a b s t r a c t

Objectives: This study aimed to validate the effectiveness of an emergency short-stay ward (ESSW) and its impact on clinical outcomes.

Methods: This retrospective observational study was performed at an urban tertiary hospital. An ESSW has been operating in this hospital since September 2017 to reduce emergency department (ED) boarding time and only targets patients indicated for admission to the general ward from the ED. Propensity-score matching was per- formed for comparison with the control group. The primary outcome was ED boarding time, and the secondary outcomes were subsequent intensive care unit admission and 30-day in-hospital mortality.

Results: A total of 7461 patients were enrolled in the study; of them, 1523 patients (20.4%) were admitted to the ESSW. After propensity-score matching, there was no significant difference in the ED boarding time between the ESSW group and the control group (P = 0.237). Subsequent ICU admission was significantly less common in the ESSW group than in the control group (P b 0.001). However, the 30-day in-hospital mortality rate did not differ significantly between the two groups (P = 0.292). When the overall hospital bED occupancy ranged from 90% to 95%, the proportion of hospitalization was the highest in the ESSW group (29%). An interaction effect test using a general linear model confirmed that the ESSW served as an effect modifier with respect to bed occupancy and boarding time (P b 0.001).

Conclusion: An ESSW can alleviate prolonged boarding time observed with hospital bed saturation. Moreover, the ESSW is associated with a low rate of subsequent ICU admission.

(C) 2019

1. Introduction

Emergency department (ED) crowding is a global healthcare issue given its unequivocal role in the deterioration of clinical outcomes [1- 3]. According to the input-throughput-output conceptual model, which is utilized extensively for systemic analyses of ED crowding, the main cause of ED crowding is the output component associated with a lack of access to inpatient beds from the ED. [4,5] As an indicator of the output component, ED boarding time is the most favored measure of hospital efficiency [4]. A prolonged ED boarding time leads to adverse clinical risks such as prolonged illness; worse outcomes in patients with stroke, cardiac disorders, and intubation; exposure to hospital-acquired

Abbreviations: ESSW, emergency short-stay ward; ED LOS, emergency department length of stay; ICU, intensive care unit; SMD, standardized mean difference; SD, standard deviation; KTAS, Korean Triage and Acuity Scale.

* Corresponding author at: Department of Emergency Medicine, Yonsei University College of Medicine, Yonsei-ro, Seodaemun-gu, Seoul 03722, Republic of Korea.

E-mail address: [email protected] (J.H. Kim).

infections; and lapses in daily medication and other routine care pro- vided by the overtaxed ED staff [6-8]. Thus, numerous studies have been conducted to develop strategies aimed at reducing the ED boarding time. The current consensus is that organized hospital-level planning outperforms specific interventions [9].

An emergency short-stay ward (ESSW) is a hospital-level strategy

that has been used to cope with prolonged ED Boarding times in our hospital since September 2017. Several studies previously introduced the concept of a short-stay unit for reducing ED crowding. Most, how- ever, were targeted to patients in the throughput stage before the deci- sion to admit or discharge [10-14], whereas our institutional ESSW is characterized by targeting patients only after a decision to admit to the general ward from the ED is reached. Patients are admitted by ED physicians for a maximum time limit of 48 h. After 48 h, patients are ex- pected to leave the ESSW according to the hospital policy. This study aimed to validate the effectiveness of the ESSW in alleviating prolonged ED boarding times and evaluate whether ESSW admission had any neg- ative influence on clinical outcomes when compared to direct admis- sion to the general ward.

https://doi.org/10.1016/j.ajem.2019.12.014 0735-6757/(C) 2019

1. Material and methods
   1. Study design and setting

This retrospective observational study used prospectively collected data. The study protocol adhered to the STROBE statement and com- plied with the tenets of the Declaration of Helsinki. The study was ap- proved by the appropriate institutional review board (approval number 4-2018-0877), and the requirement for informed consent was waived. The present study was conducted at a level 1 ED in an urban ter- tiary teaching hospital with approximately 2200 inpatient beds. Ap- proximately 100,000 patients visit this ED every year, and of these, approximately 25% of patients are subsequently admitted. The mean ED length of stay of all patients was approximately 6.2 h, and the mean ED LOS of hospitalized patients was 16.5 h.

