a b s t r a c t

Background: Emergency Department (ED) observation units (Obs Units) are prevalent in the US, but little is known regarding older adults in observation. Our objective was to describe the Obs Units nationally and obser- vation patients with specific attention to differences in care with increasing age.

Design: This is an analysis of 2010-2013 data from the National Hospital Ambulatory Medical Care Survey

(NHAMCS), a national observational cohort study including ED patients. Weighted means are presented for con- tinuous data and weighted percent for categorical data. Multivariable logistic regression was used to identify var- iables associated with placement in and admission from observation.

Results: The number of adult ED visits varied from 100 million to 107 million per year and 2.3% of patients were placed in observation. Adults >=65 years old made up a disproportionate number of Obs Unit patients, 30.6%, com- pared to only 19.7% of total ED visits (odds ratio 1.5 (95% CI 1.5-1.6), adjusting for sex, race, month, day of week, payer source, and hospital region). The overall admission rate from observation was 35.6%, ranging from 31.3% for ages 18-64 years to 47.5% for adults >=85 years old (p b 0.001). General symptoms (e.g., nausea, dizziness) and Hypertensive disease were the most common diagnoses overall. Older adults varied from younger adults in that they were frequently observed for diseases of the urinary system (ICD-9 590-599) and metabolic disor- ders (ICD-9 270-279).

Conclusions: Older adults are more likely to be cared for in Obs Units. Older adults are treated for different medical conditions than younger adults.

(C) 2018

Introduction

An inpatient hospital stay is not a benign event for older adults (adults >=65 years old), as many will experience complications such as a subsequent decline in functional status, delirium, and high mortality [1-3]. Limiting hospitalizations and reducing older adult’s length of stay in the hospital is therefore important. Emergency care providers must be judicious with hospital resources and patient needs and at- tempt to avoid admissions if medically possible [4,5]. One mechanism to achieve this goal is the use of ED observation units (Obs Units).

Obs Units are areas of the ED dedicated to the care of patients that re- quire further interventions or monitoring but do not meet the Centers for Medicare and Medicaid criteria for an inpatient stay (two midnights of

Abbreviations: ED, Emergency Department; Obs Units, Observation Units; NHAMCS, National Hospital Ambulatory Medical Care Survey; TIA, Transient Ischemic Attack.

* Corresponding author at: Department of Emergency Medicine, The Ohio State Wexner Medical Center, 750 Prior Hall, 376 W 10th Ave, Columbus, OH 43210, USA.

E-mail addresses: [email protected], @LSGeriatricEM (L.T. Southerland). @LSGeriatricEM (L.T. Southerland).

care needed) [6]. Obs Units are more efficient in obtaining testing and dis- position than Inpatient units, which decreases costs and length of stay for similar syndromes [7-9]. For older patients, observation can provide addi- tional time for the ED provider to further evaluate the patient’s home sta- tus, cognitive abilities, fall risk, and discharge safety. In addition to the standard use of observation care to obtain more testing for cardiac syn- dromes or Transient ischemic attacks , Obs Units are a suitable set- ting for focused geriatric care such as assessment by a geriatric nurse practitioner, physical therapist, or multidisciplinary geriatric team [10- 12]. Multidisciplinary geriatric assessment in an Obs Unit can not only re- duce admissions but also screen for unmet healthcare needs [10]. For ex- ample, a frail patient who comes into the ED at 3 am can be placed in observation for Physical therapy and geriatric consultation in the morn- ing. In this way, observation stay can provide the time necessary to make safe discharges and transitions of care from the ED to home, chang- ing the ED visit from a “sentinel event” to an opportunity to provide person-centered, holistic care [13,14]. This is especially important for older adults presenting after a fall, as 36-50% of these patients will have an ED revisit within 6 months [15,16].

https://doi.org/10.1016/j.ajem.2018.12.009

0735-6757/(C) 2018

Despite these possible benefits, there is minimal data on the care of older patients in observation. Prior studies of older adults in Obs Units are encouraging, but have focused on single sites [9,11,17-19]. It is un- known on a national scale what types of care older adults receive in ob- servation and whether significant numbers of older adults are cared for in these units. The National Hospital Ambulatory Medical Care Survey , collected by the Centers for Disease Control National Center for Health Statistics, includes information on Obs Units and observation visits. Data from 2009 to 2010 revealed that older age is a predictor for admission from observation with an admission rate of 49.1% for older adults [20]. However this analysis did not look at the diagnoses of these patients or the length of stay in the hospital after admission, which is one marker of whether the admission was warranted.

