Skip to main content Accessibility help
×
Home
Hostname: page-component-564cf476b6-pp5r9 Total loading time: 0.367 Render date: 2021-06-21T02:25:52.390Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "metricsAbstractViews": false, "figures": true, "newCiteModal": false, "newCitedByModal": true, "newEcommerce": true }

Predicted Effects of Stopping COVID-19 Lockdown on Italian Hospital Demand

Published online by Cambridge University Press:  18 May 2020

Jordy Bollon
Affiliation:
Department of Translational Medicine, Università del Piemonte Orientale, Novara, Italy
Matteo Paganini
Affiliation:
CRIMEDIM – Research Center in Emergency and Disaster Medicine, Università del Piemonte Orientale, Novara, Italy
Consuelo Rubina Nava
Affiliation:
Department of Economics and Political Science, University della Valle d’Aosta, Aosta, Italy
Nello De Vita
Affiliation:
Department of Translational Medicine, Università del Piemonte Orientale, Novara, Italy
Rosanna Vaschetto
Affiliation:
Department of Translational Medicine, Università del Piemonte Orientale, Novara, Italy
Luca Ragazzoni
Affiliation:
Department of Translational Medicine, Università del Piemonte Orientale, Novara, Italy CRIMEDIM – Research Center in Emergency and Disaster Medicine, Università del Piemonte Orientale, Novara, Italy
Francesco Della Corte
Affiliation:
Department of Translational Medicine, Università del Piemonte Orientale, Novara, Italy CRIMEDIM – Research Center in Emergency and Disaster Medicine, Università del Piemonte Orientale, Novara, Italy
Francesco Barone-Adesi
Affiliation:
Department of Translational Medicine, Università del Piemonte Orientale, Novara, Italy CRIMEDIM – Research Center in Emergency and Disaster Medicine, Università del Piemonte Orientale, Novara, Italy
Corresponding
Rights & Permissions[Opens in a new window]

Abstract

Objectives:

Italy has been one of the first countries to implement mitigation measures to curb the coronavirus disease 2019 (COVID-19) pandemic. There is currently a debate on when and how such measures should be loosened. To forecast the demand for hospital intensive care unit (ICU) and non-ICU beds for COVID-19 patients from May to September, we developed 2 models, assuming a gradual easing of restrictions or an intermittent lockdown.

Methods:

We used a compartmental model to evaluate 2 scenarios: (A) an intermittent lockdown; (B) a gradual relaxation of the lockdown. Predicted ICU and non-ICU demand was compared with the peak in hospital bed use observed in April 2020.

Results:

Under scenario A, while ICU demand will remain below the peak, the number of non-ICU will substantially rise and will exceed it (133%; 95% confidence interval [CI]: 94-171). Under scenario B, a rise in ICU and non-ICU demand will start in July and will progressively increase over the summer 2020, reaching 95% (95% CI: 71-121) and 237% (95% CI: 191-282) of the April peak.

Conclusions:

Italian hospital demand is likely to remain high in the next months. If restrictions are reduced, planning for the next several months should consider an increase in health-care resources to maintain surge capacity across the country.

Type
Brief Report
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 (http://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
Copyright © 2020 Society for Disaster Medicine and Public Health, Inc.

