Understanding an inclusive trauma system through characterization of admissions at level IV centers

The American Journal of Surgery (2016) -, -–-

Understanding an inclusive trauma system through characterization of admissions at level IV centers Hayden L. Smith, Ph.D.a,b, H. Clay Dean V, M.D.a, Richard A. Sidwell, M.D., F.A.C.S.a,* a

Iowa Methodist Medical Center, General Surgery Residency Program, 1415 Woodland Avenue, Suite 140, Des Moines, IA 50309-1453, USA; bDepartment of Internal Medicine, Carver College of Medicine, University of Iowa, Iowa City, IA, USA

KEYWORDS: Trauma centers; Injuries; Registries; Aged; Hip fractures; Traumatic brain injury

Abstract BACKGROUND: Level IV trauma centers are an integral part of inclusive trauma systems, although sparse data exists for these facilities. METHODS: An observational study was conducted using a Midwestern state’s inpatient data files to characterize level IV center patients. Injury and severity levels, injury mechanism and/or intent, and distances to nearest tertiary centers were determined. RESULTS: During the study year, 3,346 injured patients were admitted at level IV centers. The median distance to nearest tertiary center was 43 miles. Median patient age was 81 years, and primary injury mechanism was falls. Overall, 22% of patients had an isolated hip fracture. Of moderately injured patients, 64% had an isolated hip fracture, 8% nonisolated hip fractures, and 9% rib fractures without hip fracture. Overall, 30% of patients had a high severity of injury. CONCLUSIONS: A large number of patients were admitted to level IV trauma centers. Patients admitted tended to be elderly and have orthopedic fall injuries. Study results provide important implications for provider education, prevention efforts, need for orthopedic surgical capabilities, and necessity of capturing these data in registries. Ó 2016 Elsevier Inc. All rights reserved.

There were no relevant financial relationships or any sources of support in the form of grants, equipment, or drugs. The authors declare no conflicts of interest. Abstract presented at the 2015 Resident Trauma Papers Competition, The American College of Surgeons’ Committee on Trauma (ACS COT), Kansas City, MO, December 5–6, 2014. * Corresponding author. Tel.: 11-515-241-4076; fax: 11-515-2418040. E-mail address: [email protected] Manuscript received July 30, 2015; revised manuscript December 14, 2015 0002-9610/$ - see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2015.12.023

An inclusive trauma system uses all facilities to care for patients with varying severity of injury. The goal of the system is to match patient needs with available resources. Published data have shown trauma systems can improve patient outcomes and reduce mortality by up to 15%.1–5 Level IV trauma centers are an integral part of an inclusive trauma system. In general, these centers are expected to perform initial assessment and stabilization of patients, with many patients requiring transfer to tertiary centers. However, patients can be admitted to level IV trauma centers for definitive care. A frequent limitation

2

The American Journal of Surgery, Vol -, No -, - 2016

in published research is the inability to characterize patients who are admitted and received care at level IV centers.2,6–18 Issues arise due to the completeness and consistency of available level IV data in traditional trauma registries (eg, state trauma registries and National Trauma Data Bank). If level IV center data have been published, it has been combined with level III trauma center data19–21 or has focused on data for patients transferred from level III or IV centers to tertiary centers.2–4,7,9,10,17,18,22 Without overall data on admitted patients at level IV trauma centers, it is not possible to understand a trauma system as a whole or evaluate its complete effectiveness. The purpose of this study was to characterize injured patients admitted at level IV trauma centers in a mature inclusive trauma system.

