A nationwide 1-day point-prevalence study on rehabilitation practices for intensive care unit patients in South Korea

Abstract

Background

Early rehabilitation is important in critical care, but nationwide data on rehabilitation practices in Korean intensive care units (ICUs) are lacking. This study reports the findings of a 1-day point-prevalence survey to assess current ICU rehabilitation practices and associated ICU characteristics.

Methods

A nationwide, 1-day point-prevalence survey was conducted on December 18, 2024, across 116 ICUs with a 24-hour follow-up. Data were collected at the ICU level and included ICU characteristics, staffing and equipment, the total number of admitted patients, the number of patients receiving rehabilitation, types and frequency of interventions, and reasons for non-provision.

Results

Complete data were obtained for 1,529 patients from 116 ICUs. Among them, 374 patients (24.5%) received at least one rehabilitation intervention on the survey day. A total of 968 recorded rehabilitation interventions were delivered, of which 70.5% were physical and 23.1% respiratory rehabilitation. Most physical rehabilitation consisted of in-bed activities, while out-of-bed mobilization accounted for only 7.3% of all recorded rehabilitation interventions. Both the overall rehabilitation rate and out-of-bed mobilization were significantly higher in ICUs within larger hospitals and those with specialized rehabilitation personnel.

Conclusions

Rehabilitation was delivered to 24.5% of ICU patients present on the survey day, with out-of-bed mobilization representing a small proportion of recorded rehabilitation interventions. These findings describe rehabilitation delivery at the population level of surveyed ICU inpatients rather than uptake among medically eligible patients. Broader implementation of ICU rehabilitation may require institutional and systemic support, including dedicated rehabilitation personnel, reduced ICU workload burden, and consideration of appropriate reimbursement structures.

INTRODUCTION

Sedation and immobility, experienced by most patients admitted to intensive care units (ICUs), are well-recognized contributors to adverse sequelae such as muscle weakness, cognitive decline, and delirium [1]. In response to these complications, early mobilization and structured rehabilitation have increasingly been recognized as integral components of evidence-based ICU care.

A growing body of research highlights the benefits of early rehabilitation in critically ill populations. A comprehensive meta-analysis of 60 randomized clinical trials demonstrated that physical rehabilitation significantly improves functional outcomes at hospital discharge and reduces both ICU and overall hospital length of stay [2]. Early rehabilitation strategies-such as neuromuscular electrical stimulation, range-of-motion and breathing exercises, therapeutic positioning, sitting, and ambulation-have consistently proven to be safe and feasible in the ICU. These interventions are associated with reduced muscle weakness, shorter durations of mechanical ventilation and hospitalization, and enhanced physical and psychological recovery [3].

The accumulation of such evidence has prompted the development of clinical practice guidelines across multiple regions. Lang et al. [4] systematically reviewed 10 international guidelines and concluded that early rehabilitation is both safe and cost-effective, emphasizing the importance of standardized safety criteria, multidisciplinary team-based care, and active patient–family participation. Building on this evidence, the Japanese Society of Critical Care Medicine released comprehensive guidelines in 2023, providing detailed recommendations for initiation and discontinuation criteria, intervention frequency, post-discharge follow-up, and domain-specific therapies for swallowing, nutrition, and respiration [5]. Likewise, the European Society of Intensive Care Medicine published expert consensus guidelines recommending early mobilization within 72 hours of ICU admission, the use of neuromuscular electrical stimulation and assistive devices, and structured repositioning protocols [6].

Despite these international advances, the current state of ICU rehabilitation in Korea remains poorly defined. The number of ICU discharges in Korea continues to rise annually [7], and approximately half of ICU survivors experience persistent functional impairments across physical, cognitive, and psychological domains [8]. These trends underscore the urgent need to establish and scale rehabilitation practices for critically ill patients. Although international studies have reported encouraging implementation rates-such as 24% of mechanically ventilated patients engaging in out-of-bed mobilization in German ICUs [9] and an increase in early rehabilitation prevalence from 23.4% to 56.7% in a Japanese ICU between 2018 and 2020 [10]-comparable nationwide data from Korean ICUs are lacking. The 2020 Korean ICU Status Report indicated that roughly half of intensivists were aware of early mobilization principles; however, these findings reflect awareness rather than consistent clinical implementation [11].

To address this lack of nationwide implementation data, we conducted a nationwide 1-day point-prevalence survey to characterize ICU rehabilitation practices in Korea. Point-prevalence surveys, widely applied to monitor hospital-acquired infections [12], have recently been adapted to evaluate ICU-specific issues such as pressure injuries, delirium, and mobility practices [13-15]. This methodology allows systematic, cross-sectional assessment of prevailing clinical practices within a defined time frame, generating robust and pragmatic insights into real-world implementation. Therefore, this study aimed to determine the prevalence and characteristics of ICU rehabilitation practices in Korea, providing foundational data to inform future clinical practice and policy development for comprehensive rehabilitation of critically ill patients.

