A national retrospective cohort study of mechanical ventilator availability and its association with mortality risk in intensive care patients with COVID-19

Importance: The virulence of SARS-CoV-2 has provoked concerns about hospitals' ability to effectively care for the vast numbers of people affected. Notably, several countries reported operating near or at their intensive care capacity during the COVID-19 pandemic, although the impact of this on patient outcomes remains unclear. Objectives: To determine if there is an association between survival rates in intensive care units (ICU) and occupancy of the unit on the day of admission. Design: National retrospective observational cohort study spanning the first wave of the COVID-19 pandemic in England. Setting: 114 hospital trusts (groups of hospitals that function as a single operational unit). Participants: 4,032 adults admitted to an ICU in England between 2nd April and 1st June, 2020, with presumed or confirmed COVID-19, for whom data was submitted to the national surveillance programme and met study inclusion criteria. Interventions: N/A Main Outcomes and Measures: A Bayesian hierarchical approach was used to model the association between hospital trust level (mechanical ventilation compatible) bed occupancy, and in-hospital all-cause mortality. Results were adjusted for unit characteristics (pre-pandemic size), individual patient-level demographic characteristics (age, sex, ethnicity, time-to-ICU admission), and recorded comorbidities (obesity, diabetes, chronic respiratory disease, chronic liver disease, chronic heart disease, hypertension, immunosuppression, chronic neurological disease, chronic renal disease). Results: 38.4% (1,548) of patients admitted to an ICU died. Adjusting for patient-level factors, mortality was higher for admissions during periods of high occupancy (>85% occupancy versus the baseline of 45 - 85%) [median odds ratio (OR) 1.19 (95% posterior credible interval (PCI): 1.00 - 1.44)]. In contrast, mortality was decreased for admissions during periods of low occupancy (<45% relative to the baseline) [OR 0.75 (95% PCI: 0.62 - 0.89)]. Conclusion and Relevance: Increasing occupancy of beds compatible with mechanical ventilation, a proxy for operational strain, is associated with a higher mortality risk for individuals admitted to ICU. Public health interventions (such as expeditious vaccination programmes and non-pharmaceutical interventions) to control both incidence and prevalence of COVID-19, and therefore keep ICU occupancy low in the context of the pandemic, are necessary to mitigate the impact of this type of resource saturation.


Summary Box
What is already known on this topic: Pre-pandemic, higher occupancy of intensive care units was shown to be associated with increased mortality risk. However, there is limited data on the extent to which occupancy levels impacted patient outcomes during the first wave of COVID-19, especially in light of the mobilisation of significant additional resources.
A recent study from Belgium reported a 42% higher mortality during periods of ICU surge capacity deployment, although in the analysis surge capacity was evaluated only as a binary variable. Although, this contradicts earlier results from smaller studies in Australia and Wales, where no association between ICU occupancy and mortality was identified.

What this study adds:
The results of this study suggest that survival rates for patients with COVID-19 in intensive care settings appears to deteriorate as the occupancy of (surge capacity) beds compatible with mechanical ventilation (a proxy for operational pressure), increases. Moreover, this risk doesn't occur above a specific threshold, but rather appears linear; whereby going from 0% occupancy to 100% occupancy increases risk of mortality by 92% (after adjusting for relevant individual-level factors). Furthermore, risk of mortality based on occupancy on the date of recorded outcome is even higher; OR 4.74 (95% posterior credible interval: 3.54 -6.34). As such, this national-level cohort study of England provides compelling evidence for a relationship between occupancy and critical care mortality, and highlights the needs for decisive action to control the incidence and prevalence of COVID-19.
. CC-BY 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.

Introduction
From the first reports of a novel coronavirus (SARS-CoV-2) in late 2019, to date, global mortality associated with the resultant disease (COVID-19) has exceeded 1.7 million people. [1] The virulence of the pathogen has prompted persistent concern about the ability of health services around the world to effectively care for the vast numbers of people affected. [2] These concerns are most relevant in the context of scarce resources (e.g., mechanical ventilation) required by patients in need of high-acuity support, which is relatively common in patients with COVID-19. [3] Notably, even with the introduction of non-pharmacological interventions such as stay-at-home orders, almost a third of all hospitals in England reached 100% occupancy of their "surge" mechanical ventilation capacity (i.e., including the additional beds that were created through the re-allocation of resources) during the first wave of the pandemic. [4] England is now experiencing a second wave that is already worse than the first, with 40% more people in hospital, many hospitals overwhelmed and exhausted staff. [5] What remains unclear is whether and to what extent operating at these extremes of critical care occupancy impacted patient outcomes.
Pre-pandemic, higher occupancy levels in intensive care (which may reflect operational strain), was shown to be associated with higher mortality risk. [6] However, there is limited data on the extent to which occupancy levels impacted patient outcomes during the first wave of COVID-19. [7] A recent study from Belgium reported 42% higher mortality during periods of ICU surge capacity deployment, although in the analysis surge capacity was evaluated only as a binary variable. [8] This contradicts earlier results from smaller studies in Australia and Wales, where no association between ICU occupancy and mortality was identified. [9,10] A better understanding of how operating under such extreme circumstances affects outcomes is crucial for two reasons: firstly, to allow hospitals to adapt practice to improve outcomes and secondly, to provide policy makers with more accurate information about the potential consequences of allowing health services to be overwhelmed. As such, in this study, we sought to evaluate the extent to which mortality risk in intensive care units (ICUs) over the course of the first wave of the COVID-19 pandemic in England could be explained by differences in occupancy.
. CC-BY 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 January 13, 2021. ; https://doi.org/10.1101/2021.01.11.21249461 doi: medRxiv preprint