Since December 2019, the coronavirus disease 2019 (COVID-19) has rapidly escalated into a pandemic that is currently threatening national health systems across the world. 1 To date, with more than 100,000 confirmed cases of COVID-19 and 20,000 deaths, 2 Italy has also been one of the most affected countries, accounting for the highest death toll in Europe (22%) 3 and for 15% of the global death toll. 1 Italy has been also one of the first countries in the world to implement mitigation measures, ie, isolation, quarantine, social distancing, and community containment, in the attempt to curb the COVID-19 pandemic. On March 9, 2020, 4 a national lockdown was issued, including: (a) strict home confinement for the entire population; (b) closure of all nonessential commercial activities; (c) mobility restrictions between municipalities. As of April 17, approximately 1 month later, the number of infected subjects, intensive care unit (ICU) patients, and non-ICU patients in Italy was 106,962, 2812, and 25,786, respectively. The highest ICU (4068 beds) and non-ICU (29,010 beds) demand was recorded some weeks earlier (April 3 and 4) and then declined, while the number of infected was still on the rise, but approaching a plateau. 2 Considering that, at the beginning of 2020, there were in Italy approximately 210,000 hospital beds (3.5 beds per 1000 inhabitants) and 5000 ICU beds (0.8 beds per 1000 inhabitants), it is no surprise that the national health system was severely impaired during the early phase of the epidemic. As in other countries, there is currently a debate on when such measures should be loosened, and what can be the best strategy to manage this new phase of the pandemic. Reference Xu and Li5 In particular, there are concerns on the effects that an untimely lift of the lockdown could have on health-care systems.

Pandemics are natural disasters characterized by a slow but exponential onset and a prolonged impact. Reference Hick, Einav and Hanfling6 Even though the gradual increase of cases allows governments to progressively enact mitigation strategies and countermeasures to face an incoming wave, the sustained spread of the disease, along with a possible delayed recovery of patients, can easily lead to health-care capacity saturation and surge capability exhaustion. Reference Kelen and Mccarthy7 In particular, the COVID-19 pandemic has been posing unique challenges to emergency medical services (EMS), emergency departments (ED), and ICUs, in terms of stuff, staff, and structures. Italian ICUs are currently experiencing an unprecedented and sustained demand that is depleting resources on a local, regional, and national levels, Reference Grasselli, Pesenti and Cecconi8 thus possibly hampering further surge in capacity if a new wave of cases should occur. To forecast the demand for hospital ICU and non-ICU beds for COVID-19 patients from May to September, we developed 2 models: (1) assuming gradual easing of restrictions, (2) assuming intermittent lockdown.

METHODS

Model Definition

We used a compartmental model to predict hospital demand in Italy associated with the COVID-19 pandemic. A compartmental model divides individuals into groups (or compartments) based on their status and assumes that each individual can only be in a single compartment at a given time. In addition, over the time, an individual can transition between compartments accordingly to transition rates. Reference Blackwood and Childs9 The model was implemented using R software, and an overview of it is provided in Supplemental Figure 1 and Table 1. The input to the model is the number of infected individuals in each day of the simulation. Observed numbers of infected individuals were used until April 17 (last available observation). From April 18 onward, the future number of infected individuals under different scenarios (see hereafter) was calculated using the method proposed by Tuite and colleagues. Reference Tuite and Fisman10 Based on the number of infected individuals, the model provided the daily number of ICU and non-ICU patients, dead, and recovered. To keep the models as simple as possible, we made a series of assumptions: (i) recovered individuals remained immune from re-infection for the duration of the pandemic; (ii) individuals stopped being infectious once they were admitted to hospital (ie, we did not model transmission within health-care settings).

Calibration and Validation of the Model

The model was calibrated using the current number of infected, ICU and non-ICU patients, dead, and recovered in Italy from February 24 to March 24, which represents the training data set. These figures were obtained from the website of the Italian Civil Protection. 2 We used the R function optim() to perform a multidimensional optimization of the model, selecting the parameters’ values that minimized the mean square error between fitted and observed number of individuals in each compartment.

In the final model, mean residence times in the different compartments were reasonably consistent with those reported in the scientific literature and those observed in our University Hospital (Maggiore Hospital, Novara). Once the parameters were determined using the training data set, we evaluated the predictive accuracy of the model using the validation data set, comparing the predicted number of subjects in each compartment with the actual figures observed between March 25 and April 17. Supplemental Figure 2 shows that predicted number of subjects in the different compartments were very close to the observed ones, suggesting a satisfactory predictive accuracy of the models.

