a b s t r a c t

Objective: Among injured patients transferred from one emergency department (ED) to another, we determined factors associated with being discharged from the second ED without procedures, or admission or observation. Methods: We analyzed all patients with injury Diagnosis codes transferred between two EDs in the 2011 healthcare utilization Project State Emergency Department and State Inpatient Databases for 6 states. Multivar- iable hierarchical logistic regression evaluated the association between patient (demographics and clinical char- acteristics) and hospital factors, and discharge from the second ED without coded procedures.

Results: In 2011, there were a total of 48,160 ED-to-ED injury transfers, half of which (49%) were transferred to non-trauma centers, including 23% with major trauma. A total of 22,011 transfers went to a higher level of care, of which 36% were discharged from the ED without procedures. Relative to torso injuries, discharge without procedures was more likely for patients with soft tissue (OR 6.8, 95%CI 5.6-8.2), head (OR 3.7, 95%CI 3.1-4.6), fa- cial (OR 3.8, 95%CI 3.1-4.7), or hand (OR 3.1, 95%CI 2.6-3.8) injuries. Other factors included Medicaid (OR 1.3, 95%CI 1.2-1.5) or uninsured (OR 1.3, 95%CI 1.2-1.5) status. Treatment at the receiving ED added an additional

$2859 on average (95% CI $2750-$2968) per discharged patient to the Total charges for injury care, not including the costs of Ambulance transport between facilities.

Conclusion: Over a third of patients transferred to another ED for traumatic injury are discharged from the second ED without admission, observation, or procedures. Telemedicine consultation with sub-specialists might reduce some of these transfers.

(C) 2017

Introduction

Between 1999 and 2003, the incidence of trauma transfers increased disproportionately to total trauma volumes (34% vs 6%) [1]. While several

Abbreviations: AHA, (American Hospital Association); AIS, (Abbreviated Injury Scale); HCUP, (Healthcare Cost and Utilization Project); ISS, (Injury Severity Scale); SEDD, (State Emergency Department Database); SID, (State Inpatient Database); TIEP, (Trauma Information Exchange Program).

* Corresponding author.

E-mail addresses: [email protected] (L.N. Medford-Davis), [email protected] (D.N. Holena), [email protected] (D. Karp), [email protected] (M.J. Kallan), [email protected] (M.K. Delgado).

studies have described variation in transfer patterns for trauma patients [1-13], the reasons for these variations remain unknown, particularly for patients who are later determined to have only minor injuries. Prior authors have examined transfers with low injury severity who are discharged within two days of transfer without undergoing any major op- erating room procedures [6,8,14,15]. However, these prior studies are limited and paint an incomplete picture since all suffer from at least one of the following limitations: they are single-center [6-8,10] or single- state [4,5,15] and do not capture Regional variations in trauma care, they lack data about transferring hospitals and care prior to transfer [4, 6-8,10,14], and they do not include patients discharged from the emer- gency department (ED) prior to admission [2,9,12,14], or consider

https://doi.org/10.1016/j.ajem.2017.10.024

0735-6757/(C) 2017

procedures performed in the ED prior to discharge. Therefore it remains unknown what happens to the majority of trauma transfer patients.

Research is needed to measure the proportion and types of injured pa- tients that are transferred to another hospital and subsequently discharged without intervention. While some patients may benefit from in-person specialized consultation or observation, when EDs transfer patients with low injury severity that ultimately do not require hospitalization or proce- dural intervention, it may result in higher Costs of care and an additional burden on patients, physicians, and receiving hospitals. Treatment in a trauma center costs $5590 more per patient compared to treatment in a non-trauma center [16], and Ambulance transport costs for interfacility trauma transfer average $1863 per patient [10]. The evaluation of all trau- ma patients, even those ultimately found to have minor injuries, diverts ED staff and resources, increasing wait times and morbidity for other patients in the ED during a Trauma activation [17,18].

Our primary objective was to determine the factors associated with Discharge home from the second ED without procedures, or observation or admission after trauma transfer. A secondary outcome was the charge for hospital care at the receiving hospital for this population.

Methods

We conducted a retrospective longitudinal cohort study of injured adult patients who were transferred between hospital EDs in six states in 2011 using data from the Healthcare Cost and Utilization Project (HCUP) State Emergency Department (SEDD) and State Inpatient (SID) Databases of the Agency for Healthcare Research and Quality (AHRQ). These databases track all acute care hospitalizations (SID) and treat-and-release ED visits (SEDD) in a state. We included data from six states-California (CA), Florida (FL), Iowa (IA), Massachusetts (MA), New York (NY), and Utah (UT). These states were chosen because they report VisitLink, a unique de-identified code which allows the tracking of patients across HCUP databases within a state [19]. Since this data is publicly available, the study underwent expedited review by the IRB.

