Association of hospital and intensive care unit occupancy and non-admission to the intensive care unit decisions: a retrospective cohort study

Dear Editor,

The COVID-19 pandemic caused several waves of rapid patient surges in hospital settings, which revealed major limitations of treatment availability and intensive care unit (ICU) bed capacities worldwide. Prior to the pandemic, Sweden had the lowest ICU bed capacity per capita in a report among 14 countries in Europe, and the bed capacity has decreased since [1, 2]. We investigated the association of ICU and hospital occupancy, and non-admission to the ICU decisions for hospitalised COVID-19 patients in Stockholm.

The cohort consisted of all adult patients with COVID-19 admitted to five acute hospitals in Region Stockholm between 1 March 2020 and 31 December 2022. Data on decisions on non-admission to the ICU were obtained from electronic health records and was linked to the Swedish Intensive Care Registry, Stockholm Regional healthcare data warehouse (VAL), SmiNet, and Statistics Sweden as previously described [3]. Ethical approval was received by the Swedish Ethical Review Authority. Occupancy was estimated as: all admitted patients in hospital, COVID-19 patients in hospital, COVID-19 patients in the ICU, and total ICU patients. Modified Poisson regression models were used to investigate the risk of non-admission to the ICU decisions for each of the four occupancy variables, adjusted for age, sex, comorbidities, calendar time, care services, hospital, education level, region of birth, and household composition. A detailed description of the methods is provided in the electronic supplementary material (ESM).

The cohort comprised 20 261 individuals, out of which 2766 (13.7%) received a decision on non-admission to the ICU. Non-admission decisions were more frequent in older patients, males, and patients with comorbid conditions. Descriptive data are presented in electronic supplementary table 1.

Occupancy counts for each variable along with adjusted risk ratios for each category is presented in Fig. 1. The risk of receiving a non-admission to the ICU decision increased gradually with higher COVID-19 ICU occupancy and COVID-19 hospital occupancy. Increasing total ICU occupancy showed a weaker increased association with non-admission to the ICU decision. Total hospital occupancy showed no consistent association with increased risk of non-admission to the ICU decision.

Fig. 1.

ICU and hospital occupancy over calendar time and adjusted risk ratios for non-admission to the ICU decisions. Panel a shows the occupancy in the total ICU (represented with navy blue) and COVID-19 ICU (represented with light blue) groups on the primary y-axis, and occupancy in the total hospital (represented with green) and COVID-19 hospital (represented with red) groups on the secondary y-axis. Panel b displays the adjusted risk ratios for each category of all the occupancy variables. The four multivariable models included the occupancy variable of interest (COVID-19 hospital, total hospital, COVID-19 ICU, or total ICU), age group, sex, comorbidities, calendar time, care services, hospital, education level, region of birth, and household composition. ICU intensive care unit, CI confidence interval

The results were comparable in sensitivity analyses with a restriction to COVID-19 as main diagnosis, application of a broader definition of non-admission to the ICU, a restriction to the last decision on non-admission to the ICU within the first 72 h from admission, and application of multiple imputation (electronic supplementary table 2). Analyses restricted to patients from the second wave onwards showed similar but weaker associations.

Previous studies have shown that age, sex, comorbidities and socioeconomic factors have been associated with increased risk of decisions on non-admission to the ICU [4, 5]. In this study, limited to COVID-19 patients, we report an association between higher occupancy rates of COVID-19 ICU patients, total ICU patients, and COVID-19 hospital patients, and an increased risk of non-admission to the ICU decisions. Our findings highlight the importance of ICU bed availability in Sweden during the COVID-19 pandemic, which should be considered when preparing for future health emergencies.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

The study was supported by the Swedish Research Council (Grant Dnr 2021–04809).

Declarations

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.