Abstract
-
Purpose
This study aimed to determine the rates of intensive care unit readmission and to identify the factors influencing readmission among intensive care unit patients aged≥65 years with internal medicine conditions.
-
Methods
We retrospectively reviewed electronic medical records from a tertiary care hospital in Seoul, analyzing the characteristics of patients who were and were not readmitted between December 2020 and September 2022.
-
Results
A total of 351 patients were included. The unplanned intensive care unit readmission rate was 4.8% within 7 days and 9.1% beyond 7 days after discharge. Comorbid diabetes, higher total bilirubin levels at intensive care unit admission, lower PaO2/FiO2 ratios at discharge, and elevated Blood Urea Nitrogen (BUN) levels at discharge were associated with an increased risk of readmission within 7 days. In contrast, hypertension, prolonged intensive care unit stays, and lower hemoglobin levels at discharge were associated with readmissions occurring after 7 days.
-
Conclusion
Intensive care unit readmission among older patients is influenced by several clinical and hematological factors. Nurses should consider a patient's history of diabetes and hypertension, length of intensive care unit stay, and laboratory values-specifically total bilirubin at admission, and PaO2/FiO2 ratio, hemoglobin, and BUN levels at discharge-when making discharge decisions. These findings can inform the development of discharge guidelines.
-
Key Words: Aged; Blood Urea Nitrogen; Intensive Care Units; Length of Stay; Readmission
INTRODUCTION
Intensive Care Units (ICUs) are specialized settings staffed by highly skilled healthcare professionals who care for critically ill patients using invasive monitoring and advanced life support equipment that is not available in general wards. Increasing life expectancy and the proportion of the elderly population are leading to an increase in the incidence of chronic diseases, which in turn is increasing the number of patients requiring intensive care [
1]. Consequently, establishing clear ICU admission and discharge criteria is essential for optimal operation. Studies have indicated that patients aged 65 years and older account for 50% of ICU admissions, and survivors in this age group often require longer acute hospital care following ICU discharge [
2]. Additionally, patients with chronic conditions such as diabetes, hypertension, and chronic obstructive pulmonary disease are two to three times more likely to be admitted to the ICU than those without these conditions [
3]. These demographic and epidemiological trends highlight the need for efficient ICU resource allocation through well-defined admission and discharge criteria. The Korean Society of Critical Care Medicine provides clinical practice guidelines for ICU admission, discharge, and prioritization; however, while admission criteria are specific and include prioritization, diagnostic criteria, and objective variable models, the discharge criteria remain vague-typically defined as "when the patient's systemic condition is stabilized and the patient no longer requires ICU monitoring or management" [
4].
Patients readmitted to the ICU exhibit higher mortality rates and longer hospital stays compared to those who are not readmitted [
5,
6]. A systematic review and meta-analysis found that ICU readmissions were significantly associated with increased hospital mortality (odds ratio 2.91, 95% CI 2.37~3.58) and extended ICU and hospital stays [
6]. These readmissions contribute to poor prognoses, longer hospitalizations, and increased healthcare costs [
6]. Other countries monitor ICU readmission rates at the national level as an indicator of medical institution quality, and Korea has also used ICU readmission rates as a metric to evaluate ICU adequacy since 2014. Readmissions within 48 hours of transfer to a general ward often suggest premature discharge due to inadequate assessment of discharge readiness, underscoring the need for careful evaluation of ICU readmission rates [
7]. Furthermore, a recent systematic review and meta-analysis reported that about 40% of ICU readmissions within 48 hours were for conditions similar to the initial ICU admission, indicating that a substantial portion of early readmissions may result from premature discharge [
6]. Therefore, an objective assessment of patient status at ICU discharge is critical to identify those at high risk for readmission and to proactively prevent both premature discharge and subsequent readmission.
Several recent studies have attempted to identify potential risk factors for ICU readmission [
8-
12]. However, these risk factors are varied, and readmission cannot be predicted by a few isolated variables. They can be broadly categorized into demographic, clinical, ICU treatment, laboratory, physical, and environmental characteristics. Demographic factors, such as age, gender, body mass index, and race, have been examined, with older age consistently emerging as a significant risk factor, as studies report higher readmission or mortality rates among older patients compared to younger ones [
7,
13]. Clinical factors include admission source, comorbidities, severity scores, internal medicine conditions, emergency surgery, and the presence of at least one chronic comorbidity [
8,
13]. ICU treatment characteristics encompass interventions such as dialysis, in-hospital cardiopulmonary resuscitation, duration of ventilator support, level of consciousness, length of ICU stay, fever at discharge, respiratory rate, heart rate, early warning scores, oxygen therapy, and the use of sedatives, antipsychotics, and vasoconstrictors [
8,
13]. Laboratory characteristics involve measures such as creatinine and albumin levels at ICU admission and hyperglycemia and hemoglobin levels at discharge [
11,
12]. Additionally, physical and environmental factors include the presence of a tracheostomy tube and suction, which has been associated with a higher risk of readmission [
9,
10].
In particular, a UK cohort study reported that individuals aged 55~64 years had a 1.82-fold higher risk of death or ICU readmission, those aged 65~74 years had a 2.29-fold increase, and those aged ≥75 years had a 3.29-fold increase compared to individuals aged ≤55 years [
13]. Similarly, a large Canadian study found that older patients experienced higher in-hospital mortality, were more likely to be discharged to long-term care facilities, and incurred significantly higher mean ICU admission costs than younger patients [
14]. Nevertheless, age alone does not fully account for the high readmission rates observed among older patients. Thus, it is essential to consider a range of factors-including clinical, therapeutic, and environmental characteristics-when assessing ICU readmission risk in older patients, rather than focusing solely on age. In Korea, individuals aged ≥65 years comprised 17.5% (9.1 million) of the population in 2021, a figure projected to rise to 30.1% by 2035 [
1]. Furthermore, among those aged ≥60 years, 16.1% have one chronic condition while 78.7% have at least two chronic conditions [
3]. As the population ages and the demand for ICU care among older patients increases, it is critical to distinguish between those who do and do not require ICU care after discharge to enhance ICU management and achieve cost savings [
8]. Despite the emphasis on age as a predictor in several studies, the burden of chronic diseases, frailty, and other geriatric-specific factors necessitate a closer examination of the unique aspects of ICU readmission in older patients.
Although previous studies have frequently identified age as a predictor of ICU readmission [
7,
8,
13], there is a scarcity of research focusing specifically on the characteristics of older patients who are readmitted to the ICU. Understanding these characteristics-including the roles of multiple comorbidities and impaired recovery-in determining readmission risk is crucial for developing more effective discharge criteria for older patients. Such insights can guide interventions aimed at reducing unnecessary ICU readmissions and improving overall patient care and management. Older patients, particularly those admitted for internal medicine conditions, often present with complex multi-organ diseases that are challenging to manage in the ICU and contribute to a significant global healthcare burden due to the need for intricate treatment regimens [
3]. A large multicenter study in the United Kingdom revealed that between 1997 and 2016, the annual number of ICU admissions for older patients increased disproportionately, with the most significant rise observed in internal medical admissions among older adults [
2]. This trend can substantially affect ICU outcomes and readmission rates. Consequently, the current study aimed to identify the predictors and prognosis of ICU readmissions in older patients aged ≥65 years admitted for internal medicine conditions, thereby providing a foundation for interventions designed to reduce readmissions.
METHODS
1. Study Design
This retrospective survey study used electronic medical record data from a general hospital. We analyzed clinical, ICU treatment, laboratory, physical, and environmental characteristics at the time of ICU discharge to the general ward for patients aged ≥65 years. Patient characteristics were compared between those who were readmitted and those who were not to identify predictors of readmission.
2. Setting and Samples
This study included patients admitted to the ICU of a 2,700-bed tertiary care hospital in Seoul city, South Korea, between December 2020 and September 2022 for internal medicine-related illnesses. All patients were aged ≥65 years at the time of ICU admission and were discharged to the general ward. The age criterion was based on findings by Jones et al., which emphasize the unique physiological changes and healthcare needs of older patients in critical care settings [
2]. We excluded patients who had consented to withdraw life-sustaining treatment at ICU discharge, those with an ICU stay of less than 24 hours, and those admitted for planned surgery, as the characteristics of initial ICU admission and readmission may differ. Since the study focused on internal medicine conditions, patients admitted for surgical reasons were also excluded.
