key: cord-1030267-kjxk7mno authors: Guzzi, P. H.; Tradigo, G.; Veltri, P. title: Intensive Care Unit Resource Planning During COVID-19 Emergency at the Regional Level: the Italian case. date: 2020-03-20 journal: nan DOI: 10.1101/2020.03.17.20037788 sha: 63a39a785f7ad59953d2b48a5c384592ef266987 doc_id: 1030267 cord_uid: kjxk7mno Severe acute respiratory syndrome Covid-19 (SARS-CoV-2) has been declared a worldwide emergency and a pandemic disease by the World Health Organisation (WHO). It started in China in December 2019, and it is currently rapidly spreading throughout Italy, which is now the most affected country after China. There is great attention for the diffusion and evolution of the CoVid-19 infection which started from the north (particularly in the Lombardia region) and it is now rapidly affecting other Italian regions. We investigate on the impact of patients hospitalisation in Intensive Care Units (ICUs) at a regional and subregional granularity. We propose a model derived from well-known models in epidemic to estimate the needed number of places in intensive care units. The model will help decision-makers to plan resources in the short and medium-term in order to guarantee appropriate treatments to all patients needing it. We analyse Italian data at regional level up to March 15th aiming to: (i) support health and government decision-makers in gathering rapid and efficient decisions on increasing health structures capacities (in terms of ICU slots) and (ii) define a scalable geographic model to plan emergency and future CoVid-19 patients management using reallocating them among health structures. Finally, the here proposed model can be useful in countries where CoVid-19 is not yet strongly diffused. home quarantine for low symptoms, (ii) hospitalisation for part of them and, 23 (iii) hospitalisation in ICUs requiring respiratory support for severe ones. 24 From a clinical point of view, patients affected by COVID-19 usually 25 present symptoms similar to common flu, e.g. fever, dry cough, sickness, and 26 breathing problems. A fraction of patients do not need hospitalisation, and 27 the symptoms disappear in a short time (around 1-2 weeks). In some cases, 28 COVID-19 causes severe pneumonia, which requires respiratory support and 29 can lead to death, especially in the presence of co-morbidities such as diabetes 30 or hypertension [3] . Patients with severe pneumonia need to be treated in 31 ICUs with the use of mechanical ventilators [1] . 32 We focus on disease diffusion modelling at a regional scale in Italy and 33 3 . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.03.17.20037788 doi: medRxiv preprint also modelling ICUs Availability throughout Italian health structures. We 34 propose a prediction model at a regional scale that can be adapted and also 35 used at different scales (i.e., countries, regions or even single district) by 36 countries where COVID-19 diffusion is still growing. 2 Data analysis and model 38 We start from the analysis of epidemiological data from Wuhan city (China, which, by law, are designed to be occupied by patients for the 80% at any 60 given time. Also, these are allocated at a regional level proportionally to its 61 population and are usually managed locally (see Table 1 ). Beds of new ICU beds, the movement of people from one region to another (to 69 5 . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.03.17.20037788 doi: medRxiv preprint Figure 1 : The Compartmental Model. Our model start by considering infected people (X 0 ). A fraction of infected people presents severe symptoms and they need to be hospitalised and treated in ICUs (X 1 ). Diversely, some people may be treated at home (X 2 ) since they do not have severe complications. Treated people has a lethal outcome (X 3 ) while a hopefully large fraction of people is recovered from disease. . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.03.17.20037788 doi: medRxiv preprint Figure 2 : Here the two Italian and Chinese red zones (areas of maximal infection) are compared. On the X axis there are days and on the Y axis there are the total number of cases. The two curves are very similar showing that the initial trend of the infection follows an exponential growth, even though the Chinese government rapidly adopted stringent confinement measures. We can thus expect to observe the same initial infection evolution, before arriving to the logistic portion of the curve. . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint The Italian National Health Service is organised on a national and regional 96 scale. The central government controls the distribution of resources and 97 services are arranged at a regional scale. There are 19 regions and two 98 autonomous provinces (21 total administrative units). Therefore the ICUs is 99 availability is organised at a regional scale, established by each region. Table 100 1 summarise current ICU beds availability per administrative units. Patients Italy where restriction plans to reduce citizen mobility has been applied at 108 the regional scale. Figure 3 shows the distribution of total ICU beds versus 109 occupied ICU beds (i.e. in hospitals) for each region in Italy. Figure 3 shows 110 the infected cases for each region. 111 We compare through our model the management of beds in single regions 112 9 . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.03.17.20037788 doi: medRxiv preprint 14 . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.03.17.20037788 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . with respect to ICU beds capacity (i.e. non saturated regions, see Figure 3 ). 143 We propose to use the predicting model for these regions to early predict representation. In such case note that the government restrictions rapidly 151 adopted allow a slower diffusion of infectious. Note that in a similar way we map all data for all the 21 Italian regions. . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.03.17.20037788 doi: medRxiv preprint The emergency of COVID-19 is related to an aggressive virus that diffuses 155 rapidly and strongly stresses the resistance of health structures. We also 156 think that patients management is strictly related to the availability of hos-157 pital structure to manage such kind of diseases that require different and 158 non-standard protocols as use of respiratory devices. We think that by using 159 a scalable predictive model at regional as well as at district granularity may 160 support regional and national government in managing the emergency. We 18 . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.03.17.20037788 doi: medRxiv preprint Preparing for covid-19: early experience from an intensive care unit 201 in singapore Genomic char-204 acterisation and epidemiology of 2019 novel coronavirus: implications 205 for virus origins and receptor binding The coron-208 avirus 2019-ncov epidemic: Is hindsight 20/20? EClinicalMedicine COVID-19 and italy: what 211 next? The Lancet Covid-19: A new virus as a poten-214 tial rapidly spreading in the worldwide