Methods An observational study was conducted using retrospective 2011 inpatient hospital data from the state of Iowa in the United States. The study state had an inclusive trauma system, which had been fully functional since 2001. Institutional Review Board approval was granted for the study. Inclusion criterion required an inpatient admission at a hospital classified as a level IV trauma center. Excluded patients were: elective admissions or newborns; patients with same calendar day transfer to another facility; and patients with no trauma International Classification Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis code as defined by the National Trauma Data Bank.23 Hip fracture, rib fracture, and traumatic brain injury (TBI) status were based on ICD-9-CM code ranges: 820.00-820.99, 807.00-807.09, and 800.00-804.99 as well as 850-854.19, respectively. Abbreviated Injury Severity (AIS) scores and Injury Severity Scores (ISS) were generated for patients using ICD-9-CM lexicon via the ICDPIC 3.0 package within STATA 13.0 (StataCorp, College Station, TX, USA). Diagnosis codes for burns were not converted into AIS scores nor were codes for 958.xx (ie, Certain Early Complications of Trauma). Probability of patient mortality was also determined using the Trauma Mortality Prediction Model within the ICDPIC 3.0 package.24 Injury mechanism and intent were determined for patients using ICD-9-CM External Cause of Injury codes (E-codes) based on the Centers for Disease Control and Prevention classification matrix.25 E-codes were the only data element in the data set with missingness, with approximately 35% absent. Missingness was examined using multiple logistic regression with E-codes (missing or observed) classified as the binary dependent variable. Model results provided information toward data missing at random ((P(Ymissing, Yobserved) X), X 5 vector of covariates) and multiple imputation based on fully conditional specification of arbitrary missing

Figure 1 Flow diagram of 2011 Iowa inpatients admitted at level IV trauma centers (n 5 3,346). *Patients with only a burn or early complications of trauma diagnosis.

H.L. Smith et al.

Level IV trauma centers

3

pattern was conducted with 100 imputes for injury mechanism and intent information.26,27 Longitude and latitude data for study hospitals and level I and II trauma centers within the state and adjacent states were used to determine distances in miles (as a bird flies on an ellipsoid) between all study hospitals to the nearest level I or II trauma center. These calculations were conducted using the GEODIST function in SAS 9.4 (SAS Institute, Inc., Cary, NC, USA). Hospitals were also dichotomized into urban or rural. The American Hospital Association classifies hospitals as urban, critical access, rural, and rural referral. The critical access, rural, and rural referral hospitals were collapsed into the category of ‘‘rural.’’ Descriptive statistics were calculated for patient and hospital data. Continuous variables were presented with measures of central tendency and dispersion (ie, medians and interquartile ranges [IQR]), whereas categorical variables were presented with counts and percentages. Patients

were stratified based on ISS categories (ie, 1 to 8: minor injury, 9 to 15: moderate injury, 16 to 75: severe injury), years of age (ie, %55, 56 to 79, R80), and expiration status (ie, alive or deceased). Patient characteristics were compared within these categories using Kruskal–Wallis and/or Wilcoxon-rank sum tests for continuous by categorical data and Fisher’s exact tests for categorical by categorical data. Bivariate statistical comparisons using imputed variables were conducted with chi-square test and subsequently pooled.28 Adjusted odds ratios using imputed data were based on multiple logistic regressions and were calculated with 95% confidence intervals (CI) via jackknife variance estimation and pooled using the MIANALYZE procedure in SAS, 9.4. Statistical tests were 2-tailed with a .05 level of significance and multiplicity corrected for using the Bonferroni method. All statistics were conducted with SAS 9.4.

Results Table 1 Characteristics of inpatients defined as trauma patients admitted to Iowa level IV trauma centers in 2011 Variable

(n 5 3,346), n (%)

Median age (y) Female Race White Asian African American/black Other Unavailable/unknown Injury mechanism* Fall Motor vehicle collision Other Injury intent Unintentional Self-inflicted Assault Undetermined Traumatic brain injury Hospital classification Urban Rural Median maximum abbreviated Injury Score Median Injury Severity Score Median probability of mortality Median length of stay (d) Discharge destination Skilled care Home Deceased Hospice Other

81 (66–88) 2,168 (65) 2,872 128 76 91 179

(86) (4) (2) (3) (5)

2,642 (79) 119 (4) 585 (17) 3,285 35 18 8 233

(98) (1) (1) (0) (7)

