Abstract
Objectives Coronavirus disease (COVID-19) caused by the SARS-CoV-2 virus is spreading rapidly worldwise and threatening the collapse of national health care systems. The development of effective resource models are critical for long term health planning. The aim was to evaluate the available literature, to consider parameters affecting hospital resources, to effectively guide health policy and planning.
Design A detailed search of the literature, using Google Scholar, PubMED, MedRxiv and BioRxiv, was conducted for the time period 01-DEC-2019 to 04-MAY-2020; using appropriate keywords: resultant articles were scrutinised in detail, and appraised for reported data pertaining to hospitalization and hospital length of stay (LOS).
Results Disease presentation was described in China; 81 % mild, 14 % moderate and 5 % severe. The experience, thus far, in Europe and the USA are suggestive of a higher degree of severity. Initial reports suggest high hospitalisation and ICU admittance rates. More recent reports from the ECDC lower this estimation. Perhaps the relative age, the level of pre-existing conditions, and other health factors may be contributors to differences. Data from Irish cases suggest hospitalisation rate may be lower in parts of Europe and time dependent. Hospital LOS is described in seventeen articles, with median lengths of stay between 4 and 25 days. The evidence regarding the LOS in ICU is reported in eighteen studies, fifteen deemed relevant. The majority of studies report ICU LOS between 7 to 10 days. Many of these studies are likely skewed towards shorter stay due to study cut-off dates. Indications based on ICU LOS reported for patients continuing care suggest median ICU stay will be longer.
Conclusions These parameter estimates are key to the development of an effective health care resource model. Based on our rapid appraisal of the literature, is it essential that Europe manages mitigation measures to ensure that hospital and ICU capacity does not become overwhelmed to manage this pandemic long term.
1. INTRODUCTION
Coronavirus disease (COVID-19) caused by the SARS-CoV-2 virus is spreading rapidly across the world, threatening the collapse of many national health care systems and exerting a grave impact on the global economy. As of 9th May 2020, data compiled from the European Centre for Disease Prevention and Control (ECDC) reported 3,898,658 confirmed infections of COVID-19 cases worldwide with 274,290 deaths recorded. The SARS-CoV-2 virus, first detected in December 2019, in Wuhan China, has spread to 212 countries and territories. By 9th May 2020, European countries, including the European Union (EU)-27, the UK, Norway, Switzerland and Iceland have reported 1,259,681 cases and 147,085 deaths following infection with SARS-CoV-2 infection (ECDC, 2020).
Clinical presentations of COVID-19 range from asymptomatic to life threatening and fatal (World Health Organisation (WHO), 2020); Wu and McGoogan (2020) outlined the spectrum of disease of COVID-19 in China; 81 % were mild; 14 % severe and 5 % critical. The experience in European countries has differed, and varied among countries. The eight update Report from the ECDC (2020a) indicates that up to 30 % of diagnosed COVID-19 cases were hospitalised in some countries. Individual studies from Italy (Cereda et al., 2020; Grasselli et al., 2020) indicated a higher proprotion were accessing ICU care, with 16 and 18 % of all COVID-19 patients being cared for in an ICU setting. However, the European wide analysis report (ECDC, 2020a) suggests severe illness requiring respiratory support or ICU care has accounted for 2.4 % cases, the most recent report; the ninth update (ECDC, 2020b) lowers this esimtate (median 2 %).