Scenario Analysis

We then used our model to forecast the Italian hospital demand until September 1, 2020, under different scenarios:

In Scenario A (“intermittent lockdown”), we first assumed that from April 18 to April 30 the evolution of the pandemic follows the same trend as in the previous weeks, with a steady reduction of new cases of infection over time equivalent to an effective reproductive number (Rt) of 0.9. Then, we hypothesized that the lockdown is temporarily lifted from May 1 to May 30. Starting from May 31, a new lockdown is enforced until the end of the simulation (September 1), bringing Rt to its original value (0.9). Consequently, we assumed that 2 weeks after the lift, the number of new cases starts to increase again, in a similar manner (Rt = 1.3) as it was observed in Italy in the period before the lockdown entered into effect. A lag time of 2 wk was included to take into account the COVID-19 incubation period and the diagnostic delay after the onset of symptoms. 11

In Scenario B (“gradual relaxation of the lockdown”) we assumed that from May 1 onward, the restrictive measures are progressively reduced over time. Hence, we increased Rt by 0.1 every 30 days, up to the value of 1.3. Assumed Rt values from May 1 to September 1 are reported in Supplemental Figure 3.

We then evaluated changes in ICU and non-ICU demand under the 2 scenarios. ICU and non-ICU needs were compared with the maximum hospital bed use for COVID-19 observed before April 17.

To evaluate the uncertainty of our forecasts, we ran Monte Carlo simulations sampling the values of the parameters from uniform distributions (from -30% to +30% of each parameter’s estimated value). We generated 95% confidence intervals (CIs) by taking the 2.5% and 97.5% percentile estimates from 100 simulations.

Ethics Committee Approval

The study was based on publicly available aggregate data. No ethics committee approval was necessary.

RESULTS

Scenario Analysis

Under scenario A (“intermittent lockdown”), the assumed increase in the number of infected is predicted to translate into a rise in the demand of ICU and non-ICU beds at the beginning of June (Figure 1). The maximum demand of ICU and non-ICU beds will occur in the first weeks of July. While ICU needs will remain below the peak levels observed in April 2020 (61%; 95% CI: 40-81), the number of non-ICU demands will substantially rise and will exceed the maximum demand recorded in the early phase of the pandemic (133%; 95%CI: 94 to 171). Starting from the second part of July, bed demand will start to decrease, but non-ICU needs will still remain high until the end of August.

FIGURE 1 Observed (Circles) and Predicted (Solid Line) Number of Infected, Patients at Home, Non-ICU Hospitalized Patients, and ICU Patients Over Time. Scenario A (intermittent lockdown).

Under scenario B (“gradual relaxation of the lockdown”), a rise in the demand of ICU and non-ICU beds will start to be evident in July and will progressively increase over the summer (Figure 2). At the end of August, ICU and non-ICU demand will be 95% (95% CI: 71-121) and 237% (95% CI: 191-282) of the April peak. Differently from the previous scenario, no reduction in both ICU and non-ICU demand is predicted during the time frame covered by the simulation.

FIGURE 2 Observed (Circles) and Predicted (Solid Line) Number of Infected, Patients at Home, Non-ICU Hospitalized Patients, and ICU Patients Over Time. Scenario B (gradual relaxation of the lockdown).

DISCUSSION

In this study, we showed that an early reduction of the community containment measures currently in place in Italy could shortly translate into a substantial increase of ICU and non-ICU demand. In particular, a gradual lift of the lockdown (which is currently under discussion by the Italian government) is expected to double non-ICU admissions and to bring the number of ICU needs to the same levels observed in the early phase of the pandemic.

Our results indicate that the Italian health system should focus efforts to prepare for the next phase of the COVID-19 pandemic. Within the hospital system, critical care services are fundamental during epidemics of infectious respiratory diseases, Reference Hick, Einav and Hanfling6 as repeatedly demonstrated during H1N1, severe acute respiratory syndrome (SARS), or Middle East respiratory syndrome (MERS) outbreaks. During the early stage of an epidemic, a phased response can be useful to gradually increase treatment capacity while preserving critical care resilience and preventing the collapse of the system. Similarly, a phased deactivation of resources is of paramount importance during the recovery stage of epidemics, because the risk of new outbreaks is consistent 12 and the scarce, remaining assets should be properly managed.