Data Sources
Data on all intensive care unit (ICU) admissions across England were extracted from the COVID-19 Hospitalisation in England Surveillance System (CHESS). [11] Information on occupancy rates were extracted from the daily situation reports (i.e., 'SitRep'). 4 Both datasets are mandatory regulatory submissions for all National Health Service (NHS) acute care providers in England, and further details about them can be found in the eMethods.

Study population
All ICU admissions reported to CHESS between 2 nd April -1 st June (see eMethods for details on date selection), with presumed/confirmed COVID-19 (94.7% tested positive during admission), aged 18 -99, nonpregnant, and with valid admission and occupancy data, were eligible for inclusion (eFigure 1).

Patient and Public Involvement
No patients were involved in the design, interpretation of the results, or dissemination of this study. However, we plan to disseminate the results to patients and the public in collaboration with several media partners whom have already agreed to report on this results.

Patient-Level Data
Information extracted from CHESS about each patient comprised: administrative features (admitting trust, admission date), demographic characteristics (age, sex, ethnicity), recorded comorbidities (obesity, diabetes, asthma, other chronic respiratory disease, chronic heart disease, hypertension, immunosuppression due to disease or treatment, chronic neurological disease, chronic renal disease, chronic liver disease). Ethnicity was coded as white, Asian (Subcontinent and other), black, mixed, and other; comorbidities were recorded as binary indicator variables, with missing entries assumed to reflect the absence of a comorbidity. The appropriateness of this assumption in CHESS has been previously explored. [12] Occupancy Data Trusts are groups of geographically co-located hospitals that function as a single organisational unit within the UK's national healthcare system. Information extracted from SitRep about each trust comprised: pre-pandemic (January -March 2020) number of beds compatible with mechanical ventilation, the proportion of beds compatible with mechanical ventilation occupied on each day of the study period, and each trust's geographical region. [4] Linkage was carried out based on the trust that an individual was admitted to and the date of ICU admission in CHESS; patients in CHESS were matched via their admission date to the relevant occupancy information from the corresponding date in SitRep.
. CC-BY 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.

Outcome
The primary outcome was in-hospital all-cause mortality; all included patients were followed up until death, discharge, or transfer, and the latter two were treated as suggesting that the patient survived. All patients had a recorded outcome as of the 10 th October 2020 dated extract of CHESS used for this analysis.

Statistical Analysis
Descriptive summaries were generated as median and interquartile ranges for continuous variables, and frequency and percentage incidence for categorical factors. Exploratory analyses included: the relationship of the COVID-19 epidemic curve to bed occupancy at a national level (eFigure 2); the distribution of missingness amongst patients and trusts (eFigure 3); variation in age and comorbidity burden over the first wave (eFigure 4); the impact of modelling continuous variables either linearly, through the use of threshold functions, or via (standard cubic) splines/smooths (eMethods).
A Bayesian hierarchical approach was used to model the association between the trust, group and patient-level factors and mortality risk. Specifically, a generalised additive mixed model was utilised, with intercept and slope coefficients for population and group level effects, and a Bernoulli likelihood with logit function to link to mortality outcome. Coefficients were inferred by Markov chain Monte Carlo sampling, using Stan (CmdStan V2.25.0), with a model specified using BRMS (V2.14.4) in R (V4.0.3). [13][14][15] Further information on the Bayesian prior specification and modelling methods are reported in the eMethods.
As secondary analyses, two potential interactions were assessed: 1) baseline trust size and occupancy; 2) patient age and occupancy (results not reported due to insignificance). We also assessed the association of occupancy on the recorded outcome date with mortality, and occupancy expressed in terms of pre-pandemic ICU size.
Several sensitivity analyses were carried out: 1) filtering for different degrees of missingness of patient-level comorbidity data at trust-level; 2) adjusting for week of admission; 3) adjusting for trust and region as random effects; 4) additional patient-level factors: time from hospital admission to ICU admission, chronic liver disease and obesity; see eMethods for justifications.
. CC-BY 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 January 13, 2021. ; https://doi.org/10.1101/2021.01.11.21249461 doi: medRxiv preprint Results 4,032 individuals were included in this study following application of the inclusion/exclusion criteria (see eFigure 1), of whom 1,548 (38.4%) died. In total, 79,793 (median 13 days per patient; IQR: 6 -27) patient-days were observed, equating to a mortality rate of 19.4 per 1,000 patient days. A full summary of the recorded patient-level characteristics is reported in Table 1.