There were 337,577 inpatient admissions in the state of Iowa in 2011. Of these, 92,498 were at level IV centers. After the application of exclusion criteria, there were 3,346 injured patients included in the study sample (Fig. 1). These patients represented approximately 21% of all trauma patient data in the state for the year. Of Iowa’s 118 hospitals, 93 were level IV centers with 87 of these centers having included trauma admissions within study sample. The median number of trauma admissions per included level IV center was 33 (IQR: 13 to 71) patients. The median distance from an included center to the nearest level I or II trauma center was 43 (IQR: 29 to 59) miles, with 75% of these level IV centers being within 60 miles of a level I or II center. Eighty-two (94%) of the center hospitals were classified as rural. For the 6 excluded centers with no eligible trauma admissions, median distance to a level

753 (23) 2,593 (77) 2 4 .01 4 1,891 1,294 76 49 36

(2–3) (4–9) (.1–.02) (3–5) (57) (39) (2) (1) (1)

Counts (percentages) and medians (interquartile ranges). *Top 3 E-code subgroups in ‘‘other’’ category: overexertion (2%); struck by or against (2%); natural and/or environment (2%).

Figure 2 Age (years) distribution of patients with an injury diagnosis admitted to a hospital designated as a level IV trauma centers (n 5 3,346).

4

The American Journal of Surgery, Vol -, No -, - 2016

I or II trauma center was 35 (IQR: 30 to 47) miles, and all were classified as rural. Patients admitted at level IV centers were mostly female and Caucasian (Table 1). Median age was 81 years with 75% over aged 65 years, 50% over aged 80 years, and 25% over aged 88 years (Fig. 2). Mechanism of injury for patients was primarily falls (79%). Rural level IV centers accounted for 77% of patient data. The median calculated ISS was 4. A total of 63% had mild injury severity with an ISS of 1 to 8 (Table 2), most with an ISS of 4 or less. Thirty-four percent of admissions had moderate injury severity of ISS 9-15. Three percent of patients had a severe injury severity of ISS greater than 15. The severely injured patients tended to be admitted to hospitals in smaller

communities, male, lower proportion of Caucasian race, higher proportion of motor vehicle collisions, and more likely to have died (P , .05). Of these patients, 78% had a TBI diagnosis. Primary injuries within the ISS groups revealed that mild injury severity (ISS 1 to 8) was primarily because of low severity orthopedic injuries. Moderate injury severity (ISS 9 to 15) was also mainly because of orthopedic injuries. Of the 1,127 patients in this group, 64% had an isolated hip fracture, 8% had nonisolated hip fractures, and 9% had rib fractures without hip fracture. Overall 22% of patients in the study sample had an isolated hip fracture. The 3% of patients with a high severity of injury (ISS.15) primarily had head and/or neck injuries.

Table 2

Patient characteristics based on Injury Severity Score (ISS) category (n 5 3,346)

Variable Median miles to level I/II trauma center Hospital classification Urban Rural Median age (y) Female Race White Asian African American/black Other Unavailable/unknown Injury mechanism Fall Motor vehicle collision Other Injury intent Unintentional Self-inflicted Assault Undetermined Median maximum Abbreviated Injury Score Median Abbreviated Injury Score Head or neck Face Chest Abdomen or pelvic contents Extremities or pelvic girdle Skin and subcutaneous tissue Median length of stay (days) Median probability of mortality Discharge destination Skilled care Home Deceased Hospice Other

ISS % 8; (n 5 2,113; 63%)

ISS 9–15; (n 5 1,127; 34%)

ISS R 16; (n 5 106; 3%)

34 (23–51)

32 (19–62)*

41 (29–60)*

461 1,652 79 1,359

(22%)* (78%) (63–87)* (64%)*

278 849 83 758

(25%)† (75%) (72–89)* (67%)†

14 92 81 51

(13%)* (87%) (65–88) (48%)*,†

1,801 57 63 55 137

(85%) (3%) (3%) (3%) (6%)

