a b s t r a c t

Background: Patients with traumatic intracranial hemorrhage and mild traumatic brain injury (mTIH) receive broadly variable care which often includes transfer to a trauma center, neurosurgery consultation and ICU admis- sion. However, there may be a low risk cohort of patients who can be managed without utilizing such significant resources.

Objective: Describe mTIH patients who are at low risk of clinical or radiographic decompensation and can be safe- ly managed in an ED observation unit (EDOU).

Methods: Retrospective evaluation of patients age >= 16, GCS >= 13 with ICH on CT. Primary outcomes included clin- ical/Neurologic deterioration, CT worsening or need for neurosurgery.

Results: 1185 consecutive patients were studied. 814 were admitted and 371 observed patients (OP) were mon- itored in the EDOU or discharged from the ED after a period of observation. None of the OP deteriorated clinically. 299 OP (81%) had a single lesion on CT; 72 had mixed lesions. 120 patients had isolated subarachnoid hemor- rhage (iSAH) and they did uniformly well. Of the 119 OP who had subdural hematoma , 6 had worsening CT scans and 3 underwent burr hole drainage procedures as inpatients due to persistent SDH without new deficit. Of the 39 OP who had cerebral contusions, 3 had worsening CT scans and one required NSG admission. No patient returned to the ED with a complication. Follow-up was obtained on 81% of OP. 2 patients with SDH required burr hole procedure N 2 weeks after discharge.

Conclusions: Patients with mTIH, particularly those with iSAH, have very low rates of clinical or radiographic de- terioration and may be safe for monitoring in an emergency department observation unit.

(C) 2016

Introduction

Patients with Minor head injury make up about 6-7% of emergency department visits in the United States. The vast majority of these pa- tients have minimal injury and are discharged home without sequelae. However, the overall in-hospital mortality of Traumatic brain injury is 8% and increases to greater than 40% in patients with severe TBI, emphasizing the broad spectrum of disease severity [1,2]. The deci- sion to image a patient with head trauma has been extensively studied and evidence-based rules are available to guide this decision [3,4]. How- ever, support for the emergency physician’s Disposition decision is lack- ing, particularly in patients with traumatic intracranial hemorrhage

? Presented at the Society for Academic Emergency Medicine Scientific Assembly, San Diego, California, May 2015.

* Corresponding author at: 211 East Ontario Street, Suite 200, Chicago, IL 60611, United States.

E-mail address: [email protected] (P. Pruitt).

(TIH) and a relatively intact mental status. Historically, most patients with intracranial hemorrhage have been admitted to a trauma center and frequently to the ICU, regardless of the severity or type of hemor- rhage [5]. However, the clinical status of these patients and their under- lying pathology varies widely. There may be some subgroups which will have excellent outcomes without the resource expenditure and risks of intensive monitoring.

When Emergency Department Observation Units (EDOU) were first developed, they specialized in the evaluation of chest pain syndromes in patients at low risk for coronary ischemia [6]. The use of EDOUs has sub- sequently expanded to include patients with asthma, CHF, TIA and even minor traumatic injuries [7]. In these units, emergency providers facili- tate focused, expedited evaluations in a setting that requires less time and incurs less cost than a regular inpatient admission [8]. In the past, the use of EDOUs has been described for monitoring of patients with Traumatic brain injuries, especially patients who are on anticoagulation and have had head injury without evidence of intracranial hemorrhage

http://dx.doi.org/10.1016/j.ajem.2016.10.064

0735-6757/(C) 2016

[9]. To our knowledge, there is no existing literature about the safety of monitoring of patients with intracranial hemorrhage in observation.

We defined mild traumatic intracranial hemorrhage (mTIH) as pa- tients with radiographic hemorrhage and a preserved GCS (>= 13). We hypothesized that these patients can be safely managed in an EDOU. To evaluate the safety of this approach, we reviewed consecutive pa- tients with mTIH who were treated in our institution’s EDOU.