Study population

This study examined all consecutive patients who were admitted via this ED between October 2017 and March 2018. Patients who were in- eligible for ESSW admission (i.e., those who were admitted to the inten- sive care unit [ICU] from the ED, those admitted to the hematologic unit, and patients who requested a “do-not-resuscitate” order) were ex- cluded from the analysis.

ESSW protocol

The ESSW in our hospital has been operating since September 2017 with the intent to reduce ED boarding times. The ESSW is located phys- ically just above the ED but is administratively separated from the ED. The ESSW is a 25-bed, 3-room facility. Three nurses work during each shift, and three shifts are staffed per day, yielding a total of 12 nurses (apart from ED nurses). Patients in the ESSW are cared for by specialists, as in the general ward of this hospital. The ESSW is available to all pa- tients for whom a decision has been made to admit to the general ward from the ED. Patients who require ICU care or who are admitted to the hematology department are ineligible for ESSW admission be- cause the ESSW has no resources for these patients. The hospital policy maintains a strict ESSW administration process. If the ESSW has an available bed, the electronic medical record system informs the ED phy- sician in real time, who then selects the patient to board according to those with the longest staying time in the ED. After entering the ESSW, the patient is no longer under the control of the ED system and is managed similar to patients in other general wards. The patient can remain in the ESSW for up to 48 h as per the hospital policy. Thereafter, they should be moved to the appropriate general ward in the hospital.

Data source and collection

Data were collected from the in-house data retrieval system, which allowed us to define a clinical data model with the desired variables and conditions. In this study, we developed a clinical data model that in- cluded all consecutive patients admitted via the ED during the study pe- riod, and the specified variables were automatically coded in this model. These variables included age, sex, ED visit time, and date, Korean Triage and Acuity Scale score (level 1 indicates the most severe condition), hospital bED occupancy rate, ED Crowding index, time to disposition, ad- mitting clinical department, Insurance type, and decision by the attend- ing or residents.

Exposure and outcome variables

The exposure was ESSW admission, and the primary outcome mea- sure was the ED boarding time, defined as the time interval from the ad- mission decision to the exit from the ED. The secondary outcomes were

subsequent ICU admission after leaving the ED and 30-day mortality during hospitalization.

Statistical analysis

Categorical variables are reported as numbers and percentages, and continuous variables are expressed as medians and interquartile ranges. To control for confounders, we used the propensity-score matching method. For continuous variables, nonmatching data were compared between the groups using Student’s t-test; for categorical variables, nonmatching data were compared between the groups using the ?² test. Matching continuous data and categorical data were compared be- tween the groups using the paired t-test and McNemar’s test, respec- tively. Additionally, standardized mean differences were used to compare the balance in baseline covariates between the groups, with b0.1 defined to indicate a good balance between the groups. For propensity-score matching, we selected variables that affected the ED boarding time and study exposure, with reference to previous studies. After matching, we compared the primary outcome between the groups using the Wilcoxon signed-rank test because the Shapiro-Wilk test showed that the primary outcome variables violated normality. Differ- ences were considered statistically significant at a P value b .05. All sta- tistical analyses were performed using SAS software, version 9.4 (SAS Institute, Cary, NC, USA).

1. Results

Of the 51,676 patients who visited the ED during the study period, we examined 7461 patients. Among them, 1523 were transferred to the ESSW and included in the final analysis. All cases were matched to the control group via propensity score matching (Fig. 1). The baseline patient characteristics according to the study exposure are shown in Table 1. The Korean Triage and Acuity Scale score, hospital bed occu- pancy, arrival on regular time, admitting clinical department, insurance type, and decision by attending or residents differed significantly be- tween the ESSW and control groups. In the matching population, there were no significant differences in any variables, including those that differed before matching.