988 63 11 31 34

(88%) (6%) (1%) (2%) (3%)

83 8 2 5 8

(78%) (8%) (2%) (4%) (8%)

1,562 (74%)*,† 67 (3%) 484 (23%) 2,061 33 15 4 2 0 0 0 0 1 0 3 .01 1,029 996 33 22 33

(98%)* (2%) (1%) (0%) (1, 2)*

997 (88%)* 42 (4%) 88 (8%) 1,120 2 2 3 3

83 (79%)† 10 (9%) 13 (12%)

(99%)* (0%) (0%) (0%) (3, 3)*

104 0 1 1 4

(99%) (0%) (1%) (0%) (4, 4)*

(0, 0)* (0, 0)* (0, 0)* (0, 0)* (0, 2)* (0, 0)* (2, 5)* (.01–.01)

0 0 0 0 3 0 4 .02

(0, 0)† (0, 0)*,† (0, 0)* (0, 0)*,† (3, 3)* (0, 0)* (3, 6)*,† (.01–.02)

4 0 0 0 0 0 3 .05

(0, 4)*,† (0, 0)† (0, 3)* (0, 0)† (0, 2)* (0, 1)* (2, 6)† (.4–.10)

(48%)* (47%) (2%) (1%) (2%)

821 251 30 22 3

(73%)* (22%) (3%) (2%) (0%)

41 47 13 5 0

(39%)* (44%) (12%) (5%) (0%)

Counts (percentages) and median (interquartile ranges). Abbreviated injury score, 1 5 minor; 2 5 moderate; 3 5 serious; 4 5 severe; 5 5 critical; and 6 5 maximum. Identical superscript symbols note statistically different column groups.

H.L. Smith et al. Table 3

Level IV trauma centers

5

Patient characteristics based on deceased status (n 5 3,346)

Variable Median distance to level I/II trauma center Hospital classification Urban Rural Median age (y) Female Race White Asian African American/black Other Unavailable/unknown Injury mechanism Fall Motor vehicle collision Other Injury intent Unintentional Self-inflicted Assault Undetermined Median Injury Severity Score Median Abbreviated Injury Score Head or neck Face Chest Abdomen or pelvic contents Extremities or pelvic girdle Skin and subcutaneous tissue Median probability of mortality Median length of stay (d)

Alive (n 5 3,270) 33 (22–53)

Deceased (n 5 76)

P value

34 (24–52)

.6417 .8895

737 2,533 80 2,114

(23%) (77%) (66–88) (65%)

16 60 87 54

(21%) (79%) (81–91) (71%)

3,181 5 21 11 52

(97%) (0%) (1%) (0%) (2%)

75 0 0 0 1

(99%) (0%) (0%) (0%) (1%)

.9231

2,571 (79%) 118 (4%) 581 (17%)

71 (93%) 1 (2%) 4 (5%)

.0023

3,209 35 18 8 4 2 0 0 0 0 2 0 .02 4

76 0 0 0 9 3 0 0 0 0 2 0 .02 3

.8126

(98%) (1%) (1%) (0%) (4–9) (2–3) (0–0) (0–0) (0–0) (0–0) (0–3) (0–0) (.00–.02) (3–5)

(100%) (0%) (0%) (0%) (4–9) (2–3) (0–0) (0–0) (0–0) (0–0) (0–3) (0–0) (.01–.03) (1–4)

,.0001 .2755

.0004 .0002 ,.0001 .6760 .5767 .3903 .7650 .4050 ,.0001 .0181

Counts (percentages) and median (interquartile ranges). Abbreviated Injury Score, 1 5 minor; 2 5 moderate; 3 5 serious; 4 5 severe; 5 5 critical; and 6 5 maximum.