Methods

Study-design

This retrospective chart review was performed at an urban academic Level I trauma center with a volume of over 100 000 annual visits. Pa- tients were identified by running a query of our proprietary Emergency Department Information System (EDIS) using the International Statisti- cal Classification of Diseases and Related Health Problems (ninth edi- tion) codes for TIH (852.00-853.10, 851.00-851.90, 800.00-801.9,

803.00-804.9). All patients with a traumatic intracranial hemorrhage or Skull fracture who were age 16 years or older and had a GCS of 13 or higher on admission (adults with mild TIH), presenting between Jan- uary 1, 2009 and June 30, 2013, were included in this study. As is routine at our institution, all patients who presented with Intracranial injury re- ceived a neuroSurgical consultation. The hospital’s institutional review board approved the study.

Study protocol

All data were collected using Microsoft Access (Microsoft Corpora- tion, Redmond, WA). After the initial query of the medical records was conducted and ineligible patients were excluded, the study team ab- stracted data from physician notes, Radiology reports, laboratory data, and discharge summaries into a standardized form. Two investigators who were unblinded to the study hypothesis abstracted the data. Clini- cal and radiographic scoring criteria were standardized between the re- viewers a priori via the creation of a scoring protocol. Data were collected on a standardized data collection form.

Variables collected included age, gender, Method of arrival (walk-in, ground ambulance, Air ambulance), direct from scene versus inter- facility transfer, disposition from ED (home, floor admission, ICU admis- sion, ED observation unit), admitting service, medical history including hypertension and anticoagulant use, mechanism of injury, initial GCS, initial mental status, cranial nerve exam, motor exam, sensory exam, al- cohol intoxication, initial platelet count and international normalized ratio level (INR), initial cranial computed tomography results, follow-up CT results, and neurosurgical consultant recommendations. cranial CT data were collected from attending radiologist reports noting the presence of subdural hematoma (location, size, presence of shift), intracranial contusion (number of contusions, location, size), subarach- noid hemorrhage (location, size), epidural hematoma (location, size), and skull fracture (location, type). CT results were abstracted by a sepa- rate reviewer who was blinded to the patient and clinical information while reviewing the CT scan.

Observation status was defined as continued monitoring and evaluation for a period of greater than 6 h per neurosurgery recommen- dations or ED attending discretion. Neurosurgical consult recommenda- tions were followed, and, if a bed became available, patients were moved to the EDOU. Exclusion criteria for transfer to the observation unit or Discharge home were limited to polytrauma, spontaneous (non-traumatic) hemorrhage, GCS b 13, active treatment with intrave- nous Antihypertensive medications or ongoing correction of a coagulop- athy (with either fresh frozen plasma or platelets). Patients with coagulopathy (including warfarin, novel oral anticoagulant or aspirin use) were transferred to the observation unit once any neurosurgery treatment recommendations regarding reversal had been implement- ed. The observation unit was able to provide oral and single dose

intravenous antihypertensive and antiepileptic medications. We used an intention to treat approach such that if the patient completed the recommended period of observation in the ED and was discharged without being physically transferred to the observation unit, they were included in our study population. The final decision to admit was made by the neurosurgery and ED teams.

The electronic health record was used to determine if the patient had any follow up visits, and if, on follow-up, they noted persistent or new neurologic complaints. This review included records from our insti- tution as well as many of the other hospitals within our healthcare sys- tem, including several other Tertiary care centers and most of the region’s head injury referral centers.

Outcome measures

The primary outcomes were clinical deterioration (defined as de- crease in mental status, worsening neurologic exam or death) or the need for neurosurgery during admission. These endpoints was designed to reflect conditions that would trigger a transfer for neurosurgical con- sultation during an observation admission at a community hospital. In order to focus on patient-centered outcomes, radiographic worsening was considered as a secondary outcome separate from clinical deterio- ration. Patients who required burr-hole drainage for sub-acute and chronic subdural hematomas were classified as having a neurosurgical intervention.