MATERIALS AND METHODS

This study received exemption approval from the Institutional Review Board of Dong-A University (No. 2-1040709-AB-N-01-202409-HR-043-02, 09/12/2024). Consent to participate was obtained from all participants prior to the survey. The survey was conducted anonymously, with the only personal information collected being respondents' phone numbers, which were used solely for providing small rewards.

Study Design

This study employed a 1-day point-prevalence survey to assess the current status of rehabilitation practices in Korean ICUs across all ICU patients on the survey day.

Setting and Samples

We investigated rehabilitation practices and related characteristics of individual ICUs over a 24-hour period on a designated survey date. ICUs from secondary and tertiary hospitals across South Korea were invited to participate. One healthcare professional from each ICU—either a nurse, physician, or physiotherapist—was designated as the representative respondent. To ensure data accuracy, respondents were instructed to consult relevant team members, such as physiotherapists or charge nurses, when necessary. ICUs from institutions that declined participation were excluded from the study. A total of 120 ICUs nationwide submitted responses to the survey conducted on December 18, 2024. After excluding four duplicate entries, data from 116 ICUs were included in the final analysis (Figure 1).

Figure 1.

Flow diagram of the 1-day point-prevalence survey. ICU: intensive care unit.

Procedure

This study adopted the World Health Organization’s (WHO) point-prevalence survey methodology for hospital antibiotic use [16], with the following procedures: (1) Survey date selection: The survey date (December 18, 2024) was randomly chosen by an independent individual. Five slips, each labeled with a weekday from the third week of December, were placed in an opaque envelope, and one was drawn at random. (2) Announcement and recruitment: 1 week before the survey, an announcement was posted on the Korean Society of ICU Rehabilitation website, including a brief study overview and survey instructions via an online link. A formal letter of cooperation was also provided upon request. (3) Survey administration: The online survey link was activated at 7:00 AM on the designated date. Participants were given 3 days to submit their responses but were instructed to report data specifically for the 24-hour period from 00:00 to 23:59 on December 18. A dedicated hotline (principal investigator’s phone number) was available during the submission period to address any questions.

(4) Questionnaire development: The survey instrument was developed based on previous point-prevalence studies [9,17] and refined through consultation with ICU rehabilitation experts. It comprised 98 items covering the following domains: ICU characteristics, staffing levels, availability of protocols and equipment, number of admitted patients and those receiving rehabilitation, characteristics of rehabilitation recipients and their interventions (bedside exercise, out-of-bed mobilization, respiratory therapy and other therapeutic activities), and reasons for non-provision of rehabilitation as determined by clinical judgment within each ICU. Both closed- and open-ended questions were included. The questionnaire was pilot-tested with ICU staff from three institutions to assess clarity and relevance, and minor revisions were made accordingly.

(5) Definition of rehabilitation: For this study, “rehabilitation” was broadly defined as any physical or respiratory therapy activity documented by ICU staff, including passive range-of-motion exercises, repositioning, and assisted mobilization. Distinctions between routine nursing-based mobilization and structured rehabilitation provided by licensed therapists were not strictly enforced to reflect real-world ICU practice, and some variability in interpretation across ICUs was anticipated. Accordingly, the reported prevalence should be interpreted as a broad estimate of rehabilitation/mobilization-related care rather than therapist-delivered formal rehabilitation alone.

Statistical Analysis

Data were analyzed using IBM SPSS version 27.0 (IBM Corp.). Descriptive statistics, including frequencies, percentages, means, and standard deviations, were used to summarize ICU characteristics, patient demographics, and rehabilitation interventions. Differences in rehabilitation delivery prevalence according to ICU characteristics were examined using independent-samples t tests or one-way analysis of variance with Scheffé post hoc comparisons. Differences in out-of-bed mobilization practices were assessed using chi-square tests. All analyses were exploratory and based on cross-sectional data; therefore, no multivariate adjustments were performed. Statistical significance was defined as P<0.05.

RESULTS

ICU and Hospital Characteristics

A total of 116 ICUs across South Korea participated in the 1-day survey (Table 1). Most were located in metropolitan areas, including Seoul (30.2%), Busan (28.4%), and Daegu (12.1%). Among participating institutions, 79.3% were tertiary hospitals, and half (50.0%) had between 501 and 1,000 beds. Medical ICUs were the most common type (25.0%), followed by surgical (22.4%) and mixed (13.8%) units. Most ICUs (67.2%) operated with 11–20 beds, and 35.3% had nursing staffing grade A, corresponding to two patients per nurse.

Table 1.