Patients with an age of 80 years or greater tended to be female, be injured by falls, have higher ISS, be admitted to rural hospitals. They were not as commonly discharged to their homes (P , .05; Table 3). Overall, discharge destination based on injury level revealed 47% mild injury (ISS 1 to 8) patients were discharged home, 48% to a skilled nursing facility, and 2% died. Of patients with moderate injury (ISS 9 to 15), 22% were discharged home, 73% discharged to skilled care facility, and 3% died. Patients in the severely injured group (ISS .15) had 44% discharged to home, 41% to skilled care facility, and 12% died. Patients that died were older and had higher AIS for the head and neck region (Table 4). Patients with a fall mechanism of injury had a crude higher percentage of death than those patients with a nonfall mechanism (3% vs , 1%). Although, these differences were not significant (adjusted odds ratios; 2.54, 95% CI: .42, 15.46) when controlling for age (years) and TBI status; adjusted odds of death increased 1.29 (95% CI: 1.26, 1.33) times for every 5 year increase in age and patients with TBI diagnosis had a 2.87 (95% CI: 1.48,

5.56) times greater adjusted odds of death than patients without a TBI diagnosis.

Comments More than 20% of trauma patients in the study state were admitted and received definitive care at a level IV trauma center. The median number of admissions per included hospital was 33, with 6 excluded hospitals having no admitted trauma patients. Most hospitals were within 60 miles of the nearest level I or II trauma center. Median patient age was over 80 years and primary mechanism of injury was falls. Overall injury level for patients was low with an ISS of 4. Patients with a greater injury level were likely to have a hip fracture, and the 3% of patients with the highest injury level were likely to have a TBI. To the authors’ knowledge, this is the first study to specifically characterize trauma admissions at level IV centers in a mature trauma system. Only a few other studies

The American Journal of Surgery, Vol -, No -, - 2016

6 Table 4

Patient characteristics based on age (n 5 3,346)

Variable Median distance to level I/II trauma center Hospital classification Urban Rural Median age (y) Female Race White Asian African American/black Other Unavailable/unknown Injury mechanism Fall Motor vehicle collision Other Injury intent Unintentional Self-inflicted Assault Undetermined Median Injury Severity Score Injury Severity Score category Score 1–8 Score 9–15 Score 16–75 Median maximum Abbreviated Injury Score Head or neck Face Chest Abdomen or pelvic contents Extremities or pelvic girdle Skin and subcutaneous tissue Median probability of mortality Median length of stay (d) Discharge destination Skilled care Home Deceased Hospice Other

%55 (n 5 479) 32 (8–53)

Patient age, y; 56–79; (n 5 1,087) 32 (8–53)*

135 344 42 1,995

(28%)* (72%) (29–50)* (41%)*

284 803 71 668

444 0 12 10 13

(93%)*,† (0%) (2%) (2%) (3%)

1,060 3 6 1 17

184 (38%)* 69 (14%) 226 (48%) (89%)*,† (7%) (3%) (1%) (1–8)*

361 99 19 2 0 0 0 0 2 0 .01 2

(75%)* (21%) (4%) (1–2)* (0–0) (0–0)*,† (0–0) (0–0)*,† (0–2)* (0–1)*,† (.00–.02)* (1–4)*,†

52 404 1 0 22

(11%)* (84%) (0%) (0%) (5%)

1,081 2 2 2 4

34 (24–60)*

(26%)† (74%) (64–76)* (61%)*

334 1,446 87 1,305

(19%)*,† (81%) (84–91)* (73%)*

(98%)* (0%) (1%) (0%) (2%)

1,752 2 3 0 23

(98%)† (0%) (0%) (0%) (1%)

848 (78%)* 37 (3%) 202 (19%)

426 33 15 5 4

R80 (n 5 1,780)

1,610 (90%)* 13 (1%) 157 (9%)

(100%)* (0%) (0%) (0%) (4–9)*

1,778 0 1 1 5

(100%)† (0%) (0%) (0%) (4–9)*

721 335 31 2 0 0 0 0 2 0 .01 4

(66%)* (31%) (3%) (2–3)* (0–0) (0–0)* (0–0) (0–0)* (0–3)* (0–0)* (.01–.02)* (3–5)*