Data analysis

Continuous variables were summarized by mean and standard devi- ation (SD) and categorical variables were summarized by count and percentage. The risk of deterioration was compared using chi-squared tests and summarized using odds ratios with 95% confidence intervals (CIs) in the univariate analysis. A difference was considered statistically significant at a p value b 0.05. Inter-rater reliability was assessed on the first 100 charts, with kappa being 0.95 for the 3 outcome variables and

0.91 for the remainder of the clinical variables.

Results

1185 patients met our initial inclusion criteria. 814 were admitted to the hospital directly from the ED and were therefore excluded from our analysis. This group consisted of patients with multisystem trauma who were admitted to the trauma service per our hospital protocol or pa- tients with more Severe head injuries who were admitted to the neuro- surgical service at the discretion of the neurosurgery team. The remaining 371 observed patients (OP) had isolated head trauma. Of these, 239 were transferred to the EDOU while 132 were discharged di- rectly from the ED after a period of observation without being moved to the EDOU. Patient demographics are summarized in Table 1. A break- down of the different types of lesion and their frequency is summarized in Table 2. Compared to the OPs, the admitted patients tended to have lower GCS and an increased rate of anticoagulant/antiplatelet drug use. OPs were more likely to have a single type of intracranial lesion (74.3% vs. 24%, p b 0.0001). For patients who were transferred to the ob- servation unit, there was a subsequent admission rate of 2.9% if patients with epidural hematoma were excluded, or a rate of 5% if EDH patients were included. Outpatient follow-up was obtained via medical record review in 81% of OP.

299 (81%) of OP had a single type of lesion on CT, while the remain- ing 72 had multiple types of lesion. The 120 patients with isolated sub- arachnoid hemorrhage (iSAH), comprising 32% of the study population, did uniformly well. No patient with iSAH had clinical or radiographic de- terioration throughout their period of observation. Two patients with iSAH required admission to the hospital, however both admissions were unrelated to head trauma: one patient was admitted for operative management of a radius fracture, the other was admitted to treat atrial