Table 2 presents the study outcomes before and after matching. Be- fore matching, the boarding time of the ESSW group was significantly longer than that of the control group, but this difference was no longer significant after matching. With respect to clinical outcomes, after matching, the number of subsequent ICU admissions was significantly lower in the ESSW group than in the control group, whereas the 30- day mortality did not differ significantly between the two groups. Dur- ing matching, hospital bed occupancy had a strong impact on the boarding time, and thus, a subsequent subgroup analysis was per- formed using this variable. Fig. 2 shows the changes in the proportions of patients transferred to the ESSW among all patients admitted via the ED according to an increase in hospital bed occupancy. Among all patients admitted via the ED, the proportion of patients entering the ESSW increased as the bed occupancy increased up to 95%. ESSW admis- sion was the highest (29%) at an overall hospital bed occupancy of 90%- 95%.

Fig. 3 presents changes in boarding times between the two groups according to the change in hospital bed occupancy in the matching co- hort. When the bed occupancy was b75%, the boarding time of the con- trol group was significantly shorter than that of the ESSW group. When the occupancy rate was above 90%, the boarding time of the ESSW group was shorter than that of the control group, although this difference was not statistically significant. We conducted an interaction effect test of the general linear model to determine whether ESSW admission af- fected the increase in boarding time with increasing bed occupancy. The results confirmed that the ESSW served as an effect modifier rela- tive to bed occupancy and boarding time (P b 0.001).

Fig. 1. Flowchart of patient inclusion. ED, emergency department; ICU, intensive care unit; ESSW, emergency short-stay ward.

1. Discussion

The present study showed that a high bed occupancy rate increased the proportion of inpatients entering the ESSW. Although the ESSW did not significantly reduce the overall boarding time, we observed a more gradual increase in boarding time according to hospital bed occupancy

in the ESSW group. The crude boarding time in the control group was lower than that in the ESSW group before matching probably because the boarding time in the latter group was essentially longer than that in the former group. Specifically, ED physicians in this study were ex- pected to select patients with long boarding times for transfer to the ESSW, as per the hospital policy. The proportion of inpatients admitted