Therefore, we reviewed the latest NHAMCS data, 2010-2013, with specific attention to the association between age and Observation status. Secondary outcomes include the reason for observation, rates of hospi- tal admission from observation, total length of stay in observation and admission length of stay.

Methods

Study design

This is a secondary analysis of existing NHAMCS data, which was col- lected by the Center for Disease Control, National Center for Health Sta- tistics (CDC). As this is a publically available dataset, this study was exempt from Institutional Review. The NHAMCS is an annual, national probability sample of visits made to non-federal, general, and short- stay hospitals across the United States. Descriptions of this dataset have been previously published [21]. Data from the latest available years (2010-2014) were accessed and analyzed. The data from 2014 had differently assigned variables for observation which made it unable to be analyzed with the prior 4 years, therefore we limited this analysis to the years 2010-2013.

Study setting and population

NHAMCS uses standardized data collection and probability estima- tions detailed in prior literature [22]. ED visits were defined as undergo- ing ED observation if they had an ED observation disposition (variables “OBSHOS” if admitted from observation and “OBSDIS” if discharged). These variables were included in all study years. The estimate of preva- lence of ED observation units was made using the question “Does your ED have a physically separate observation or clinical decision unit?” Only years 2010 and 2011 were included as they were the only years with both the questions and ED weights available at time of analysis.

Data analysis

SAS 9.4 (SAS Institute Inc., Cary, NC) was used for data management and all data analyses were conducted using STATA 14 (StataCorp, College Station, TX). All analyses used survey procedures with weights and strata as provided in the NHAMCS data sets, and included all records in the data files to obtain the correct sample variance estimates. Estimates consid- ered unreliable by standard NHAMCS criteria (relative standard error of 30% of more or based on b30 records) are not reported. Weighted means are presented for continuous data and weighted percent for cate- gorical data to produce national estimates. Our data were compared to the Emergency Department Summary tables from the CDC for the respec- tive year as a double check for our computations of patients placed in observation.

Diagnosis International Classification of Diseases 9 (ICD-9) codes were classified based on the first 3 numerals without decimals (see Supple- mental Table 1). All other variables presented are as defined in the NHAMCS documentation or above. As ED weights have not yet been re- leased for calendar years 2012 and 2013, only visit-level data are

presented. Differences among subgroups were compared using a two- tailed t-test (p b 0.05). Logistic regression analysis was done to determine the significance of the association between age and Hospital admission rates and age and hospital length of stay. The model controlled for sex, race/ethnicity, and hospital characteristics (region, metropolitan status and ownership).

Results

Over 2010-2013, the number of adult ED visits varied from 100 mil- lion to 107 million per year and 2.3% of patients were placed in observa- tion. There were 10,225,371 weighted adult Obs Unit visits, or approximately 2.56 million per year (Table 1). Patients were mostly fe- male (55.6%) and Caucasian (65.1%). Most patients were community dwelling, although the rate of residence in an extended care facility in- creased to 23.2% for patients >=85 years old. In 2010 and 2011, 20.9% (95% confidence interval: 16.5-26.2%) of EDs had Obs Units. The num- ber of EDs with Obs Units was not included for years 2012 and 2013.

Over the 4 year period, approximately 3.13 million (95% CI 2.6 million-3.6 million) older adults were cared for in Obs Units; this repre- sents 782,000 per year. Older adults were also assigned to observation at a consistently higher percentage than younger adults (Table 1). Older adults made up 30.6% of Obs Unit patients, despite comprising only 19.7% of ED patients during this timeframe (odds ratio 1.5 (95% CI 1.5-1.6) for placement in observation for age >=65 years, adjusting for pa- tient sex and race, visit month and day of week, payer source, and hospital region of country). Additionally, adults in the 65-74 years age group had the longest average stay in observation, 26.0 h, compared to 18.0 h for adults 18-64 years old.

Disposition from observation status also varied with age (Fig. 1). This trend persisted even when controlling for sex, race, and hospital charac- teristics (region, metropolitan status and ownership) (p b 0.001). Over- all admission rate for those 65 years and older was 44.2%, compared to 31.3% for patients 18-64 years old. Assuming these patients would have required admission if an Obs Unit was unavailable, these units prevented an average 436,000 admissions of older adults per year.