Table 1 Observed patient demographics.
	Admitted	Discharged from ED	ED observation unit	p (home vs	Totals observed	p
	N = 814	N = 132	N = 239	EDOU)	patients	(admitted vs
						total observed)
Demographic/history Age (range)	60.1 (16-101)	50.6 (17-94)	58.4 (18-95)	b 0.0001a
Sex (M)	497 (61.1%)	88 (66.7%)	166 (69.4%)	0.57	254 (68.5%)	0.01
Hypertension	313 (38.5%)	35 (26.5%)	94 (39.3%)	0.01	129 (34.8%)	0.22
GCS 15	634 (77.9)	124 (94%)	218 (91.2%)	0.35	342 (92.2%)	b 0.0001
alcohol/drug use	191 (23.5%)	26 (19.6%)	29 (12.1%)	0.05	55 (14.8%)	0.0007
Any coagulopathy	409 (50.2)	34 (25.8%)	85 (35.6%)	0.05	119 (32.1%)	b 0.0001
Aspirin	262 (32.2%)	24 (18.1%)	68 (28.4%)	0.03	92 (24.8%)	0.01
Clopidogrel	42 (5.2%)	4 (3%)	5 (2.1%)	0.73b	9 (2.4%)	0.002
Warfarin	105 (12.9%)	6 (4.5%)	12 (5%)	0.84	18 (4.9%)	b 0.0001
Mechanism of injury
Fall	567 (69.6%)	87 (65.9%)	157 (65.6%)	1.0	244 (65.8%)	0.18
Assault	96 (11.8%)	30 (22.7%)	28 (11.7%)	0.05	58 (15.6%)	0.07
Motor vehicle crash	89(10.9%)	9 (6.8%)	14 (5.9%)	0.72	23 (6.2%)	0.01
Pedestrian struck	43 (5.3%)	3 (2.3%)	16 (6.7%)	0.06	19 (5.1%)	0.92
Bicyclist struck	6 (0.7%)	2 (1.5%)	16 (6.7%)	0.03	18 (4.9%)	b 0.0001
Motor cycle crash	11 (1.4)	1 (0.7%)	2 (0.8%)	1.0b	3 (0.8%)	0.57b
Unknown/other	2 (0.2%)	0 (0%)	6 (2.5%)	0.09b	6 (1.6%)	0.01b
Demographics Age (range)	60.1 (16-101)	50.6 (17-94)	58.4 (18-95)	b 0.0001a
Sex (M)	497 (61.1%)	88 (66.7%)	166 (69.4%)	0.57	254 (68.5%)	0.01
History of hypertension	313 (38.5%)	35 (26.5%)	94 (39.3%)	0.01	129 (34.8%)	0.22
GCS 15	634 (77.9)	124 (94%)	218 (91.2%)	0.35	342 (92.2%)	b 0.0001
Alcohol/Drug Use	191 (23.5%)	26 (19.6%)	29 (12.1%)	0.05	55 (14.8%)	0.0007
Coagulopathy
Any	409 (50.2)	34 (25.8%)	85 (35.6%)	0.05	119 (32.1%)	b 0.0001
Aspirin	262 (32.2%)	24 (18.1%)	68 (28.4%)	0.03	92 (24.8%)	0.01
Clopidogrel	42 (5.2%)	4 (3%)	5 (2.1%)	0.73b	9 (2.4%)	0.002
Warfarin	105 (12.9%)	6 (4.5%)	12 (5%)	0.84	18 (4.9%)	b 0.0001
Mechanism of injury
Fall	567 (69.6%)	87 (65.9%)	157 (65.6%)	1.0	244 (65.8%)	0.18
Assault	96 (11.8%)	30 (22.7%)	28 (11.7%)	0.05	58 (15.6%)	0.07
Motor vehicle crash	89(10.9%)	9 (6.8%)	14 (5.9%)	0.72	23 (6.2%)	0.01
Pedestrian struck	43 (5.3%)	3 (2.3%)	16 (6.7%)	0.06	19 (5.1%)	0.92
Bicyclist struck	6 (0.7%)	2 (1.5%)	16 (6.7%)	0.03	18 (4.9%)	b 0.0001
Motor cycle crash	11 (1.4)	1 (0.7%)	2 (0.8%)	1.0b	3 (0.8%)	0.57b
Unknown/other	2 (0.2%)	0 (0%)	6 (2.5%)	0.09b	6 (1.6%)	0.01b

Otherwise chi square comparisons. GCS, Glasgow Coma Score. Proportions of patients in each column provide the basis for comparisons.

^a t-test.

^b Fisher Exact test.

fibrillation. During subsequent follow-up visits, multiple patients with SAH complained of headaches, mental fogginess and concentration dif- ficulties. No SAH patients required surgical intervention on follow-up.

Of the 169 OP who had subdural hematomas (SDH), 6 (3.6%) had worsening CT scans and 3 of those (1.8%) underwent a burr-hole drain- age procedure during their index hospitalization. All of these interven- tions were planned; the decision to operate was made days prior to the procedure due to Persistent symptoms. As with the subarachnoid patients, none of these patients deteriorated clinically. Two patients in the SDH group required delayed surgical management after discharge due to persistent symptoms and ongoing presence of SDH on CT scan; both surgical interventions occurred greater than two weeks after the initial injury.

Of the 39 OP with cerebral contusions, 5 had worsening CT scans. As with the SAH and SDH groups, none of the observed patients had a clin- ical deterioration. Two of those patients required admission to the neu- rosurgery service. Neither patient required a surgical intervention and both were discharged from the hospital after an inpatient stay. Two pa- tients with contusion reported persistent headaches on their follow-up visit. One patient had a new subdural hematoma on a repeat visit 2.5 months later; there was intervening trauma noted and so this was likely unrelated to the initial traumatic injury.