Worldwide, age and co-morbidities have been demonstrated as major risk factors for the development of severe disease (Chen et al., 2020; ECDC, 2020a; 2020b; COVID-19 Surveillance Group, 2020; Gang et al., 2020; Paranjpe et al., 2020; Qian et al., 2020; Richardson et al., 2020; Russell et al., 2020; Wu and McGoogan, 2020). Children are implicated as possible asymptomatic carriers of SARS-CoV-2 virus and a number of articles report both symptomatic and asymptomatic infection in children (Chan et al., 2020; Hu et al., 2020; Korean Society of Infectious Diseases et al., 2020; Liao et al., 2020; Qiu et al., 2020; Tao et al., 2020; Wu and McGoogan 2020). Many, but not all (Korean Society of Infectious Diseases 2020), studies indicate a higher proportion of cases are male (Chen et al., 2020; ECDC, 2020a; 2020b; Fan et al., 2020; Linton et al., 2020; COVID-19 Surveillance Group, 2020; ICNARC, 2020a; 2020b); and recent risk assessments conducted by the ECDC (ECDC, 2020a; 2020b) indicates that males are at higher risk of death compared with their female contempories. Close contact with a case is a risk factor (WHO, 2020), and the basis of public health advice includes social distancing and staying at home apart from essential activity (ECDC, 2020a). Bi et al. (2020) reports that household contacts (Odds Ratio 6) and those travelling with a case were at higher risk of infection (OR 7) than other close contacts. Other factors, including smoking (Vardavas and Nikitara, 2020) have been highlighted as risk factors for severe infection. There are indications too that infection rate may be altered by blood-group (Fan et al., 2020; Zhao et al., 2020).
The prevention of infection among health care workers is of critical importance to the management of this pandemic. Of serious concern to policymakers is the clear evidence indicating that both healthcare workers and hospitalised persons are at particularly high risk of exposure (WHO 2020; Istituto Superiore di Sanità 2020). Indeed some of the work conducted early in the epidemic reported hospital-associated transmission of virus was suspected in 41% of patients (Wang et al., 2020; n= 138). Of concern too, many healthcare authorities are bringing retired medics and nurses back into service. The placement of these people considered vulnerable due to their age must be considered. At a national level, the latest work by Stier et al. (2020) emphasizes population density, with attack rates increasing with city size.
Preparedness is key and resource matching to clinical requirement essential. Previous work (Rhodes et al., 2012) suggests a large variation in availability of critical care bed in the EU / EEA area, ranging from 29.2 in Germany to 4.2 critical care beds per 100,000 population in Portugal. Perhaps this is key to understanding the difference in coping between different EU countries. Clearly too, differences will exist between cultures, few western governments will seek to use hospitalisation as a tool for quarantine. As the infection is spreading across the world; this rapid review paper considers parameters affecting hospital resources resulting from infection with SARS-CoV-2 beyond the evidence published from China. The objective is to evaluate the current scientific literature from across the developed world with the overall aim to influence health policy and support evidence-based decision-making. The proportion of patients hospitalized due to infection with SARS-CoV-2 infection is discussed. The proportion of patients who required critical or ICU care is outlined. Time spent hospitalised; both in a general and an ICU setting is considered. For those who have been admitted to an ICU, the length of time pre ICU, the length of time in ICU and the length of time post ICU to discharge or death was collated from available literature. These estimates are central to national resource modelling, which are being developed to aid policymakers to ensure optimal resourcing during this pandemic.
2. MATERIALS AND METHODS
2.1 Initial study appraisal
This paper follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses—Extension for Scoping Reviews (PRISMA-ScR) checklist (Tricco et al., 2018). A comprehensive search of the scientific literature and national and international government reports was conducted by contributing researchers for the time period 1st December 2019 to 4th May 2020. The search engines Google Scholar, PubMED, MedRxiv and BioRxiv were used, using the following keywords: (“Novel coronavirus” OR “SARS-CoV-2” OR “2019-nCoV” OR “COVID-19”) AND (“length of stay” OR “duration of stay” OR “hospital stay” OR “ICU”). No restriction on language was imposed as long as the abstract was available in English. References within these publications were also searched as additional possibilities for inclusion. Inclusion criteria for ‘the proportion of cases hospitalised’ were papers that reported the proportion of hospitalised and proportion of those hospitalised accessing higher level care including critical care and intensive care unit (ICU) care. Inclusion criteria for ‘length of stay’ (LOS) included articles that reported length of hospital stay, length of stay in ICU, length of stay prior to admission to ICU, and articles that referenced length of stay post discharge from ICU to a general hospital setting. Each article was assessed briefly for suitability for inclusion in this study. Papers that did not report length or duration of time in hospital or in ICU were discarded. The latest governmental reports from the Health Protection Surveillance Centre (HPSC), Ireland, and from the Intensive Care National Audit and Research Centre (ICNARC, 2020a), UK were consulted after the 4th May cut-off.