Table 1

Patient characteristics before and after matching

Variables	Before matching				After matching
	Control (n = 5938)	ESSW (n = 1523)	SMD		Control (n = 1523)	ESSW (n = 1523)	SMD
Age (years), mean +- SD	61.0 +- 18.0	60.5 +- 16.9	0.026		60.0 +- 17.8	60.5 +- 16.9	0.040
Sex (male, %)	2779 (46.8)	641 (42.1)	0.095		623 (40.9)	641 (42.1)	0.054
KTAS (n, %) 1	192 (3.2)	22 (1.4)	0.232		22 (1.4)	22 (1.4)	0.052
2	854 (14.4	132 (8.7)			124 (8.1)	132 (8.7)
3	2105 (35.5)	544 (35.7)			531 (34.9)	544 (35.7)
4	2541 (42.8)	744 (48.9)			759 (49.8)	744 (48.9)
5	246 (4.1)	81 (5.3)			87 (5.7)	81 (5.3)
Bed occupancy (%, mean +- SD)	87.9 +- 7.6	91.2 +- 4.7	0.522		91.1 +- 4.7	91.2 +- 4.7	0.007
Crowding index (mean +- SD)	54.6 +- 14.8	54.8 +- 14.4	0.012		54.5 +- 15.1	54.8 +- 14.4	0.027
Disposition time (min, mean +- SD)	314.4 +- 204.4	313.5 +- 212.8	0.005		315.6 +- 208.1	313.5 +- 212.8	0.014
Arrival on regular time to the ESSW (n, %)	1775 (29.9)	672 (44.1)	0.298		663 (43.5)	672 (44.1)	0.012
Admitting department (n, %) Cardiology	757 (12.8)	59 (3.9)	0.587		63 (4.1)	59 (3.9)	0.092
Internal medicine	2928 (49.3)	1064 (69.9)			1074 (70.5)	1064 (69.9)
General surgery	648 (10.9)	197 (12.9)			184 (12.1)	197 (12.9)
Orthopedic surgery	387 (6.5)	14 (0.9)			20 (1.3)	14 (0.9)
Neurology	170 (2.9)	57 (3.7)			49 (3.2)	57 (3.7)
Neurosurgery	367 (6.2)	46 (3.0)			43 (2.8)	46 (3.0)
Other departments Insurance type (n, %) Korea Medicaid type I	681 (11.5) 247 (4.2)	86 (5.7) 99 (6.5)	0.179		90 (5.9) 100 (6.6)	86 (5.7) 99 (6.5)	0.041
Korea Medicaid type II	19 (0.3)	7 (0.5)			5 (0.3)	7 (0.5)
National health insurance	5546 (93.4)	1409 (92.5)			1409 (92.5)	1409 (92.5)
Motor vehicle insurance	52 (0.9)	2 (0.1)			2 (0.1)	2 (0.1)
no insurance	11 (0.2)	2 (0.1)			1 (0.1)	2 (0.1)
International insurance	63 (1.1)	4 (0.3)			6 (0.4)	4 (0.3)
Decision by attending (n, %)	4193 (70.6)	1208 (79.3)	0.202		1207 (79.3)	1208 (79.3)	0.003

ESSW, Emergency short-stay ward; SMD, standardized mean difference; SD, standard deviation; KTAS, Korean Triage and Acuity Scale.

Table 2

Comparison of outcomes in each group

Variables	Before matching				After matching
	Control (n = 5938)	ESSW (n = 1523)	P-value		Control (n = 1523)	ESSW (n = 1523)	P-value
Boarding time (minutes, median [IQR])	206.0 (569.0)	323.0 (895.5)	b0.001		329 (957)	323 (895.5)	0.237
ED LOS (minutes, median IQR])	558.0 (739.5)	671.0 (992.5)	b0.001		679 (1050.5)	671 (992)	0.163
Subsequent ICU admission (n, %)	304 (5.1)	37 (2.4)	b0.001		75 (4.9)	37 (2.4)	b0.001
30-day mortality (n, %)	153 (2.6)	32 (2.1)	0.287		41 (2.7)	32 (2.1)	0.292

ESSW, emergency short-stay ward; ED LOS, emergency department length of stay; ICU, intensive care unit.

Boarding time and ED LOS were analyzed using the Wilcoxon rank-sum test and are expressed as medians (interquartile ranges; IQRs). Subsequent ICU admission and 30-day mortality were analyzed using the chi-square test and McNemar’s test.

to the ESSW increased as bed occupancy increased, and the increase in boarding time was more gradual in the ESSW group than in the control group. This means that when the bed occupancy is low, the need for ESSW decreases because patients are admitted directly to the appropri- ate wards without long boarding times. The role of the ESSW in improv- ing the output component of ED crowding is better demonstrated when a high bed occupancy delays access to inpatient beds from the ED. Ac- cording to the queuing theory, bottlenecks in flow are formed when a system operates above 85%-90% capacity, thus decreasing efficiency [15]. Because a long boarding time generally occurs because of a bottle- neck in patient flow from the ED to the ward, the ESSW could effectively reduce boarding times when hospitals are completely occupied.

