a b s t r a c t

Background: Prolonged emergency department (ED) length of stay (LOS) is associated with poorer clinical out- comes and patient experience. At our community hospital, trauma patients were experiencing extended ED LOS incommensurate with their clinical status. Our objective was to determine if operational modifications to pa- tient flow would reduce the LOS for trauma patients.

Method: We conducted a retrospective chart review of admitted trauma patients from January 1, 2015 to June 30, 2016 to study two interventions. First, a communication intervention [INT1], which required the ED provider to directly notify the trauma service, was studied. Second, a bed intervention [INT2], which reserved two temporary beds for trauma patients, was added. The primary outcome was the average ED LOS change across three time pe- riods: (1) Baseline data [BASE] collected from January 1, 2015 to June 30, 2015, (2) INT1 data collected from July 1, 2015 to October 18, 2015, and (3) INT2 data collected from October 19, 2015 to June 30, 2016. Data was ana- lyzed using descriptive statistics, two-sample t-tests, and multivariate linear regression.

Results: A total of 777 trauma patients were reviewed, with 151, 150 and 476 reviewed during BASE, INT1, and INT2 time periods, respectively. BASE LOS for trauma patients was 389 min. After INT1, LOS decreased by

74.35 min (+-31.92; p b 0.0001). After INT2 was also implemented, LOS decreased by 164.56 min (+-22.97; p b

0.0001) from BASE LOS.

Conclusion: Direct communication with the trauma service by the ED provider and reservation of two temporary beds significantly decreased the LOS for trauma patients.

(C) 2018

Introduction

Inefficiencies in emergency department (ED) patient care or longer ED length of stay times have been linked to overcrowding. In overcrowding, ED physical or personnel capacity is exceeded by the number of patients waiting to be seen, undergoing assessment and treatment, and waiting for departure [1]. Both external and internal fac- tors contribute to an ED environment vulnerable to patient overflow and inefficiency. External elements include federal laws such as the Emergency medical treatment and Active Labor Act that mandates pa- tient care regardless of insurance or legal status, a lack of effective pre- ventative care, and inappropriate ED utilization for convenience care [2-5]. There is also the demographic component of increasingly older ED patients who often present with atypical, high acuity conditions on top of pre-existing comorbid illnesses that necessitate heavier resource

* Corresponding author at: 27405 Greenfield Rd Apt 7, Southfield, MI 48076-3641, United States of America.

E-mail address: [email protected] (D. Huang).

use along with a higher risk of post-ED discharge incidents [4-6]. Inter- nally, inefficient output, or access block, whereby admitted ED patients are not transferred to inpatient (IP) beds efficiently, contribute signifi- cantly to overcrowding. These inefficiencies may be compounded by the throughput of patients from ED arrival to admission or discharge. The effectiveness of these processes is a function of simpler elements, ranging from inadequate staffing and hospital bed shortages, to more complex considerations, such as the local patient population served and intra-hospital department dynamics [1,7,8]. Unfortunately, despite Affordable Care Act changes in the healthcare system designed to im- prove preventative care and insurance levels, some studies have shown that ED visits may be continuing to increase despite reforms, highlighting the continued significance of access block, throughput, and overcrowding in the ED [9-14].

Of Clinical concern, the phenomenon of ED overcrowding is linked to adverse outcomes related to cardiovascular disease, ineffective antibi- otic administration, pain management, and lower patient and staff sat- isfaction as well as general care quality [15-19]. These injurious ramifications extend to trauma patients, manifesting as increased

https://doi.org/10.1016/j.ajem.2018.08.021

0735-6757/(C) 2018

detrimental outcomes related to poor pain management, increased in- fection rates, and increases in trauma mortality resulting from ambu- lance diversion linked to ED overflow [20-22]. Various local interventions have been implemented to ameliorate overcrowding. For example, chest pain units, rapid assessment zones, clinical decision units (CDU), and lean principles that remove inefficient work processes have demonstrated varied benefits to ED patient care, such as reducing mortality and LOS [1,23-25]. Despite some improvements, the continu- ing demand for effective ED services underscores the imperative need to develop and study simple as well as affordable methods of facilitating care.