The most common Diagnostic codes were for symptoms (Table 2). The ICD-9 category of General Symptoms (ICD780-789) includes syn- cope, dizziness, fever, tachycardia, and vomiting. Hypertensive disor- ders and ischemic heart disease were also common diagnoses. Older adults varied from younger adults in that they were also placed in Ob- servation for diseases of the urinary system (ICD-9 590-599) and meta- bolic disorders (ICD-9 270-279). See Supplemental Tables 1 and 2 for further breakdown of the ICD codes used and the rates of placement for different ICD codes.

Discussion

Obs Units care for over 2.56 million adult visits per year, and a higher than expected amount (30.6%) are older adults. Similar to past analyses, we found that age is an independent risk factor for a longer length of stay in observation and for hospital admission from the Obs Unit [17,18,20]. While the admission rate from observation is higher for older adults, Obs Unit care is still effective at avoiding full admission for over half of older adults, with an estimated 436,000 avoided admissions per year.

Since older adults comprise almost a third of the patients in these units, Obs Unit staff may want to consider how to optimize their care. Prior studies suggest that staffing Obs Units with geriatric-trained per- sonnel or offering multidisciplinary assessments avoids full hospital ad- missions, perhaps by identifying and managing issues such as delirium [11,12,23]. This is recommended by the Geriatric ED Guidelines [24]. For example, geriatric specific protocols can be used to address underly- ing needs such as risk for falls, polypharmacy, and cognitive deficits. Physical therapists can provide great insight into fall risk and ways to improve ambulation safety [10,25]. Protocols that focus on safe

Table 1

Demographics of adult patients placed in observation in calendar years 2010-2013. Data are weighted. Observation length of stay and admission rates are significantly higher for any of the older age groups as compared to 18-64 year olds (p b 0.001).

All ED Patients n = All Observation Unit Patients

527,362,430		Alla n= 10,225,371 (1.9%)	18-64 years n = 6,217,770 (1.9%)	65-74 years n = 1,204,642 (3.5%)	75-84 years n = 1,215,588 (4.1%)	>=85 years n = 706,105 (4.1%)
Female	55.4%	55.6%	55.7%	52.5%	60.3%	64.3%
Race
White	62.2%	65.1%	62.6%	65.1%	73.9%	78.3%
Black/African American	19.8%	23.0%	27.0%	20.3%	10.5%	13.1%
Asian	1.5%	1.4%	1.0%	1.6%	2.7%	1.2%
Native Hawaiian/Pacific Islander	0.3%	0.4%	0.4%	0.7%	0.2%	0.0%
American Indian/Alaska Native	0.7%	0.5%	0.3%	2.4%	0.2%	0.0%
More than one race reported	0.4%	0.2%	0.1%	0.2%	0.1%	0.2%
Unknown/missing	14.9%	9.5%	8.7%	9.6%	12.4%	7.2%
Residence
Private residence	92.7%	89.1%	91.7%	91.0%	80.0%	68.8%
Extended care facility	1.9%	4.1%	1.2%	3.8%	11.2%	23.2%
Homeless	0.6%	0.7%	1.0%	0.0%	0.0%	0.0%
Other	1.2%	1.6%	1.5%	2.4%	2.9%	1.0%
Unknown/missing	3.7%	4.5%	4.6%	2.8%	6.0%	6.9%
Observation length of stay (hours)	–	20.0	18.0	26.0	22.2	24.1
Admission rate from Observation	–	35.6%	31.3%	38.2%	48.3%	47.5%
a 881,266 visits with missing age.

transitions of care to home often use Case management or home health needs assessments prior to discharge [26-28].

In addition to protocols, Obs Units can be made more geriatric friendly with additional equipment or physical layout changes. Provid- ing assist devices such as reading glasses, hearing amplifiers, walkers, canes, and high rise toilet seats can assist with comfort, communication, and mobility. This can also reduce delirium as hearing and visual im- pairments are risk factors for delirium [29]. In this way, the Obs Unit can become an area of high quality geriatric care.

This study also found a higher admission rate from observation than the oft-quoted goal of 20% [30,31]. The overall admission rate was 35.6% which is slightly decreased from the 2009-2010 rate of 40.4% but almost double the 18% rate in the 2007 data. [20,32] This increased admission rate from the 2000s to the 2010s is interesting as the total percentage of patients assigned to observation has not changed significantly (2.1% in 2007, 1.87% in 2008, and 1.95% to 2.50% in years 2010-2013) [32,33].