We identified all patients ages 15 and above (referred to as ‘adults’ in this manuscript since many adult trauma hospitals will admit patients older than 14) who presented primarily to an ED with previously vali- dated ICD-9 diagnosis codes for traumatic injury [14]. Patients who died in the ED after transfer and who were transferred a second time from the receiving hospital were excluded, as were patients from 39 re- ceiving hospitals that did not report any procedure codes. Transfers were identified based on a disposition of “transfer to short-term hospi- tal” from the first encounter in the SEDD followed by an encounter orig- inating in the ED in a different hospital within one day of the first ED arrival in the SEDD (if the patient was discharged from the receiving ED without admission) or SID (if the patient was admitted to the receiv- ing hospital). These two records were linked using VisitLink to create a continuous episode of care.

For our primary analyses, we only included patients who were trans- ferred to a higher level of care, as other types of transfer are more likely to be driven by patient preferences, insurance, and health system factors [2]. Therefore in these analyses we excluded transfers to non-trauma centers (Level 4 or undesignated) and transfers to a lower level of care (for example, from a Level 1 to Level 2 or undesignated trauma center). The primary outcome of interest was a disposition of discharge from the receiving hospital ED without a therapeutic procedure performed there after transfer (“discharge without procedures”) (Fig. 1). Thera- peutic procedures were coded in HCUP and included laceration repair, wound care, fracture management such as reduction, casting, or splinting, Critical care procedures such as intubation, mechanical venti- lation, or central line and chest tube insertion, and any operations per- formed in the operating room, but excluded imaging tests, laboratory tests, and medication administration. Patients who had observation ser- vices coded at either the transferring or receiving hospital were also ex-

cluded from the definition of discharge without procedure.

All patient demographic and injury characteristics were derived from the First visit in SEDD. We tabulated age, sex, race, primary payer, and the median Household income of the zip code where the pa- tient resided. ICDPIC, a validated software program, was used to map in- jury diagnosis codes to valid measures of injury mechanism and severity [20,21]. Injury severity included overall Injury Severity Score (ISS), Ab- breviated injury scales (AIS) by body region, and pattern of most severe injury. Injury pattern was classified by the Barell injury classification by body region for the most severe injury listed [21]. In cases of patients with two or more injury diagnosis codes with equivalent severity, the first listed was used. Each Barell code was then sequentially assigned to one of nine categories based on the specialist needed to care for it: burns, eye injuries, facial injuries, hand injuries, spine injuries, or head injuries. Contusions, sprains, strains, and lacerations of any body region including the face, hand, head, or back were classified as skin and soft tissue injuries. Major internal abdominothoracic injuries such as pneu- mothorax, sternal fractures, diaphragmatic injuries, and injuries of the abdominal or pelvic organs were classified as torso injuries.

hospital characteristics from the 2011 American Hospital Associa- tion (AHA) Annual Survey [22] and trauma center designation (Levels 1 through 4 or non-designated) from the 2010 American Trauma Socie- ty Trauma Information Exchange Program (TIEP) [23] were matched using the hospital’s AHA identification number from the HCUP data. The driving distance and driving time between hospitals was deter- mined based on the two hospitals’ geographic latitude and longitude and using the Google Maps Application Programming Interface. We coded hospitals as metropolitan (counties in metro areas), rural-adja- cent to metro (counties adjacent to a metro area), or rural (counties not adjacent to a metro area) using the hospital’s county and the United States Department of Agriculture (USDA) Economic Research Service Rural-Urban Continuum Codes [14]. Hospital characteristics for the transferring and receiving hospitals included trauma center designa- tion, annual ED visit volume, number of Hospital beds, total surgical op- erations, metropolitan or rural location, teaching status, and total charges for each patient’s care.

Descriptive statistics were used to describe the patient, injury, and hospital characteristics of the trauma transfer population, total charges for hospital care, trauma transfer patterns, and the prevalence of differ- ent types of injury and procedures. Univariate testing of explanatory variables and the outcome of discharge without procedures was tested by chi-squared or One-way ANOVA for categorical variables and t-tests for continuous variables.

We used a multivariable hierarchical logistic regression model to es- timate the association between injury pattern and the odds of discharge without procedures adjusted for patient demographics, mechanism of injury, injury severity, and both transferring and receiving hospital characteristics. We modeled random intercepts for the transferring hos- pital and adjusted for state-level fixed effects. All variables included had b 4% missing data. STATA software version 13.1 was used to conduct all analyses. Analytic code is included in Appendix 1.