At our community hospital, a multidisciplinary trauma team tasked with optimizing ED care of injured patients identifiED throughput inef- ficiencies where our trauma patients were experiencing extended ED LOS incommensurate with their clinical status. This extended LOS was independent of access block. This prompted the implementation of a two part intervention to address the vulnerability. The primary objec- tive of this study focuses on whether two quality improvements:

(1) streamlining physician communication and consult placement with trauma providers and (2) reserving two temporary beds in the CDU for trauma patients, would effectively reduce components of as well as the total LOS for trauma patients in the ED at our community hospital. In addition, a secondary objective of the study focuses on a sub- group analysis considering differential impacts of the interventions based on demographic and disease severity variables.

Materials and methods

Study design, setting, and participants

We conducted a retrospective analysis of ED CDU trauma patient re- cords from January 1, 2015 to June 30, 2016. Inclusion criteria included adult ED trauma patients (age >=18) who were to be admitted to the trauma service through the CDU. Patients with Hip fractures and those requiring a higher level of care than a regular medical floor were ex- cluded as they would follow a different flow process. Patients moving directly to an operating room or to intensive care unit monitoring were also excluded.

Four time points were collected (occurring in the following order): ED arrival, emergency physician consultation to physician assistant (PA), admit decision, and ED departure (or time of placement into a CDU bed). Total LOS was defined as: time from ED “Arrival” to “Depar- ture” (see Fig. 1). Time data was collected in two registries: electronic medical record and ED Data Cube.

In the pre-intervention communication protocol, the EP would initi- ate a trauma consult by placing an order into the electronic medical re- cord for the trauma team PA. The ED secretary would then call the EP to ensure that the PA was notified. The PA would then conduct an elec- tronic medical record review and see the patient, initiating additional studies if appropriate and deciding whether to admit the patient to trauma. If admitted, the patient would be discharged or transferred to a bed (if available) in the CDU. The patient then completes their course of care under observation or IP status. Observation status is designated for patients who do not meet acute care criteria for IP admission.

A multidisciplinary trauma team identified various inefficiencies in the original protocol: (1) the ED secretary may miss the new consult order due to preoccupation with another task, (2) the EP and patient must wait for the PA decision to admit, (3) and the EP may also order tests and be unaware of an admit decision or new tests ordered by the PA. (4) As there may be no beds available in the CDU, a patient requiring admission may have to linger in the ED, unnecessarily extending patient waiting time.

Interventions

The first intervention was the communication intervention [INT1], which was implemented on July 1, 2015 and required the EP to directly notify the trauma service (via phone). INT1 simplified the admit process by removing the ED secretary’s role as a communication intermediary between the EP, who provides the initial consultation, and the PA. Direct notification established uninterrupted communication between the EP and PA in order to facilitate coordination of care and more transparent decision making (see Fig. 1). INT1 affected consultation to admit deci- sion time. The second intervention was the bed intervention [INT2], which was implemented on October 19, 2015 and reserved two tempo- rary beds in the CDU for trauma patients. INT2 affected admit decision to ED departure time.

Outcomes and statistical analysis

Our primary outcome was total LOS change across three time pe- riods: (1) Baseline data [BASE] collected from January 1, 2015 to June 30, 2015, (2) INT1 data collected from July 1, 2015 to October 18,

2015, and (3) INT2 data collected from October 19, 2015 to June 30, 2016. To evaluate the effect of INT1 and INT2 on total LOS, changes to LOS after INT1 but before INT2 and after both INT1 and INT2 were com- pared with the LOS before both INT1 and INT2 (BASE). Additionally, pre- and post-INT1 consult to admit decision wait times were examined to evaluate the individual effect of INT1. Pre- and post-INT2 admit decision