Our analysis also shows that non-ICU demand could become a rate-determining step for the health systems, because of the length of stay in the wards and the step-down of patients from the ICU. The former could be partly due to the commonly used discharge criteria for COVID-19 cases. The European Centre for Disease Prevention and Control (ECDC) currently recommends clinical resolution and 2 negative tests for discharge. 13 However, ECDC recognizes that, in the context of sustained widespread transmission (as it is currently the case in Italy), hospital discharge should be also based on other factors, such as the existing capacity of the health-care system, laboratory diagnostic resources, and the current epidemiological situation. 13 In particular, the earlier discharge from hospital of mild cases, if clinically appropriate, may be considered, provided that they are placed into home care or another type of community care and periodically evaluated. 13 Telemedicine could be also useful to foster early discharge of patients, thus helping to keep non-ICU demand under control.

Our model is based on several assumptions and has limitations. In particular, we assumed that the trend of non-ICU and ICU admission rates in the next months will remain similar to what we observed so far. Thus, for example, an improvement in the community-based management of COVID-19 patients would translate into a reduction of hospital demand. Moreover, we did not take into account the effect that the gradual depletion of susceptible persons from the population would have on our estimates. However, considering that the prevalence of infected persons is still low in Italy (<10%), this simplification should not have substantially affected our results. Reference Signorelli, Scognamiglio and Odone14

In conclusion, our results suggest that Italian hospital demand is likely to remain high in the next months if restrictions are reduced, which seems likely to occur. Given the cuts recently suffered by the Italian National Health System, Reference Armocida, Formenti and Ussai15 planning for the next few months should consider an increase in health-care resources to maintain surge capacity across the country. The available assets should be deployed to the most struggling parts of the country with a certain grade of flexibility over time, taking also into account the immunity status of the population.

Supplementary material

To view supplementary material for this article, please visit https://doi.org/10.1017/dmp.2020.157

Author Contributions

J.B., C.R.N., and F.B.A. conceived and designed the experiments. J.B., C.R.N., and F.B.A. analyzed the data. J.B., M.P., C.R.N., N.D.V., S.V., F.D.C., L.R., and F.B.A. wrote the study.

All authors approved the final version of the manuscript.