There were 5 patients who were admitted to the EDOU with epidu- ral hematomas. All five of these patients were eventually admitted to

the neurosurgery service. However, none of them had any clinical dete- rioration or significant radiographic worsening during their admission.

Discussion

Our research demonstrates the feasibility of an observation pathway for patients with mTIH, and shows that patients with mTIH who have been evaluated by a neurosurgeon can be safely observed in settings with lower resource utilization. Additionally, these data can help ED, hospital and trauma leaders develop observation protocols for patients with mTIH. The current management of patients with mTIH varies widely, especially in terms of patient disposition from the ED. Nishijima at al. demonstrated wide variability in triage decisions, showing that they were more closely tied to institutional preference than patient fac- tors [5]. Their work also showed that mTIH patients who are admitted to the ICU only rarely undergo a critical care intervention. Additionally, it has been shown that if they are at low risk of decompensation, patients prefer admission to less intensive monitoring environments [10]. Sever- al studies have examined whether patients with mTIH need to be ad- mitted to the ICU or can be admitted to the floor [11,12]. However, ours was the first to examine these patients in an EDOU setting.

The use of an EDOU for the management of stable trauma patients has been described in the literature multiple times, beginning in the late 1980s [7,13]. Outside of head injured patients, observation

Table 2 CT findings and clinical outcomes.
	Admitted	Home from ED	ED observation	p	Total observed	p
	N = 814	N = 132	N = 239	(home vs EDOU)	N = 371	(admitted vs total
						observed)
All CT Lesions EDH	27 (3.3%)	0	5(2.1%)	0.11a	5	0.05
SDH	334 (41%)	52 (39.4%)	118 (49.4%)	0.06	170	0.12
SAH	216 (26.5%)	60 (45.4%)	122 (51%)	0.30	182	b0.0001
Contusion	139 (17.1%)	25 (18.9%)	46 (19.2%)	0.92	71	0.39
Skull fracture	94 (11.6%)	30 (22.7%)	24 (10%)	0.0009	54	0.15
Linear	74 (78.7%)	26	12		38
Basilar	15 (16%)	1	10		11
Depressed	5 (5.3%)	3	2		15
Isolated CT lesions
Any	198 (24.3%)	108 (81.8%)	174 (72.8%)	0.05	282 (76%)	b0.0001
EDH	2 (0.25%)	0 (0%)	5 (1.3%)	0.31a	5	0.03+
SDH	98 (12%)	40 (30.3%)	76 (31.8%)	0.76	116	b0.0001
SAH	63 (7.7%)	40 (30.3%)	75 (56.8%)	0.82	115	b0.0001
Contusion	30 (3.7%)	13 (9.8%)	17 (12.9%)	0.35	30	0.0001
Skull Fracture	5 (0.6%)	15 (11.4%)	3 (2.3%)	b 0.0001	18	b0.0001
Linear	3	11	1		12
Basilar	2	1	1		2
Depressed	0	3	1		4
Outcomes Clinical deterioration	15 (1.8%)		0		0	0.004a
Worsening CT	27 (3.4%)		11 (4.6%)		11	0.86a
Neurosurgery	33 (4.1%)		3 (1.3%)		3	0.0002a
Composite outcome (%)	75 (9.2%)		14 (5.9%)		14	0.0001
Medical admission			4		6
Trauma/neurosurgery admission			8
Follow up (%)		111 (84.1%)	190 (79.5%)	0.28
Neurosurgery		1	0
Post traumatic seizure		2	3
Concussive symptoms		8	16

^a Fisher exact test. Patients who were discharged home had no evidence of deterioration in the ED. Statistical tests were only performed on clinically relevant proportions.

protocols have been described for patients with cardiac contusion and blunt Abdominal injury. In a single center prospective study, 259 trauma-activated patients (with multiple different types of injury) who were admitted to an EDOU were followed at 30 days; there was an acceptable Hospital admission rate and there were no adverse out- comes [9]. Many head injury observation protocols utilize a prolonged period of observation in lieu of radiographic testing, especially in pedi- atric populations [14]. Prior to this point, there have been no described observation protocols for patients with confirmed intracranial hemor- rhage on CT scan.

