key: cord-0704553-44rfv4vi authors: Lane, E. A.; Barrett, D. J.; Casey, M.; McAloon, C. G.; Collins, A. B.; Hunt, K.; Byrne, A. W.; McEvoy, D.; Barber, A.; Griffin, J. M.; Wall, P.; More, S. J. title: Country differences in hospitalisation, length of stay and admission to Intensive Care Units due to SARS-CoV-2 infection: a rapid review of available literature date: 2020-05-16 journal: nan DOI: 10.1101/2020.05.12.20099473 sha: ffe12e10b62a3e309c03c6a11f5e5874bb02513a doc_id: 704553 cord_uid: 44rfv4vi 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. latest work by Stier et al. (2020) emphasizes population density, with attack rates increasing 130 with city size. Preparedness is key and resource matching to clinical requirement essential. Previous work 133 (Rhodes et al., 2012) suggests a large variation in availability of critical care bed in the EU / 134 EEA area, ranging from 29.2 in Germany to 4.2 critical care beds per 100,000 population in 135 Portugal. Perhaps this is key to understanding the difference in coping between different EU 136 countries. Clearly too, differences will exist between cultures, few western governments will 137 seek to use hospitalisation as a tool for quarantine. As the infection is spreading across the 138 world; this rapid review paper considers parameters affecting hospital resources resulting CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 4 th May 2020. The search engines Google Scholar, PubMED, MedRxiv and BioRxiv were used, 157 using the following keywords: ("Novel coronavirus" OR "SARS-CoV-2" OR "2019-nCoV" OR 158 "COVID-19") AND ("length of stay" OR "duration of stay" OR "hospital stay" OR "ICU"). No 159 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 CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 screened by three members of the team (KH, SM and LL) Eighteen studies relating to hospital LOS and fifteen studies or reports relating to LOS in ICU 189 were selected for detailed scrutiny as they each reported summary estimates with a 190 corresponding variance. The Italian report (COVID-19 Surveillance Group, 2020) was the only 191 report without an estimate of variance that was considered for detailed scrutiny, as the 192 study population was significant (n = 25,452). CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 Protection Surveillance Centre daily epidemiological reports. These may be accessed a: Preliminary estimates of severity were collated based on the analysis of data from EU and 228 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 230 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 232 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 cases only, severe illness was reported in 3,567 of 38,960 (9.2 %) cases from 19 countries, 234 median 15 % (IQR 3.8 to 35 %). Death occurred in 1,005 of 9,368 (11 %) of hospitalized cases 235 from 21 countries (3.9 %, IQR 0 to 13 %). Age-specific risk increased among those 60 years 236 and older, thus consideration to age profile within a country is important. The latest report 237 on 23 rd April reported increased rate of hospitalisation with a pooled estimate of 42 % . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 In Ireland, the proportion of COVID-19 test positives cases (confirmed cases) hospitalised is 260 reducing with time ( Figure 2 CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 The literature pertaining to length of ICU stay due to COVID-19 is reported in Table 3, 286 including eighteen studies and reports. Figure 4 illustrates the median (and IQR) ranges of 287 length of stay in ICU for relevant articles. In this figure, fifteen studies are presented; Three 288 case studies were excluded due to small numbers of relevant cases (Chen et al., 2020; 289 Kujawski et al., 2020; Young et al., 2020) . Median lengths of stay in ICU between 7 to 10 days 290 were observed in studies based in China (Cao et al., 2020; Fan et al., 2020; Yang et al., 291 20202; Zhou et al., 2020) . European studies conducted by Graselli et al. (2020) . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. to ICU care and very high mortality rates (Argenziano et al, 2020; Cummings et al., 2020; 333 Lewnard et al., 2020; Myers et al., 2020; Paranjpe et al., 2020) . The cause of the differences 334 in disease presentation between countries is not clear. Perhaps differences in case 335 definition, the relative age of the population, the level of pre-existing conditions within the 336 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 experience in Italy, thus far, suggests that must carefully manage mitigation measures to 338 ensure ICU capacity is not reached or breached. However, another major influencer on 339 hospitalisation rate will be the definition of a