Results

After exclusions there were 48,160 injured adult transfer patients (Fig. 1). Most (79.5%) initially presented to a non-trauma center. Of these, 48.7% were then transferred to another non-trauma center (Table 1). Only 77.0% of patients with Major injuries and 62.8% of pa- tients with moderate injuries were transferred to a designated Level 1 or 2 trauma center (Table A.1). The rates of discharge from the second ED without procedures varied by state from 28.6% in Utah to 59.3% in Massachusetts (Table A.2).

Our primary analysis cohort consisted of 22,011 transfers to a higher level of care (45.7% of all injured adult transfer patients; Fig. 1). Patients in this cohort were treated at 791 transferring hospitals and 141 receiv- ing hospitals across 6 states. Transferring hospital trauma transfer pa- tient volumes ranged from 1 (36 hospitals) to 214 patients (median

Fig. 1. Patient population flowchart.

19; IQR 7-38) while receiving hospital trauma transfer patient volumes ranged from 1 (2 hospitals) to 1050 patients (median 74; IQR 24-214). At the receiving ED among those transferred for a higher level of care, 10,380 patients (47.2%) were discharged. Only 2346 (22.6%) discharged patients received a therapeutic procedure at the receiving hospital prior to discharge. The most frequently used therapeutic proce- dure codes among discharged patients were for fracture management (39.0%) and laceration repair (32.0%), while just 19.7% of all therapeutic codes were for complex Wound management such as burns or lacera- tions of the hands or face, or for procedures typically performed in an

operating room.

After excluding discharged patients who received any therapeutic procedure or observation services, 8019 were discharged without pro- cedures, representing 36.4% of all higher level of care transfer patients and 77.3% of those discharged after transfer (Fig. 1). Table 2 describes the characteristics of the patient population, their injuries, and the transferring and receiving hospitals.

After adjustment for patient, transferring and receiving hospital factors, all injury patterns had higher odds of discharge without

procedures than torso injuries except for burns, which were not signif- icantly different (Fig. 2). The highest odds of discharge without proce- dures were for skin or soft tissue injuries (OR 6.77, 95%CI 5.60-8.18), facial injuries (OR 3.83, 95%CI 3.14-4.68), head injuries (OR 3.74, 95%CI 3.07-4.55), and hand injuries (OR 3.14, 95%CI 2.56-3.84). Increas- ing age over 34, female sex (OR 1.24, 95%CI 1.15-1.34), Medicaid (OR 1.33, 95%CI 1.19-1.49), and being uninsured (OR 1.32, 95%CI 1.19- 1.46) were also significantly associated with being discharged without procedures.

On the other hand, discharge without procedures was less likely with additional comorbidities (OR 0.89, 95%CI 0.86-0.92), higher injury severity scores (OR 0.89, 95%CI 0.88-0.0.90), injury by laceration (OR 0.53, 95%CI0.46-0.61), and being transferred from further away (OR 0.96, 95%CI 0.94-0.97 for every additional 10 min of driving time).

Among the approximately two-thirds of patients for whom charge data was available, the mean charges for care of patients discharged without procedures were $6076 (95% CI $5896-$6256),

$2859 (95% CI $2750-$2968) of which came from the receiving hospital ED (Table 3). Total charges for the second ED visit for the 65.5% of

Table 1

Trauma transfer patterns by trauma center designation in 2011, N = 48,160.

Initial hospital	Receiving hospital					Totals (initial hospital designation)
	Not a trauma center	Level 1	Level 2	Level 3	Level 4
Not a trauma center	18,651 (48.7)a	11,670 (30.5)	7734 (20.2)	229 (0.6)	19 (0.1)a	38,303 (79.5)
Level 1	2358 (67.7)a	865 (24.9)a	228 (6.6)a	27 (0.8)a	3 (0.1)a	3481 (7.2)
Level 2	2256 (55.9a)	995 (24.7)	684 (17.0)a	73 (1.8)a	25 (0.6)a	4033 (8.4)
Level 3	355 (31.8)a	227 (20.3)	265 (23.7)	212 (19.0)a	59 (5.3)a	1118 (2.3)
Level 4	131 (10.7)a	183 (14.9)	441 (36.0)	267 (21.8)	203 (16.6)a	1225 (2.5)
Totals (receiving hospital designation)	23,751 (49.3)	13,940 (29.0)	9352 (19.4)	808 (1.7)	309 (0.6)
a Excluded from further analysis.

Table 2

Patient and injury characteristics of trauma transfers by discharge without procedures in 2011.