Fig. 1. Diagram depicting movement of emergency department trauma patients destined for the clinical decision unit . Patients are placed into the observation unit (OBS) or designated as inpatient (IP). The communication intervention, intervention 1 (INT1), and the bedding intervention (INT2), are depicted where they affect the flow process.

to departure wait times were examined to evaluate the individual effect of INT2. Categorical variables were analyzed with a Chi-Square Test and continuous variables were analyzed with a two-sample t-test. Statistical significance was defined as p b 0.05. Multivariate linear regression de- termined if significant differences existed before and after the interven- tions while adjusting for subgroup variables (Age Category, Trauma Level, Injury Severity Score (ISS), Gender, Ethnicity, and Primary Spoken Language). The ISS is a standardized medical score used to assess sever- ity of trauma, incorporating data such as injury location, mortality, mor- bidity, and hospitalization time [26]. In this study, an ISS score <=8 was considered minor whereas a score N8 was considered not minor.

To obtain our secondary outcomes, an interaction analysis was per- formed to determine how INT1 and INT2 impacted the various sub- groups. Subgroup analyses were planned a priori. Inclusion criteria for each subgroup were based on an empirical assumption that these vari- ables could differ in the investigated time intervals as well as be differ- entially impacted by the interventions.

Results

In 2016 our community hospital had an annual ED census of 98,324 visits and there was a total of 2830 trauma patients seen by the trauma team. The ED had a capacity of 65 beds with 18 ED observation beds and an ED medical admission rate of 26.1%. The overall IP licensed bed occu- pancy rate was 78.3% (total non-CDU beds: 520; CDU beds: 31), and the total occupancy rate was 92% (licensed IP and short stay, observation beds designated for patients). A total of 777 trauma patients were reviewed during the study period, with 151, 150 and 476 reviewed dur- ing the time periods of BASE, INT1 only, and both INTs, respectively (see Table 1). There were no changes to the total number of physician and nursing care providers throughout the time intervals studied. Addition- ally, there were no other ED changes related to the trauma patient pro- tocol outside of the interventions evaluated in this investigation.

Primary outcome

Total LOS decreased with statistical significance due to INT1 and INT2. Adjusted for subgroup differences, after INT1 was introduced the LOS decreased by 19.1% from BASE (see Table 2). After INT2 was also im- plemented, the LOS decreased by 42.3% from BASE. The decrease in LOS

Table 2

Linear regression results for evaluating intervention effects on total LOS, communication intervention effect on consult to admit time, and bed intervention effect on admit decision to departure time.

Linear regression results for evaluating intervention effects on wait times.

Minutes (95% CI)

Intervention effect on total LOS

After communication and before bed intervention -74.33 (-106.27,

-42.43)?

After communication and bed intervention -164.55 (-187.52,

-141.59)?

Before communication and bed intervention Reference group Communication intervention effect on consult to

admit time

After communication intervention -32.24 (-45.35, -19.13)?

Before communication intervention Reference group Bed intervention effect on admit to departure time

After bed intervention -73.00 (-87.39, -58.62)?

Before bed intervention Reference group

LOS, Length of Stay.

* p value b 0.05.

was independent of age and ISS, except for those patients age >=85, where INT1 did not result in a statistical difference.

With regards to individual analyses of each INT, before INT1 the ad- justed average consultation to admit time was 112 min, which de- creased to an adjusted average of 80 min following INT1 for an average decrease of 28.8% (see Table 2). Prior to INT2, the adjusted av- erage admit decision to departure time was 151 min, which decreased to an adjusted average of 78 min afterwards for an average decrease of 48.3%.

Secondary outcome

With regards to INT1 effect on age, there were statistically significant decreases in consult to admit decision wait time for participants ages 18-64 and 65-84. There were no significant changes in wait times for those aged >=85 (see Table 3). For INT2, all age groups saw statistically significant decreases in admit decision to departure times. Both inter- ventions had statistically significant reductions in wait times regardless of primary language spoken. With regards to ISS, both INTs had statisti- cally significant reductions in wait times independent of ISS. INT1 had a 24.0% greater reduction for minor ISS scores compared to non-minor ISS

Table 1

A total of 777 trauma patients were reviewed during the study period, with 151, 150 and 476 reviewed during the three time periods of baseline, only communication intervention im- plemented, and both communication and bed interventions implemented, respectively. Patient numbers divided based on subgroups are shown.