Characteristics of participating ICUs

Characteristics	Category	No. (%)	Mean±SD
Location	Seoul	35 (30.2)	-
	Busan	33 (28.4)
	Daegu	14 (12.1)
	Gyeongnam	9 (7.8)
	Ulsan	7 (6.0)
	Daejeon	6 (5.2)
	Jeonbuk	6 (5.2)
	Gyeonggi	4 (3.4)
	Chungnam	2 (1.7)
Type of hospital	Tertiary	92 (79.3)	-
	Secondary	24 (20.7)
Number of hospital beds	>1,000	47 (40.5)	1,105.9±573.7
	501–1,000	58 (50.0)
	<501	11 (9.5)
Type of ICU	Medical	29 (25.0)	-
	Surgical	26 (22.4)
	Mixed	16 (13.8)
	Cardiovascular	14 (12.1)
	Neurological	12 (10.3)
	Neonatal/pediatric	8 (6.9)
	Trauma	7 (6.0)
	Emergency	4 (3.4)
Number of ICU beds	>20	24 (20.7)	17.5±6.5
	11–20	78 (67.2)
	<11	14 (12.1)
ICU nursing staffing level (nurse to patient ratio)	Grade S (1:1.5)	24 (20.7)	-
	Grade A (1:2)	41 (35.3)
	Grade 1 (1:2.5)	40 (34.5)
	≥Grade 2 (1:3)	11 (9.5)
PM&R physician	Yes	27 (23.3)	Number of ICUs in charge per physician: 4.0±1.5
	No	89 (76.7)
ICU rehabilitation coordinator	Yes	20 (17.2)	Number of ICUs in charge per coordinator: 6.4±2.5
	No	96 (82.8)
Physiotherapist	Yes	26 (22.4)	Number of ICUs in charge per physiotherapist: 3.8±1.7
	No	90 (77.6)
Occupational therapist	Yes	14 (12.1)	Number of ICUs in charge per occupational therapist: 5.5±1.3
	No	102 (87.9)
Rehabilitation device and equipment availablea)	Neuromuscular electrical stimulation	48 (41.4)
	Low-frequency electrical stimulation	19 (16.4)
	Upper body ergometer	22 (19.0)
	Bedside cycle ergometer	49 (42.2)
	Positioning assist device	12 (10.3)
	Active stand aids	34 (29.3)
	Powered standing assist device	5 (4.3)
	Tilt table	9 (7.8)
	Walker	53 (45.7)
	High-frequency chest wall oscillation device	96 (82.8)
	Cough assist device	51 (44.0)
	None of the above	10 (8.6)
ICU rehabilitation protocols	Yes	42 (36.2)	-
	No	74 (63.8)

ICU: intensive care unit; SD: standard deviation; PM&R: physical medicine and rehabilitation.

^a)Multiple responses allowed.

Only 23.3% of ICUs reported assignment of a physical medicine and rehabilitation (PM&R) physician, 17.2% had access to a rehabilitation coordinator, 22.4% to a physiotherapist, and 12.1% to an occupational therapist. These professionals were not exclusively dedicated to the ICU but served multiple units. On average, one PM&R physician covered 4.0±1.5 ICUs, rehabilitation coordinators 6.4±2.5, physiotherapists 3.8±1.7, and occupational therapists 5.5±1.3.

Of the 116 ICUs, 106 (91.4%) reported having at least one type of rehabilitation equipment. The most common devices were chest wall oscillators (82.8%), walkers (45.7%), cough-assist devices (44.0%), and bedside cycle ergometers (42.2%). In contrast, tilt tables (7.8%) and powered standing-assist devices (4.3%) were rarely available. Only 36.2% of ICUs had a formal rehabilitation protocol.

Rehabilitation Delivery Prevalence and Patient Characteristics

At 10:00 AM on the survey day, 1,529 patients were admitted across the participating ICUs. Of these, 374 (24.5%) received at least one rehabilitation intervention between 00:00 and 23:59 on December 18. Examination of level of consciousness during rehabilitation, as assessed using the Richmond Agitation-Sedation Scale (RASS), revealed that 31.6% of patients had a RASS ≤–2 (moderate to deep sedation), whereas 29.7% had a score of 0, reflecting an alert, calm state. Among rehabilitation recipients, 36.1% were receiving sedative medications, and 17.4% were documented as experiencing delirium.

Regarding respiratory support, 51.6% of patients were on mechanical ventilation, 21.7% on oxygen therapy via nasal cannula, and 14.7% on high-flow nasal cannula. In addition, 15.2% (n=57) had chest tubes, 14.4% (n=54) were undergoing continuous renal replacement therapy, and 6.7% (n=25) had intracranial pressure monitoring or an external ventricular drain in place (Table 2).

Table 2.