Demographic characteristic	Total pop N = 22,011 N (% of population)	Discharged without procedures N (% of characteristic) N = 8019	Admitted, observed, or underwent procedures N (% of characteristic) N = 13,992
Patient characteristics Age (years)a	46 (29-63)	36 (26-50)	52 (32-72)
15-34	7561 (34.4)	3747 (49.6)	3814 (50.4)
35-54	6568 (29.9)	2825 (43.0)	3743 (57.0)
55-64	2635 (12.0)	776 (29.5)	1859 (70.5)
65 +	5191 (23.6)	646 (12.4)	4545 (87.6)
Sex
Male	13,492 (61.6)	4904 (36.4)	8588 (63.7)
Female	8401 (38.4)	3060 (36.4)	5341 (63.6)
Race/ethnicity
White, non-Hispanic	14,794 (69.0)	4668 (31.6)	10,126 (68.5)
Black, non-Hispanic	2501 (11.7)	1284 (51.3)	1217 (48.7)
Hispanic	3214 (15.0)	1485 (46.2)	1729 (53.8)
Other	940 (4.4)	364 (38.7)	576 (61.3)
Primary insurance
Private	5866 (26.7)	1925 (32.8)	3941 (67.2)
Medicare	5644 (25.7)	1025 (18.2)	4619 (81.8)
Medicaid	3443 (15.7)	1830 (53.1)	1613 (46.9)
Uninsured	5364 (24.4)	2600 (48.5)	2764 (51.5)
Other	1628 (7.4)	603 (37.0)	1025 (63.0)
Median household income of home zip codeb
b$39,000	4283 (28.0)	1605 (37.5)	2678 (62.5)
$39,000-47,999	4593 (30.0)	1504 (32.8)	3089 (67.2)
$48,000-63,999	3804 (24.9)	1254 (33.0)	2550 (67.0)
>=$64,000	2608 (17.1)	862 (33.1)	1746 (66.9)
Median number of comorbid chronic conditionsa	0 (0-1)	0 (0-1)	0 (0-2)
Mechanism of injury
Motor vehicle crash	2774 (12.6)	813 (29.3)	1961 (70.7)
Fall	7907 (35.9)	2044 (25.9)	5863 (74.1)
Assault	2275 (10.3)	1030 (45.3)	1245 (54.7)
gunshot wound	313 (1.4)	79 (25.2)	234 (74.8)
Laceration/Stab wound	1419 (6.5)	533 (37.6)	886 (62.4)
Pattern of most severe injury
Skin and soft tissue	5094 (24.1)	3369 (66.1)	1725 (33.9)
Orthopedic	4378 (20.8)	1137 (26.0)	3241 (74.0)
Head	4004 (19.0)	773 (19.3)	3231 (80.7)
Hand	2096 (9.9)	909 (43.4)	1187 (56.6)
Face	2088 (9.9)	1008 (48.3)	1080 (51.7)
Torso	1756 (8.3)	187 (2.4)	1569 (11.8
Spine	1420 (6.7)	272 (19.2)	1148 (80.9)
Ophthalmic	221 (1.1)	88 (39.8)	133 (60.2)
Burns	44 (0.2)	8 (18.2)	36 (81.8)
Overall injury severity scorea	4 (1-9)	2 (1-4)	4 (2-9)
Minor: ISS 0-8	15,740 (74.6)	7285 (46.3)	8455 (53.7)
Moderate: ISS 9-15	3354 (15.9)	352 (10.5)	3002 (89.5)
Severe: ISS 15-75	2008 (9.5)	114 (5.7)	1894 (94.3)
Serious head injury: AIS >= 3	2816 (12.8)	292 (10.4)	2524 (89.6)
Serious chest injury: AIS >= 3	862 (3.9)	40 (4.6)	822 (95.4)
Serious Abdominal injury: AIS >= 3	181 (0.8)	10 (5.5)	171 (94.5)
Serious extremity injury AIS >= 3	1220 (5.5)	49 (4.0)	1171 (96.0)
Transferring hospital characteristics Driving distance between hospitals (miles)a	19.8 (8.9-44.0)	12.2 (5.7-28.7)	25.7 (11.0-48.8)
Driving time between hospitals (minutes)a	28 (16-49)	22 (13-38)	33 (19-56)
Trauma center (level 2)	995 (4.5)	460 (46.2)	535 (53.8)
Annual ED visit volume (thousands)a	31.8 (18.4-51.0)	33.9 (20.3-58.2)	28.9 (16.3-48.6)
Number of hospital bedsa	196 (110-322)	223 (120-354)	182 (100-295)
Total surgical operations (thousands)a	5.9 (3.4-9.0)	6.6 (4.0-10.4)	5.7 (3.2-8.5)
Rural/urban location
Metropolitan	18,581 (84.7)	7264 (39.1)	11,317 (60.9)
Rural-adjacent to metro	2646 (12.1)	603 (22.8)	2043 (77.2)
Rural-not adjacent	704 (3.2)	111 (15.8)	593 (84.2)
Teaching hospital	5380 (24.5)	2464 (45.8)	2916 (54.2)
Receiving hospital characteristics
Trauma center (level 1 or 2)	21,515 (97.8)	7799 (36.3)	13,716 (63.8)
Annual ED visit volume (thousands)a	61.6 (51.4-88.9)	75.5 (57.1-97.7)	60.1 (46.0-87.5)
Number of hospital bedsa	563 (367-812)	576 (368-815)	563 (367-812)
Total surgical operations (thousands)a	14.5 (11.1-25.9)	14.1 (11.9-25.3)	14.5 (10.7-26.4)