1,031 693 56 2 0 0 0 0 2 0 .01 4

(58%)* (39%) (3%) (2–3)* (0–0) (0–0)† (0–0) (0–0)† (0–3)* (0–0)† (.01–.02)* (3–6)†

532 522 15 6 7

(49%)* (48%) (1%) (1%) (1%)

1,302 368 60 43 7

(73%)* (21%) (3%) (3%) (0%)

Counts (percentages) and median (interquartile ranges). Abbreviated Injury Score, 1 5 minor; 2 5 moderate; 3 5 serious; 4 5 severe; 5 5 critical; and 6 5 maximum. Identical superscript symbols note statistically different column groups.

have included data on level IV admissions, and none differentiated between admissions at level III and IV centers.19,21 In the presented study it was revealed that there is a large number of injured people admitted at level IV trauma centers. Although level IV centers generally transfer patients for definitive care, more than 3,000 injured patients were admitted to these hospitals during the study year. For comparison, 1 of the 2 level I trauma centers in the state averaged 1,550 trauma admissions annually during a comparable period.29,30 It should be noted that,

individually, each level IV center did not admit many patients, but collectively these patients represented a significant sum, reinforcing the need to understand the patient population and their needs. Study results provided reasonable data to support the effectiveness of an inclusive trauma system. Undertriage, defined as severely injured patients not being transferred to higher levels of care, did not appear to be a significant problem in the state’s trauma system. Field triage criteria published by the Centers for Disease Control and Prevention

H.L. Smith et al.

Level IV trauma centers

target an undertriage rate of less than 5%.31 In the presented study, only 3% of patients admitted at level IV centers had an ISS greater than 15. Overall mortality at the level IV centers appears favorable at 2%. The presented study was not able to address specific reasons why patients were admitted to level IV centers. It did not appear, however, to be related to the distance to higher levels of care. Seventy-five percent of the level IV hospitals in the state were within 60 miles of a verified level I or II trauma center. The median age for patients admitted at level IV trauma centers was 81 years. Overall 75% were older than 65 years and 25% were older than 88 years. This underscores the growing problem of injury in the geriatric population and demonstrates the need for geriatric trauma education for providers at smaller centers. Education should reflect the specific issues related to injury in the geriatric population such as altered physiology, anticoagulation medication and reversal strategies, elder abuse, rehabilitative measures, and end-of-life considerations. Furthermore, 79% of injuries were due to falls, creating orthopedic and head injury. Level IV trauma centers should use this information to help focus injury prevention efforts in their community on falls in the elderly. Twenty-two percent of patients had isolated hip fractures. Among the moderately injured patients (ISS 9 to 15), more than 70% had hip fractures. Of note, orthopedic surgery capability was not a requirement for level trauma IV centers within the state. This information reveals the critical need for orthopedic surgeons at centers and how it may be a major determinant of whether definitive injury care can be provided at the level IV facility or a transfer to higher level of care is necessary. Study results showed that comprehensive data from level IV centers should be included in standard trauma registries. Effective performance improvement requires an understanding of all elements of a trauma system, and it is an oversight to discount this population as insignificant. One common concern is that level IV trauma centers lack the personnel to perform trauma data submission in a timely manner. With a median of 33 trauma admissions per year, 50% of centers would average less than 3 patient charts to abstract and submit each month. ‘‘Resources for Optimal Care of the Injured Patient 2014’’ includes trauma registry data submission as a criterion for verification as a level IV hospital.32 This task does not appear to be overly burdensome, and states should adopt and enforce similar criteria in their respective trauma system. These data are also critical in assisting trauma systems to develop evidence-based guidelines regarding which patients are appropriate for admission at level IV centers and which are better served by transfer to tertiary centers. Urban located centers accounted for 6% of level IV centers and 23% of reviewed trauma patient data. Although the presented study did not assess the appropriateness of admissions at level IV centers, severely injured patients (ISS R 16) were less likely to be admitted at the urban level IV centers. Future research should better examine the