Patient volume stratified by demographics and subgroup in each time interval.

	Baseline (n = 151)	Communication intervention only (n = 150)	Both communication and bed interventions (n = 476)
Age Categories Age 18-64	46 (30.5%)	58 (39%)	173 (36.3%)
Age 65-84	68 (45.0%)	47 (31%)	180 (37.8%)
Age 85+	37 (24.5%)	45 (30%)	123 (25.9%)
Trauma Level Trauma consult	148 (98.0%)	136 (91%)	441 (92.6%)
Trauma level 1/2 activation	3 (2.0%)	14 (9%)	35 (7.4%)
Injury Severity Score Minor (ISS <= 8)	109 (72.2%)	114 (76%)	316 (66.4%)
Not minor (ISS N 8)	42 (27.8%)	36 (24%)	160 (33.6%)
Gender Male	56 (37.1%)	69 (46%)	178 (37.4%)
Female	95 (62.9%)	81 (54%)	298 (62.6%)
Ethnicity Not Hispanic/Latino	121 (80.1%)	123 (82%)	387 (81.3%)
Arabic or Middle Eastern Descent	12 (8.0%)	13 (9%)	27 (5.7%)
Other/not answered/unavailable Language	18 (11.9%)	14 (9%)	62 (13.0%)
English	134 (88.7%)	135 (90%)	437 (91.8%)
Other/unknown	17 (11.3%)	15 (10%)	39 (8.2%)
ISS, Injury Severity Score.

Table 3

Interaction analysis for evaluating differential effects of each intervention on various subgroups of the study population. Times refer to consult to admit decision time for the communi- cation intervention and admit decision to departure for the bed intervention.

Interaction analysis for evaluating differential intervention effects on subgroups.

	Communication intervention Minutes (95% CI)		Bed intervention Minutes (95% CI)
Age Categories
Age 18-64 Age 65-84 Age 85+	-48.52 (-75.38, -21.67) -35.28 (-53.76, -16.79) -6.38 (-28.26, 15.50)		-64.80 (-87.61, -41.99) -89.22 (-114.36, -64.09) -60.76 (-85.82, -35.69)
Trauma Level
Trauma level 1/2 activation	-25.26 (-101.84, 51.33)		-120.34 (-228.68, -12.00)
Trauma consult	-32.41 (-45.71, -19.10)		-69.98 (-83.56, -56.40)
Injury Severity Score Category Minor (ISS <= 8)	-34.35 (-50.26, -18.44)	-63.22 (-78.30, -48.15)
Not minor (ISS N 8)	-26.99 (-49.45, -4.52)	-97.15 (-129.68, -64.63)
Gender
Male	-46.31 (-67.88, -24.74)	-81.12 (-106.72, -55.52)
Female	-23.82 (-40.03, -7.60)	-67.62 (-83.86, -51.37)
Ethnicity
Arabic or Middle Eastern	-17.57 (-60.39, 25.25)	-88.96 (-145.70, -32.23)
Other/no answer/unavailable	-28.56 (-62.46, 5.34)	-79.57 (-116.70, -42.44)
Not Hispanic/Latino	-34.15 (-47.26, -21.05)	-70.69 (-86.48, -54.89)
Primary Spoken Language
English	-34.04 (-46.47, -21.61)	-70.70 (-85.80, -55.60)
Other/unknown	-17.00 (-53.14, 19.16)	-95.52 (-141.70, -49.34)

ISS, Injury Severity Score.

scores. This was reversed in INT2, which showed a 42.3% greater reduc- tion for non-minor ISS scores compared to minor ISS scores.

