Introduction

Intensive care unit (ICU) hospitalizations may be one of the major burdens on the healthcare system. Eligibility criteria for ICU treatment are well defined in most cases, but in practice, the number of hospitalizations in ICUs may vary depending on such factors as financial resources, human resources, medical resources, organization of the country’s health care, and other factors. In a study from the USA, based on hospital discharge data in the State Inpatient Database for Maryland and Washington States, the proportion of hospitalized patients admitted to an ICU across hospitals ranged from 3% to 55% [1]. Hospitalizations in ICUs concern emergencies associated with a sudden threat to health or life. Respiratory failure was reported to be one of the most common reasons for ICU admissions and hospitalizations [2]. Another study revealed the incidence of sepsis in ICUs at 58 per 100,000 person-years, 41.9% of whom died in hospital [3]. Factors found to be important for the prognostic outcome of sepsis in patients hospitalized in ICUs include gender and selected clinical factors and diseases [4].

High mortality rates were reported to be related to ICU care, being 19.0% in the hospital and 21.7% within 6 months [5]. ICU readmission was associated with an increased severity of illness scores during the same hospitalization in adult patients [6]. Another study based on a large dataset of 77,132 ICU survivors reported that 89% of patients returned home, but after 1 year, 77% of patients were still at home and 17% of them had died [7]. Compared to non-ICU patients, ICU patients showed an increased hospital length of stay and mortality. ICU patients were more frequently readmitted to hospital than controls, and mortality up to 1 year in this group was also higher in ICU patients [8]. Compared with late readmission, early readmission was associated with decreased mortality [9]. Readmission to the ICU during the same hospitalization may not be a rare problem. In a study from Brazil, out of 5,779 ICU patients, 10% were readmitted to the ICU during the same hospitalization [10]. A significant proportion of patients hospitalized in ICUs died during hospitalization. Infection at ICU discharge, ICU readmission, age, length of hospital stay, and higher SAPS III values increased the risk of death in ICU survivors [11]. ICU and in-hospital mortality may be reported with considerable heterogeneity. In one study, mortality of older patients (≥ 75 years) admitted to the ICU ranged from 1% to 51%, and in-hospital mortality from 10% to 76% [12]. In a study from India, the most common cause of death among elderly patients was found to be sepsis, followed by pneumonia [13]. In another report, end-of-life care expenditures for the older population with cancer were highly concentrated until the last month [14]. Hospitalizations in ICUs involve costs of care during hospitalization and may be related to costs of post-hospital care. Critical care survivors had greater 1-year post-discharge healthcare resource utilization than non-critical inpatients, including 68% of those with longer hospital stays [15]. In a recent study including data on 19,119 older adults aged 65 years or more who had received high-intensity care at least once and died in the intensive care unit in South Korea, the annual cost of high-intensity care at the end of life increased steadily from 2016 to 2019 [16]. In Poland, ICU admissions are guided by uniform qualification criteria recommended by the Polish Society of Anaesthesiology and Intensive Therapy (PTAiIT). Guidelines for ICU admission in Poland in the form of priorities were first published in 2012 [17] and recently modified [18].

There is limited information on ICU hospitalizations in Poland based on a nationwide database registry [19, 20]. The aim of the study was to describe long-term trends in ICU hospitalizations and related fatality rates in Poland.

Material and methods

Our study is a population-based, retrospective analysis of hospital discharge records of patients hospitalized, and these hospitalizations included a stay or stays in ICUs. Data were obtained from the National Institute of Public Health, National Institute of Hygiene – National Research Institute in Warsaw, and they covered the period from 2012 to 2021. All hospitals in Poland, except psychiatric facilities, are legally required to send discharge data to the institute. The data are anonymous and include information on hospitalizations, dates of admission and discharge, sex, date of birth, and place of residence. The hospital morbidity database used in this study was previously prepared and cleaned by NIZP-PZH (National Institute of Public Health – National Institute of Hygiene) as part of its statutory activities, and no other exclusion criteria were specified in our study. In 11,557 hospitalization records (1.6% of all records), inaccuracies or missing data were noted when reporting the patient’s place of residence. In 971 hospitalization records (0.1% of all records) inaccuracies or missing data were noted when reporting the date of birth or age. The inclusion criterion was a stay in the ICU (department code – 4260) for the entire period or part of the hospitalization. Records were excluded if a patient was hospitalized in the cardiology ICU department (code – 4106). Other studies also separated cardiovascular and medical ICUs [21, 22]. Information on the study presented in this paper was submitted to the local bioethics committee.