7 role of these urban level IV centers, and their capabilities to provide definitive care for the less severely injured patients. Such examinations could help provide recommendations and minimize overburdening resources at proximal level I and II centers. With no direct source for level IV center data, an alternative approach was used to create the study dataset. Inpatient hospital data and documented diagnoses were used. These data were formatted and mapped into AIS scores and ISS values via an analytical package. The mapping software has been shown accurate and potentially beneficial when used with nontruncated lists of ICD-9-CM codes.33,34 Its application with secondary data placed a significant level of reliance on the accuracy of trauma ICD-9-CM codes. The inpatient administrative data source was an established state data set, which was highly complete, standardized, and had no restrictions on the number of reportable diagnosis codes. Although, an unavoidable limitation of the dataset was no available information on patient vitals or clinical notes, which could have been used to better understand individual encounters.

Conclusions Study results revealed that in a mature inclusive trauma system, a large number of patients were admitted annually to level IV trauma centers. Very few patients were severely injured and mortality appeared low, demonstrating evidence toward the effectiveness of trauma systems. In particular, data showed patients admitted to level IV trauma centers in the study state tended to be elderly and had sustained orthopedic injuries primarily from falls. This has important implications regarding provider education, prevention efforts, and the need for orthopedic surgical capability at level IV centers. At last, study results supported the inclusion of these data in trauma registries for the comprehensive evaluation of trauma systems.

Acknowledgments The authors would like to thank James P. Halsey for his assistance on this project.

References 1. Hill AD, Fowler RA, Nathens AB. Impact of interhospital transfer on outcomes for trauma patients: a systematic review. J Trauma 2011;71: 1885–901. 2. Rivara FP, Koepsell TD, Wang J, et al. Outcomes of trauma patients after transfer to a level 1 trauma center. J Trauma 2008;64:1594–9. 3. Utter GH, Maier RV, Rivara FP, et al. Inclusive trauma systems: do they improve triage or outcomes of the severely injured? J Trauma 2006;60:529–37. 4. Lansink KW, Leenen LP. Do designated trauma systems improve outcome? Curr Opin Crit Care 2007;13:686–90. 5. Celso B, Tepas J, Langland-Orban B, et al. A systematic review and meta-analysis comparing outcome of severely injured patients treated

The American Journal of Surgery, Vol -, No -, - 2016

8

6.

7.

8. 9.

10.

11. 12.

13.

14.

15.

16.

17.

18.

19.

in trauma centers following the establishment of trauma systems. J Trauma 2006;60:371–8. Sampalis JS, Denis R, Fre´chette P, et al. Direct transport to tertiary trauma centers versus transfer from lower level facilities: impact on mortality and morbidity among patients with major trauma. J Trauma 1997;43:288–96. Sorensen MJ, von Recklinghausen FM, Fulton G, et al. Secondary overtriage: the burden of unnecessary interfacility transfers in a rural trauma system. JAMA Surg 2013;148:763–8. Sihler KC, Hansen AR, Torner JC, et al. Characteristics of twicetransferred, rural trauma patients. Prehosp Emerg Care 2002;6:330–5. Nirula R, Maier R, Moore E, et al. Scoop and run to the trauma center or stay and play at the local hospital: hospital transfer’s effect on mortality. J Trauma 2010;69:595–601. Helling TS, Davit F, Edwards K. First echelon hospital care before trauma center transfer in a rural trauma system: does it affect outcome? J Trauma 2010;69:1362–6. McCrum ML, McKee J, Lai M, et al. ATLS adherence in the transfer of rural trauma patients to a level I facility. Injury 2013;44:1241–5. Billeter AT, Miller FB, Harbrecht BG, et al. Interhospital transfer of blunt multiply injured patients to a level 1 trauma center does not adversely affect outcome. Am J Surg 2014;207:459–66. Haas B, Gomez D, Zagorski B, et al. Survival of the fittest: the hidden cost of undertriage of major trauma. J Am Coll Surg 2010;211: 804–11. Thakur NA, Plante MJ, Kayiaros S, et al. Inappropriate transfer of patients with orthopaedic injuries to a level I trauma center: a prospective study. J Orthop Trauma 2010;24:336–9. Crichlow RJ, Zeni A, Reveal G, et al. Appropriateness of patient transfer with associated orthopaedic injuries to a level I trauma center. J Orthop Trauma 2010;24:331–5. Goldfarb CA, Borrelli Jr J, Lu M, et al. A prospective evaluation of patients with isolated orthopedic injuries transferred to a level I trauma center. J Orthop Trauma 2006;20:613–7. Ciesla DJ, Sava JA, Street 3rd JH, et al. Secondary overtriage: a consequence of an immature trauma system. J Am Coll Surg 2008;206: 131–7. Osen HB, Bass RR, Abdullah F, et al. Rapid discharge after transfer: risk factors, incidence, and implications for trauma systems. J Trauma 2010;69:602–6. Garwe T, Cowan LD, Neas B, et al. Survival benefit of transfer to tertiary trauma centers for major trauma patients initially presenting to nontertiary trauma centers. Acad Emerg Med 2010;17:1223–32.