Characteristics of patients undergoing rehabilitation

Characteristics	Category	No. (%)
Level of consciousness (RASS)	≥+2	11 (2.9)
	1	31 (8.3)
	0	111 (29.7)
	–1	99 (26.5)
	≤–2	118 (31.6)
	Not responded	4 (1.0)
Sedation treatment	Yes	135 (36.1)
	No	229 (61.2)
	Stop for rehabilitation	1 (0.3)
	Not responded	9 (2.4)
Delirium	Yes	65 (17.4)
	No	303 (81.0)
	Not responded	6 (1.6)
Oxygen therapy	Room air	29 (7.8)
	Nasal cannula	81 (21.7)
	HFNC	55 (14.7)
	Ventilator	193 (51.6)
	Others	14 (3.7)
	Not responded	2 (0.5)
Special treatmenta)	ECMO	12 (3.2)
	CRRT	54 (14.4)
	ICP monitoring or EVD drain	25 (6.7)
	Chest tube	57 (15.2)
	Others	2 (0.5)

RASS: Richmond Agitation-Sedation Scale; HFNC: high-flow nasal cannula; ECMO: extracorporeal membrane oxygenation; CRRT: continuous renal replacement therapy; ICP: intracranial pressure; EVD: external ventricular drain.

^a)Multiple responses allowed.

Rehabilitation Characteristics and Interventions

Rehabilitation characteristics and interventions are summarized in Table 3. Among patients who received rehabilitation, 69.8% participated in a single session, and 72.5% underwent sessions lasting less than 30 minutes. Most (89.8%) had a documented rehabilitation prescription, and physiotherapists served as the primary providers for 81.0%. In total, 968 recorded rehabilitation interventions were delivered to 374 patients (Figure 2). Physical rehabilitation accounted for most recorded interventions (n=682), with the majority categorized as in-bed activities (n=611), including passive range-of-motion (26.2%), active range-of-motion (11.6%), and neuromuscular electrical stimulation (9.7%). Out-of-bed mobilization accounted for 71 recorded rehabilitation interventions (7.3%), and included sitting at the edge of the bed (n=35, 3.6%), standing (n=22, 2.3%), and walking (n=14, 1.4%).

Table 3.

Characteristics of rehabilitation provided

Characteristics	Category	No. (%)
Number of rehabilitation sessions on the survey day	1	261 (69.8)
	2	56 (15.0)
	≥3	57 (15.2)
Time per session	<30 min	271 (72.5)
	30–60 min	82 (21.9)
	>60 min	20 (5.3)
	Not responded	1 (0.3)
Rehabilitation prescription	Yes	336 (89.8)
	No	38 (10.2)
Primary provider	Nurse	35 (9.4)
	Rehabilitation coordinator	3 (0.8)
	Physiotherapist	303 (81.0)
	Occupational therapist	3 (0.8)
	Jointly provided	30 (8.0)
Rehabilitation contenta) (n=968)
Physical (n=682)
Bedside exercise (n=611)
Electrical stimulation		94 (9.7)
Passive ROM		254 (26.2)
Active ROM		112 (11.6)
Cycle ergometer training		61 (6.3)
Strength training		90 (9.3)
Out-of-bed mobilization (n=71)
Sitting at the edge of the bed		35 (3.6)
Standing		22 (2.3)
Walking		14 (1.4)
Respiratory (n=224)
Coughing training		129 (13.3)
Respiratory muscle strengthening training		89 (9.2)
Ventilator weaning		6 (0.6)
Others (n=62)
Activities of daily living training		7 (0.7)
Swallowing rehabilitation		48 (5.0)
Cognitive rehabilitation		7 (0.7)

ROM: range of motion.

^a)Multiple responses allowed.

Figure 2.

Rehabilitation interventions provided to critically ill patients on the survey day. This figure illustrates the content of rehabilitation delivered to critically ill patients in the intensive care unit (ICU) on the point-prevalence survey day. The bar graph shows the numbers of each rehabilitation intervention performed and reported by participating ICUs. PROM: passive range of motion; AROM: active range of motion; ES: electrical stimulation; ST: strength training; CET: cycle ergometer training; SIT: sitting at the edge of the bed; CT: coughing training; RMST: respiratory muscle strengthening training; VW: ventilator weaning; ADLT: activities of daily living training; SR: swallowing rehabilitation; CR: cognitive rehabilitation.

Respiratory rehabilitation comprised 224 recorded rehabilitation interventions, with cough training being the most frequent category (13.3%). The “other” category included 62 recorded interventions, including swallowing training (n=48, 5.0%), activities of daily living training (n=7, 0.7%), and cognitive rehabilitation (n=7, 0.7%).

Differences in Rehabilitation Delivery Prevalence by ICU Characteristics

Differences in rehabilitation delivery prevalence according to ICU characteristics are presented in Table 4. Rehabilitation delivery prevalence was significantly higher in ICUs within tertiary hospitals (t=2.41, P=0.018) and in institutions with more than 1,000 beds (F=4.27, P=0.016). ICUs with nursing staffing grade S also showed significantly higher rehabilitation delivery prevalence than those with grade 2 or lower (F=3.46, P=0.019) (Figure 3).

Table 4.