A previous study evaluating ICU readmissions for issues directly related to the initial ICU stay reported a mean readmission time of 7 days [
15]. Several studies have used a 7-day threshold to differentiate early from late readmissions, associating this cutoff with significant differences in patient outcomes and resource utilization [
12,
16]. Early readmissions within 7 days are typically related to premature discharge or inadequate post-ICU care, whereas later readmissions may result from new complications or exacerbation of chronic conditions [
16]. In this study, all ICU readmissions during the hospitalization period were included, and patients were categorized into three groups: readmission within 7 days, readmission after 7 days, and non-readmission.
The sample size was calculated using the G*Power 3.1.9.7 program (Universität Düsseldorf, Düsseldorf, Germany). Based on a previous study where the ICU readmission rate was 9.6% [
15] and the odds ratio for pH among readmission-associated factors was 1.70 (⍺=.05, power=0.80) in multiple logistic regression analysis, the minimum sample size was determined to be 314. Accounting for an anticipated dropout rate of 15%, consistent with previous research on critically ill patients [
17], 362 patients were targeted. A total of 370 patients were initially included; however, 16 patients were excluded due to surgical admissions and 3 patients were excluded because of missing data. Ultimately, data from 351 patients were analyzed.
This study was reviewed and approved by the Institutional Review Board of Asan Medical Center (AMC IRB 2022-1551). All collected data were anonymized and restricted to the research team.
4. Measurements/Instruments
1) Variables
We examined ICU readmission rates and post-discharge outcomes, including 28-day mortality and the acquisition of a Physician Order for Life-Sustaining Treatment (POLST) within 28 days. Additionally, we analyzed demographic, clinical, ICU treatment, laboratory, physical, and environmental characteristics associated with ICU readmission, as informed by the literature.
(1) Demographic characteristics
Demographic data included the patient's gender and age.
(2) Clinical characteristics
Clinical information was obtained from electronic medical records and included the admission source, primary diagnosis at hospitalization, diagnosis upon ICU admission, type and number of comorbidities, Charlson Comorbidity Index (CCI) [
18], and Acute Physiology and Chronic Health Evaluation (APACHE) IV score [
19]. The APACHE IV score, which ranges from 0 to 286, assesses the severity of illness at ICU admission, with higher scores indicating greater severity. Comorbidities, including hypertension, were identified based on recorded diagnoses and corresponding diagnostic codes at admission.
(3) ICU treatment characteristics
ICU treatment characteristics encompassed the length of hospital stay and the length of stay during the initial ICU admission. We also recorded the use of advanced life support devices (such as ventilators, hemodialysis, continuous renal replacement therapy, and extracorporeal membrane oxygenation), duration of ventilator use, the occurrence of cardiac arrest, and the presence of mycobacteria or multidrug-resistant bacteria during the ICU stay. Patient characteristics at discharge included the National Early Warning Score (NEWS) [
20] for early prediction of acute deterioration, the total Glasgow Coma Scale (GCS) [
21], presence of a tracheostomy tube, use of oxygen therapy, high-flow nasal cannula, or bi-level positive airway pressure, Fraction of Inspired Oxygen (FiO
2, %), presence of a drainage tube, and urine output on the day preceding discharge. Drug-related variables were assessed based on the use of continuous infusion sedatives and vasopressors at discharge.
(4) Laboratory characteristics
Respiratory function was evaluated using the ratio of arterial oxygen partial pressure to fractional inspired oxygen (PaO2/FiO2 ratio). Hematologic function was assessed via hemoglobin and hematocrit levels. Hepatobiliary function was measured using aspartate transaminase, alanine transaminase, total bilirubin, and albumin levels. Renal function was determined by measuring Blood Urea Nitrogen (BUN), creatinine, sodium, potassium, and the Glomerular Filtration Rate (GFR). Endocrine function was assessed using hemoglobin A1C levels, and immune function was measured by neutrophil counts and C-reactive Protein (CRP) levels. Laboratory values were collected from the records closest to the ICU admission date (both before and after admission) and from those closest to the ICU discharge date (within 7 days prior to discharge).
(5) Physical and environmental characteristics
Physical characteristics included the Richmond Agitation-Sedation Scale (RASS) [
22], which ranges from -5 (unarousable) to +4 (combative) to measure sedation; the Confusion Assessment Method for the ICU (CAM-ICU) [
23], which provides a binary assessment for delirium; the Critical Care Non-Verbal Pain Scale (CNPS) [
24] to assess pain at discharge (scoring from 0 to 12); and the Braden scale for pressure ulcer risk assessment at discharge (ranging from 6, indicating high risk, to 23, indicating low risk). Environmental characteristics related to nursing care included diet type, the use of oral or tracheal suction, and the application of physical restraints at discharge.
Following IRB approval, data were requested and obtained from the hospital's medical records team. The electronic records of patients aged ≥65 years, discharged to the general ward between December 2020 and September 2022 after an ICU admission for internal medicine conditions, were reviewed. All data were directly extracted by a single researcher using a case record sheet designed to capture patient conditions and nursing needs from ICU admission to ICU discharge. Admission data were collected from the records nearest the ICU admission date, while discharge data were obtained from values recorded at approximately 8:00 a.m. This time was chosen because most ICU discharge decisions are made then, ensuring consistency in measurements. The timing of these values is important, as certain physiological parameters may influence discharge decisions.
6. Data Analysis
All statistical analyses were performed using SPSS Statistics version 27 (IBM Corp, Armond, NY).
Descriptive statistics were employed to compare demographic, clinical, ICU treatment, laboratory, physical, and environmental characteristics between the non-readmission group and both the readmission within 7 days and readmission after 7 days groups.
Continuous variables that met normality assumptions (assessed via skewness and kurtosis) were expressed as means with standard deviations and compared using independent t-tests. Variables that did not meet normality assumptions were presented as medians, and comparisons were made using the non-parametric Mann-Whitney U test. Categorical variables were reported as frequencies and percentages and were analyzed using the x2 test or Fisher's exact test when more than 20% of cells had an expected frequency of ≤5.
Binary logistic regression was performed for variables associated with readmission at a significance level of
p<.10 in the initial analyses. This threshold was chosen to capture potentially important predictors that might not reach the conventional
p<.05 level, thereby allowing for a broader identification of variables relevant to clinical practice [
25]. Statistical significance was set at
p<.05.
RESULTS
1. ICU Readmission Rates and Post-discharge Outcomes
A total of 351 patients were included in the study. Among these, 17 patients (4.8%) were readmitted to the ICU within 7 days, forming the "readmission within 7 days" group. An additional 32 patients (9.1%) were readmitted after 7 days, comprising the "readmission after 7 days" group, while 302 patients were not readmitted and formed the non-readmission group.
Within 28 days after ICU discharge, 47 patients (13.4%) died. Specifically, 40 patients (13.2%) in the non-readmission group, 1 patient (5.9%) in the readmission within 7 days group, and 6 patients (18.8%) in the readmission after 7 days group died, although these differences were not statistically significant.
Overall, 77 patients (21.9%) obtained a POLST within 28 days of ICU discharge: 69 patients (22.8%) in the non-readmission group, 4 patients (23.5%) in the readmission within 7 days group, and 4 patients (12.5%) in the readmission after 7 days group.
2. Comparison of Characteristics among the Non-Readmission, Readmission within 7 days, and Readmission after 7 days groups
We compared the non-readmission group to both the readmission within 7 days and readmission after 7 days groups across demographic, clinical, ICU treatment, laboratory, physical, and environmental characteristics relevant to readmission.
1) Demographic and clinical characteristics
Table 1 presents a comparison of the demographic and clinical characteristics of older patients discharged from the ICU. Overall, 64.1% of the patients were male, and the mean age was 74.7±7.1 years. There were no statistically significant differences in demographic characteristics among the non-readmission, readmission within 7 days, and readmission after 7 days groups. Hypertension was the most common comorbidity (55.3%), followed by malignancy (42.5%). However, the readmission after 7 days group had a significantly higher incidence of hypertension (78.1%) compared to the non-readmission group (52.3%;
x2=7.78,
p=.005). Additionally, the average APACHE IV score on admission was significantly lower in the non-readmission group (84.6±21.8) compared to the readmission within 7 days group (94.9±20.5, t=-2.17,
p=.030) and the readmission after 7 days group (94.5±20.3, t=-2.84,
p=.005).