Table 2 (continued)

Demographic characteristic	Total pop N = 22,011 N (% of population)	Discharged without procedures N (% of characteristic) N = 8019	Admitted, observed, or underwent procedures N (% of characteristic) N = 13,992
Rural/urban location Metropolitan	21,606 (98.2)	7914 (36.6)	13,692 (63.4)
Rural-adjacent to metro	5 (0.02)	4 (80.0)	1 (20.0)
Rural-not adjacent	400 (1.8)	101 (25.3)	299 (74.8)
Teaching hospital	16,832 (76.5)	6582 (39.1)	10,250 (60.9)

AIS: Abbreviated Injury Scale.

^a Results for continuous variables presented as Median (IQR).

^b Data not available in California.

patients discharged without procedures for which charges were avail- able were $31.9 million.

Discussion

This is a large study of nearly 50,000 trauma transfer patients across six states with diverse geographies and demographics, and is the first to tie together patient, hospital, and visit data from nearly 800 hospitals, both transferring and receiving, trauma and non-trauma-designated, to provide a robust picture of ED-to-ED trauma transfers and rates of discharge from the receiving ED without procedures. We have several novel findings. Nearly half (49%) of trauma transfers never went to a designated trauma center including a surprisingly high proportion

(23%) of patients with severe injuries, and that 36% of patients trans- ferred to a higher level of care were discharged from the receiving ED without procedural intervention. After adjustment of patient and hospi- tal factors, discharge without procedures was more likely for subspe- cialty amenable injuries. We identified socioeconomic disparities as evidenced by patients with Medicaid or who were uninsured had higher odds of experiencing a transfer resulting in discharge without proce- dure. This phenomenon poses a significant burden on patients and hos- pitals as the treatment at the receiving ED added an additional $2859 on average (95% CI $2750-$2968) per patient to the total charges for trau- ma care, not including the costs of ambulance transport between facili- ties, and a quarter of patients had to travel more than an hour’s drive from their initial ED.

Fig. 2. Multivariable analysis of factors associated with discharge without procedures after trauma transfer.

Table 3 Total charges for care.
	All patients Mean +- SD	Discharge without procedures Mean +- SD	Admitted, observed, or underwent procedures Mean +- SD
Initial hospital	$5487 +- $6410	$3197 +- $4155	$6675 +- $7022
Missing N (%)	6409 (29.1)	2690 (33.5)	3719 (26.6)
Receiving hospital	$45,183 +- $100,255	$2859 +- $4054	$61,850 +- $114,104
Missing N (%)	3323 (15.1)	2739 (34.2)	584 (4.2)
Total	$36,477 +- $65,819	$6076 +- $6651	$52,016 +- $76,224
Missing N (%)	6490 (29.5)	2769 (34.5)	3721 (26.6)

We observed similar injury patterns associated with discharge with- out procedures as those found in prior, smaller studies, including higher rates for head [1,5-7], face [5], and orthopedic injuries [1,7]. It has been shown that growth in transfer rates for orthopedic and head injuries has outpaced growth in their overall incidence [1]. Studies examining or- thopedic transfers in particular have found that 52% could have been managed by a general community orthopedist on call at the transferring hospital [24], and 27% were due to lack of orthopedic coverage at trans- ferring hospitals [25]. Our study also found high rates of discharge with- out procedures for hand injuries. While hand injuries have not been factors in prior studies of trauma transfers, a review of hand transfer cases found that 66% of these transfers did not require emergent treat- ment by a specialist not available at the referring facility [26].

All of these injury patterns suggest the potential benefit of enhanc- ing access to sub-specialty care and expertise at non-trauma center hos- pitals, specifically neurosurgeons, facial surgeons, Orthopedic surgeons, hand surgeons, and ophthalmologists. Nationwide 74% of EDs have dif- ficulty maintaining full-time on-call coverage, with difficulties more pronounced for neurosurgical (75%), plastic (81%), and hand (80%) cov- erage [27], with only 7% of hospitals in one state offering 24/7 hand spe- cialist availability [28], and on-call coverage has continued to worsen over time [29]. Low reimbursement from underinsured emergency and trauma patients, increased Medical malpractice liability for emer- gent cases, and a desire to improve lifestyle by limiting unpredictable on-call hours all contribute to poor on-call coverage for these specialists [30], and this lack of coverage has created a current standard of care that requires transfer by ambulance to obtain specialty follow-up care.