Differences in prevalence of rehabilitation practices according to ICU characteristics

Variable	Category	Rehabilitation prevalence (%), mean±SD	t or F(P)	Post-hoc test
Location	Seoul	36.8±28.3	2.23 (0.071)
	Busan	30.5±33.1
	Daegu	20.0±22.0
	Gyeongnam	14.3±24.9
	Othersa)	20.4±23.2
Type of hospital	Tertiary	30.9±30.5	2.41 (0.018)
	Secondary	15.5±13.1
Number of hospital beds	>1,0001	36.6±31.5	4.27 (0.016)	1>2
	501–1,0002	22.6±26.1
	<5013	16.7±15.1
Type of ICU	Medical	28.5±30.0	1.35 (0.250)
	Surgical	28.0±25.7
	Mixed	23.1±25.2
	Cardiovascular	20.3±18.8
	Neurological	45.9±37.6
	Othersb)	23.9±30.1
Number of ICU beds	>20	21.8±22.3	2.03 (0.136)
	11–20	27.2±28.9
	<11	40.7±32.9
ICU nurse to patient ratio	Grade S (1:1.5)1	42.2±31.8	3.46 (0.019)	1>4
	Grade A (1:2)2	24.8±29.6
	Grade 1 (1:2.5)3	26.0±25.4
	≥Grade 2 (1:3)4	12.8±14.1
PM&R physician	Yes	45.0±29.8	3.80 (<0.001)
	No	22.5±26.1
ICU rehabilitation coordinator	Yes	47.5±30.6	3.59 (<0.001)
	No	23.6±26.4
Physiotherapist	Yes	48.2±31.4	4.50 (<0.001)
	No	21.8±24.8
Occupational therapist	Yes	45.9±27.9	2.62 (0.010)
	No	25.2±27.8
ICU rehabilitation protocols	Yes	39.5±28.5	3.52 (0.001)
	No	21.0±26.4

Different superscript numbers (1–3 or 1–4) indicate significant differences among groups according to one-way analysis of variance followed by Scheffé post hoc tests (P<0.05).

ICU: intensive care unit; SD: standard deviation; PM&R: physical medicine and rehabilitation.

^a)Ulsan, Daejeon, Jeonbuk, Gyeonggi, Chungnam;

^b)Neonatal/pediatric, trauma, emergency.

Figure 3.

Comparison of out-of-bed mobilization performance and average rehabilitation prevalence by nurse-to-patient ratio. Clustered bar and line chart showing the number of intensive care units (ICUs) by nurse-to-patient ratio groups (bars); the line indicates the average prevalence of rehabilitation practices within each group.

Similarly, rehabilitation delivery prevalence was significantly higher in ICUs with access to specialized personnel, including PM&R physicians (t=3.80, P<0.001), rehabilitation coordinators (t=3.59, P<0.001), physiotherapists (t=4.50, P<0.001), and occupational therapists (t=2.62, P=0.010) (Figure 4). ICUs with formal rehabilitation protocols also demonstrated significantly higher rehabilitation delivery prevalence than those without (t=3.52, P=0.001).

Figure 4.

Comparison of out-of-bed mobilization performance and average rehabilitation prevalence according to the availability of specialized personnel. Clustered bar and line chart showing out-of-bed mobilization performance (%) by the presence of specialized personnel (bars); the line indicates the average prevalence of rehabilitation practices. SP: specialized personnel; OBM: out-of-bed mobilization; PM&R: physical medicine and rehabilitation; ICU: intensive care unit.

Differences in Out-of-Bed Mobilization by ICU Characteristics

On the survey day, 22 ICUs (19.0%) implemented at least one form of out-of-bed mobilization, including sitting at the edge of the bed, standing, or walking (Table 5). Among these, 20 ICUs (17.2%) provided sitting-edge-of-bed exercises, 14 (12.1%) facilitated standing, while only seven (6.0%) implemented walking interventions. Significant regional variation was observed: out-of-bed mobilization was most common in ICUs located in Seoul (40.0%), followed by Gyeongnam (22.2%), Daegu (14.3%), and Busan (12.1%) (χ²=17.20, P=0.002). Implementation was higher in institutions with >1,000 beds (29.8%) than in those with 501–1,000 beds (13.8%) (χ²=7.17, P=0.028).

Table 5.