2) ICU treatment characteristics
Table 2 compares the ICU treatment characteristics among the groups. There were no statistically significant differences in treatment characteristics during the ICU stay or at discharge between the non-readmission group and the readmission within 7 days group. In contrast, significant differences were observed when comparing the non-readmission group with the readmission after 7 days group.
The median length of hospital stay was 13.0 days (interquartile range (IQR): 8.0~22.0) for the non-readmission group and 29.5 days (IQR: 17.3~57.8) for the readmission after 7 days group (Z=-4.89, p<.001). Similarly, the median ICU stay was 10.0 days (IQR: 6.0~16.0) in the non-readmission group versus 14.5 days (IQR: 10.0~37.5) in the readmission after 7 days group (Z=-3.49, p<.001). The duration of ventilator use also differed significantly, with a median of 8.0 days (IQR: 4.0~13.0) in the non-readmission group and 14.0 days (IQR: 6.3~33.3) in the readmission after 7 days group (Z= -2.94, p=.003). Additionally, patient characteristics at discharge-such as the GCS score (Z=-2.44, p=.014), the presence of a tracheostomy tube (x2=8.94, p=.003), receipt of oxygen therapy (x2=3.94, p=.047), and the presence of a drainage tube (x2=6.09, p=.014) differed significantly between the non-readmission group and the readmission after 7 days group.
3) Laboratory characteristics
When comparing the non-readmission group with the readmission within 7 days group, the total bilirubin level was significantly higher in the readmission within 7 days group [median 1.0 mg/dL (IQR: 0.6~2.4) vs. 0.6 mg/dL (IQR: 0.4~1.1); Z=-1.97, p=.049]. In contrast, the BUN level was significantly higher in the readmission within 7 days group (t=-2.12, p=.035). At ICU discharge, BUN levels remained significantly higher in the readmission within 7 days group compared to the non-readmission group (Z=-3.12, p=.002).
In the comparison between the non-readmission group and the readmission after 7 days group, discharge laboratory results showed that hemoglobin (t=4.69,
p<.001) and hematocrit levels (t=3.38,
p=.001) were significantly lower in the readmission after 7 days group (
Table 3).
4) Physical and environmental characteristics
Table 4 compares the physical and environmental characteristics across groups. No statistically significant differences were observed between the non-readmission group and the readmission within 7 days group.
However, when comparing the non-readmission group with the readmission after 7 days group regarding diet at discharge, 48.7% of patients in the non-readmission group received enteral feeding, 36.4% received oral feeding, and 14.9% were fasting (including those receiving total parenteral nutrition). In contrast, in the readmission after 7 days group, 74.2% received enteral feeding, 16.1% received oral feeding, and 9.7% were fasting-a difference that was statistically significant (x2=7.47, p=.024). Additionally, 33.1% of patients in the non-readmission group versus 62.5% in the readmission after 7 days group received oral or tracheal suction at discharge, which was also statistically significant (x2=10.86, p=.001).
3. Predictive Factors for ICU Readmission
Logistic regression analysis was performed using the demographic, clinical, ICU treatment, laboratory, physical, and environmental characteristics associated with readmission. The results are presented in
Table 5.
Variables that differed significantly between the non-readmission group and the readmission within 7 days group (p≤.10) were entered as independent variables. The analysis revealed that having diabetes was associated with a 3.32-fold higher risk of readmission within 7 days (odds ratio [OR]=3.32, 95% confidence interval [CI]=1.11~9.96, p=.032). Additionally, each 1 mg/dL increase in total bilirubin level at ICU admission was associated with a 1.18-fold increase in the risk of readmission (OR=1.18, 95% CI=1.01-1.37, p=.036). A 1-unit increase in the PaO2/FiO2 ratio at discharge was associated with a decreased risk of readmission (OR=0.99, 95% CI=0.99~1.00, p=.026), while a 1 mg/dL increase in the BUN level at discharge was associated with an increased risk (OR=1.02, 95% CI=1.01~1.04, p=.011).
The variables that revealed a significant difference between the non-readmission group and the readmission after 7 days group at a significance level of ≤0.10 were categorized as independent variables. For variables that significantly affected readmission after 7 days, patients with hypertension had a 3.06-fold higher risk of readmission than those without hypertension (OR=3.06, 95% CI=1.14~8.22, p=.027), and each additional day of ICU stay was associated with a 1.03-fold higher risk of readmission (OR=1.03, 95% CI=1.01~1.05, p=.002). Finally, each 1 g/dL increase in hemoglobin level at ICU discharge was associated with a 0.33-fold decrease in the risk of readmission (OR=0.33, 95% CI=0.12~0.92, p=.033).
DISCUSSION
In this study, the ICU readmission rate among older patients (aged≥65 years) was 4.8% within 7 days of discharge and 9.1% beyond 7 days. For comparison, a study in the United States reported a 9.1% readmission rate for trauma patients aged ≥65 years [
10], and a study in China found a 7-day readmission rate of 16.6% for older patients (aged ≥60 years) [
11]. These variations in readmission rates may be due to differences in healthcare systems, ICU admission and discharge policies, and the specific characteristics of the older populations studied. The higher readmission rate after 7 days (9.1%) compared to within 7 days (4.8%) suggests a sustained risk of readmission. This pattern may reflect the complex interplay of factors such as underlying comorbidities, functional decline, and post-intensive care syndrome (PICS), which can emerge over time in older patients. A recent study found that 49.7% of patients discharged from the ICU were positive in at least one PICS domain at 3 months post-discharge, including 31.6% with mental health issues, 25.1% with cognitive impairment, and 19.8% with physical disability [
26]. Additionally, only about half of older ICU survivors experienced functional recovery after a critical illness, with a median recovery time of 3 months [
27]. These findings underscore the need to consider long-term outcomes and the delayed onset of post-ICU complications when managing older patients after a critical illness.
In this study, diabetes, elevated total bilirubin levels on admission, lower PaO
2/FiO
2 ratios at discharge, and higher BUN levels at discharge were significantly associated with ICU readmission within 7 days. Specifically, the presence of diabetes increased the likelihood of early readmission by 3.32 times. This finding is consistent with previous studies reporting that hyperglycemia (e.g., >180 mg/dL) is a risk factor for readmission and negatively affects outcomes [
5]. Diabetes management in older patients is particularly challenging due to age-related changes in glucose metabolism, a higher risk of hypoglycemia, and the coexistence of multiple comorbidities [
28]. These factors likely contribute to the increased readmission risk observed, highlighting the need for thorough assessment and ongoing diabetes management following ICU discharge, with special attention to the unique needs of older patients. Furthermore, each 1 mg/dL increase in total bilirubin level at ICU admission was associated with a 1.18-fold increase in the risk of readmission within 7 days, a finding that aligns with previous research identifying chronic liver disease as a significant risk factor for ICU readmission [
5]. Older patients are particularly susceptible to liver dysfunction due to age-related changes in liver structure and function, as well as an increased vulnerability to drug-induced liver injury [
29]. This association underscores the importance of evaluating liver function, as patients with compromised liver function may require extended ICU management-a factor that should be carefully considered during discharge planning for older patients. A 1-unit increase in the PaO
2/FiO
2 ratio at ICU discharge was associated with a 0.99-fold decrease in the risk of readmission within 7 days. Given that older patients often have reduced lung function and decreased respiratory muscle strength, resulting in lower PaO
2/FiO
2 ratios, they may be more vulnerable to respiratory complications [
30]. This finding emphasizes the need for careful respiratory assessment and appropriate support in this population. Additionally, a 1 mg/dL increase in BUN level at discharge was associated with a 1.02-fold increased risk of readmission within 7 days. Notably, 17.4% of discharged patients received hemodialysis and 19.7% required continuous renal replacement therapy in the ICU, suggesting that-even with appropriate renal replacement therapy and clinical stabilization-elevated BUN levels remain a significant risk factor. This may be particularly relevant in older patients, who are more prone to acute kidney injury and have a higher prevalence of chronic kidney disease [
31]. The age-related decline in renal function likely contributes to elevated BUN levels and an increased risk of readmission.