Recently, most major hospitals have implemented a geographic bed plan wherein wards are zoned for specific clinical areas. However, this plan contributes to the inefficient use of inpatient bed resources. Partic- ularly, many hospitals operate geographic bed plans focused on elective surgery because of economic motivation. This strategy leads to insuffi- cient bed resources for ED patients, resulting in an outflow obstruction from the ED. [16,17] Therefore, the ESSW was operated with a pooling bed plan in contrast to the geographic bed plan. The flexibility of re- source utilization in the pooling bed policy allowed more boarding pa- tients to be hospitalized. In addition, the decision for admission to the ESSW is made by ED physicians and not by staff in the administration department. Although there might be an available bed in the ward, it is difficult to assign patients immediately to available beds due to ad- ministrative aspects [6]. Therefore, the ESSW, which is intended to ac- commodate only patients admitted through the ED, helped to improve the efficiency of patient flow because patients were transferred from the ED to the ESSW as soon as a bed was vacated. The 48-hour limit

on an ESSW stay also maximized the function of the ESSW as a buffer space to alleviate ED crowding.

The ESSW, as a hospital-level strategy for ED crowding, is distinct from the ED-based short stay unit or inpatient hallway suggested in pre- vious reports [18-21]. The former units are usually designed for addi- tional diagnostic evaluations, pre-discharge treatment, or observations for patient safety before discharge from the ED. [13,14,19] In contrast, the ESSW only targets Boarding patients after a decision is made to admit and is operated using medical resources separate from that of ED. Thus, the initial stabilization can be completed as the patient is ad- mitted to the general ward, and a clinically identical level of care can be provided after entering the ESSW. Patients treated in hallways report a significantly lower level of satisfaction than those treated in general wards [20,21]. Moreover, the medical care given by physicians is disrupted in hallway settings by a lack of privacy and confidentiality, leading to adverse clinical outcomes [22]. In this study, the proportion of patients subsequently admitted to the ICU after admission was lower in the ESSW group than in the control group. Moreover, there was no significant difference in mortality between the two groups.

This study also had some limitations. First, it was conducted retro- spectively, and some confounders may have been unidentified. Second, we did not confirm the long-term effect of the ESSW because the anal- ysis was conducted using data spanning a relatively short period. Fi- nally, this study was conducted at a single tertiary hospital, thus limiting its generalizability. Further large-scale research and long-term analyses are needed to clearly elucidate the effect of the ESSW on ED crowding.

In conclusion, the ESSW, which targets only patients indicated for admission from the ED, can reduce prolonged boarding time during

Fig. 2. Proportion of patients entering the ESSW among those entering the ED. ESSW, emergency short-stay ward.

Fig. 3. Comparison of changes in boarding times between the two groups according to the change in hospital bed occupancy in the matching cohort. ESSW, emergency short-stay ward.

hospital bed saturation. Thus, the ESSW can be an effective strategy to address ED crowding caused by delayed admission to a general ward.

Financial support

This research received no specific grant from any funding agency, commercial or not-for-profit sectors.

Declaration of competing interest

None.