To perform most statistical analyses, Statistica (TIBCO Software Inc, version 13) [23] and WINPEPI [24] were used. For continuous variables means, SD, medians and IQR were computed, respectively. For nominal variables, counts, and percentages were analyzed. The data were anonymised, but date of birth, place of residence, and gender were the basis for estimating annual frequencies of hospitalization. Data from Statistics Poland (the national census) were used as denominators [25]. To assess trends, linear regression was used. The trend analysis included the numerical values of events as a dependent variable and the size of a specific population in each year of the study period as an independent variable. In the case of selected variables such as age and length of stay (LOS), a trend analysis in linear regression was performed based on average values calculated for each year in the study period. A two-sided p-value lesser than 0.05 was considered to be statistically significant.

Results

Discussion

Trends in ICU hospitalizations

Based on the hospital registry, the mean annual hospitalization incidence was estimated to be 1.9 per 1,000, and this rate seems to be relatively low in comparison to data from other countries. In a large study from Denmark, the number of ICU patients per 1000 person-years for the 5-year period was 4.3 patients, and it ranged from 3.7 to 5.1 patients per 1000 person-years in the five regions of Denmark, and from 2.8 to 23.1 patients per 1000 person-years in the 98 municipalities [26]. In a study from Korea based on ICU admissions from August 2009 to September 2014, the ICU admission rate was 744.6 per 100,000 person-years (869.5 per 100,000 person-years in men and 622.0 per 100,000 person-years in women) [27]. A significant predominance of male patients was also observed in our study, as presented in Table I. In another study on 1425 patients, 780 (54.7%) were male [28].

The study presented in this paper showed a statistically significant upward trend in the frequency of hospitalizations, as presented in Table I. The growing trend in hospitalizations may result both from the organization of health care and the aging of the Polish population [29]. The elderly were the main groups hospitalized in ICUs, as reported in Table III. This may result both from the organization of health care and multi-morbidity among the elderly, which increases with age and increases the health needs of this group of patients. Hospitalization incidence rates per 1,000 by age range are presented in Table III. In 2021, compared to 2012, an increase in the frequency of hospitalizations was observed in most cases, except for patients aged 0–19 and patients aged 80 years and over. In a study from Texas, United States, which was based on ICU hospitalizations of patients with dementia aged 65 years or older, increasing comorbidity burden, rising severity of illnesses, and increasing use of health care resources were reported [30].

The mean age in our study was relatively high, at 61.6 years, and the trend of mean age was significantly increasing during the study period. Additionally, as shown in Table I, in each year during the study period, male predominance was observed in the study group and among the total number of deaths, as well as deaths during the stay in the ICU. Data from another study on a large group of 223,129 patients admitted to ICUs in Australia and New Zealand showed a lower mean age of 59.2 years and 41.7% of female patients [31]. Older age and male sex were also reported to be risk factors for ICU admission and death. In a recent meta-analysis of 59 studies when hospitalized with COVID-19, men more often developed severe COVID-19 disease and more often required intensive care admission, ultimately resulting in death more often. Additionally, in this study it was reported that patients aged 70 years and above affected by COVID-19 were more likely to require ICU admission and experience in-hospital mortality compared with patients younger than 70 years [32].

In another meta-analysis based on 229 studies including over 10 million patients, it was reported that men had a higher risk for COVID-19 infection, hospitalization, disease severity, ICU admission, and death [33].

In our study, the mean LOS was on a stable level, with small fluctuations when analyzing the length of hospitalization stay. However, the mean ICU stay was increasing. The mean LOS seems to be long when compared to data from other countries. In one study on healthcare-associated infections in ICUs, the observed LOS was 11.66 days [34]. In another study on 108,302 ICU patients, the reported LOS was 9 days [8]. Yet another study from Australia and New Zealand showed the ICU LOS of 3.6 days [31]. Another study revealed the median ICU and hospital length of stay of 4 and 13 days, respectively [27]. Long ICU stay I associated with a significant financial burden. However, premature discharges appear to have a significant impact on mortality in the hospital care of critically ill patients [35].

In our study, the average percentage of patients staying in ICUs for 10 days or more was 27.8%, which may be relatively high. In another study, patients staying in the ICU for 10 days or more accounted for only 5% of all ICU patients. It was also highlighted that these patients’ outcomes were markedly worse than the outcomes of patients staying in the ICU for less than 10 days [36]. LOS plays an important role not only in patients’ outcomes. It should be highlighted that the economic burden of ICU-acquired pneumonia may be mainly related to an increased length of stay of surgical patients and patients with mid-range severity scores at admission [37].

In our study, during one hospitalization, one stay in the ICU was recorded in 94.5% of cases, and this may be related to the long ICU stay observed in our study. In the analyzed 10-year period of the study, not more than one hospitalization with ICU stay was found in 87.6% of patients, which can be considered a satisfactory result. It was reported that readmission to the ICU was associated with poor clinical outcomes, increased length of ICU and hospital stay, and increased costs. In one study based on data of 5,779 patients admitted to the ICU, 10% of patients were readmitted to the ICU during the same hospitalization [10]. In another study, in a year after the ICU care, 10% were readmitted to the ICU [5].