2.2 Study appraisal
All articles and reports that described the proportion of COVID-19 positive cases hospitalised, the proportion of cases that were admitted to higher level care including critical or ICU care, the length of hospital stay or the duration of hospital stay or length of time in critical care or ICU were scrutinised in detail, and appraised for reported data and quality of the scientific evidence. Parameter estimates for the length of stay in non-ICU or in an ICU setting from the relevant articles were recorded and evaluated.
For quality control, studies were (1) selected from search terms outlined above and initially screened by three members of the team (KH, SM and LL), with parameters identified and recorded; (2) reviewed and supplemented by manual search by LL, again with parameters identified and recorded. (iii) Finally, the review was then internally reviewed by an additional three members of the team (SM, DB, MC), and cross-referenced with other parameter synthesis documents being developed by the group (all authors). Details of the resultant list (n = 179), which included published and pre-printed articles, were entered in an Excel file.
Eighteen studies relating to hospital LOS and fifteen studies or reports relating to LOS in ICU were selected for detailed scrutiny as they each reported summary estimates with a corresponding variance. The Italian report (COVID-19 Surveillance Group, 2020) was the only report without an estimate of variance that was considered for detailed scrutiny, as the study population was significant (n = 25,452).
2.3 Source data and management
Data tabulating the daily number of notified cases and deaths for all countries and Territories worldwide was downloaded in Excel format from the European Centre for Disease Prevention and Control (ECDC) and may be accessed at: https://www.ecdc.europa.eu/en/publications-data/download-todays-data-geographic-distribution-covid-19-cases-worldwide. Data was imported into STATA version 14 (STATA Corporation, USA). Cumulative cases and deaths were generated for each calendar day for countries of interest, and expressed as a percentage of cumulative deaths to cumulative cases. The crude mortality rate was expressed as a percentage over time, from day of outbreak, defined as the day on which the tenth case was reported, in selected European Countries.
Data regarding the epidemiological situation in Ireland was extracted from the Health Protection Surveillance Centre daily epidemiological reports. These may be accessed a: https://www.hpsc.ie/a-z/respiratory/coronavirus/novelcoronavirus/casesinireland/epidemiologyofcovid-19inireland/. Data were entered into Excel, then imported into STATA 14 to generate the trends in cumulative cases, percentage of known cases hospitalised and the percentage of confirmed cases admittance into ICU over time.
3. RESULTS
3.1 HOSPITALISATION RATE
A summary of available scientific literature from the outbreaks worldwide relating to the proportion of COVID-19 cases requiring hospitalisation and the need for stepped up medical care, including critical care and ICU, by country, is presented in Table 1. Data from 33 studies or reports are presented.
A summary of published articles and reports relating to hospitalisation following COVID-19 infection, by country.