Discussion

Previous studies have shown that ED overcrowding results from multiple factors, ranging from an increasingly older patient population to inefficiencies in Patient throughput and output. Importantly, trauma patients are particularly susceptible to the injurious consequences of overcrowding, with detrimental outcomes related to increased infec- tions as well as trauma mortality [20-22]. Capacity interventions, such as chest pain Observation Units and CDUs, as well as lean interventions cutting unnecessary throughput steps have demonstrated the potential to reduce overcrowding [1,23-25]. Prior studies have also emphasized that lean intervention success may be limited to hospital areas with simpler flow processes and may be more effective for non-admitted, low acuity patients, such as those seen in ED CDUs [7,25,27,28]. Our re- sults corroborate prior evidence on the nature and effectiveness of throughput interventions in an ED setting less compromised by access block.

This investigation focused on a primary objective of whether a communication-based lean intervention and a capacity bed reservation can effectively reduce total LOS for CDU trauma patients in the ED at our community hospital. The results show that both the communication streamlining and additional bed placement have statistically significant and independent effects on reducing LOS for CDU trauma patients. There were sequential reductions in total LOS as each intervention was implemented, indicating that the lean intervention did not produce a discernible downstream bottleneck, a common concern and conse- quence of lean interventions [7]. Notably, the combined total reduction in LOS after both INTs were implemented was greater than the com- bined individual INT1 and INT2 wait time reductions of their affectED flow intervals (see Table 1). There are several possible explanations for this discrepancy. Firstly, the Hawthorne effect, whereby care pro- viders are influenced by their awareness of being observed, may result in increasED efficiency independent of the interventions themselves. In addition, the total LOS includes the patient arrival to consult time in- terval, which was not analyzed in the current study as there was no spe- cific intervention targeting this interval (see Fig. 1). The discrepancy in the total LOS may have resulted from this portion of the throughput.

Despite the possibility of observation effects, there were statistically sig- nificant and distinct wait time decreases averaged over relatively long intervals for the two specific intervals studied. Overall, in an ED setting less compromised by access block, these results underscore the feasibil- ity of effectively implementing both a communication-based lean inter- vention along with a capacity intervention in a relatively quick, sequential fashion.

The secondary objective investigated any differential impact of the interventions based on demographic variables. In contrast to INT2, which had a statistically significant reduction in wait times regardless of age group, INT1 had no effect on wait times for the >=85 age group (see Table 2). The more complex nature of older ED patients, often re- quiring additional testing and more interdisciplinary teamwork, may complicate the effectiveness of any individual communication interven- tion [4,6]. The decreased effectiveness of INT1 on more complex pa- tients was also reflected in smaller wait time reductions for non- minor ISS patients. However, this was reversed in INT2, where addi- tional bedding was more effective for the higher ISS group. These find- ings indicate that considering both direct, fluid communications as well as capacity interventions may facilitate better care for more com- plex, older patients.

When comparing wait time reductions based on language spo- ken, both interventions had statistically significant reductions re- gardless of the whether English was the primary language of the patient. This observation may indicate that potential communication issues between the patients and the providers do not affect the established communication between the EP and the PA. Of note, wait time reductions in LOS due to INT1 were not significant for patients selecting Arabic/Middle Eastern and the Other/no answer/ unavailable option as their ethnicity (see Table 2). However, unlike language spoken, healthcare providers may not be able to discern pa- tient ethnicity, reducing the likelihood of provider and intervention- related bias. Though the possibility of provider bias and potential ef- fects on patient care is prevalent in the healthcare field, the nature of communication and capacity interventions as well as throughput protocols is likely unaffected as demonstrated by significant reductions in wait times independent of language spoken [29-32]. Considering demographic and injury severity variables in project im- provement studies may help to tailor future interventions to more specific patient populations.