Telemedicine has been shown to decrease length of stay and transfer rates and to save money in trauma care [31], and offers a potential solu- tion to reduce rates of discharge without procedures. It has been validat- ed for facial [32,33], head [34,35], hand [32], ophthalmic [36], pediatric orthopedic [37], and burn injuries [38]. For many of these conditions, re- view of imaging studies and a focused physical exam determine dispo- sition. For some injuries, such as partial finger amputations or complex facial lacerations, patients may be transferred for procedures to be per- formed by specialists. However a prior study revealed that 65% of proce- dures performed in discharged trauma transfers were minor orthopedic or wound care management procedures which could be performed by a general practitioner [7], and our study also found that only 27% of the discharged patients who had Therapeutic procedures performed had a specialized or complex procedure that would be out-of-scope for a gen- eral emergency physician.

It was surprising to find that half of trauma transfers are not sent to a designated level 1 or 2 trauma center. Possible reasons are that patients have isolated specialty injuries and the appropriate specialist is avail- able in a non-trauma hospital, or an effort by hospitals to keep patients within their affiliated networks irrespective of trauma designation. A prior study found that privately insured patients with major trauma were less likely to be transferred at all, presumably due to higher reim- bursement for their care [2], and this phenomenon may play a role in transfers to non-trauma centers within the same hospital system. While the higher use of trauma centers among patients who were not discharged home without procedures in this study suggest appropriate

triage of more severe patients to trauma centers, the finding that 23% of patients with major traumatic injuries do not go to a trauma center is concerning given the benefit of trauma center care for major trauma [39,40], and requires further investigation.

This study has several limitations: it does not compare injured pa- tients who were discharged from the first ED to those who were trans- ferred, although this comparison has been made previously [2]. It also does not compare injured patients who were directly admitted upon transfer to those who were transferred to a second ED, although this is likely a sicker population. We assume that patients who were discharged without therapeutic procedures or observation had positive outcomes, but we do not track patients beyond the initial discharge to determine whether discharge was appropriate.

While HCUP data is well validated, there are limitations to adminis- trative data including the accuracy of data entry and the availability of all predictor variables. Patients were identified using validated ICD-9 codes for trauma, but we cannot ensure that another problem or diag- nosis was not the primary reason for transfer, such as a patient who fell and suffered Minor injury due to a Cardiac syncope or stroke. How- ever, we did analyze the prevalence of all diagnosis codes occurring in the population at rates N 0.05% (data available upon request) and found no ICD-9 codes for stroke, acute myocardial infarction, or cardiac arrest, and <= 2% prevalence of codes for seizure, coronary artery disease, and cardiac dysrhythmias. We also analyzed all procedure codes and found that just 130 patients (0.3%) underwent diagnostic or interven- tional cardiology procedures, 182 (0.4%) underwent psychiatric evalua- tions, and 73 (0.2%) underwent pregnancy-related diagnostic or interventional procedures. In addition, the cohort of interest was deter- mined based on whether or not a procedure occurred after transfer, but some procedures may be undercoded in the data which could lead to an overestimate of the observed prevalence of discharge after transfer without procedural intervention.

We report on the charges associated with transfer for injury, focusing on charges for the second ED visit as potentially avoidable. However, these are charge data and do not reflect true costs of care. Charges for transport are not available either, and it is unknown what percent of patients used an ambulance, helicopter, or private vehicle for transport.

Finally, while our data presents all visits from six states, the number of visits in small rural states (Iowa and Utah) constitute a smaller pro- portion of the sample than visits from California, New York, and Florida, giving more weight to trends from these larger states. We adjusted the model with state-level fixed effects to mitigate this issue. Transfers of patients to or from hospitals across state lines, which may be more prev- alent in the rural states with limited trauma and academic centers, also would not be captured in this study since longitudinal identifiers are not consistent across state lines.

In conclusion, more than a third of injured patients undergoing ED- to-ED transfer to a higher level of trauma care are transferred to another ED and discharged without procedural intervention. Increasing access to hand, facial, neurosurgical, and ophthalmology specialists might re- duce some of these transfers and the associated financial and logistical burdens on patients, physicians, and hospitals.

