ones.

At the beginning of the epidemic, preliminary Italian data about the outbreak show a dramatic situation, including higher mortality rates Reference Onder, Rezza and Brusaferro5 and a higher proportion of severely ill patients treated with invasive ventilation in the ICUs Reference Grasselli, Zangrillo and Zanella1 compared to those of China. Understanding factors that affect the high ICU admission rates and the high mortality rates is of primary importance to improve health care resources planning.

The present study aims to explore whether a relationship exists between the immediate sharp increase in ICU admissions and the mortality rates in Italy, with a focus on 2 Italian regions, Lombardia and Veneto.

Although Lombardia and Veneto are the 2 Italian regions where the virus first spread, the evolution of the epidemic in Lombardia was different from the evolution of the epidemic in Veneto. Soon after the starting of the outbreak, it has become clear that the proportions of COVID-19 patients hospitalized and those of patients admitted to the ICUs were markedly different among these 2 Italian regions. 6 Such differences are probably the result of 2 different approaches of health care governance and outbreak management. The Italian National Health Service (NHS) provides universal coverage free of charge by tax funding that is, in-hospital care and general practitioners consultation, or at a minimal charge, that is, drugs purchase and outpatient visits. However, the NHS is regionally based, meaning that the management of the regional health care systems falls within the competence of each Italian region, in a common framework established by the national government. Consequently, slightly different strategies have been put forward to organize the regional health care systems during the COVID-19 outbreak. Reference Cavarretta, Biondi-Zoccai, Frati and Versaci7 For these reasons, these 2 Italian regions would serve as a paradigm to understand how different approaches in the management of the regional ICU systems may affect mortality.

Methods

The source of data for COVID-19 ICU admissions, deaths, and swabs was the Italian Civil Protection Department of the Council of Ministers Presidency (https://github.com/pcm-dpc/COVID-19 ), while the resident population data have been retrieved from the National Italian Statistics Institute (www.istat.it). The time series of ICU admissions have been considered until April 22, 2020. In Italy, the regional policies for the diagnosis of COVID-19 infection are heterogeneous, providing different levels of inclusion of the general population. For this reason, the regional estimates of ICU per 100 000 inhabitants were computed using a weighted ratio estimator.

Statistical Analysis

The weekly total number of deaths and ICU admissions were reported in absolute values and over 100 000 inhabitants. The data were displayed for Italy, Lombardia, and Veneto.

The ICU admissions (ie, daily variation in ICU admissions) over 100 000 inhabitants, in comparison with the number of deaths (daily variation in deaths) over 100 000 inhabitants, were represented in a plot together with a Local Polynomial Regression Smoothing curve (LOESS). Reference Chambers and Hastie8

The forward lagged (0, 3, 7, 14 days) daily variations in the number of deaths according to the number of days after the outbreak started and the daily increases in ICU admissions were estimated via the Ordinary Least Square (OLS) method. The non-linearity of the effects was modeled using a Restricted Cubic Spline (RCS) approach. Reference Harrell9 The estimated effects were reported according to different increases in ICU occupation, together with relative SE and 95% CI.

The model estimates were reported according to different variations in ICU daily admissions:

  1. 1. The effects of daily death variations for Italy were evaluated for an increase of ICU admissions ranging from 0 to 50, 50 to 100, and 100 to 250. The time effect was estimated from the 11th to the 30th epidemic day.

  2. 2. An increase in ICU admissions ranging from 0 to 5, 5 to 10, and 10 to 20 was instead considered for Veneto. The time effect was computed from the 9th to the 26th epidemic day.

  3. 3. For Lombardia, the effect on daily death variations was evaluated for an increase of ICU occupations ranging from 0 to 20, 20 to 40, and 40 to 60. The time effect was estimated from the 11th to the 30th epidemic day.

Computations were performed using R 3.5.210 with rms Reference Harrell11 package.

Results

Table 1 shows descriptive data on weekly COVID-19 ICU admissions and the number of deaths from the beginning of the outbreak to April 21. The ICU admission rates in Italy showed an increasing trend until April 7, with 45.91 ICU admissions per 100 000 inhabitants in the week between April 1 and April 7. Conversely, the mortality rates increased until the last observation week included in the study (April 14-21). Lombardia and Veneto showed the same pattern observed for Italian data. However, both ICU admissions and mortality rates were found to be markedly higher in Lombardia compared to Veneto in all of the weeks considered.