Limitations and future studies

The complexity of interdepartmental dynamics in the hospital and ED variability amplify the challenges in implementing evidence-based ED efficiency interventions. As with many other lean studies, this study is based on a before-after single site model without concurrent control comparisons sites. This limits randomization and obscures whether any increases in efficiency can be attributed to the studied in- terventions as opposed to external or internal Policy changes, such as performance incentives and other hospital lean interventions that may indirectly cause reductions in LOS. The results may also not be easily generalizable due to differences in staff profile and use of a CDU, which are likely to vary widely among hospital EDs. Additionally, this study cannot rule out an intervention-independent, additional reduc- tion in wait times resulting from the Hawthorne effect. Lastly, the study assumes that patient variability factors such as seasonally varying cases will not take longer to process. This assumption is made because the interventions only involve patient processing as opposed to specific treatments that would be affected by seasonal variations. Future inves- tigations should focus on these limitations by replicating the study with multiple ED sites with a comparison control site.

Financial elements should be explored as even simple lean inter- ventions may be associated with significant costs. Expenditures as- sociated with external consultants, data analysis, staff training, and time investment may counter any cost savings attributed to a re- duced LOS, compromising the impetus to improve efficiency [33- 35]. In our investigation, comparing the calculated patient-hours daily before either intervention and after both interventions demon- strates a benefit of around 5 patient-hours daily, or a 42% reduction in patient-hours daily compared to baseline. Although the interven- tions focused on trauma patients, the improvement in throughput can leave secondary gains throughout the ED. These benefits can manifest as increased patient satisfaction, increased bed availability leading to reducED boarding, and Improved care provider availability [7,8]. Due to the interconnected nature of ED processes, future stud- ies should also focus on evaluating changes in clinical outcomes and identifying possible downstream effects of the interventions that can manifest insidiously in other sections of the ED and hospital depart- ments [7].

Conclusion

Direct communication with the trauma service by the ED provider and the reservation of two temporary trauma beds significantly de- creased the ED LOS for trauma patients. These findings have both imme- diate and long term impacts. At the immediate level, these findings underscore the ability of simple throughput interventions to reduce LOS, which may improve patient experience and mitigate the deleteri- ous consequences of overcrowding [15-19]. On a wider scale, this study reinforces the effectiveness of evidence-based throughput strate- gies that are increasingly crucial in addressing the seismic demographic changes of an increasingly older, more complex patient population in the ED and beyond [4-6].

Declarations of interest

None.

Supplemental

As of June 9, 2018, these interventions have remained in the system at our community hospital due to their continued effectiveness and the results of the initially reported data.

Acknowledgements

Patrick Karabon, statistician: assistance with statistical analysis and consultation.

This research did not receive any specific grant from funding agen- cies in the public, commercial, or not-for-profit sectors.