20. Newgard CD, McConnell KJ, Hedges JR, et al. The benefit of higher level of care transfer of injured patients from nontertiary hospital emergency departments. J Trauma 2007;63:965–71. 21. Staudenmayer K, Lin F, Mackersie R, et al. Variability in California triage from 2005 to 2009: a population- based longitudinal study of severely injured patients. J Trauma Acute Care Surg 2014;76:1041–7. 22. Rogers FB, Osler TM, Shackford SR, et al. Study of the outcome of patients transferred to a level 1 hospital after stabilization at an outlying hospital in a rural setting. J Trauma 1999;46:328–33. 23. ACS NTDB National Trauma Data Standard: Data Dictionary 2015 Admissions. Available at: http://www.ntdsdictionary.org/dataElements/ documents/NTDSDataDictionary-2015Admissions_040215.pdf. Accessed April 5, 2016. 24. Glance LG, Osler TM, Mukamel DB, et al. TMPM-ICD9: a trauma mortality prediction model based on ICD-9-CM codes. Ann Surg 2009;249:1032–9. 25. Recommended framework for presenting injury mortality data. MMWR Recomm Rep 1997;46:1–30. 26. Allison P. Missing Data. Thousand Oaks, CA: Sage Publications, Inc.; 2001. 27. Berglund P, Heeringa S. Multiple Imputation of Missing Data Using SAS. Cary, NC: SAS Institute; 2014. 28. Li K, Meng XL, Raghunathan TE, et al. Significance levels from repeated p-values with multiply-imputed data. Statist Sinica 1991;1: 65–92. 29. Smith HL, Chrischilles E, Janus TJ, et al. Clinical indicators of midface fracture in patients with trauma. Dent Traumatol 2013;29: 313–8. 30. Landeen C, Smith HL. Examination of pneumonia risks and risk levels in trauma patients with pulmonary contusion. J Trauma Nurs 2014;21: 41–9. 31. Sasser SM, Hunt RC, Sullivent EE, et al. Guidelines for field triage of injured patients. Recommendations of the National Expert Panel on field triage. MMWR Recomm Rep 2009;58:1–35. 32. American College of Surgeons Committee on Trauma. Resources for Optimal Care of the Injured Patient 2014. 6th ed. Chicago, IL: American College of Surgeons Committee on Trauma; 2014. 33. Clark DE, Turner MO, Hahn DR. ICDPIC Version 3.0. Available at: http://fmwww.bc.edu/repec/bocode/t/trauma.html. Accessed July 15, 2015. 34. Di Bartolomeo S, Tillati S, Valent F, et al. ISS mapped from ICD-9CM by a novel freeware versus traditional coding: a comparative study. Scand J Trauma Resusc Emerg Med 2010;18:17.