suspect case and the extent of COVID-19 340 testing within the community; the detection of a greater number of milder cases, diagnosed 341 within the community will ultimately influence the hospitalisation rate. The influence of 342 community testing is becoming evident in Ireland; as April progresses, it is clear in each 343 subsequent HPSC report (2020a; 2020b; 2020c; 2020d) that the rate of hospitalisation and 344 proportion of patients admitted to ICU is reducing. This is evidence perhaps that a broader 345 case suspect definition and a wider testing approach will impact the overall rate of Superiore di Sanità, 2020) reporting the median hospital stay of 5 days for cases that did not 356 require ICU level of care. However, no range or confidence interval is described by this 357 report. Similar to the shorter duration of hospital stay in Italy, a more recent study 358 conducted in the USA, which examined hospital records of 5,700 sequential admissions to 359 New York hospitals, reported a short length of stay, median 4.1 days (IQR 2.3 to 6.8). The CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. . https://doi.org/10.1101/2020. 05.12.20099473 doi: medRxiv preprint admitted to hospital for further care, indicating perhaps the pressure on bed availability. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. Length of stay post ICU is illustrated in a figure in some small case series; and as such, were 408 data were not extracted for reporting or discussion in this review. Limitations associated with any rapid review are pertinent to this study. A narrow aim was 412 set, namely to determine rate of hospitalisation and the length of stay in hospital due to 413 COVID-19 infection in countries across the world now at the centre of the pandemic. The . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. The length of stay in hospital outside of the ICU setting, due to COVID-19 is highly variable in 435 the literature, with median estimates varying from 4 to 25 days. Studies diverged in their 436 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 438 required to test negative prior to discharge or where there was pressure to free up 439 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. . https://doi.org/10.1101/2020.05.12.20099473 doi: medRxiv preprint overcrowded hospitals. Indeed it is likely that some countries used hospitalisation to ensure 440 quarantine. As the pandemic spreads rapidly throughout the world, more and more evidence regarding 443 the length of stay in ICU following COVID-19 is reported. Many of these reports indicate a 444 median ICU length of stay between 7 to 10 days; however a number of these studies report 445 length of stay based only on resolved cases, and as such duration of stay it is likely skewed 446 towards a short stay due to bias towards those who have completed their ICU stay and are 447 now discharged or have died. Thus the upper estimates of the range should be considered 448 for use in resource models. It is apparent too that the data generated by each individual 449 country during the initial months of the outbreak will be critical to estimate the parameters 450 used for any national resource model. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 surveillance, investigations and control measures, January 2020. Euro CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 Group, 2020) indicates a median of 8 days (including ICU stay for deceased patients); 795 no estimate of variance was reported. The UK report separates survivors from those 796 who have died in care, the authors also indicate bias towards shorter lengths of stay 797 as this cohort has completed their ICU stay. 798 799 800 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. . . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 Date . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 16, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2020. . https://doi.org/10.1101/2020.05.12.20099473 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2020. . https://doi.org/10.1101/2020.05.12.20099473 doi: medRxiv preprint Table 3 . A summary of published articles and reports relating to length of stay in an ICU setting following COVID-19 infection, by country. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2020 . . https://doi.org/10.1101 N = 94; severe China n = 39; severe pneumonia China = 24; 50 % died USA; Bhatraju et la., 2020; n = 12; Survived or continuing care USA Contituing care Italy; Graselli et al. 2020b; n = 405 Died Italy; Graselli et al. 2020b; n = 256 Survivors Italy; Graselli et al. 2020b; n = 1581; All Italy; COVID-19 Surveillance Group = 101; Died (all died) China; Cao et al., 2020; n = 44; Died China; Cao et al., 2020; n = 199; RCT China