Mortality risk increases linearly with admission and date-of-outcome specific occupancy
The fully adjusted OR for mortality (Figure 3), using occupancy on the day of admission coded as a continuous linear variable ranging from 0 to 1 was 1.92 (95% PCI: 1.36 -2.67). Moreover, using the bed occupancy from each individuals' outcome date identified an even larger association (full model specification in eTable 2), OR 4.74 (95% PCI: 3.54 -6.34).

Larger ICUs experience exaggerated impacts of extremes of mechanical ventilator occupancy rates
Although pre-pandemic number of beds did not substantially alter the OR of occupancy as a sensitivity analysis (eTable 1), it did appear independently informative. An increase in pre-pandemic size by 25 beds was associated with a 25% decrease in risk of mortality (eTable 3). The introduction of an interaction term between prepandemic size and occupancy identified that larger ICUs experience exaggerated impacts of extremes of mechanical ventilator occupancy rates (eFigure 8).
. CC-BY 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 January 13, 2021. ; https://doi.org/10.1101/2021.01.11.21249461 doi: medRxiv preprint . CC-BY 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.

Discussion
The results of this study highlight a potential major impact of operational pressure on patient survival during the first wave of the COVID-19 pandemic. Survival rates for patients with COVID-19 in intensive care settings appears to deteriorate as the occupancy of beds compatible with mechanical ventilation (a proxy for operational pressure), [17] increases. These observations are consistent with the aforementioned Belgian study, [8] except that our results suggest a linear association rather than single step increase at a specific threshold. Moreover, the results might partially explain the decreased mortality rate seen in the latter half of the first wave in the UK, [18] where occupancy rates were much lower than at the peak. [4] Our findings also corroborate previous reports of an association between larger ICUs and lower COVID-19 mortality. [19] However, we additionally observe an interaction with (pre-pandemic) unit size, whereby larger ICUs experience more exaggerated impacts from both higher and lower (surge) occupancy rates. It is unclear from our data what is driving this heterogeneity.

Strengths and Limitations
The strengths of this study are the national cohort of patient-level data with near-perfect capture of admissions, [20] coupled with a rigorous modelling method (eTable 4 & eMethods). Limitations include a lack of physiological data, limiting our ability to adjust for differences in severity upon admission. Moreover, the characterisation of operational strain as a function of surge occupancy likely fails to fully reflect the complexity of using non-specialist staff and other resource allocation issues present when 'creating' new ICU beds (see eTable 5 using an alternative definition of occupancy based on baseline capacity; mortality risk given this linear continuous factor was 1.08 (95% PCI: 0.98 -1.26). Finally, we lack clear 30-day outcome data for discharged and transferred individuals, and thus were forced to model under a naïve assumption that these individuals survived.

Implications for Policy Makers and Clinicians
In summary, our study highlights the importance of public health interventions (such as expeditious vaccination programmes and non-pharmacological interventions), to control both incidence and prevalence of COVID-19, and therefore actively manage ICU occupancy, as there is evidence of direct harm to patients as a consequence of saturation. This is especially relevant given the identification of a new strain of COVID-19 with a potentially increased risk of transmission, [21] coupled with observations that second wave-related operational pressures (including bed occupancy rates) in England have exceeded levels seen during the first wave of the pandemic, [5] suggesting immediate and decisive action is necessary.
. CC-BY 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 January 13, 2021.

Data Access Statement
Data cannot be shared publicly as it was collected by Public Health England (PHE) as part of their statutory responsibilities, which allows them to process patient confidential data without explicit patient consent. Data utilised in this study were made available through an agreement between the University of Warwick and PHE.
Individual requests for access to CHESS data are considered directly by PHE (contact via covid19surv@phe.gov.uk).

Role of the funding source
This study received no direct funding. The individual authors' funder had no role in the design of the study, the analysis, or the formulation of the manuscript.

Transparency Statement
The senior authors (SJV, BAM) had full access to all data and had final responsibility for the decision to submit for publication. SJV and BAM affirm that the manuscript is an honest, accurate, and transparent account of the . CC-BY 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 January 13, 2021. ; https://doi.org/10.1101/2021.01.11.21249461 doi: medRxiv preprint study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as originally planned have been explained.
. CC-BY 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 January 13, 2021. ; https://doi.org/10.1101/2021.01.11.21249461 doi: medRxiv preprint . CC-BY 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 January 13, 2021. ; https://doi.org/10.1101/2021.01.11.21249461 doi: medRxiv preprint  The full posterior distribution of the odds ratio (OR) for mortality given low occupancy 0 -45% (Top; Green), and high occupancy 85 -100% (Bottom; Red) are presented. The PCIs presented are equally tailed credibility intervals for the posterior odds ratio distributions. The outer (less saturated) interval is the 95% PCI, and the inner (more saturated) interval shows the 90% PCI.  Occupancy was specified without multiplying out by 100 (i.e., 20% = 0.20), therefore the odds ratio is for a step from 0% to 100% (i.e., 0.0 to 1.0). Exact values are tabulated below. . CC-BY 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)

Posterior Credible Intervals
The copyright holder for this preprint this version posted January 13, 2021.