In this study, we observed that hypertension, prolonged ICU stays, and lower hemoglobin levels at discharge were significantly associated with ICU readmission 7 days postdischarge. Patients with hypertension were 3.06 times more likely to be readmitted after 7 days compared to those without hypertension. These findings align with previous research in older patients with coronary artery disease, in which hypertension was identified as a major risk factor for unplanned readmissions [
11]. Chronic hypertension induces structural changes in blood vessels-such as arterial stiffening and endothelial dysfunction-that can impair organ perfusion and oxygenation [
32]. This vascular remodeling may affect multiple organ systems, especially in older patients with limited physiological reserves. Additionally, hypertension often coexists with comorbidities such as diabetes and chronic kidney disease, further complicating recovery and increasing vulnerability to acute illnesses [
33]. Thus, effective management of hypertension in ICU patients, particularly during transitions to general wards and after discharge, is crucial. We found that each additional day of ICU stay increased the risk of readmission after 7 days by a factor of 1.03, a result consistent with international studies showing that the length of the initial ICU stay influences readmission rates [
12]. Prolonged ICU stays are linked to various complications that can adversely affect patient outcomes, particularly among older individuals. Furthermore, extended ICU stays have been associated with a higher risk of developing Post-Intensive Care Syndrome (PICS), with one recent study reporting that up to 57.5% of ICU survivors experience at least one PICS-related problem within a year, disproportionately affecting older adults due to reduced physiological resilience and a higher prevalence of comorbidities [
26]. Accordingly, a more comprehensive approach to discharge planning and management is necessary for patients requiring long-term ICU treatment. Our study further identified that hemoglobin levels at discharge are a significant factor in predicting ICU readmission after 7 days. Specifically, for every 1 g/dL increase in hemoglobin, there was a corresponding 0.33-fold decrease in the risk of readmission. In critically ill patients-especially older adults-anemia affects more than just oxygen-carrying capacity; it also impairs nutritional status, immune function, and overall metabolic processes, creating a multifaceted vulnerability that compromises post-ICU recovery. For instance, one study reported a 19.3% prevalence of anemia in the Russian population aged 65 years and older, noting that anemia in the older adults was more strongly associated with decreased muscle strength and elevated CRP levels [
34]. Low hemoglobin levels can impair protein synthesis and muscle regeneration, which are critical for older patients with limited physiological reserves [
34]. Moreover, persistent anemia is accompanied by elevated inflammatory markers, and a prolonged inflammatory state after critical illness may weaken the immune system and increase susceptibility to infections that could necessitate readmission. These findings underscore that severe anemia (hemoglobin <7 mg/dL) can profoundly impact overall health and recovery capacity, emphasizing the importance of a comprehensive discharge plan that addresses both nutritional support and hemoglobin management.
This study confirms that the length of ICU stay, along with various physiological and clinical indicators, are critical risk factors in the discharge decision-making process. Overlooking these factors can increase the risk of readmission in older patients. A systematic assessment and monitoring of these indicators is essential for developing effective strategies to prevent readmission. Recent studies have highlighted structured interventions-such as checklists to identify necessary adjustments prior to discharge, multimodal communication tools, enhanced family engagement and preparation, post-discharge monitoring by critical care team members, periodic follow-up visits, and nurse-led transitional care programs that promote knowledge exchange among nursing teams-as effective strategies for optimizing discharge planning [
35]. Integrating these strategies can better support patients transitioning from the ICU to the general ward, thereby reducing the risk of readmission and promoting optimal recovery. Many healthcare organizations have well-defined ICU admission criteria, yet discharge criteria remain vague and are often adjusted based on medical judgment, bed availability, and staffing considerations. Therefore, discharge decisions should be grounded in a thorough evaluation of the patient's clinical condition, incorporating the identified risk factors for readmission. These findings underscore the importance of actively managing such risk factors during discharge planning and can serve as a basis for improving ICU discharge protocols, ultimately enhancing the quality of healthcare and patient safety for older adults.
However, this study has some limitations. It was conducted at a single institution, which may limit the generalizability of the findings. In addition, the retrospective design relied on data extracted from electronic medical records rather than direct patient observation, introducing potential limitations and errors in the data collection process. Furthermore, the scarcity of studies directly comparing the readmission rates of older patients with those of the general adult population limits our ability to fully understand the differences that a comparative analysis might reveal. Future research should validate these findings through multicenter studies and directly compare readmission rates between older and younger adult patients to clarify any differences. Moreover, the development and evaluation of targeted interventions to reduce readmissions will be an important area for future research. By identifying key factors associated with readmission rates both within and beyond 7 days post-discharge in older patients, this study lays the foundation for individualized management strategies aimed at preventing readmissions and improving quality of life. In clinical practice, these indicators can be used to assess the risk of readmission at ICU discharge and serve as objective criteria in making discharge decisions for older patients.
CONCLUSION