Prior studies have examined the safety of traumatic subarachnoid hemorrhage in patients with preserved mental status, however never in a specific clinical environment. The Canadian Head CT investigators posited that there were some injuries so minor they did not require ad- ditional evaluation; this is the first study to examine that cohort. One retrospective review found no difference in hospital length of stay or mortality when comparing patients with isolated Traumatic SAH to those with concussion [15]. A second study looked at patients with iso- lated traumatic SAH and preserved mental status, finding that no pa- tients had clinical deterioration or need for neurosurgical intervention [16]. Based on this, we feel that the subset of our cohort with subarach- noid hemorrhage is in line with prior research. There is less data avail- able for patients with isolated SDH or contusion.

While prior studies have shown that patients with mTIH are at low risk for acute decompensation, ours is the first study examining patients with mild TIH who were observed either in some combination of an ED and an EDOU. Based on our data, patients with traumatic hemorrhage on head CT and a preserved GCS have a very low to absent risk of clinical deterioration. Further, there were no incidents of poor outcome attrib- uted to delayed recognition of a change in clinical status or imaging. Therefore, we feel that our data supports monitoring in a less resource intensive setting, like an EDOU, for these patients.

Our study group was broad. There were no specific requirements for the type of lesion, and while patients with SAH did exceptionally well, all lesions (except EDH) had an acceptable admission rate. Importantly, we did not exclude patients with medical comorbidities, including those on anticoagulation or those with hypertension; in fact these patients made up a substantial proportion of our cohort. However, these patients must be carefully selected. Consistent with our protocols, coagulopa- thies should be reversed prior to transfer to the EDOU. Additionally, blood pressures should be able to be managed with Oral medications if blood pressure control is mandated by local neurosurgical practice.

In an era of cost consciousness and limited resources, it is the re- sponsibility of emergency physicians to reduce utilization where safe and feasible. While critically ill patients can individually use up a large amount of resources, the bulk of expenditures are related to the larger group of patients who have minor injuries but still require monitoring [17]. ED Observation units have been shown to be revenue-neutral-to- positive for a variety of diagnoses [18]. Additionally, as the transition away from a fee-for-service model incentivizes reducing admissions and limiting healthcare expenditures, observation may prove an attrac- tive way to manage these low-risk patients in a cost-effective manner.

Limitations

General applicability of this study may be limited due to several fac- tors. First, all patients in this study were seen at a single, large, resource- rich Level I trauma center. Given that our institution in a regional refer- ral center for these injuries, there is a large amount of institutional expe- rience managing traumatic intracranial pathology; the EDOU nurses and providers frequently care for this type of patient. However, this re- search may serve as a foundation to generate protocols allowing similar care in settings without the same expertise.

Further, this was a retrospective study. A multidisciplinary protocol guided the disposition of patients with non-operative TIH, though the final decision to place the patient in the EDOU was made by the ED at- tending physician. There may have been additional clinical factors con- sidered when making these decisions. Some patients may have been admitted to the hospital based on ED physician and neurosurgeon judgement who would otherwise have fallen into our analysis. There was unfortunately no way to capture this decision making in a retro- spective study.