European Centre for Disease Prevention and Control (ECDC updates)
The European Centre for Disease Prevention and Control (ECDC) collates the laboratory-confirmed cases reported to the European Surveillance System (TESSy) and thus far has published nine update reports as the situation unfolds across Europe (ECDC, 2020a; 2020b). Preliminary estimates of severity were collated based on the analysis of data from EU and EEA countries and from the UK. As expected some data were incomplete or missing in TESSy. The eight update (ECDC, 2020a) indicates that 48,755 of 152,375 (32 %) were hospitalised based on data from 26 European countries; with a country median 28 % (IQR 14 to 63 %). Severe presentations that required ICU and or respiratory support accounted for 2,859 of 120,788 (2.4 %) cases from 16 countries; median 1.4 % (IQR 0 to 33 %). Among hospitalized cases only, severe illness was reported in 3,567 of 38,960 (9.2 %) cases from 19 countries, median 15 % (IQR 3.8 to 35 %). Death occurred in 1,005 of 9,368 (11 %) of hospitalized cases from 21 countries (3.9 %, IQR 0 to 13 %). Age-specific risk increased among those 60 years and older, thus consideration to age profile within a country is important. The latest report on 23rd April reported increased rate of hospitalisation with a pooled estimate of 42 % (160,485 of 381,410 cases); based on data from 19 countries, however the country median was 28 % (IQR 16 to 39 %). Severe presentations that required ICU and or respiratory support accounted for 5,456 of 220,412 (2 %) cases with data from 14 countries; median 2 % (IQR 0 to 4 %). Again, hospitalisation increased with age, and for males from 30 years. The changes over time in the crude mortality rate for all notified cases since day of outbreak, defined as the day on which the tenth case was notified, for selected European countries is presented in Figure 1.
The crude mortality rate (CMR) expressed as the percentage of notified deaths to cases over time, from number of days lapsed since the notification of the tenth case for 12 selected European countries. Data tabulating the notified cases and deaths from all countries is available from the European Centre for Disease Control and Prevention; https://www.ecdc.europa.eu/en/publications-data/download-todays-data-geographic-distribution-covid-19-cases-worldwide
USA Studies
Admission rates to ICU in the USA varied between 4.9 to 11.5 % (CDC COVID-19 Response Team, 2020) report. Myers et al. (2020) indicated 8.7 % ICU admission in a study conducted in California in March. While Lewnard et al., (2020) observed a higher ICU admission rate (25.6 %) also in California. A number of studies conducted in New York outlined ICU admission rates of 14.2 % (Richards et al., 2020); 22 % (Cummings et al., 2020); 23.6 % (Argenizano et al., 2020); and 36% (Paranjpe et al., 2020). Of significance, are the mortality rates in hospitalised groups (and specifically not ICU only groups) reported in a number of US studies: 15.4 % Lewnard et al., (2020); 15.6 % Myers et al., 2020; 21 %; Richards et al., 2020; 21 %; Argenziano et al., 2020; and 29 % Paranjpe et al 2020.
In Ireland, the proportion of COVID-19 test positives cases (confirmed cases) hospitalised is reducing with time (Figure 2). Daily epidemiology reports compiled by the Health Protection Surveillance Centre (HPSC, 2020a) with data to March 30th reports a hospitalisation rate of 27.9 % (834 of 2,990 cases hospitalised) and admission to ICU in 4.2 % of cases (126 of 2,990). The hospitalisation rate reduced over time to 17.5 % (1,968 of 11,261) by April 13th (HPSC, 2020b); 13.5 % (2,669 of 9,723) by 29nd April (HPSC, 2020c); and 13.1 (2,878 of 21,908 cases by 4th May (HPSC, 2020d). The rate of admission to ICU reduced, concomitantly, from 4.2% of cases (126 of 2,990; HPSC 2020a) to 1.7 % (373 of 21,908; HPSC 2020d) of cases by May 4th. The reports collated by the HPSC (HPSC Epi Team, 2020) in Ireland indicate that proportion of hospitalisations and transfer to ICU increased occurred with age.
The cumulative number of notified cases (grey bars) in Ireland between 30th March and 23rd April; the percentage of COVID positive cases that were hospitalised (black line); and those that were admitted to ICU (maroon line) for the same period. Data extracted from publicly available data published by the Health Protection Surveillance Centre (HPSC), Ireland.
3.2 LENGTH OF STAY (LOS)
3.2.1 Length of stay in hospital (Non-ICU)
A summary of the literature pertaining to length of hospital stay in non-ICU settings due to COVID-19, by country, is reported in Table 2. The overall length of hospital stay is reported in eighteen studies or reports, with overall median lengths of stay between four days and twenty-five days. Figure 3 illustrates the median (and IQR) ranges of hospital length of stay for relevant articles.
A summary of published articles and reports relating to length of stay in an non ICU settings following COVID-19 infection, by country.