Looking at the reasons for observation placement, this increased admis- sion rate may be due to the diversity and Complexity of patients. The most common diagnostic category in this dataset was not chest pain, but General Symptoms, (ICD780-789), which includes syncope, fever, tachycardia, and vomiting among others. These symptoms can be

80 6

70 5

4.9

4.9

5.2

4

Disposition percentage

60

Hospital LOS (days)

50 4

40 3

30 2

20

10 1

0 0

18-64 65-74 75-84 85+

Age (years)

Admit Discharge Hospital LOS if admitted

Fig. 1. The bar graphs display the weighted disposition of Obs Unit adult patients by percentage. The red line demonstrates the weighted average hospital length of stay in days among observation patients admitted to the hospital by age, demonstrating that the majority of people admitted from Obs Units have a hospital length of stay of at least 4 days.

indicative of a broad range of underlying illnesses. This is not surprising as Obs Units are typically caring for patients without a clear diagnosis or who require further testing to make a diagnosis. Since older adults may present with atypical symptoms and increased diagnostic uncertainty, observation can be helpful to elucidate the cause. Or it may be that ED providers are reluctant to give a specific diagnosis and prefer these broader diagnoses describing symptoms rather than causes.

Interestingly, the oldest patients (>=85 years old) are observed for dif- ferent reasons than the younger old (65-84 year olds). Diseases of the urinary system are a diagnostic code in 16% of visits in >=85 year olds and b6% of 65-84 year olds. The oldest age group is also more frequently placed in observation after fractures (ICD9 codes 805-809). This could relate to needing more assistance with ambulation assessments and home health care after a fracture, or it could be due to the higher risk of fractures with older age. More information on the type of care these patients are receiving (e.g., physical therapy assessments, durable med- ical equipment provided, case management) would be helpful to under- stand how Obs Units are providing care to complex older patients.

A newly identified trend from this data analysis is the use of obser- vation status for intoxication (ICD-9 codes 303-305). The use of obser- vation services for Intoxicated patients is not well described in the literature. One article mentions an Obs Unit as a feasible place for mon- itoring of intoxicated patients, but there are no studies of patients in Obs Units for this purpose [34]. This is surprising as this was the 4th most common diagnosis category for younger adults. However, diagnoses are not mutually exclusive, so intoxication could also be a secondary or tertiary diagnosis and not the main reason for observation.

The 2010-2013 data does differ from prior years in the estimated number of Obs Units (21%). The NHAMCS data from 2001 to 2008 demon- strated that 34% of EDs have an Obs Unit [33]. This older estimate is more consistent with data published by the Emergency Department Benchmarking Alliance, a consortium of 1200 EDs, which reported in 2015 that 35% of their EDs serving over 40,000 patient visits a year had Obs Units [35]. This fluctuation in the estimated number of Obs Units may be also due to some of the sampling limitations of the NHAMCS dataset, given that larger EDs are more likely to have an Obs Unit than smaller EDs.

Limitations

NHAMCS uses weighted percentages which can be biased by the hospital sampling process [36]. While the NHAMCS study uses a strati- fication algorithm to sample from a variety of hospitals (e.g., urban,

Table 2

The top 10 weighted diagnoses among adults in Emergency Department observation units, as grouped by ICD-9 codes for the calendar years 2010-2013. Percentage and number of total weighted diagnoses in that age group are listed. Diagnostic codes are not mutually exclusive and therefore total number of diagnoses is greater than the total number of patient visits.