This study found that among older patients (aged ≥65 years), the unplanned ICU readmission rate was 4.8% within 7 days and 9.1% after 7 days of discharge. Specifically, the presence of diabetes, higher total bilirubin levels at ICU admission, lower PaO2/FiO2 ratios, and higher BUN levels at discharge were associated with an increased risk of readmission within 7 days. In contrast, hypertension, a longer ICU stay, and lower hemoglobin levels at ICU discharge were associated with a higher risk of readmission beyond 7 days. Accordingly, when making ICU discharge decisions for older patients admitted with internal medicine conditions, it is essential to assess for comorbid diabetes and hypertension, evaluate the length of ICU stay, and monitor total bilirubin at admission as well as BUN, PaO2/FiO2 ratio, and hemoglobin levels at discharge. These indicators should be used to objectively assess patient condition at ICU discharge to prevent readmission and serve as practical, field-applicable discharge criteria for identifying patients at high risk for ICU readmission.
-
CONFLICTS OF INTEREST
The authors declared no conflict of interest.
-
AUTHORSHIP
Study conception and/or design acquisition - CS, LH, HKS, and KSR; analysis- CS, LH, HKS, and KSR; interpretation of the data- SHC, HML; and drafting or critical revision of the manuscript for important intellectual content- CS, LH, HKS, and KSR; Literature search- CS and LH.
Table 1.Comparison of Demographic and Clinical Characteristics between the Readmission and Non-readmission Groups (N=351)
|
Variables |
Total (n=351) |
Non-readmission (n=302) |
Readmission within 7 days (n=17) |
t or x2 (p) |
Readmission after 7 days (n=32) |
t or x2 (p) |
|
n (%) or M±SD |
n (%) or M±SD |
n (%) or M±SD |
n (%) or M±SD |
|
Gender |
|
|
|
|
|
|
|
Male |
225 (64.1) |
197 (65.2) |
10 (58.8) |
0.29 (.590) |
18 (56.2) |
1.02 (.313) |
|
Female |
126 (35.9) |
105 (34.8) |
7 (41.2) |
|
14 (43.8) |
|
|
Age (year) |
74.7±7.1 |
74.5±7.0 |
74.8±7.5 |
-0.17 (.861) |
75.9±8.1 |
-1.02 (.306) |
|
Admission source |
|
|
|
|
|
|
|
Ward |
181 (51.6) |
155 (51.3) |
9 (52.9) |
0.07 (.966) |
17 (53.1) |
0.14 (.934) |
|
Emergency room |
169 (48.1) |
146 (48.4) |
8 (47.1) |
|
15 (46.9) |
|
|
Outpatient department |
1 (0.3) |
1 (0.3) |
0 (0.0) |
|
0 (0.0) |
|
|
Main diagnosis of hospital admission |
|
|
|
|
|
|
|
Pulmonary disease |
147 (41.9) |
129 (42.7) |
9 (52.9) |
8.65 (.372) |
9 (28.1) |
7.35 (.499) |
|
Immune disease |
49 (14.0) |
37 (12.3) |
5 (29.4) |
|
7 (21.9) |
|
|
Gastrointestinal disease |
42 (12.0) |
36 (11.9) |
1 (5.9) |
|
5 (15.6) |
|
|
Hematologic disease |
33 (9.4) |
28 (9.3) |
2 (11.8) |
|
3 (9.4) |
|
|
Renal disease†
|
15 (4.3) |
14 (4.6) |
0 (0.0) |
|
1 (3.1) |
|
|
Cerebrovascular disease |
11 (3.1) |
10 (3.3) |
0 (0.0) |
|
1 (3.1) |
|
|
Cardiovascular disease |
10 (2.8) |
10 (3.3) |
0 (0.0) |
|
0 (0.0) |
|
|
Endocrine disease |
8 (2.3) |
6 (2.0) |
0 (0.0) |
|
2 (6.3) |
|
|
Others |
36 (10.2) |
32 (10.6) |
0 (0.0) |
|
4 (12.5) |
|
|
Diagnosis upon ICU admission |
|
|
|
|
|
|
|
Respiratory |
207 (59.0) |
184 (60.9) |
9 (52.9) |
5.76 (.330) |
14 (43.8) |
11.48 (.043) |
|
Sepsis |
68 (19.4) |
52 (17.2) |
4 (23.5) |
|
12 (37.5) |
|
|
Cardiovascular |
24 (6.8) |
20 (6.6) |
3 (17.7) |
|
1 (3.1) |
|
|
Neurological |
22 (6.3) |
19 (6.3) |
0 (0.0) |
|
3 (9.4) |
|
|
Renal |
18 (5.1) |
18 (6.0) |
0 (0.0) |
|
0 (0.0) |
|
|
Gastrointestinal |
12 (3.4) |
9 (3.0) |
1 (5.9) |
|
2 (6.2) |
|
|
Comorbidity‡
|
|
|
|
|
|
|
|
Hypertension |
194 (55.3) |
158 (52.3) |
11 (64.7) |
0.99 (.319) |
25 (78.1) |
7.78 (.005) |
|
Malignancies |
149 (42.5) |
131 (43.4) |
6 (35.3) |
0.43 (.512) |
12 (37.5) |
0.41 (.523) |
|
Diabetes mellitus |
129 (36.8) |
108 (35.8) |
10 (58.8) |
3.67 (.055) |
11 (34.4) |
0.02 (.876) |
|
Cardiovascular disease |
107 (30.5) |
92 (30.5) |
7 (41.2) |
0.86 (.353) |
8 (25.0) |
0.41 (.521) |
|
Pulmonary disease |
87 (24.8) |
72 (23.8) |
5 (29.4) |
0.27 (.569)†
|
10 (31.3) |
0.86 (.354) |
|
Cerebrovascular disease |
62 (17.7) |
51 (16.9) |
2 (11.8) |
0.31 (.748)†
|
9 (28.1) |
2.48 (.115) |
|
Renal disease†
|
56 (16.0) |
46 (15.2) |
5 (29.4) |
2.41 (.164)†
|
5 (15.6) |
0.00 (＞.999)†
|
|
Gastrointestinal disease |
40 (11.4) |
37 (12.3) |
2 (11.8) |
0.00 (＞.999)†
|
1 (3.1) |
2.39 (.150)†
|
|
Others†
|
76 (21.7) |
66 (21.9) |
5 (29.4) |
0.53 (.548)†
|
5 (15.6) |
0.67 (.413) |
|
Number of comorbidities |
2.6±1.4 |
2.5±1.4 |
3.1±1.5 |
-1.75 (.082) |
2.7±1.2 |
-0.67 (.505) |
|
Charlson comorbidity index |
12.7±5.1 |
12.6±5.0 |
11.9±5.4 |
0.57 (.566) |
13.7±5.5 |
-1.15 (.249) |
|
APACHE IV score at admission |
84.8±21.8 |
83.2±21.6 |
94.9±20.5 |
-2.17 (.030) |
94.5±20.3 |
-2.84 (.005) |
Table 2.Characteristics related to ICU Treatment between the Readmission and Non-readmission Groups (N=351)
|
Variables |
Total (n=351) |
Non-readmission (n=334) |
Readmission within 7 days (n=17) |
t or Z or x2 (p) |
Readmission after 7 days (n=17) |
t or Z or x2 (p) |
|
M±SD or Median (IQR) or n (%) |
M±SD or Median (IQR) or n (%) |
M±SD or Median (IQR) or n (%) |
M±SD or Median (IQR) or n (%) |
|
Hospital length of stay (days)‡
|
14.0 (8.0~23.0) |
13.0 (8.0~22.0) |
12.0 (6.5~25.5) |
-0.25 (.799) |
29.5 (17.3~57.8) |
-4.89 (＜.001) |
|
ICU length of stay (days)‡
|
10.0 (6.0~17.0) |
10.0 (6.0~16.0) |
9.0 (5.0~17.0) |
-0.01 (.995) |
14.5 (10.0~37.5) |
-3.49 (＜.001) |
|
During ICU |
|
|
|
|
|
|
|
Use of devices |
|
|
|
|
|
|
|
Ventilator |
283 (80.6) |
242 (80.1) |
13 (76.5) |
0.14 (.756)†
|
28 (87.5) |
1.01 (.314) |
|
CRRT |
69 (19.7) |
61 (20.2) |
3 (17.6) |
0.07 (＞.999)†
|
5 (15.6) |
0.38 (.537) |
|
HD |
61 (17.4) |
53 (17.5) |
3 (17.6) |
0.00 (＞.999)†
|
5 (15.6) |
0.08 (.785) |
|
ECMO |
20 (5.7) |
16 (5.3) |
1 (5.9) |
0.01 (＞.999)†
|
3 (9.4) |
0.90 (.409)†
|
|
Ventilator duration (days)‡
|
8.0 (4.0~14.0) |
8.0 (4.0~13.0) |
9.0 (3.5~17.0) |
-0.28 (.777) |
14.0 (6.3~33.3) |
-2.94 (.003) |
|
History of cardiac arrest in ICU |
13 (3.7) |
9 (3.0) |
1 (5.9) |
0.45 (.426)†
|
3 (9.4) |
3.42 (.097)†
|
|
Positive signal of culture |
269 (76.6) |
228 (75.5) |
15 (88.2) |
1.44 (.379)†
|
26 (81.3) |
0.53 (.468) |
|
Multidrug-resistant bacteria, Positive |
102 (29.1) |
84 (27.8) |
4 (23.5) |
0.15 (＞.999)†
|
14 (43.8) |
3.54 (.060) |
|
At discharge |
|
|
|
|
|
|
|
NEWS |
4.5±2.3 |
4.5±2.3 |
3.9±2.2 |
0.97 (.330) |
5.0±2.3 |
-1.07 (.288) |
|
GCS score†
|
13.0 (10.0~15.0) |
13.0 (10.0~15.0) |
14.0 (10.5~15.0) |
-1.00 (.315) |
10.0 (9.0~13.8) |
-2.44 (.014) |
|
Tracheostomy |