Differences in out-of-bed performance by ICU characteristics

Variable	Category	Sitting edge of bed	Standing	Walking	Out-of-bed mobilization
					Yes	No	χ2 (P)
Total		20 (17.2)	14 (12.1)	7 (6.0)	22 (19.0)	94 (81.0)	-
Location	Seoul	14 (40.0)	9 (25.7)	5 (14.3)	14 (40.0)	21 (60.0)	17.20 (0.002)
	Busan	2 (6.1)	2 (6.1)	1 (3.0)	4 (12.1)	29 (87.9)
	Daegu	2 (14.3)	1 (7.1)	1 (7.1)	2 (14.3)	12 (85.7)
	Gyeongnam	2 (22.2)	2 (22.2)	0	2 (22.2)	7 (77.8)
	Othersa)	0	0	0	0	25 (100.0)
Type of hospital	Tertiary	18 (19.6)	14 (15.2)	7 (7.6)	20 (21.7)	72 (78.3)	2.23 (0.240)
	Secondary	2 (8.3)	0	0	2 (8.3)	22 (91.7)
Number of hospital beds	>1,000	14 (29.8)	9 (19.1)	4 (8.5)	14 (29.8)	33 (70.2)	7.17 (0.028)
	501–1,000	6 (10.3)	5 (8.6)	3 (5.2)	8 (13.8)	50 (86.2)
	<501	0	0	0	0	11 (100.0)
Type of ICU	Medical	4 (13.8)	2 (6.9)	0	4 (13.8)	25 (86.2)	17.26 (0.004)
	Surgical	11 (42.3)	7 (26.9)	6 (23.1)	12 (46.2)	14 (53.8)
	Mixed	1 (6.3)	1 (6.3)	1 (6.3)	1 (6.3)	15 (93.8)
	Cardiovascular	2 (14.3)	2 (14.3)	0	2 (14.3)	12 (85.7)
	Neurological	1 (8.3)	1 (8.3)	0	2 (16.7)	10 (83.3)
	Othersb)	1 (5.3)	1 (5.3)	0	1 (5.3)	18 (94.7)
Number of ICU beds	>20	5 (20.8)	4 (16.7)	1 (4.2)	5 (20.8)	19 (79.2)	6.45 (0.040)
	11–20	11 (14.1)	6 (7.7)	5 (6.4)	11 (14.1)	67 (85.9)
	<11	4 (28.6)	4 (28.6)	1 (7.1)	6 (42.9)	8 (57.1)
ICU nurse to patient ratio	Grade S (1:1.5)	9 (37.5)	6 (25.0)	4 (16.7)	9 (37.5)	15 (62.5)	10.27 (0.016)
	Grade A (1:2)	8 (19.5)	6 (14.6)	3 (7.3)	9 (22.0)	32 (78.0)
	Grade 1 (1:2.5)	3 (7.5)	2 (5.0)	0	4 (10.0)	36 (90.0)
	≥Grade 2 (1:3)	0	0	0	0	11 (100.0)
PM&R physician	Yes	8 (29.6)	5 (18.5)	4 (14.8)	8 (29.6)	19 (70.4)	2.60 (0.158)
	No	12 (13.5)	9 (10.1)	3 (3.4)	14 (15.7)	75 (84.3)
ICU rehabilitation coordinator	Yes	8 (40.0)	5 (25.0)	4 (20.0)	8 (40.0)	12 (60.0)	6.96 (0.023)
	No	12 (12.5)	9 (9.4)	3 (3.1)	14 (14.6)	82 (85.4)
Physiotherapist	Yes	12 (46.2)	8 (30.8)	4 (15.4)	12 (46.2)	14 (53.8)	16.12 (<0.001)
	No	8 (8.9)	6 (6.7)	3 (3.3)	10 (11.1)	80 (88.9)
Occupational therapist	Yes	6 (42.9)	4 (28.6)	2 (14.3)	6 (42.9)	8 (57.1)	5.91 (0.026)
	No	14 (13.7)	10 (9.8)	5 (4.9)	16 (15.7)	86 (84.3)
ICU rehabilitation protocols	Yes	15 (35.7)	10 (23.8)	5 (11.9)	15 (35.7)	27 (64.3)	12.02 (0.001)
	No	5 (6.8)	4 (5.4)	2 (2.7)	7 (9.5)	67 (90.5)

Values are presented as number (%).

ICU: intensive care unit; PM&R: physical medicine and rehabilitation.

^a)Ulsan, Daejeon, Jeonbuk, Gyeonggi, Chungnam;

^b)Neonatal/pediatric, trauma, emergency.

By ICU type, surgical ICUs had the highest rate of out-of-bed mobilization (46.2%), followed by neurological (16.7%) and cardiovascular (14.3%) units (χ²=17.26, P=0.004). Smaller ICUs (<11 beds) showed higher implementation rates (42.9%) than larger ICUs (>20 beds, 20.8%) (χ²=6.45, P=0.040).

Out-of-bed mobilization was also more frequent in ICUs with nursing staffing grade S (37.5%) than in those with grade A (22.0%) (χ²=10.27, P=0.016) (Figure 3). Higher implementation rates were observed in ICUs with rehabilitation coordinators (40.0%; χ²=6.96, P=0.023), physiotherapists (46.2%; χ²=16.12, P<0.001), and occupational therapists (42.9%; χ²=5.91, P=0.026) (Figure 4). ICUs with established rehabilitation protocols reported significantly higher rates of out-of-bed mobilization (35.7%) than those without (χ²=12.02, P=0.001).