All >=18 years	18-64 years	65-74 years	75-84 years	>=85 years
1 Symptoms (780-789)	Symptoms (780-789)	Symptoms (780-789)	Symptoms (780-789)	Symptoms (780-789)
49.1% (n = 5,023,392)	47.5% (n = 2,955,498)	58.3% (n = 702,564)	61.9% (n = 752,297)	51.3% (n = 362,454)
Hypertensive disease	Hypertensive disease	Other forms of heart disease	Hypertensive disease	Other diseases of urinary
2 (401-405)	(401-405)	(420-429)	(401-405)	system (590-599)
7.7% (n = 791,806)	7.5% (n = 464,069)	8.9% (n = 107,250)	12.8% (n = 156,118)	16.4% (n = 115,578)
3 Other metabolic and immunity	Ischemic heart disease	Other metabolic and immunity	Other metabolic and immunity	Other forms of heart disease
disorders (270-279)	(410-414)	disorders (270-279)	disorders (270-279)	(420-429)
7.5% (n = 767,277)	6.4% (n = 397,381)	8.2% (n = 98,368)	11.0% (n = 133,679)	15.5% (n = 109,624)
4 Ischemic heart disease	Psychoactive substance	Diseases of other endocrine glands	Ischemic heart disease	Other metabolic and immunity
(410-414)	(303-305)	(249-259)	(410-414)	disorders (270-279)
6.1% (n = 626,478)	6.2% (n = 386,397)	7.6% (n = 91,984)	10.3% (n = 125,628)	15.0% (n = 105,747)
5 Chronic obstructive pulmonary disease and allied conditions (490-496) 6.0% (n = 611,661)	Chronic obstructive pulmonary disease and allied conditions (490-496) 5.9% (n = 368,583)	Hypertensive disease (401-405) 7.6% (n = 91,411)	Other forms of heart disease (420-429) 9.6% (n = 116,625)	Hypertensive disease (401-405) 11.4% (n = 80,208)
6 Other diseases of urinary	Other metabolic and immunity	Other diseases of urinary system	Cerebrovascular disease	Ischemic heart disease
system (590-599)	disorders (270-279)	(590-599)	(430-438)	(410-414)
5.4% (n = 555,988)	5.7% (n = 355,531)	5.8% (n = 69,502)	5.9% (n = 71,361)	6.2% (n = 43,900)
7 Other forms of heart disease (420-429) 4.7% (n = 477,993)	Other diseases of urinary system (590-599) 4.9% (n = 302,312)	Ischemic heart disease (410-414) 4.8% (n = 58,049)	Pneumonia and influenza (480-488) 5.3% (n = 64,156)	Cerebrovascular disease (430-438) 5.8% (n = 41,238)
8 Psychoactive substance	Other diseases of digestive	Chronic obstructive pulmonary	Nephritis, Nephrotic syndrome	Fracture of neck and trunk
(303-305)	system (570-579)	disease and allied conditions	and nephrosis (580-589)	(805-809)
4.1% (n = 416,450)	4.2% (n = 263,435)	(490-496) 4.8% (n = 57,811)	5.1% (n = 61,392)	5.2% (n = 36,590)
9 Diseases of other endocrine	Pain (338)	Persons with potential health	Chronic obstructive pulmonary	Chronic obstructive pulmonary
glands (249-259) 3.7% (n = 375,675)	3.9% (n = 242,124)	hazards from personal/family history (V10-V19) 4.2% (n = 50,444)	disease and allied conditions (490-496) 4.8% (n = 58,125)	disease and allied conditions (490-496) 5.1% (n = 36,263)
10 Other diseases of digestive	Diseases of other endocrine	Nephritis, nephrotic syndrome	Other diseases of digestive	Nonspecific abnormal findings
system (570-579)	glands (249-259)	and nephrosis (580-589)	system (570-579)	(790-796)
3.5% (n = 362,604)	3.5% (n = 219,538)	3.8% (n = 46,358)	4.4% (n = 53,996)	4.6% (n = 32,142)

rural) there are limitations with generalizing the data to predict na- tional trends. Lack of consistency in coding and use of diagnostic codes is another limitation. For example, a patient who was evaluated for chest pain may be given a variety of diagnostic codes. This contributed to our choice to evaluate diagnoses by ICD-9 grouping rather than by specific diagnoses. Therefore we recommend that the data in Table 2 be used just to identify trends and not to try to estimate the exact num- ber of patients seen for a specific diagnosis such as angina or urinary tract infections. Additionally, changes in the coding variables over time can create swings in the data from year to year. This was the cause of our decision not to include 2014 in this analysis, as the Obs unit variables were coded differently and the data did not align with the prior years.

Conclusions

In conclusion, the NHAMCS database suggests that about 2% or 2.56 million adult ED patients per year are cared for in ED Obs Units. Older adults make up a disproportionate number of these patients, and have higher admission rates and observation lengths of stay. Obs Units may be an Ideal setting to target quality improvement processes of geriatric care.

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2018.12.009.