References

1. Singer AJ, Thode Jr HC, Viccellio P, Pines JM. The association between length of emer- gency department boarding and mortality. Acad Emerg Med 2011;18(12):1324-9.
2. Kulstad EB, Sikka R, Sweis RT, Kelley KM, Rzechula KH. ED overcrowding is associ- ated with an increased frequency of Medication errors. Am J Emerg Med 2010;28 (3):304-9.
3. Liu SW, Thomas SH, Gordon JA, Hamedani AG, Weissman JS. A pilot study examining undesirable events among emergency department-boarded patients awaiting inpa- tient beds. Ann Emerg Med 2009;54(3):381-5.
4. Solberg LI, Asplin BR, Weinick RM, Magid DJ. Emergency department crowding: con- sensus development of potential measures. Ann Emerg Med 2003;42(6):824-34.
5. Fatovich DM, Nagree Y, Sprivulis P. access block causes emergency department overcrowding and Ambulance diversion in Perth, Western Australia. Emerg Med J 2005;22(5):351-4.
6. Rabin E, Kocher K, McClelland M, Pines J, Hwang U, Rathlev N, et al. Solutions to emergency department “boarding” and crowding are underused and may need to be legislated. Health Aff (Millwood) 2012;31(8):1757-66.
7. Cass D. Once upon a time in the emergency department: a cautionary tale. Ann Emerg Med 2005;46(6):541-3.
8. Mitra B, Cameron PA, Archer P, Bailey M, Pielage P, Mele G, et al. The association be- tween time to disposition plan in the emergency department and in-hospital mor- tality of general Medical patients. Intern Med J 2012;42(4):444-50.
9. Chang AM, Cohen DJ, Lin A, Augustine J, Handel DA, Howell E, et al. Hospital strate- gies for reducing emergency department crowding: a mixed-methods study. Ann Emerg Med 2018;71(4):497-505.e4.
10. Baugh CW, Venkatesh AK, Bohan JS. Emergency department Observation Units: a clinical and financial benefit for hospitals. Health Care Manage Rev 2011;36(1): 28-37.
11. Ross MA, Naylor S, Compton S, Gibb KA, Wilson AG. Maximizing use of the emer- gency department observation unit: a novel hybrid design. Ann Emerg Med 2001; 37(3):267-74.
12. Strom C, Mollerup TK, Kromberg LS, Rasmussen LS, Schmidt TA. Hospitalisation in an emergency department short-stay unit compared to an internal medicine depart- ment is associated with fewer complications in older patients — an observational study. Scand J Trauma Resusc Emerg Med 2017;25(1):80.
13. Shetty AL, Teh C, Vukasovic M, Joyce S, Vaghasiya MR, Forero R. Impact of emergency department discharge stream short stay unit performance and hospital bed occu- pancy rates on access and patient flowmeasures: a single site study. Emerg Med Australas 2017;29(4):407-14.
14. Galipeau J, Pussegoda K, Stevens A, Brehaut JC, Curran J, Forster AJ, et al. Effectiveness and safety of short-stay units in the emergency department: a systematic review. Acad Emerg Med 2015;22(8):893-907.
15. Proudlove NC, Gordon K, Boaden R. Can good bed management solve the over- crowding in accident and emergency departments? Emerg Med J 2003;20(2): 149-55.
16. McManus ML, Long MC, Cooper A, Mandell J, Berwick DM, Pagano M, et al. Variabil- ity in surgical caseload and access to intensive care services. Anesthesiology 2003;98 (6):1491-6.
17. Vermeulen MJ, Ray JG, Bell C, Cayen B, Stukel TA, Schull MJ. Disequilibrium between admitted and discharged hospitalized patients affects emergency department length of stay. Ann Emerg Med 2009;54(6):794-804.
18. Viccellio A, Santora C, Singer AJ, Thode Jr HC, Henry MC. The association between transfer of emergency department boarders to inpatient hallways and mortality: a 4-year experience. Ann Emerg Med 2009;54(4):487-91.
19. Perimal-Lewis L, Ben-Tovim DI, Li JY, Hakendorf PH, Thompson CH. Emergency de- partment lengths of stay: characteristics favouring a delay to the admission decision as distinct from a delay while awaiting an inpatient bed. Intern Med J 2014;44(4): 384-9.
20. Garson C, Hollander JE, Rhodes KV, Shofer FS, Baxt WG, Pines JM. Emergency depart- ment patient preferences for boarding locations when hospitals are at full capacity. Ann Emerg Med 2008;51(1):9-12 [e1-3].
21. Viccellio P, Zito JA, Sayage V, Chohan J, Garra G, Santora C, et al. Patients overwhelm- ingly prefer inPatient boarding to emergency department boarding. J Emerg Med 2013;45(6):942-6.
22. Stoklosa H, Scannell M, Ma Z, Rosner B, Hughes A, Bohan JS. Do EPs change their clin- ical behaviour in the hallway or when a companion is present? A cross-sectional survey. Emerg Med J 2018;35(7):406-11.