96 (27.4) |
76 (25.2) |
4 (23.5) |
0.02 (＞.999)†
|
16 (50.0) |
8.94 (.003) |
|
Oxygen therapy§
|
284 (80.9) |
239 (79.1) |
15 (88.2) |
0.82 (.540)†
|
30 (93.8) |
3.94 (.047) |
|
HFNC |
91 (25.9) |
77 (25.5) |
7 (41.2) |
2.04 (.163)†
|
7 (21.9) |
0.20 (.653) |
|
BiPAP |
26 (7.4) |
21 (7.0) |
1 (5.9) |
0.03 (＞.999)†
|
4 (12.5) |
1.29 (.280)†
|
|
FiO2 (%) |
36.7±13.0 |
36.7±13.3 |
40.4±12.1 |
-1.11 (.269) |
34.8±9.8 |
1.04 (.305) |
|
Sedatives |
54 (15.7) |
49 (16.6) |
1 (5.9) |
1.20 (.483)†
|
4 (12.5) |
0.35 (.554) |
|
Vasopressors |
22 (6.3) |
18 (6.0) |
2 (11.8) |
0.92 (.289)†
|
2 (6.5) |
0.01 (＞.999)†
|
|
Drainage tube |
108 (30.8) |
87 (28.8) |
5 (29.4) |
0.00 (＞.999)†
|
16 (50.0) |
6.09 (.014) |
|
Urine output, mL/day |
1,923.6±1,180.6 |
1,938.3±1,206.0 |
1,754.5±1,001.7 |
0.62 (.538) |
1,875.2±1,038.4 |
0.29 (.776) |
Table 3.Comparison of Laboratory Characteristics between the Readmission and Non-readmission Groups (N=351)
|
Variables |
Total (n=351) |
Non-readmission (n=302) |
Readmission within 7 days (n=17) |
t or Z or x2 (p) |
Readmission after 7 days (n=32) |
t or Z or x2 (p) |
|
M±SD or Median (IQR) |
M±SD or Median (IQR) |
M±SD or Median (IQR) |
M±SD or Median (IQR) |
|
At admission |
|
|
|
|
|
|
|
Respiratory system |
|
|
|
|
|
|
|
PaO2/FiO2 ratio |
226.4±133.5 |
227.0±133.0 |
182.8±94.1 |
1.35 (.178) |
244.0±154.0 |
-0.68 (.498) |
|
Hematologic system |
|
|
|
|
|
|
|
Hemoglobin, g/dL |
10.1±2.5 |
10.1±2.5 |
10.2±2.8 |
-0.15 (.881) |
9.7±2.5 |
0.75 (.457) |
|
Hct, % |
30.6±7.5 |
30.6±7.5 |
30.5±8.5 |
0.07 (.942) |
29.8±7.7 |
0.58 (.565) |
|
Hepatobiliary system |
|
|
|
|
|
|
|
AST, IU/L†
|
32.0 (22.0~57.0) |
31.0 (21.0~58.3) |
29.0 (27.0~57.5) |
-0.47 (.638) |
41.0 (22.3~56.8) |
-0.70 (.485) |
|
ALT, IU/L†
|
20.0 (12.0~38.0) |
19.0 (12.0~38.0) |
20.0 (15.0~44.5) |
-0.53 (.593) |
20.5 (12.0~34.3) |
-0.02 (.985) |
|
Total bilirubin, g/dL†
|
0.7 (0.4~1.2) |
0.6 (0.4~1.1) |
1.0 (0.6~2.4) |
-1.97 (.049) |
0.6 (0.4~1.1) |
-0.23 (.818) |
|
Albumin, g/dL |
2.5±0.6 |
2.5±0.6 |
2.4±0.4 |
0.72 (.471) |
2.5±0.6 |
0.18 (.854) |
|
Renal system |
|
|
|
|
|
|
|
BUN, mg/dL |
36.4±26.2 |
35.4±25.4 |
48.9±27.2 |
-2.12 (.035) |
39.0±31.6 |
-0.74 (.461) |
|
Creatinine, mg/dL†
|
1.1 (0.7~2.2) |
1.1 (0.7~2.1) |
2.0 (0.9~3.5) |
-1.79 (.073) |
1.0 (0.7~2.5) |
-0.28 (.778) |
|
GFR, mL/min/1.73 m2
|
55.5±30.3 |
56.2±30.0 |
43.1±32.2 |
1.74 (.083) |
55.5±31.1 |
0.12 (.905) |
|
Sodium, mmol/L |
136.4±6.3 |
136.4±6.4 |
136.5±6.0 |
-0.02 (.987) |
136.4±5.2 |
0.01 (.996) |
|
Potassium, mmol/L |
4.2±0.8 |
4.2±0.8 |
4.0±0.9 |
1.32 (.189) |
4.3±1.0 |
-0.54 (.589) |
|
Endocrine system |
|
|
|
|
|
|
|
HbA1C, %†
|
6.5 (5.8~7.5) |
6.5 (5.8~7.4) |
7.2 (5.9~8.8) |
-0.79 (.431) |
6.4 (5.4~10.5) |
0.00 (＞.999) |
|
Immune system |
|
|
|
|
|
|
|
Neutrophil, %†
|
85.9 (76.6~91.1) |
85.9 (76.5~91.1) |
91.3 (76.8~92.9) |
-1.85 (.064) |
85.4 (77.4~88.6) |
-0.77 (.440) |
|
CRP, mg/dL |
12.6±10.7 |
12.1±10.4 |
16.9±13.0 |
-1.83 (.069) |
14.4±11.8 |
-1.15 (.250) |
|
At discharge |
|
|
|
|
|
|
|
Respiratory system |
|
|
|
|
|
|
|
PaO2/FiO2 ratio |
295.2±108.8 |
298.3±110.4 |
245.6±95.3 |
1.87 (.062) |
291.0±95.6 |
0.35 (.728) |
|
Hematologic system |
|
|
|
|
|
|
|
Hemoglobin, g/dL |
9.2±1.7 |
9.3±1.7 |
9.2±2.1 |
0.35 (.724) |
8.5±0.9 |
4.69 (＜.001) |
|
Hematocrit, % |
28.0±4.8 |
28.2±4.9 |
27.5±6.1 |
0.62 (.537) |
26.3±2.9 |
3.38 (.001) |
|
Hepatobiliary system |
|
|
|
|
|
|
|
AST, IU/L†
|
25.0 (19.0~38.0) |
25.0 (18.8~37.3) |
26.0 (20.5~47.0) |
-0.56 (.573) |
33.5 (19.0~46.5) |
-1.10 (.273) |
|
ALT, IU/L†
|
23.0 (14.0~39.0) |
22.5 (14.0~38.3) |
21.0 (17.0~47.0) |
-0.69 (.491) |
24.5 (13.0~48.8) |
-0.08 (.934) |
|
Total bilirubin, mg/dL†
|
0.6 (0.4~1.1) |
0.6 (0.4~1.1) |
0.7 (0.4~1.1) |
-0.06 (.949) |
0.6 (0.4~1.1) |
-0.35 (.724) |
|
Albumin, g/dL |
2.5±0.4 |
2.5±0.4 |
2.6±0.4 |
-0.58 (.563) |
2.5±0.4 |
-0.02 (.986) |
|
Renal system |
|
|
|
|
|
|
|
BUN, mg/dL†
|
27.0 (19.0~40.0) |
26.0 (18.0~39.0) |
41.0 (26.5~65.0) |
-3.12 (.002) |
29.5 (23.0~47.0) |
-1.68 (.093) |
|
GFR, ml/min/1.73m2
|
66.2±28.2 |
66.7±28.1 |
54.6±32.7 |
1.71 (.087) |
67.1±26.4 |
-0.06 (.951) |
|
Creatinine, mg/dL†
|
0.8 (0.6~1.5) |
0.8 (0.6~1.4) |
1.7 (0.6~2.3) |
-1.37 (.172) |
0.8 (0.6~1.3) |
-0.12 (.908) |
|
Sodium, mmol/L |
137.5±4.9 |
137.4±4.7 |
137.8±8.4 |
-0.19 (.851) |
138.3±4.3 |
-1.05 (.293) |
|
Potassium, mmol/L†
|
3.9 (3.5~4.2) |
3.9 (3.6~4.2) |
4.1 (3.4~4.3) |
-0.50 (.619) |
3.9 (3.5~4.4) |
-0.30 (.767) |
|
Endocrine system |
|
|
|
|
|
|
|
HbA1C, %†
|
5.8 (5.4~6.9) |
5.8 (5.5~7.0) |
5.9 (5.2~6.7) |
-0.19 (.846) |
5.8 (5.1~9.2) |
-0.19 (.846) |
|
Immune system |
|
|
|
|
|
|
|
Neutrophil, %†
|
80.7 (70.5~86.8) |
80.2 (70.0~86.6) |
85.0 (77.2~89.8) |
-1.35 (.176) |
81.7 (73.2~87.5) |
-0.33 (.740) |
|
CRP, mg/dL†
|
3.2 (1.1~6.8) |
3.2 (1.1~6.8) |
3.7 (1.3~7.7) |
-0.56 (.576) |
3.7 (1.4~6.5) |
-0.12 (.903) |
Table 4.Comparison of Physical and Environmental Characteristics between the Readmission and Non-readmission Groups (N=351)
|
Variables |
Total (n=351) |
Non-readmission (n=302) |
Readmission within 7 days (n=17) |
t or Z or x2 (p) |
Readmission after 7 days (n=32) |
t or Z or x2 (p) |
|
M±SD or Median (IQR) or n (%) |
M±SD or Median (IQR) or n (%) |
M±SD or Median (IQR) or n (%) |
M±SD or Median (IQR) or n (%) |
|
RASS at discharge |
-0.7±1.2 |
-0.6±1.2 |
-0.9±1.1 |
1.13 (.261) |
-1.0±1.5 |
1.34 (.188) |
|
CAM-ICU at discharge, Positive |
126 (39.3) |
108 (38.7) |
4 (23.6) |
0.58 (.446) |
14 (50.0) |
1.36 (.244) |
|
Pressure injury at discharge |
122 (34.8) |
105 (34.8) |
5 (29.4) |
0.20 (.651) |
12 (37.5) |
0.10 (.758) |
|
Braden scale at discharge |
15.6±2.2 |
15.7±2.2 |
15.6±2.2 |
0.03 (.974) |
15.0±1.9 |
1.55 (.122) |
|
Pain at discharge |
30 (8.5) |
28 (9.3) |
1 (5.9) |
0.22 (＞.999)†
|
1 (3.1) |
1.38 (.336)†
|
|
CNPS at discharge‡
|
0.0 (0.0~0.0) |
0.0 (0.0~0.0) |
0.0 (0.0~0.0) |
-0.89 (.373) |
0.0 (0.0~0.0) |
-1.21 (.225) |
|
NRS at discharge‡
|
0.0 (0.0~0.0) |
0.0 (0.0~0.0) |
0.0 (0.0~0.0) |
-0.24 (.811) |
0.0 (0.0~0.0) |
-0.61 (.541) |
|
Diet at discharge |
|
|
|
|
|
|
|
Enteral feeding |
177 (50.6) |
147 (48.7) |
7 (41.2) |
2.57 (.277) |
23 (74.2) |
7.47 (.024) |
|
Oral feeding |
120 (34.3) |
110 (36.4) |
5 (29.4) |
|
5 (16.1) |
|
|
NPO |
53 (15.1) |
45 (14.9) |
5 (29.4) |
|
3 (9.7) |
|
|
Suction (oral or tracheal) at discharge |
125 (35.6) |
100 (33.1) |
5 (29.4) |
0.10 (.752) |
20 (62.5) |
10.86 (.001) |
|
Use of physical restraint at discharge |
72 (20.5) |
63 (20.9) |
1 (5.9) |
2.25 (.211)†
|