Impact statement

We certify that this work is novel clinical research that reports on an area of clinical care that has had little investigation- older adults in ob- servation units. We report the breakdown for rates of use of observation and quality metrics for these units in addition to the difference in rea- sons for use (diagnostic codes) for younger versus older adults on a na- tional level. This information has never been reported prior and may be

helpful for Emergency Medicine physicians and hospital administrators evaluating care in their observation units.

Presentations

This information was presented as an abstract at the 2018 Annual Meeting of the American Geriatrics Society, Orlando, FL.

Financial support/conflicts of interest

The authors have no conflicts of interest to report. This work was supported by the National Institutes of Health [JMC, R01AG05081] and the Emergency Medicine Foundation/Emergency Medicine Residents’ Association Resident Research Grant [KMH]. The sponsors had no role in the design, analysis, or manuscript preparation for this study.

Author contributions

LTS, KMH, and JMC designed the study. LTS and KMH performed the statistical analysis. All authors participated in analysis and data inter- pretation and manuscript preparation.

References

1. Avelino-Silva TJ, Farfel JM, Curiati JA, Amaral JR, Campora F, Jacob-Filho W. Compre- hensive geriatric assessment predicts mortality and adverse outcomes in hospital- ized older adults. BMC Geriatr 2014;14:129.
2. Boyd CM, Landefeld CS, Counsell SR, Palmer RM, Fortinsky RH, Kresevic D, et al. Re- covery of activities of daily living in older adults after hospitalization for acute med- ical illness. J Am Geriatr Soc 2008;56(12):2171-9.
3. Fick DM, Steis MR, Waller JL, Inouye SK. Delirium superimposed on dementia is as- sociated with prolonged length of stay and poor outcomes in hospitalized older adults. J Hosp Med 2013;8(9):500-5.
4. Pines JM, Mullins PM, Cooper JK, Feng LB, Roth KE. National trends in emergency de- partment use, care patterns, and quality of care of older adults in the United States. J Am Geriatr Soc 2013;61(1):12-7.
5. Kocher KE, Dimick JB, Nallamothu BK. Changes in the source of unscheduled hospi- talizations in the United States. Med Care 2013;51(8):689-98.
6. Medicare benefit policy manual. CMS. http://www.cms.gov/Regulations-and- Guidance/Guidance/Manuals/downloads/bp102c06.pdf; 2014.
7. Baugh CW, Liang LJ, Probst MA, Sun BC. National cost savings from observation unit management of syncope. Acad Emerg Med 2015;22(8):934-41.
8. 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.
9. Nahab F, Leach G, Kingston C, Mir O, Abramson J, Hilton S, et al. Impact of an emer- gency department observation unit transient ischemic attack protocol on length of stay and cost. J Stroke Cerebrovasc Dis 2012;21(8):673-8.
10. Southerland LT, Vargas AJ, Nagaraj L, Gure TR, Caterino JM. An emergency depart- ment observation unit is a feasible setting for multidisciplinary geriatric assessments in compliance with the Geriatric EMergency department guidelines. Acad Emerg Med 2018;25(1):76-82.
11. Foo CL, Siu VW, Tan TL, Ding YY, Seow E. Geriatric assessment and intervention in an emergency department observation unit reduced re-attendance and hospitalisation rates. Australas J Ageing 2012;31(1):40-6.
12. Pareja-Sierra T, Hornillos-Calvo M, Rodriguez-Solis J, Sepulveda-Moya DL, Bassy-Iza N, Martinez-Peromingo FJ, et al. Implementation of an emergency department ob- servation unit for elderly adults in a university-affiliated hospital in Spain: a 6- year analysis of data. J Am Geriatr Soc 2013;61(9):1621-2.
13. Lamantia MA, Platts-Mills TF. Bending the curve of health trajectories for older adults discharged from the emergency department. Ann Emerg Med 2017;69(4): 434-6.
14. American Geriatrics Society Expert Panel on Person-Centered C. Person-centered care: a definition and essential elements. J Am Geriatr Soc 2016;64(1):15-8.
15. Liu SW, Obermeyer Z, Chang Y, Shankar KN. Frequency of ED revisits and death

    among older adults after a fall. Am J Emerg Med 2015;33(8):1012-8.

    Sri-On J, Tirrell GP, Bean JF, Lipsitz LA, Liu SW. Revisit, subsequent hospitalization, re- current fall, and death within 6 months after a fall among elderly emergency depart- ment patients. Ann Emerg Med 2017;70(4):516-21 [e2].