Table 1. Number of cases, deaths, and ICU admissions according to weeks. The number of weekly events is reported both as absolute values and over 100 000 inhabitants. The data are reported for Italy (Panel 1), Veneto (Panel 2), and Lombardia (Panel 3)

A direct relationship has been shown between the sharp increase of ICU admissions and mortality rates (Figure 1 and Figure S1, Supplementary Material). Furthermore, the analysis of the forward lagged (0, 3, 7, 14 days) daily variations in the number of deaths according to the increases in ICU daily admissions showed that an increase in the daily number of ICU admissions resulted in a significantly higher number of deaths after 3, 7, and 14 days. The most pronounced effect was detected after 7 days, with 250 deaths (95% CI: 108.1-392.8) for the highest increase in the ICU admissions, from 100 to 200 (Table 2, Figure S2).

Figure 1. ICU admissions (daily variation in ICU admission) over 100 000 inhabitants versus death (daily variation in deaths) over 100 000 inhabitants.

Table 2. Estimated forward lagged (0, 3, 7, 14 days) daily variation in the number of deaths according to the days since the outbreak started and daily increases in ICU occupations. The estimated effects (effect daily deaths) are reported according to different ICU occupation increases (identified in the columns labeled high and low) together with SE and 95% CI

The analysis of Veneto data did not detect any significant effect in the forward lagged (0, 3, 7, 14 days) daily variations in the number of deaths. Conversely, in line with Italian data, Lombardia data showed that the daily increase in the ICU admissions resulted in a significant increase in the daily variation of the number of deaths, especially after 7 and 14 days. An increase in ICU admissions from 0 to 20 resulted in 112.9 deaths (95% CI: 28.56-197.2) after 7 days and an increase in ICU admissions from 20 to 40 resulted in 74 new deaths (95% CI: 28.29-119.7) after 7 days.

Discussion

The results of the present study show a strong relationship between the immediate sharp increase of ICU admissions and mortality rates in Italy. These results are confirmed by the analysis of data of 2 Italian regions, Lombardia and Veneto, where the epidemic outbreak has started. Lombardia presented both higher ICU admissions and higher mortality compared to Veneto and showed a significant effect of the daily increase in the ICU admissions on the forward lagged daily variation in the number of deaths, especially after 7 and 14 days. This is consistent with the literature showing that the median ICU length of stay is 9 days (interquartile range, 6-13), Reference Grasselli, Zangrillo and Zanella1 and mortality continued to increase after the peak of ICU admissions.

Literature has already shown that the ICU bed shortage resulted in higher COVID-19 mortality, especially at the beginning of the outbreak. Reference Lin12 The value added by the present study is that the daily variation itself of the ICU admission is associated with higher mortality rates. The results of this study may have an important impact on the organization of health care systems during the COVID-19 outbreak. They highlight the need for preventing the sharp increase in ICU admissions. Such findings could be explained by the fact that the hospital resources available were not able to face the quick increase of ICU admissions, likely affecting the quality of care. As an example, the ICU capacity has been increased soon after the start of the outbreak, Reference Grasselli, Pesenti and Cecconi13 but trained personnel have not been increased to treat the increase of critically ill patients.

Several factors could affect the immediate and sharp increase of ICU admissions, including both non-modifiable (eg, the age structure of the population, comorbidities, the severity of the underlying disease) and modifiable factors. Among modifiable factors, the virus transmission speed could be affected by the adoption of containment measures at the community level, which have been proven to be effective. Reference Gregori, Azzolina and Lanera14 However, ICU admissions may also depend on the actions taken by the health care decision-makers. An analysis of the testing strategies for COVID-19 used by Lombardia and Veneto shows that the testing strategy is associated with hospital admission, in favor of a wide testing strategy, including also mild/asymptomatic subjects, as the one adopted by the Veneto region. Reference Berchialla, Giraudo and Fava15,Reference Lorenzoni, Lanera and Azzolina16 Such findings could be explained by the fact that testing also asymptomatic and mildly symptomatic patients allows for a prompt quarantine for those subjects and for better clinical monitoring that would be helpful in preventing the worsening of symptoms. Finally, the ICU admission rates would also be affected by the organization of the health care system itself. As an example, the Lombardia region’s health care organization is characterized by the shortage of community health facilities, resulting in a physiologically higher hospital admission rate.

Conclusion

This study shows a clear relationship between the immediate sharp increase in ICU admissions and mortality rates in Italy. These results would serve as a warning for the scientific community and the health care decision-makers in order to prevent a quick and out-of-control saturation of the ICU beds in case of a relapse of the COVID-19 outbreak.