Funding source

DNH and MKD’s time was supported by the NIH NHLBI Career Devel- opment Program in emergency research (Grant NIH K12HL109009). LMD’s time was supported by the RWJF Clinical Scholars Program.

Presentations

The work was presented as a lighting oral abstract at SAEM 2016.

Disclosures

LMD, DNH, DK, MJK, and MKD report no conflicts of interest.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2017.10.024.

References

1. Esposito TJ, Crandall M, Reed RL, Gamelli RL, Luchette FA. socioeconomic factors, medicolegal issues, and trauma patient transfer trends: is there a connection? J Trauma 2006;61(6):1380-6 [discussion 1386-1388].
2. Delgado MK, Yokell MA, Staudenmayer KL, Spain DA, Hernandez-Boussard T, Wang NE. Factors associated with the disposition of severely injured patients initially seen at non-trauma center emergency departments: disparities by insurance status. JAMA Surg 2014;149(5):422-30.
3. MacKenzie EJ, Rivara FP, Jurkovich GJ, Nathens AB, Frey KP, Egleston BL, et al. A na- tional evaluation of the effect of trauma-center care on mortality. N Engl J Med 2006; 354(4):366-78.
4. Vogel LD, Vongsachang H, Pirrotta E, Holmes JF, Sherck J, Newton C, et al. Variations in pediatric trauma transfer patterns in Northern California pediatric trauma centers (2001-2009). Acad Emerg Med 2014;21(9):1023-30.
5. Con J, Long D, Sasala E, Khan U, Knight J, Schaefer G, et al. Secondary overtriage in a statewide rural Trauma system. J Surg Res 2015;198(2):462-7.
6. Sorensen MJ, von Recklinghausen FM, Fulton G, Burchard KW. Secondary overtriage: the burden of unnecessary interfacility transfers in a rural trauma system. JAMA Surg 2013;148(8):763-8.
7. Tang A, Hashmi A, Pandit V, Joseph B, Kulvatunyou N, Vercruysse G, et al. A critical analysis of secondary overtriage to a level I trauma center. J Trauma Acute Care Surg 2014;77(6):969-73.
8. Ciesla DJ, Sava JA, Street 3rd JH, Jordan MH. Secondary overtriage: a consequence of an immature trauma system. J Am Coll Surg 2008;206(1):131-7.
9. Babu MA, Nahed BV, Demoya MA, Curry WT. Is trauma transfer influenced by factors other than medical need? An examination of insurance status and transfer in pa- tients with mild head injury. Neurosurgery 2011;69(3):659-67 [discussion 667].
10. Bible JE, Kadakia RJ, Kay HF, Zhang CE, Casimir GE, Devin CJ. How often are interfacility transfers of spine injury patients truly necessary? Spine J 2014; 14(12):2877-84.
11. Newgard CD, McConnell KJ, Hedges JR. Variability of trauma transfer practices among non-tertiary care hospital emergency departments. Acad Emerg Med 2006; 13(7):746-54.
12. Koval KJ, Tingey CW, Spratt KF. Are patients being transferred to level-I trauma cen- ters for reasons other than medical necessity? J Bone Joint Surg Am 2006;88(10): 2124-32.
13. Nathens AB, Maier RV, Copass MK, Jurkovich GJ. Payer status: the unspoken triage criterion. J Trauma 2001;50(5):776-83.
14. Osen HB, Bass RR, Abdullah F, Chang DC. Rapid discharge after transfer: risk factors, incidence, and implications for trauma systems. J Trauma 2010;69(3):602-6.
15. Mohr NM, Harland KK, Shane DM, Miller SL, Torner JC. Potentially avoidable pediat- ric interfacility transfer is a costly burden for rural families: a cohort study. Acad Emerg Med 2016;23(8):885-94.
16. Newgard CD, Staudenmayer K, Hsia RY, Mann NC, Bulger EM, Holmes JF, et al. The cost of overtriage: more than one-third of low-risk injured patients were taken to major trauma centers. Health Aff (Millwood) 2013;32(9):1591-9.
17. Fishman PE, Shofer FS, Robey JL, Zogby KE, Reilly PM, Branas CC, et al. The impact of trauma activations on the care of emergency department patients with potential acute coronary syndromes. Ann Emerg Med 2006;48(4):347-53.
18. Smith DC, Chapital A, Burgess Uperesa BM, Smith ER, Ho C, Ahana A. Trauma activa- tions and their effects on non-trauma patients. J Emerg Med 2011;41(1):90-4.
19. Steiner C, Barrett M, Hunter K. Hospital readmissions and multiple emergency de- partment visits, in selected states, 2006-2007: statistical brief #90. Healthcare cost and utilization project (HCUP) statistical briefsedn. Rockville (MD). ; 2006.