Barriers to ICU Rehabilitation

On the survey day, rehabilitation was not provided to 1,155 patients, and a total of 1,254 barrier responses were collected, as multiple responses per patient were allowed. Reported barriers to ICU rehabilitation are shown in Figure 5. The most frequently reported reason was inappropriate patient condition (n=360, 31.2%), followed by absence of a rehabilitation prescription (n=301, 26.1%). Other barriers included insufficient rehabilitation personnel (n=256, 22.2%), excessive clinical workload (n=205, 17.7%), and lack of appropriate equipment (n=132, 11.4%).

Figure 5.

Barriers to intensive care unit rehabilitation. Horizontal bar chart illustrates the barriers to intensive care unit rehabilitation as reported by participating units among patients who did not receive rehabilitation on the survey day. Multiple responses were permitted.

DISCUSSION

This nationwide 1-day point-prevalence survey found that 374 (24.5%) of 1,529 patients admitted to 116 ICUs received rehabilitation on the survey day. Among the recorded rehabilitation interventions, approximately 70% were physical rehabilitation interventions, and most consisted of in-bed activities. Out-of-bed mobilization accounted for 7.3% of recorded rehabilitation interventions. ICUs located in tertiary hospitals and those staffed with specialized rehabilitation personnel demonstrated higher rehabilitation delivery prevalence, and ICUs staffed with specialized rehabilitation personnel also demonstrated higher out-of-bed mobilization rates. To our knowledge, this is the first nationwide survey of ICU rehabilitation practices in South Korea; as such, this study provides foundational data for future clinical and policy development. These findings may also serve as reference data for countries where ICU rehabilitation remains underdeveloped or inconsistently implemented.

The overall rehabilitation delivery prevalence of 24.5% observed in this study was substantially lower than that reported in Brazil (87.4%) [15]. Similarly, the rate of out-of-bed mobilization among Korean ICU patients was far below previous reports-24% in Germany [9], 39.9% in Brazil [15], and 41% in the United Kingdom [17]-suggesting a relatively conservative approach to ICU rehabilitation in Korea. Nevertheless, it is encouraging that rehabilitation was provided even to patients with complex clinical conditions; more than half had RASS scores of –1 or lower, approximately one-third were receiving sedatives, over half required mechanical ventilation, and some were undergoing extracorporeal membrane oxygenation or continuous renal replacement therapy.

The limited rehabilitation delivery prevalence and intensity of ICU rehabilitation in Korea may partly reflect structural barriers related to ICU rehabilitation delivery. In this study, absence of a rehabilitation prescription was the second most frequently cited barrier to rehabilitation delivery. In the Korean healthcare system, rehabilitation prescriptions are often influenced by the availability of corresponding reimbursement pathways [18]. Notably, rehabilitation delivery prevalence was numerically higher in neurological ICUs, which may be partly related to the presence of a specific reimbursement code for central nervous system rehabilitation (MM105). MM105 is one of the most frequently used and highest-cost rehabilitation-related codes in Korea’s insurance claims database [19]. Because out-of-bed mobilization requires additional time, equipment, and skilled personnel, the availability of reimbursement pathways may facilitate broader implementation of ICU rehabilitation. However, as of the time of writing of this manuscript, no specific billing code exists for ICU rehabilitation in Korea [20]. Therefore, developing reimbursement structures tailored to critically ill patients may help reduce some of the structural barriers to ICU rehabilitation implementation.

ICUs with access to PM&R physicians, rehabilitation coordinators, physiotherapists, and occupational therapists showed significantly higher rehabilitation delivery prevalence, while out-of-bed mobilization rates were significantly higher in ICUs with access to rehabilitation coordinators, physiotherapists, and occupational therapists. These findings underscore the importance of multidisciplinary collaboration in promoting effective ICU rehabilitation. Increasing the availability of dedicated rehabilitation personnel may help enhance both the quantity and quality of services. For instance, a Japanese ICU reported a substantial increase in rehabilitation provision—from 23.4% to 56.7%—along with reductions in mechanical ventilation duration and ICU length of stay following the introduction of a dedicated therapist and formal rehabilitation protocol [10]. Despite the fact that approximately 80% of participating ICUs in our study were located in tertiary hospitals, only about 20% had access to PM&R physicians or physiotherapists, and even fewer to rehabilitation coordinators (17.2%) or occupational therapists (12.1%). Each specialist was responsible for an average of five to six ICUs, reflecting a severe shortage of rehabilitation staff.