The median (maroon square) and 25th (blue diamond) and 75th percentile (green diamond) for eight relevant articles and reports describing the median (and IQR) length of stay in hospital due to COVID-19 infection. Two studies (Linton et al., 2020; Yuan et al., 2020) reported means (maroon circle) and confidence intervals (diamonds); while one study (Liu et al., 2020) reported median (square) and range (X) for length of stay following COVID-19 infection. The Italian report (COVID-19 Group, 2020) indicates a median of 5 days; no estimate of variance was reported. A subgroup of the Valente-Acosta (2020) article is greyed, as only 4 remained in the group.
3.2.2 Length of stay in ICU
The literature pertaining to length of ICU stay due to COVID-19 is reported in Table 3, including eighteen studies and reports. Figure 4 illustrates the median (and IQR) ranges of length of stay in ICU for relevant articles. In this figure, fifteen studies are presented; Three case studies were excluded due to small numbers of relevant cases (Chen et al., 2020; Kujawski et al., 2020; Young et al., 2020). Median lengths of stay in ICU between 7 to 10 days were observed in studies based in China (Cao et al., 2020; Fan et al., 2020; Yang et al., 20202; Zhou et al., 2020). European studies conducted by Graselli et al. (2020) in Italy, Ceruti et al (2020) in Switzerland, and the ICNARC report (2020a) report similar LOS between 7 to 10 days. Earlier reports from the ICNARC (2020b) indicated shorter LOS for ICU patients, indicating the bias towards shorter stay as fewer patients had progress through their complete stay. However median LOS outlined by studies based in North America were more variable and reported lengths of stay between 3.5 to 23 days for various groups (Figure 3).
The median (blue diamond) and 25th (maroon square) and 75th percentile (green diamond) for eight relevant articles and reports describing the median (and IQR) length of stay in an ICU due to COVID-19 infection. The Italian report (COVID-19 Group, 2020) indicates a median of 8 days (including ICU stay for deceased patients); no estimate of variance was reported. The UK report separates survivors from those who have died in care, the authors also indicate bias towards shorter lengths of stay as this cohort has completed their ICU stay.
A summary of published articles and reports relating to length of stay in an ICU setting following COVID-19 infection, by country.
3.2.3 Length of hospital stay prior to transfer to ICU
Length of hospital stay prior to transfer to ICU due to COVID-19, by country, is presented in Table 4. Four studies made reference to time in hospital prior to ICU transfer, including two small case series (Young et al., 2020; Kujawski et al., 2020). A long median hospital stay prior to transfer to ICU was described by Fan et al. (2020; median 8 days, IQR 4 to 11 days). The UK ICNARC report (ICNARC 2020a) details a median length of stay of one day (IQR 0 to 3 days); indicating the short duration before stepped up care need is identified in COVID-19 patients.
A summary of published articles and reports relating to length of hospital stay prior to transfer to ICU following COVID-19 infection, by country.
3.2.4 Length of hospital post ICU
No eligible studies outlining the length of stay post discharge from ICU were discovered.