8 (25.0) |
0.30 (.586) |
Table 5.Multiple Multivariate Logistic Regression Model for Readmission within 7 Days (A) and Readmission after 7 Days (B)
|
Variables |
Univariate analysis |
Multivariate analysis |
|
OR |
95% CI |
p
|
OR |
95% CI |
p
|
|
A. Readmission within 7 days |
|
|
|
|
|
|
|
Diabetes mellitus |
2.57 |
0.95~6.93 |
.063 |
3.32 |
1.11~9.96 |
.032 |
|
Number of comorbidities |
1.34 |
0.96~1.86 |
.085 |
1.10 |
0.65~1.85 |
.735 |
|
APACHE IV score at admission |
1.02 |
1.00~1.04 |
.034 |
1.01 |
0.98~1.03 |
.731 |
|
Total bilirubin at admission |
1.15 |
1.00~1.32 |
.051 |
1.18 |
1.01~1.37 |
.036 |
|
BUN at admission |
1.02 |
1.00~1.03 |
.040 |
1.01 |
0.98~1.05 |
.425 |
|
GFR at admission |
0.99 |
0.97~1.00 |
.089 |
0.98 |
0.95~1.02 |
.452 |
|
Creatinine at admission |
1.08 |
0.88~1.32 |
.460 |
0.75 |
0.41~1.34 |
.328 |
|
Neutrophil at admission |
1.03 |
0.98~1.08 |
.189 |
1.02 |
0.97~1.07 |
.397 |
|
CRP at admission |
1.04 |
1.00~1.08 |
.074 |
1.03 |
0.98~1.09 |
.232 |
|
PaO2/FiO2 ratio at discharge |
0.99 |
0.99~1.00 |
.062 |
0.99 |
0.99~1.00 |
.026 |
|
BUN at discharge |
1.02 |
1.00~1.03 |
.013 |
1.02 |
1.01~1.04 |
.011 |
|
GFR at discharge |
0.99 |
0.97~1.00 |
.094 |
1.01 |
0.98~1.05 |
.570 |
|
Nagelkerke=.173, Hosmer-Lemeshow test: x2=3.92, p=.864 |
|
B. Readmission after 7 days |
|
|
|
|
|
|
|
Diagnosis upon ICU admission_respiratory (ref) |
|
|
0.138 |
|
|
.528 |
|
Sepsis |
3.03 |
1.32~6.96 |
0.009 |
2.22 |
0.70~7.10 |
.177 |
|
Cardiovascular |
0.66 |
0.08~5.26 |
0.692 |
0.73 |
0.04~11.92 |
.827 |
|
Neurologic |
2.08 |
0.55~7.87 |
0.283 |
4.02 |
0.77~21.08 |
.10 |
|
Renal |
0.00 |
0.00~0.00 |
0.998 |
0.00 |
0.00~0.00 |
.999 |
|
Gastrointestinal |
2.92 |
0.57~14.84 |
0.196 |
1.01 |
0.08~12.64 |
.995 |
|
Hypertension |
3.25 |
1.37~7.75 |
.008 |
3.06 |
1.14~8.22 |
.027 |
|
APACHE IV score at admission |
1.02 |
1.01~1.04 |
.006 |
1.01 |
0.98~1.03 |
.660 |
|
Hospital length of stay (days) |
1.02 |
1.00~1.03 |
.036 |
1.00 |
0.98~1.01 |
.430 |
|
ICU length of stay (days) |
1.04 |
1.02~1.06 |
＜.001 |
1.03 |
1.01~1.05 |
.002 |
|
Ventilator duration (days) |
1.01 |
1.00~1.02 |
.104 |
1.00 |
0.98~1.02 |
.728 |
|
History of cardiac arrest in ICU |
3.37 |
0.86~13.14 |
.080 |
1.81 |
0.19~16.96 |
.605 |
|
Multidrug-resistant bacteria, Positive |
2.02 |
0.96~4.24 |
.064 |
0.86 |
0.31~2.40 |
.774 |
|
GCS score at discharge |
0.89 |
0.79~1.00 |
.041 |
1.01 |
0.79~1.28 |
.944 |
|
Tracheostomy at discharge |
2.97 |
1.42~6.23 |
.004 |
0.63 |
0.10~3.89 |
.623 |
|
Oxygen therapy at discharge |
3.95 |
0.92~16.99 |
.065 |
4.15 |
0.45~37.90 |
.207 |
|
Drainage tube at discharge |
2.47 |
1.18~5.16 |
.016 |
2.26 |
0.86~5.94 |
.098 |
|
Hemoglobin at discharge |
0.64 |
0.46~0.88 |
.006 |
0.33 |
0.12~0.92 |
.033 |
|
Hematocrit at discharge |
0.90 |
0.81~0.99 |
.027 |
1.36 |
0.97~1.90 |
.071 |
|
BUN at discharge |
1.01 |
1.00~1.02 |
.208 |
1.00 |
0.98~1.03 |
.707 |
|
Diet _ Enteral feeding (ref.) |
|
|
.033 |
|
|
.245 |
|
Oral feeding |
0.29 |
0.11~0.79 |
.015 |
0.31 |
0.06~1.54 |
.152 |
|
NPO |
0.43 |
0.12~1.49 |
.181 |
0.45 |
0.09~2.21 |
.322 |
|
Suction (oral or tracheal) at discharge |
3.37 |
1.58~7.16 |
.002 |
1.33 |
0.28~6.35 |
.719 |
|
Nagelkerke=.178, Hosmer-Lemeshow test: x2=5.07 p=.750 |
REFERENCES
- 1. Statistics Korea. Population projections for Korea (2020~2070) [Internet]. Daejeon: Statistics Korea; 2020. [cited 2025 January 28]. Available from: https://kostat.go.kr/board.es?mid=a20108080000&bid=11748&act=view&list_no=416209
- 2. Jones A, Toft-Petersen AP, Shankar-Hari M, Harrison DA, Rowan KM. Demographic shifts, case mix, activity, and outcome for elderly patients admitted to adult general ICUs in England, Wales, and Northern Ireland. Critical Care Medicine. 2020;48(4):466-74. https://doi.org/10.1097/ccm.0000000000004211
- 3. Tavares JL, Cohen MA, Silberman S, Popham L. Chronic inequities: measuring disease cost burden older adults in the U.S.A.: health and retirement study analysis [Internet]. National Council on Aging; [cited 2025 January 28]. Available from: https://assets-us-01.kc-usercontent.com/ffacfe7d-10b6-0083-2632-604077fd4eca/34789367-444d-46d0-b38c-3c1348b1ea52/2022-Chronic_Inequities_Measuring_Burden.pdf
- 4. The Korean Society of Critical Care Medicine. 2022 Guidelines for intensive care unit admission, discharge, and triage [Internet]. The Korean Society of Critical Care Medicine; [cited 2025 January 28]. Available from: https://www.ksccm.org/html/?pmode=BBBS0006700004&smode=view&seq=2755
- 5. Laytin AD, Sims CA. Risk factors for unplanned ICU readmission among trauma patients: age matters. Critical Care Explorations. 2022;4(10):e0778. https://doi.org/10.1097/cce.0000000000000778
- 6. Vollam S, Dutton S, Lamb S, Petrinic T, Young JD, Watkinson P. Out-of-hours discharge from intensive care, in-hospital mortality and intensive care readmission rates: a systematic review and meta-analysis. Intensive Care Medicine. 2018;44(7):1115-29. https://doi.org/10.1007/s00134-018-5245-2
- 7. Health Insurance Review & Assessment Service. Health insurance review & assessment service data 2019 [Internet]. HIRA; [cited 2025 January 28]. Available from: https://www.hira.or.kr/bbsDummy.do?pgmid=HIRAA020045020000/&brdScnBltNo=4&brdBltNo=2312&pageIndex=1&pageIndex2=1
- 8. Woldhek AL, Rijkenberg S, Bosman RJ, van der Voort PH. Readmission of ICU patients: a quality indicator? Journal of Critical Care. 2017;38:328-34. https://doi.org/10.1016/j.jcrc.2016.12.001
- 9. Son YJ, Kim GO, Lee YM, Oh M, Choi J. Predictors of early and late unplanned intensive care unit readmission: a retrospective cohort study. Journal of Nursing Scholarship. 2021;53(4):400-7. https://doi.org/10.1111/jnu.12657
- 10. Agoubi LL, Reimel BA, Maine RG, O'Connell KM, Maier RV, McIntyre LK. Intensive care unit readmission in injured older adults: modifiable risk factors and implications. Journal of Trauma and Acute Care Surgery. 2024;96(5):813-9. https://doi.org/10.1097/ta.0000000000004203
- 11. Song X, Tong Y, Luo Y, Chang H, Gao G, Dong Z, et al. Predicting 7-day unplanned readmission in elderly patients with coronary heart disease using machine learning. Frontiers in Cardiovascular Medicine. 2023;10:1190038. https://doi.org/10.3389/fcvm.2023.1190038
- 12. Shi K, Ho V, Song JJ, Bechler K, J HC. Predicting unplanned 7-day intensive care unit readmissions with machine learning models for improved discharge risk assessment. AMIA Joint Summits on Translational Science Proceedings. 2022;2022:446-55.