A notable limitation of this study is the potential conceptual overlap between “mobilization” and “rehabilitation” during data collection. In some ICUs, passive range-of-motion exercises or repositioning-often performed by nurses or mobilization teams-may have been recorded as rehabilitation, even when not constituting structured, goal-directed therapy by licensed rehabilitation professionals. This definitional ambiguity may have led to an overestimation of rehabilitation prevalence and limits the ability to distinguish between general mobilization and formal rehabilitation. However, rehabilitation interventions were presented as selectable items in the survey, and participating ICUs were able to seek real-time clarification during data collection. These measures may have improved consistency to some extent; however, misclassification cannot be excluded. Importantly, even if routine nursing-based mobilization was included by some institutions, only 24.5% of ICU patients received rehabilitation on the survey day, and out-of-bed mobilization accounted for only a small proportion of recorded rehabilitation interventions (7.3%), highlighting ongoing challenges in the implementation of mobilization and rehabilitation in real-world ICU settings. Future studies should adopt more rigorous operational definitions and clearly categorize interventions at the data collection stage to improve validity and comparability.

Although earlier literature identified lack of staff awareness and negative attitudes as major barriers to ICU rehabilitation [21,22], more recent evidence suggests that these issues are improving. Instead, workforce shortages and excessive clinical workload have become the dominant barriers [23]. Both were frequently cited in this study, and 9.4% of patients received rehabilitation solely from nurses without specialized training. Expanding the pool of trained rehabilitation professionals and alleviating staff workload could enable broader and more effective implementation of ICU rehabilitation. Under current workforce constraints, establishing multidisciplinary mobility teams that include rehabilitation specialists alongside ICU nurses may be a pragmatic approach. Such team-based models, as recommended in recent clinical guidelines, have been associated with improved functional outcomes and reduced healthcare costs [24,25].

Nearly half of Korean ICU survivors develop post-intensive care syndrome (PICS), characterized by long-term physical, psychological, and cognitive impairments [8]. These patients are at heightened risk of readmission and death within six months of discharge, with healthcare expenditures more than ten times higher than those of other hospitalized patients during the same period [26]. These findings highlight the vulnerability of ICU survivors and the importance of early rehabilitation to mitigate adverse outcomes. Developing a dedicated reimbursement system and investing in specialized rehabilitation personnel could facilitate early rehabilitation, thereby reducing the incidence of PICS and lowering long-term healthcare costs.

This study has several limitations. First, participation was voluntary, introducing potential selection bias. ICUs with greater interest in or stronger infrastructure for rehabilitation may have been more likely to participate, potentially leading to overestimation of rehabilitation prevalence. In addition, ICU staff with greater interest in rehabilitation or relatively lighter workloads may have been more likely to respond to the survey. Although ICUs from diverse hospital types and geographic regions were included, these factors may have limited the representativeness of the sample. Therefore, the findings should be interpreted with caution when considering broader generalization. Second, the survey was conducted at the ICU level only; patient-level data such as demographics, diagnoses, and ICU length of stay were not collected. In addition, standardized clinical criteria to identify patients medically appropriate for rehabilitation (e.g., patient severity or hemodynamic stability) were not applied, and decisions regarding rehabilitation provision relied on clinical judgment by each ICU team. Therefore, this study was limited in its ability to precisely quantify patients who did not receive rehabilitation due to medically inappropriate conditions. Furthermore, differences in patient severity and case mix between hospital types were not adjusted for, which limits direct comparisons of rehabilitation prevalence across institutions. Future studies should apply standardized clinical criteria to classify rehabilitation candidates and assess reasons for non-provision. Finally, self-reported data may have introduced reporting bias or inter-rater variability, particularly in classifying mobilization versus formal rehabilitation. Future research should consider obtaining institutional review board approval at each participating center to allow patient-level data collection, enabling more detailed analyses of patient characteristics, intervention effects, and outcomes.

In conclusion, this nationwide 1-day point-prevalence survey found that only one in four ICU patients received rehabilitation services. Most interventions were limited to in-bed physical activities, while out-of-bed mobilization was notably infrequent. Rehabilitation prevalence-including both overall and out-of-bed mobilization rates-was higher in ICUs within larger hospitals and those staffed with dedicated rehabilitation personnel. The limited implementation of ICU rehabilitation in Korea appears related to workforce shortages and associated staff workload. To improve access to and quality of rehabilitation for critically ill patients, future policy efforts should consider appropriate health insurance reimbursement structures together with multidisciplinary, team-based rehabilitation models. These strategies may facilitate earlier and more effective rehabilitation, with the potential to improve patient outcomes and reduce long-term healthcare burdens.

KEY MESSAGES

▪ In this nationwide point-prevalence study of 116 intensive care units (ICUs) across South Korea, only 24.5% of patients received rehabilitation on the survey day; among recorded rehabilitation interventions, passive range-of-motion exercises (26.2%) were the most frequently provided.

▪ Out-of-bed mobilization was significantly more frequent in ICUs with dedicated rehabilitation personnel (e.g., ICU rehabilitation coordinators, physiotherapists, or occupational therapists).

▪ The most frequently reported barriers were patients’ inappropriate clinical condition and the absence of a rehabilitation prescription, highlighting the urgent need for standardized protocols and adequate staffing to expand early mobilization in critically ill patients.