Supplementary material

For supplementary material accompanying this paper visit https://doi.org/10.1017/dmp.2021.314

References

Grasselli, G, Zangrillo, A, Zanella, A, et al. Baseline characteristics and outcomes of 1591 patients infected with SARS-CoV-2 admitted to ICUs of the Lombardy region, Italy. JAMA. 2020;323(16):1574-1581. doi: 10.1001/jama.2020.5394 CrossRefGoogle ScholarPubMed
Zhuang, Z, Cao, P, Zhao, S, et al. The shortage of hospital beds for COVID-19 and non-COVID-19 patients during the lockdown of Wuhan, China. Ann Transl Med. 2021;9(3):200. doi: 10.21037/atm-20-5248 CrossRefGoogle ScholarPubMed
Harari, S, Vitacca, M. COVID-19 spread: the Italian case. Respir Med Res. 2020;78:100771.Google ScholarPubMed
Sebastiani, G, Massa, M, Riboli, E. Covid-19 epidemic in Italy: evolution, projections and impact of government measures. Eur J Epidemiol. 2020;35(4):341-345.CrossRefGoogle ScholarPubMed
Onder, G, Rezza, G, Brusaferro, S. Case-fatality rate and characteristics of patients dying in relation to COVID-19 in Italy. JAMA. 2020;323(18):1775-1776. doi: 10.1001/jama.2020.4683 Google ScholarPubMed
COVID-19 – Italia. R-UBESP. Date unknown. Accessed March 28, 2020. https://r-ubesp.dctv.unipd.it/shiny/covid19ita/ Google Scholar
Cavarretta, E, Biondi-Zoccai, G, Frati, G, Versaci, F. Veneto’s successful lesson for a world shocked by COVID-19: think globally and act locally. J Cardiothorac Vasc Anesth. 2020;34(9):2346-2348.CrossRefGoogle ScholarPubMed
Chambers, JM, Hastie, TJ, eds. Statistical models in S. Chapman & Hall; 1992.Google Scholar
Harrell, FE. Jr Regression modeling strategies: with applications to linear models, logistic and ordinal regression, and survival analysis. Springer; 2015.CrossRefGoogle Scholar
The R Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing. Published 2015. Accessed 30 Oct 2021. https://www.R-project.org/ Google Scholar
Harrell, FEJ. rms: regression modeling strategies. R package version 4.1-3. Published online 2014. http://CRAN.R-project.org/package=rms Google Scholar
Lin, Z. Analysis of hospital resource availability and COVID-19 mortality across the United States. J Hosp Med. 2021;16(4):211-214.Google Scholar
Grasselli, G, Pesenti, A, Cecconi, M. Critical care utilization for the COVID-19 outbreak in Lombardy, Italy: early experience and forecast during an emergency response. JAMA. 2020;323(16):1545-1546. doi: 10.1001/jama.2020.4031 CrossRefGoogle ScholarPubMed
Gregori, D, Azzolina, D, Lanera, C, et al. A first estimation of the impact of public health actions against COVID-19 in Veneto (Italy). J Epidemiol Community Health. Published online May 4, 2020. doi:10.1136/jech-2020-214209CrossRefGoogle Scholar
Berchialla, P, Giraudo, MT, Fava, C, et al. To swab or not to swab? The lesson learned in Italy in the early stage of the COVID-19 pandemic. Appl Sci. 2021;11(9):4042.CrossRefGoogle Scholar
Lorenzoni, G, Lanera, C, Azzolina, D, et al. Is a more aggressive COVID-19 case detection approach mitigating the burden on ICUs? Some reflections from Italy. Crit Care. 2020;24(1):175. doi: 10.1186/s13054-020-02881-y CrossRefGoogle ScholarPubMed
Figure 0

Table 1. Number of cases, deaths, and ICU admissions according to weeks. The number of weekly events is reported both as absolute values and over 100 000 inhabitants. The data are reported for Italy (Panel 1), Veneto (Panel 2), and Lombardia (Panel 3)

Figure 1

Figure 1. ICU admissions (daily variation in ICU admission) over 100 000 inhabitants versus death (daily variation in deaths) over 100 000 inhabitants.

Figure 2

Table 2. Estimated forward lagged (0, 3, 7, 14 days) daily variation in the number of deaths according to the days since the outbreak started and daily increases in ICU occupations. The estimated effects (effect daily deaths) are reported according to different ICU occupation increases (identified in the columns labeled high and low) together with SE and 95% CI

Supplementary material: PDF

Lorenzoni et al. supplementary material

Figures S1 and S2

Download Lorenzoni et al. supplementary material(PDF)
PDF 164.5 KB