- 13. Fabes J, Seligman W, Barrett C, McKechnie S, Griffiths J. Does the implementation of a novel intensive care discharge risk score and nurse-led inpatient review tool improve outcome? a prospective cohort study in two intensive care units in the UK. BMJ Open. 2017;7(12):e018322. https://doi.org/10.1136/bmjopen-2017-018322
- 14. Chin-Yee N, D'Egidio G, Thavorn K, Heyland D, Kyeremanteng K. Cost analysis of the very elderly admitted to intensive care units. Critical Care. 2017;21(1):109. https://doi.org/10.1186/s13054-017-1689-y
- 15. Jo YS, Lee YJ, Park JS, Yoon HI, Lee JH, Lee CT, et al. Readmission to medical intensive care units: risk factors and prediction. Yonsei Medical Journal. 2015;56(2):543-9. https://doi.org/10.3349/ymj.2015.56.2.543
- 16. Jeong BH, Na SJ, Lee DS, Chung CR, Suh GY, Jeon K. Correction: Readmission and hospital mortality after ICU discharge of critically ill cancer patients. PLoS One. 2019;14(6):e0218196. https://doi.org/10.1371/journal.pone.0218196
- 17. Wang CH, Lin HC, Chang YC, Maa SH, Wang JS, Tang WR. Predictive factors of in-hospital mortality in ventilated intensive care unit: a prospective cohort study. Medicine (Baltimore). 2017;96(51):e9165. https://doi.org/10.1097/md.0000000000009165
- 18. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. Journal of Chronic Diseases. 1987;40(5):373-83. https://doi.org/10.1016/0021-9681(87)90171-8
- 19. Zimmerman JE, Kramer AA, McNair DS, Malila FM. Acute physiology and chronic health evaluation (APACHE) IV: hospital mortality assessment for today's critically ill patients. Critical Care Medicine. 2006;34(5):1297-310. https://doi.org/10.1097/01.Ccm.0000215112.84523.F0
- 20. Smith GB, Prytherch DR, Meredith P, Schmidt PE, Featherstone PI. The ability of the national early warning score (NEWS) to discriminate patients at risk of early cardiac arrest, unanticipated intensive care unit admission, and death. Resuscitation. 2013;84(4):465-70. https://doi.org/10.1016/j.resuscitation.2012.12.016
- 21. Bastos PG, Sun X, Wagner DP, Wu AW, Knaus WA. Glasgow coma scale score in the evaluation of outcome in the intensive care unit: findings from the acute physiology and chronic health evaluation III study. Critical Care Medicine. 1993;21(10):1459-65. https://doi.org/10.1097/00003246-199310000-00012
- 22. Sessler CN, Gosnell MS, Grap MJ, Brophy GM, O'Neal PV, Keane KA, et al. The richmond agitation-sedation scale: validity and reliability in adult intensive care unit patients. American Journal of Respiratory and Critical Care Medicine. 2002;166(10):1338-44. https://doi.org/10.1164/rccm.2107138
- 23. Ely EW, Inouye SK, Bernard GR, Gordon S, Francis J, May L, et al. Delirium in mechanically ventilated patients: validity and reliability of the confusion assessment method for the intensive care unit (CAM-ICU). Journal of the American Medical Association. 2001;286(21):2703-10. https://doi.org/10.1001/jama.286.21.2703
- 24. Choi EH, Kim JH, Choi KO, Yoo JS, Kim MS, Kim PJ, et al. Development of a pain assessment tool for critically ill patients. Journal of Korean Clinical Nursing Research. 2012;18(1):136-48. https://doi.org/10.22650/JKCNR.2012.18.1.136
- 25. Bursac Z, Gauss CH, Williams DK, Hosmer DW. Purposeful selection of variables in logistic regression. Source Code for Biology and Medicine. 2008;3:17. https://doi.org/10.1186/1751-0473-3-17
- 26. Kang J, Lee MH. Incidence rate and risk factors for post-intensive care syndrome subtypes among critical care survivors three months after discharge: a prospective cohort study. Intensive and Critical Care Nursing. 2024;81:103605. https://doi.org/10.1016/j.iccn.2023.103605
- 27. Ferrante LE, Pisani MA, Murphy TE, Gahbauer EA, Leo-Summers LS, Gill TM. Factors associated with functional recovery among older intensive care unit survivors. American Journal of Respiratory and Critical Care Medicine. 2016;194(3):299-307. https://doi.org/10.1164/rccm.201506-1256OC
- 28. ElSayed NA, Aleppo G, Aroda VR, Bannuru RR, Brown FM, Bruemmer D, et al. 13. Older adults: standards of care in diabetes-2023. Diabetes Care. 2023;46(Suppl 1):S216-29. https://doi.org/10.2337/dc23-S013
- 29. Radonjić T, Dukić M, Jovanović I, Zdravković M, Mandić O, Popadić V, et al. Aging of liver in its different diseases. International Journal of Molecular Sciences. 2022;23(21):https://doi.org/10.3390/ijms232113085
- 30. Schneider JL, Rowe JH, Garcia-de-Alba C, Kim CF, Sharpe AH, Haigis MC. The aging lung: physiology, disease, and immunity. Cell. 2021;184(8):1990-2019. https://doi.org/10.1016/j.cell.2021.03.005
- 31. Tam-Tham H, Ravani P, Zhang J, Weaver RG, Quinn RR, James MT, et al. Association of initiation of dialysis with hospital length of stay and intensity of care in older adults with kidney failure. JAMA Network Open. 2020;3(2):e200222. https://doi.org/10.1001/jamanetworkopen.2020.0222
- 32. Guzik TJ, Nosalski R, Maffia P, Drummond GR. Immune and inflammatory mechanisms in hypertension. Nature Reviews Cardiology. 2024;21(6):396-416. https://doi.org/10.1038/s41569-023-00964-1
- 33. Huang Y, Meng L, Liu C, Liu S, Tao L, Zhang S, et al. Global burden of disease attributable to high systolic blood pressure in older adults, 1990-2019: an analysis for the global burden of disease study 2019. European Journal of Preventive Cardiology. 2023;30(10):917-27. https://doi.org/10.1093/eurjpc/zwac273
- 34. Turusheva A, Frolova E, Korystina E, Zelenukha D, Tadjibaev P, Gurina N, et al. Association between anemia, physical performance, dependency and mortality in older adults in the North-west region of Russia. Journal of Aging Research & Clinical Practice. 2015;4(1):34-43.
- 35. Österlind J, Gerhardsson J, Myrberg T. Critical care transition programs on readmission or death: a systematic review and meta-analysis. Acta Anaesthesiologica Scandinavica. 2020;64(7):870-83. https://doi.org/10.1111/aas.13591
E-SUBMISSION