Since there was not a clear protocol for observation, the length of ob- servation varied substantially between patients. All patients were ob- served for a minimum of 6 h based on our neurosurgery protocol, but some were observed for longer periods, up to 24-36 h. Given a lack of clinical deterioration at all observation times, we do not feel that this substantially impacted our results. Additionally, there were no stan- dardized criteria for duration or necessity of reimaging. All patients who were observed had a second CT scan performed. However, subse- quent scans were at the discretion of the treating team. As interventions were rare and there were no clinical deteriorations, we do not believe that this caused significant under-reporting of worsening scans.

In addition, there were patients placed in observation status who

remained in the emergency department due to bed availability or other factors. It is possible that there may be an unmeasured difference in care between these patients and those who physically moved to the EDOU.

Given that we included only patients who were observed while in the emergency department and discharged directly, patients who were admitted to the neurosurgery service from the ED after a period of observation but not transferred to the EDOU would not have been captured in our analysis. This limitation occurred because patient data was collected at discharge or transfer to the observation unit. This may have created selection bias towards better outcomes among the patients who were observed in the ED. In an attempt to better describe these groups, we have presented data about the patients observed in the ED separately in Tables 1 and 2.

Finally, we did not attempt to get follow-up on observed patients who did not return for a subsequent clinic or ED visit. There may have been some patients in the 19% lost to follow up who had significant out- comes such as clinical deterioration. Based on the absence of deteriora- tion in the larger group, and the large percentage of regional head trauma visits captured by our electronic medical record, this is unlikely.

Future directions

This is the largest cohort to date of patients with mTIH who have been monitored in an observation unit. Our study setting was a tertiary care hospital, and all patients had neurosurgeons actively involved in their care. Future research should focus on identifying a cohort of pa- tients who are safe to monitor at community hospitals without transfer to tertiary centers. Telemedicine could allow these patients to undergo remote neurosurgical evaluation. Subgroup analysis of patients with mTIH will help to determine which clinical and radiographic factors identify these low-risk individuals.

Conclusion

Patients with isolated traumatic intracranial hemorrhage and pre- served GCS have a very low rate of clinical or radiographic decompensa- tion. In a tertiary care facility, these patients were safely monitored in an

ED observation unit; there were no clinical decompensations or emer- gent interventions needed. The few neurosurgical interventions in the studied group were limited to patients with SDH and were planned days to weeks in advance. The EDOU may be a safe way to monitor these patients with lower costs and resource utilization.

Acknowledgements

Dr Pruitt is a National Research Service Award postdoctoral fellow supported by Agency for Healthcare Research and Quality (AHRQ)T[HYPHEN]32 HS 000078 (PI: Jane L. Holl, MD, MPH). AHRQ

was not involved in the design or execution of this research.