20. Greene NH, Kernic MA, Vavilala MS, Rivara FP. Validation of ICDPIC software injury severity scores using a large regional trauma registry. Inj Prev 2015;21(5):325-30.
21. Barell V, Aharonson-Daniel L, Fingerhut LA, Mackenzie EJ, Ziv A, Boyko V, et al. An introduction to the Barell body region by nature of injury diagnosis matrix. Inj Prev 2002;8(2):91-6.
22. Fuller DA, Sinclair E, Geller J, Quanbeck C, Snook J, The Treatment and Advocacy Center. Going, going, gone: trends and consequences of eliminating state psychiatric beds; 2016.
23. Torrey EF, Fuller DA, Geller J, Jacobs C, Ragosta K. The treatment advocacy center. No room at the inn: trends and consequences of closing public psychiatric. Hospitals 2012:2005-10.
24. Thakur NA, Plante MJ, Kayiaros S, Reinert SE, Ehrlich MG. Inappropriate transfer of patients with orthopaedic injuries to a level I trauma center: a prospective study. J Orthop Trauma 2010;24(6):336-9.
25. Crichlow RJ, Zeni A, Reveal G, Kuhl M, Heisler J, Kaehr D, et al. Appropriateness of pa- tient transfer with associated orthopaedic injuries to a level I trauma center. J Orthop Trauma 2010;24(6):331-5.
26. Friebe I, Isaacs J, Mallu S, Kurdin A, Mounasamy V, Dhindsa H. Evaluation of appro- priateness of patient transfers for hand and microsurgery to a level I trauma center. Hand (N Y) 2013;8(4):417-21.
27. Rao MB, Lerro C, Gross CP. The shortage of on-call surgical specialist coverage: a na- tional survey of emergency department directors. Acad Emerg Med 2010;17(12): 1374-82.
28. Mueller MA, Zaydfudim V, Sexton KW, Shack RB, Thayer WP. Lack of emergency hand surgery: discrepancy between elective and emergency hand care. Ann Plast Surg 2012;68(5):513-7.
29. Rudkin SE, Langdorf MI, Oman JA, Kahn CA, White H, Anderson CL. The worsening of ED on-call coverage in California: 6-year trend. Am J Emerg Med 2009;27(7): 785-91.
30. Chang G, Weiss A, Kosowsky JM, Orav EJ, Smallwood JA, Rauch SL. Characteristics of adult Psychiatric patients with stays of 24 hours or more in the emergency depart- ment. Psychiatr Serv 2012;63(3):283-6.
31. Duchesne JC, Kyle A, Simmons J, Islam S, Schmieg Jr RE, Olivier J, et al. Impact of tele- medicine upon rural trauma care. J Trauma 2008;64(1):92-7 [discussion 97-98].
32. Paik AM, Granick MS, Scott S. plastic surgery telehealth consultation expedites emer- gency department treatment. J Telemed Telecare 2016.
33. Fonseca AS, Goldenberg DC, Stocchero GF, Luiz AV, Gemperli R. Validation of video- conference with smartphones in telemedicine facial trauma care: analysis of concor- dance to on-site evaluation. Ann Plast Surg 2015.
34. Stippler M, Yonas H. Introduction of telemedicine in traumatic brain injury care. In: Timmons SD, editor. Neurotrauma & critical care news. American Association of Neurological Surgeons; Fall; 2013.
35. Moya M, Valdez J, Yonas H, Alverson DC. The impact of a telehealth web-based solu- tion on neurosurgery triage and consultation. Telemed J E Health 2010;16(9):945-9.
36. Blackwell NA, Kelly GJ, Lenton LM. Telemedicine ophthalmology consultation in re- mote Queensland. Med J Aust 1997;167(11-12):583-6.
37. Zennaro F, Grosso D, Fascetta R, Marini M, Odoni L, Di Carlo V, et al. Teleradiology for remote consultation using iPad improves the use of health system human resources for paediatric fractures: prospective controlled study in a tertiary care hospital in Italy. BMC Health Serv Res 2014;14:327.
38. Saffle JR, Edelman L, Theurer L, Morris SE, Cochran A. Telemedicine evaluation of acute burns is accurate and cost-effective. J Trauma 2009;67(2):358-65.
39. Newgard CD, McConnell KJ, Hedges JR, Mullins RJ. The benefit of higher level of care transfer of injured patients from nontertiary hospital emergency departments. J Trauma 2007;63(5):965-71.
40. Garwe T, Cowan LD, Neas B, Cathey T, Danford BC, Greenawalt P. Survival benefit of transfer to Tertiary trauma centers for major trauma patients initially presenting to nontertiary trauma centers. Acad Emerg Med 2010;17(11):1223-32.