16. Caterino JM, Hoover EM, Moseley MG. Effect of advanced age and vital signs on ad- mission from an ED observation unit. Am J Emerg Med 2013;31(1):1-7.
17. Ross MA, Compton S, Richardson D, Jones R, Nittis T, Wilson A. The use and effective- ness of an emergency department observation unit for elderly patients. Ann Emerg Med 2003;41(5):668-77.
18. Zdradzinski MJ, Phelan MP, Mace SE. Impact of frailty and sociodemographic factors on hospital admission from an emergency department observation unit. Am J Med Qual 2017 May/Jun;32(3):299-306.
19. Napoli AM, Mullins PM, Pines JM. Predictors of hospital admission after ED observa- tion unit care. Am J Emerg Med 2014;32(11):1405-7.
20. CDC/National Center for Health Statistics. Ambulatory health care data: about the ambulatory health care surveys. https://www.cdc.gov/nchs/ahcd/about_ahcd.htm; 2017, Accessed date: 25 September 2017.
21. CDC/National Center for Health Statistics. Ambulatory health care data. https:// www.cdc.gov/nchs/ahcd/index.htm; 2017, Accessed date: 2 November 2017.
22. Southerland LT, Vargas AJ, Nagaraj L, Gure TR, Caterino JM. An emergency depart- ment observation unit is a feasible setting for multidisciplinary geriatric assessments in compliance with the geriatric emergency department guidelines. Acad Emerg Med 2018 Jan;25(1):76-82. https://doi.org/10.1111/acem.13328 Epub 2017 Nov 24.
23. Carpenter CR, Bromley M, Caterino JM, Chun A, Gerson LW, Greenspan J, et al. Opti- mal older adult emergency care: introducing multidisciplinary geriatric emergency department guidelines from the American College of Emergency Physicians, American Geriatrics Society, emergency nurses Association, and Society for Aca- demic Emergency Medicine. Ann Emerg Med 2014;63(5):e1-3.
24. Plummer L, Sridhar S, Beninato M, Parlman K. Physical therapist practice in the emergency department observation unit: descriptive study. Phys Ther 2015;95(2): 249-56.
25. Richards S, Coast J. Interventions to improve access to health and social care after discharge from hospital: a systematic review. J Health Serv Res Policy 2003;8(3): 171-9.
26. Rosted E, Wagner L, Hendriksen C, Poulsen I. Geriatric nursing assessment and inter- vention in an emergency department: a pilot study. Int J older people Nurs 2012;7 (2):141-51.
27. Hwang U, Dresden SM, Rosenberg MS, Garrido MM, Loo G, Sze J, et al. Geriatric emergency department innovations: transitional care nurses and hospital use. J Am Geriatr Soc 2018;66(3):459-66.
28. Han JH, Zimmerman EE, Cutler N, Schnelle J, Morandi A, Dittus RS, et al. Delirium in older emergency department patients: recognition, risk factors, and psychomotor subtypes. Acad Emerg Med 2009;16(3):193-200.
29. Komindr A, Baugh CW, Grossman SA, Bohan JS. Key operational characteristics in emergency department observation units: a comparative study between sites in the United States and Asia. Int J Emerg Med 2014;7(1):6.
30. Ross MA, Granovsky M. History, principles, and policies of observation medicine. Emerg Med Clin North Am 2017;35(3):503-18.
31. Wiler JL, Ross MA, Ginde AA. National study of emergency department observation services. Acad Emerg Med 2011;18(9):959-65.
32. Venkatesh AK, Geisler BP, Gibson Chambers JJ, Baugh CW, Bohan JS, Schuur JD. Use of observation care in US emergency departments, 2001 to 2008. PLoS One 2011;6(9): e24326.
33. Wheatley MA. Additional conditions amenable to observation care. Emerg Med Clin North Am 2017;35(3):701-12.
34. JJ A. Emergency department survey shows spike in volume, structural changes, pa- tient boarding concerns. http://www.acepnow.com/article/2015-emergency- department-survey-shows-spike-volume-structural-changes-patient-boarding- concerns/; 2015. 2016 [accessed 11/2/2017].
35. McCaig LF, Burt CW. Understanding and interpreting the National Hospital Ambula- tory Medical Care Survey: key questions and answers. Ann Emerg Med 2012;60(6): 716-21 [e1].