References

1. Fu TS, Jing R, McFaull SR, Cusimano MD. Recent trends in hospitalization and in- hospital mortality associated with traumatic brain injury in Canada: a nationwide, population-based study. J Trauma Acute Care Surg 2015;79:449-54. http://dx.doi. org/10.1097/TA.0000000000000733.
2. Dawes AJ, Sacks GD, Cryer HG, Gruen JP, Preston C, Gorospe D, et al. Compliance with Evidence-based guidelines and interhospital variation in mortality for patients with severe traumatic brain injury. JAMA Surg 2015;150:965. http://dx.doi.org/10.1001/ jamasurg.2015.1678.
3. Stiell IG, Wells GA, Vandemheen K, Clement C, Lesiuk H, Laupacis A, et al. The Cana- dian CT Head rule for patients with minor head injury. Lancet 2001;357:1391-6 (doi:S014067360004561X [pii]).
4. Haydel MJ, Preston CA, Mills TJ, Luber S, Blaudeau E, DeBlieux PM. Indications for computed tomography in patients with minor head injury. N Engl J Med 2000; 343:100-5. http://dx.doi.org/10.1056/NEJM200007133430204.
5. Nishijima DK, Haukoos JS, Newgard CD, Staudenmayer K, White N, Slattery D, et al. Variability of ICU use in adult patients with minor traumatic intracranial hemor- rhage. Ann Emerg Med 2013;61:509-17. http://dx.doi.org/10.1016/j. annemergmed.2012.08.024 e4.
6. Ross MA, Aurora T, Graff L, Suri P, O’Malley R, Ojo A, et al. State of the art: emergency department observation units. Crit Pathw Cardiol 2012;11:128-38. http://dx.doi. org/10.1097/HPC.0b013e31825def28.
7. Conrad L, Markovchick V, Mitchiner J, Cantrill SV. The role of an emergency depart- ment observation unit in the management of trauma patients. J Emerg Med 1985;2: 325-33. http://dx.doi.org/10.1016/0736-4679(85)90285-9.
8. Rydman RJ, Isola ML, Roberts RR, Zalenski RJ, McDermott MF, Murphy DG, et al. Emergency department observation unit versus hospital inpatient care for a chronic asthmatic population: a randomized trial of health status outcome and cost. Med Care 1998;36:599-609.
9. Holly J, Bledsoe J, Black K, Robbins R, Davis V, Bossart P, et al. Prospective evaluation of an ED observation unit protocol for Trauma activation patients. Am J Emerg Med 2012;30:1402-6. http://dx.doi.org/10.1016/j.ajem.2011.10.012.
10. Wu C, Melnikow J, Dinh T, Holmes JF, Gaona SD, Bottyan T, et al. Patient admission preferences and perceptions. West J Emerg Med 2015;16:707-14. http://dx.doi. org/10.5811/westjem.2015.7.27458.
11. Schaller B, Evangelopoulos DS, Muller C, Martinolli L, Pouljadoff MP, Zimmermann H, et al. Do we really need 24-h observation for patients with minimal brain injury and small Intracranial bleeding? The Bernese trauma unit protocol. Emerg Med J 2010;27:537-9. http://dx.doi.org/10.1136/emj.2009.073031.
12. Nishijima DK, Sena M, Galante JM, Shahlaie K, London J, Melnikow J, et al. Deri- vation of a clinical decision instrument to identify adult patients with mild trau- matic intracranial hemorrhage at low risk for requiring ICU admission. Ann Emerg Med 2014;63:448-56. http://dx.doi.org/10.1016/j.annemergmed.2013.

    11.003 e2.

    Welch RD. Management of traumatically injured patients in the emergency depart- ment observation unit. Emerg Med Clin North Am 2001;19:137-54.

13. Kuppermann N, Holmes JF, Dayan PS, Hoyle JD, Atabaki SM, Holubkov R, et al. Iden- tification of children at Very low risk of clinically-important brain injuries after head trauma: a prospective cohort study. Lancet 2009;374:1160-70. http://dx.doi.org/10. 1016/S0140-6736(09)61558-0.
14. Levy AS, Orlando A, Hawkes AP, Salottolo K, Mains CW, Bar-Or D. Should the man- agement of isolated traumatic subarachnoid hemorrhage differ from concussion in the setting of mild traumatic brain injury? J Trauma-Injury Infect Crit Care 2011; 71:1199-204. http://dx.doi.org/10.1097/TA.0b013e31822067fc.
15. Quigley MR, Chew BG, Swartz CE, Wilberger JE. The clinical significance of isolated traumatic subarachnoid hemorrhage. J Trauma Acute Care Surg 2013;74:581-4. http://dx.doi.org/10.1097/TA.0b013e31827d6088.
16. MacKenzie EJ, Morris JA, Smith GS, Fahey M. Acute hospital costs of trauma in the United States: implications for regionalized systems of care. J Trauma 1990;30: 1096-101 [discussion 1101-3].
17. Baugh CW, Venkatesh AK, Bohan JS. Emergency department observation units: a clinical and financial benefit for hospitals. Health Care Manage Rev 2011;36: 28-37. http://dx.doi.org/10.1097/HMR.0b013e318f3c035.