4. DISCUSSION
4.1 HOSPITALISATION RATE
The relative variation in disease severity is very clear in the published literature from China; 81 % mild presentation, 14 % moderate presentation with 5 % severe or critical presentations that require high-level on-going care in critical care beds or within the ICU environment (Wu and Grogan, 2020). Nevertheless, the experience, thus far, in Europe and the USA would suggest some countries are experiencing a higher degree of severity of COVID-19 being reported. Early indications from Italy suggested hospitalisation rates of 30 % and admission to ICU between 16 to 18 % (Cereda et al., 2020; Graselli et al., 2020). While the eight update report from the ECDC (ECDC 2020a) restates this but suggests based on data from 16 EU countries that respiratory and or ICU support is needed in 2.4 % of cases; the latest ECDC report (ninth Update; ECDC 2020b) revises this slightly country specific median 2 % (IQR 0 to 4 %). Studies conducted in the USA too report high rates of hospitalisation (CDC COVID-19 Response Team, 2020), high proportion of patients admitted to ICU care and very high mortality rates (Argenziano et al, 2020; Cummings et al., 2020; Lewnard et al., 2020; Myers et al., 2020; Paranjpe et al., 2020). The cause of the differences in disease presentation between countries is not clear. Perhaps differences in case definition, the relative age of the population, the level of pre-existing conditions within the population, and other health factors may be contributing to these differences. Given the experience in Italy, thus far, suggests that must carefully manage mitigation measures to ensure ICU capacity is not reached or breached. However, another major influencer on hospitalisation rate will be the definition of a suspect case and the extent of COVID-19 testing within the community; the detection of a greater number of milder cases, diagnosed within the community will ultimately influence the hospitalisation rate. The influence of community testing is becoming evident in Ireland; as April progresses, it is clear in each subsequent HPSC report (2020a; 2020b; 2020c; 2020d) that the rate of hospitalisation and proportion of patients admitted to ICU is reducing. This is evidence perhaps that a broader case suspect definition and a wider testing approach will impact the overall rate of hospitalisation and admission to ICU.
4.2 LENGTH OF STAY (LOS)
4.2.1 Length of stay in hospital (Non-ICU)
The overall length of hospital stay varied between 4 and 25 days among studies; two of these studies were case series based on the observations of 10 (Lo et al., 2020) or 12 (Kujawski et al., 2020) patients, a further article (Linton et al., 2020) was based on data extracted from publically available datasets. The studies are very diverse in their reporting and presentation of hospital length data. The largest study is an Italian report (Instituto Superiore di Sanità, 2020) reporting the median hospital stay of 5 days for cases that did not require ICU level of care. However, no range or confidence interval is described by this report. Similar to the shorter duration of hospital stay in Italy, a more recent study conducted in the USA, which examined hospital records of 5,700 sequential admissions to New York hospitals, reported a short length of stay, median 4.1 days (IQR 2.3 to 6.8). The hospital length of stay varies dramatically among different studies. Perhaps differences in hospital capacity, overcrowding, government policy, funding source and other factors might influence length of stay. Of interest Richardson et al. (2020) indicates 45 patients were re-admitted to hospital for further care, indicating perhaps the pressure on bed availability. Argenziano et al. (2020) outlines the critical situation faced in some Emergency Departments in New York and describes the death of 14.7 % of ED only patients; patients who died in the emergency room prior to being admitted to hospital. Given the large variation in length of stay and the difference between continents; the studies relating to hospital LOS outside an ICU setting were not appropriate to use to calculate a central tendency.
4.2.2 Length of stay in hospital (ICU)
Studies from China (Cao et al., 2020; Fan et al., 2020; Yang et al., 2020; Zhou et al., 2020) report median ICU stays of 7 to 10 days. Similar median lengths of stay have been reported in Italy. The Italian report (COVID-19 Surveillance Group, 2020) dated 29th April 2020 describes median LOS of 8 days if admitted to an ICU setting (n = 25,452); however, actual length of time in ICU and proportion or number of patients cared for in ICU were not specified. Important too, to note, this report while including a large number of cases, only includes those who have died, and hence will not reflect the possible longer stay of survivors. Another Italian study conducted by Grasselli et al. (2020b) reports the median length of ICU stay was 9 days (IQR 6 to 13 days) for 1,581 patients. A Swiss study, too, outlined the median length of stay was 9 days for 41 patients followed to discharge (Ceruti et al., 2020). It is noticeable that length of ICU stay is more variable in US studies. In particular studies that breakdown length of stay into those completed stay compared to patients continuing care. It is evident that care must be taken with interpretation, continuing care groups are reported to have longer stays; median LOS of 11.4 (Petrilli et al., 2020); 14 (Bhatraju et al., 2020); and 23 days (Argenziano et al., 2020) are outlined. A smaller case series conducted by Valente-Acosta in Mexico highlights this continuing care group; this study describes a case series of 33 patients admitted to ICU in Mexico city, the median LOS of 23 days is indicative that LOS in ICU can be significant; although their continuing care group is only n = 4. The authors of the UK ICNARC (2020a) report clearly highlight the influence of bias towards shorter stay; with both survivor and non-survivor groups reported to have a median LOS of 8 days (ICNARC, 2020a). Indeed earlier reports from INCARC (2020b) reported shorter LOS of ICU patients, as fewer persons had progressed through to completion of treatment. It is evident that study cutoff, and short follow ups, have influenced reported ICU LOS; it is clear that the influence of continuing care groups will lengthen length of ICU stay somewhat in time.