References

1. Lo S, Choi K, Wong E, et al. Effectiveness of Emergency Medicine Wards in reducing length of stay and overcrowding in emergency departments. Int Emerg Nurs 2014; 22(2):116-20. https://doi.org/10.1016/j.ienj.2013.08.003.
2. Akbari A, Mayhew A, Al-Alawi M, et al. Interventions to improve outpatient referrals from primary care to secondary care. Cochrane Database Syst Rev 2008. https://doi. org/10.1002/14651858.cd005471.pub2.
3. Ross M, Hockenberry J, Mutter R, Barrett M, Wheatley M, Pitts S. Protocol-driven emergency department observation units offer savings, shorter stays, and reduced admissions. Health Aff 2013;32(12):2149-56. https://doi.org/10.1377/hlthaff.2013. 0662.
4. Pines J, Mullins P, Cooper J, Feng L, Roth K. National trends in emergency department use, care patterns, and quality of care of older adults in the United States. J Am Geriatr Soc 2013;61(1):12-7. https://doi.org/10.1111/jgs.12072.
5. Howard M, Davis B, Anderson C, Cherry D, Koller P, Shelton D. Patients’ perspective on choosing the emergency department for nonurgent medical care: a qualitative study exploring one reason for overcrowding. J Emerg Nurs 2005;31(5):429-35. https://doi.org/10.1016/j.jen.2005.06.023.
6. Latham L, Ackroyd-Stolarz S. Emergency department utilization by older adults: a descriptive study. Can Geriatr J 2014;17(4). https://doi.org/10.5770/cgj.17.108.
7. Kreindler S. Six ways not to improve patient flow: a qualitative study. BMJ Qual Saf 2016. https://doi.org/10.1136/bmjqs-2016-005438 (bmjqs-2016-005438).
8. Rabin E, Kocher K, McClelland M, et al. Solutions to emergency department ‘boarding’ and crowding are underused and may need to be legislated. Health Aff 2012;31(8):1757-66. https://doi.org/10.1377/hlthaff.2011.0786.
9. Smulowitz P, O’Malley J, Yang X, Landon B. Increased use of the emergency depart- ment after health care reform in Massachusetts. Ann Emerg Med 2014;64(2): 107-115.e3. https://doi.org/10.1016/j.annemergmed.2014.02.011.
10. Taubman S, Allen H, Wright B, Baicker K, Finkelstein A. Medicaid increases emergency-department use: evidence from Oregon’s Health Insurance Experiment. Science 2014;343(6168):263-8. https://doi.org/10.1126/science.1246183.
11. Dresden S, Powell E, Kang R, McHugh M, Cooper A, Feinglass J. Increased emergency department use in Illinois after implementation of the patient protection and afford- able care act. Ann Emerg Med 2017;69(2):172-80. https://doi.org/10.1016/j. annemergmed.2016.06.026.
12. Cummings S. Increased Primary care access: can it reduce emergency department overcrowding? Acad Emerg Med 2005;12(Supplement 1):151. https://doi.org/10. 1197/j.aem.2005.03.426.
13. Hernandez-Boussard T, Morrison D, Goldstein B, Hsia R. Relationship of Affordable Care Act implementation to emergency department utilization among young adults. Ann Emerg Med 2016;67(6):714-720.e1. https://doi.org/10.1016/j.annemergmed. 2015.11.034.
14. Akosa Antwi Y, Moriya A, Simon K, Sommers B. Changes in emergency department use among young adults after the Patient Protection and Affordable Care Act‘s De- pendent Coverage Provision. Ann Emerg Med 2015;65(6):664-672.e2. https://doi. org/10.1016/j.annemergmed.2015.01.010.
15. Pines J, Pollack C, Diercks D, Chang A, Shofer F, Hollander J. The association between emergency department crowding and adverse cardiovascular outcomes in patients with chest pain. Acad Emerg Med 2009;16(7):617-25. https://doi.org/10.1111/j. 1553-2712.2009.00456.x.
16. Fee C, Weber E, Maak C, Bacchetti P. Effect of emergency department crowding on time to antibiotics in patients admitted with community-acquired pneumonia. Ann Emerg Med 2007;50(5):501-509.e1. https://doi.org/10.1016/j.annemergmed. 2007.08.003.
17. Hwang U, Richardson L, Livote E, Harris B, Spencer N, Sean Morrison R. Emergency department crowding and decreased quality of pain care. Acad Emerg Med 2008; 15(12):1248-55. https://doi.org/10.1111/j.1553-2712.2008.00267.x.