References

World Health Organization. Coronavirus disease 2019 (COVID-19) situation report – 91. https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports/. Accessed April 21, 2020.Google Scholar
Italian Civil Protection. Coronavirus disease 2019 (COVID-19) Italian situation map. http://opendatadpc.maps.arcgis.com/apps/opsdashboard/index.html#/b0c68bce2cce478eaac82fe38d4138b1. Accessed April 21, 2020.Google Scholar
World Health Organization – Regional Office for Europe. COVID-19 situation in the WHO European Region, data as of: 21 April 2020, 10:00 (CET). https://who.maps.arcgis.com/apps/opsdashboard/index.html#/ead3c6475654481ca51c248d52ab9c61. Accessed April 21, 2020.Google Scholar
The Prime Minister of the Italian Republic – Decree March 9th, 2020. Urgent Measures of Containment and Management of Epidemiological Emergency COVID-19, further dispositions. https://www.gazzettaufficiale.it/eli/id/2020/03/09/20A01558/sg. Accessed April 13, 2020.Google Scholar
Xu, S, Li, Y. Beware of the second wave of COVID-19. Lancet. 2020;395(10233):1321-1322. doi: 10.1016/S0140-6736(20)30845-X CrossRefGoogle ScholarPubMed
Hick, JL, Einav, S, Hanfling, D, et al. Surge capacity principles: care of the critically ill and injured during pandemics and disasters: CHEST consensus statement. Chest. 2014;146(4 Suppl):e1S-e16S.CrossRefGoogle ScholarPubMed
Kelen, GD, Mccarthy, ML. The science of surge. Acad Emerg Med. 2006;13(11):1089-1094. doi: 10.1197/j.aem.2006.07.016 CrossRefGoogle ScholarPubMed
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. doi: 10.1001/jama.2020.4031 Google Scholar
Blackwood, JC, Childs, LM. An introduction to compartmental modeling for the budding infectious disease modeler. Lett Biomath. 2018;5(1):195-221. doi: 10.1080/23737867.2018.1509026 CrossRefGoogle Scholar
Tuite, AR, Fisman, DN. Reporting, epidemic growth, and reproduction numbers for the 2019 novel coronavirus (2019-nCoV) epidemic. Ann Intern Med. 2020;172(8):567-568. doi: 10.7326/M20-0358 CrossRefGoogle ScholarPubMed
World Health Organization. Report of the WHO-China joint mission on coronavirus disease 2019 (COVID-19). https://www.who.int/docs/default-source/coronaviruse/who-china-joint-mission-on-covid-19-final-report.pdf. Accessed March 29, 2020.Google Scholar
US Department of Homeland Security. Pandemic influenza preparedness, response, and recovery guide for critical infrastructure and key resources. September 19, 2006. https://training.fema.gov/programs/emischool/el361toolkit/assets/cikr_pandemicinfluenzaguide.pdf. Accessed April 20, 2020.Google Scholar
European Centre for Disease Prevention and Control (ECDC). Discharge criteria for confirmed COVID-19 cases – When is it safe to discharge COVID-19 cases from the hospital or end home isolation? https://www.ecdc.europa.eu/sites/default/files/documents/COVID-19-Discharge-criteria.pdf. Accessed April 19, 2020.Google Scholar
Signorelli, C, Scognamiglio, T, Odone, A. COVID-19 in Italy: impact of containment measures and prevalence estimates of infection in the general population. Acta Biomed. 2020;91(3-S):175-179. doi: 10.23750/abm.v91i3-S.9511 Google ScholarPubMed
Armocida, B, Formenti, B, Ussai, S, et al. The Italian health system and the COVID-19 challenge. Lancet Public Health. 2020;5(5):e253. doi: 10.1016/S2468-2667(20)30074-8 CrossRefGoogle ScholarPubMed
Figure 0

FIGURE 1 Observed (Circles) and Predicted (Solid Line) Number of Infected, Patients at Home, Non-ICU Hospitalized Patients, and ICU Patients Over Time. Scenario A (intermittent lockdown).

Figure 1

FIGURE 2 Observed (Circles) and Predicted (Solid Line) Number of Infected, Patients at Home, Non-ICU Hospitalized Patients, and ICU Patients Over Time. Scenario B (gradual relaxation of the lockdown).

Supplementary material: File

Bollon et al. supplementary material

Figure S2

Download Bollon et al. supplementary material(File)
File 30 KB
Supplementary material: File

Bollon et al. supplementary material

Table S1

Download Bollon et al. supplementary material(File)
File 9 KB
Supplementary material: File

Bollon et al. supplementary material

Figure S3

Download Bollon et al. supplementary material(File)
File 67 KB
Supplementary material: File

Bollon et al. supplementary material

Figure S1

Download Bollon et al. supplementary material(File)
File 58 KB
You have Access
Open access
3
Cited by

Send article to Kindle

To send this article to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

Note you can select to send to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Predicted Effects of Stopping COVID-19 Lockdown on Italian Hospital Demand
Available formats
×

Send article to Dropbox

To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

Predicted Effects of Stopping COVID-19 Lockdown on Italian Hospital Demand
Available formats
×

Send article to Google Drive

To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

Predicted Effects of Stopping COVID-19 Lockdown on Italian Hospital Demand
Available formats
×
×

Reply to: Submit a response

Please enter your response.

Your details

Please enter a valid email address.

Conflicting interests

Do you have any conflicting interests? *