4.2.4 Length of hospital stay prior to transfer to ICU
A long hospital stay (median 8 days, IQR 4 to 11 days) was reported by Fan et al., (2020) prior to transfer to ICU, some of whom were diagnosed based on clinical findings. Given the timing of the study so early in the outbreak and the lack of consistency regarding the diagnosis, these data should be discounted when considering future resources. Only the UK ICNARC report (2020a) details useful data pertaining to length of hospital in Europe. Median length of hospital stay pre admission to ICU for laboratory confirmed COVID-19 patients (n = 8,233) was 1 day (IQR 0 to 3 days), indicating the short duration before stepped up care need is identified in COVID-19 patients.
Length of stay post ICU is illustrated in a figure in some small case series; and as such, were data were not extracted for reporting or discussion in this review.
5. STUDY LIMITATIONS
Limitations associated with any rapid review are pertinent to this study. A narrow aim was set, namely to determine rate of hospitalisation and the length of stay in hospital due to COVID-19 infection in countries across the world now at the centre of the pandemic. The sources of the literature may be limited by the time-limited constraint of gathering relevant literature (less than five weeks). However, this rapid approach was necessary, given the aim was to provide evidence-based estimates of length of hospital stay to steer policy and provide parameter estimates for utilisation within a hospital resource model. Inclusion of articles in this review was assessed based on the subject being reported in the article and if the numbers of cases were considered relevant. Another shortcoming was some of the available literature is in pre-print form and only undergoing peer review.
6. CONCLUSIONS
Differences in the severity of case presentations have been described between continents; and Ireland should base its expectation on those cases observed in Europe thus far in the pandemic. Advice from the ECDC suggests hospitalisation rates of up to 42 % can be expected. Additionally, we must expect increased level of care requirements, to critical care and ICU levels, in up to 3 % of cases to allow for spare capacity, as mitigation measures are relaxed in our jurisdiction. There is no doubt that the case definition in use and the availability of community testing will determine the hospitalisation rate within a country. In countries where community testing has been significant, we can expect the rate of hospitalisation and the proportion requiring ICU level care to be lower. Preliminary data in Ireland certainly supports this assertion.
The length of stay in hospital outside of the ICU setting, due to COVID-19 is highly variable in the literature, with median estimates varying from 4 to 25 days. Studies diverged in their inclusion criterion. It is likely that the aim of hospitalisation varied among countries. Discharge criteria is divergent in different jurisdictions, whether it related to a patient required to test negative prior to discharge or where there was pressure to free up overcrowded hospitals. Indeed it is likely that some countries used hospitalisation to ensure quarantine.
As the pandemic spreads rapidly throughout the world, more and more evidence regarding the length of stay in ICU following COVID-19 is reported. Many of these reports indicate a median ICU length of stay between 7 to 10 days; however a number of these studies report length of stay based only on resolved cases, and as such duration of stay it is likely skewed towards a short stay due to bias towards those who have completed their ICU stay and are now discharged or have died. Thus the upper estimates of the range should be considered for use in resource models. It is apparent too that the data generated by each individual country during the initial months of the outbreak will be critical to estimate the parameters used for any national resource model.
Data Availability
Data can be made available on request.
7. CONFLICT OF INTEREST STATEMENT
The authors do not have a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper.
8. FUNDING
All authors were employed through their home institutions. No additional funding was used.