18. Pines J, Hollander J. Emergency department crowding is associated with poor care for patients with severe pain. Ann Emerg Med 2008;51(1):1-5. https://doi.org/10. 1016/j.annemergmed.2007.07.008.
19. Pines J, Iyer S, Disbot M, Hollander J, Shofer F, Datner E. The effect of emergency de- partment crowding on patient satisfaction for admitted patients. Acad Emerg Med 2008;15(9):825-31. https://doi.org/10.1111/j.1553-2712.2008.00200.x.
20. Carr B, Kaye A, Wiebe D, Gracias V, Schwab C, Reilly P. Emergency department length of stay: a major risk factor for pneumonia in intubated blunt trauma patients. J Trauma 2007;63(1):9-12. https://doi.org/10.1097/ta.0b013e31805d8f6b.
21. Begley C, Chang Y, Wood R, Weltge A. Emergency department diversion and trauma mortality: evidence from Houston, Texas. J Trauma 2004;57(6):1260-5. https://doi. org/10.1097/01.ta.0000135163.60257.a6.
22. Hwang U, Richardson L, Sonuyi T, Morrison R. The effect of emergency department crowding on the management of pain in older adults with hip fracture. J Am Geriatr Soc 2006;54(2):270-5. https://doi.org/10.1111/j.1532-5415.2005.00587.x.
23. Roberts M, Baird W, Kerr P, O’Reilly S. Can an emergency department-based Clinical Decision Unit successfully utilize alternatives to emergency hospitaliza- tion? Eur J Emerg Med 2010;17(2):89-96. https://doi.org/10.1097/mej. 0b013e32832f05bf.
24. Vermeulen M, Stukel T, Guttmann A, et al. Evaluation of an emergency department lean process improvement program to reduce length of stay. BMC Health Serv Res 2014;14(Suppl. 2):P110. https://doi.org/10.1186/1472-6963-14-s2-p110.
25. Rezaei F, Yarmohammadian M, Haghshenas A, Tavakoli N. Overcrowding in emer- gency departments: a review of strategies to decrease future challenges. J Res Med Sci 2017;22(1):23. https://doi.org/10.4103/1735-1995.200277.
26. Kuo S, Kuo P, Chen Y, Chien P, Hsieh H, Hsieh C. Comparison of the new exponential injury severity score with the injury severity score and the New Injury Severity Score in trauma patients: a cross-sectional study. PLoS One 2017;12(11):e0187871. https://doi.org/10.1371/journal.pone.0187871.
27. Hoot N, Aronsky D. Systematic review of emergency department crowding: causes, effects, and solutions. Ann Emerg Med 2008;52(2):126-136.e1. https://doi.org/10. 1016/j.annemergmed.2008.03.014.
28. Asplin B, Magid D, Rhodes K, Solberg L, Lurie N, Camargo C. A conceptual model of emergency department crowding. Ann Emerg Med 2003;42(2):173-80. https:// doi.org/10.1067/mem.2003.302.
29. Chen E, Shofer F, Dean A, et al. gender disparity in analgesic treatment of emergency department patients with acute abdominal pain. Acad Emerg Med 2008;15(5): 414-8. https://doi.org/10.1111/j.1553-2712.2008.00100.x.
30. Pletcher M, Kertesz S, Kohn M, Gonzales R. Trends in opioid prescribing by race/eth- nicity for patients seeking care in US emergency departments. JAMA 2008;299(1). https://doi.org/10.1001/jama.2007.64.
31. Todd K, Ducharme J, Choiniere M, et al. Pain in the emergency department: results of the pain and emergency medicine initiative (PEMI) multicenter study. J Pain 2007;8

    (6):460-6. https://doi.org/10.1016/j.jpain.2006.12.005.

    Green C, Anderson K, Baker T, et al. The unequal burden of pain: confronting racial and Ethnic disparities in pain. Pain Med 2003;4(3):277-94. https://doi.org/10. 1046/j.1526-4637.2003.03034.x.

32. Bayley M, Schwartz J, Shofer F, et al. The Financial burden of emergency department congestion and Hospital crowding for chest pain patients awaiting admission. Ann Emerg Med 2005;45(2):110-7. https://doi.org/10.1016/j.annemergmed.2004.09. 010.
33. McConnell K, Richards C, Daya M, Weathers C, Lowe R. Ambulance diversion and lost hospital revenues. Ann Emerg Med 2006;48(6):702-10. https://doi.org/10.1016/j. annemergmed.2006.05.001.
34. Pines J, Batt R, Hilton J, Terwiesch C. The financial consequences of lost demand and reducing boarding in hospital emergency departments. Ann Emerg Med 2011;58 (4):331-40. https://doi.org/10.1016/j.annemergmed.2011.03.004.