- Email: [email protected]
Outcomes after Motor Vehicle Trauma: Transfers to Level I Trauma Centers Compared with Direct Admissions
The Journal of Emergency Medicine, Vol. -, No. -, pp. 1–7, 2017 Ó 2017 Elsevier Inc. All rights reserved. 0736-4679/$ - see front matter
http://dx.doi.org/10.1016/j.jemermed.2017.04.001
Original Contributions
OUTCOMES AFTER MOTOR VEHICLE TRAUMA: TRANSFERS TO LEVEL I TRAUMA CENTERS COMPARED WITH DIRECT ADMISSIONS Aleksandr Rozenberg, MD,* Timothy Danish, MD,† Viktor Y. Dombrovskiy, MD, PHD,‡ and Todd R. Vogel, MD, MPH§ *Department of Radiology, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania, †Department of Surgery, Division of Thoracic Surgery, Vanderbilt University Medical Center, Nashville, Tennessee, ‡Department of Surgery, Rutgers-Robert Wood Johnson Medical School, New Brunswick, New Jersey, and §Department of Surgery, Division of Vascular Surgery, University of Missouri School of Medicine, Columbia, Missouri Reprint Address: Aleksandr Rozenberg, MD, Department of Radiology, Thomas Jefferson University Hospital, 1087 Main Building, 132 S. 10th Street, Philadelphia, PA 19107
, Abstract—Background: The multilevel designation system given to U.S. trauma centers has proven useful in providing injury-level-appropriate care and guiding field triage. Despite the system, patients are often transferred to Level I trauma centers for higher-level care/specialized services. Objectives: The objective of this study is to assess whether there is a difference in outcomes of patients transferred to Level I centers compared with direct admissions. Methods: The Nationwide Inpatient Sample was queried to identify patients involved in motor vehicle accidents, using International Classification of Diseases, Ninth Revision, Clinical Modification E-codes. Patients that were admitted to Level I trauma centers were identified using American College of Surgeons or American Trauma Society designations. Results: There were 343,868 patients that met inclusion criteria. Of these patients, 29.2% (100,297) were admitted to Level I trauma centers, 5.7% (5691) of which were identified as trauma transfers. The lead admitting diagnosis for transfers was pelvic fracture (11.5%). Caucasians were 2.62 times as likely to be transferred as African-Americans (confidence interval 2.32–2.97), and 3.71 times as likely as Hispanics (confidence interval 3.25–4.23). Despite transfer patients having higher adjusted severity scores and higher adjusted risk of mortality, there were no differences in mortality (p = 0.95). Conclusions: Nationally, trauma transfers do not have an increase in mortality when compared with directly admitted patients, despite a higher adjusted severity of illness and
higher adjusted risk of mortality. All rights reserved.
Ó 2017 Elsevier Inc.
, Keywords—motor vehicle accidents; Level I trauma center; Trauma transfer; Direct admission; Outcomes
INTRODUCTION The evolution of trauma systems has created a field of study evaluating their organization and potential benefit to patients, society, and the health care delivery system. Such systems should theoretically offer trauma-centertriaged patients a distinct survival advantage, as well as being cost effective (1). Studies of trauma systems that have a multilevel designation system as well as a field triage algorithm suggest that certain groups of severely injured patients have decreased mortality when transferred to a Level I trauma center (2,3). Despite the fact that triage systems exist and continue to be developed, trauma patients are still transferred regularly from other hospital facilities to high-level centers for a variety of reasons, including injury severity and availability of specialized services (4). Additionally, Level I center transfers seem to be increasing in number; however, current research does not definitively support the decision to
RECEIVED: 11 January 2017; FINAL SUBMISSION RECEIVED: 2 April 2017; ACCEPTED: 5 April 2017 1
2
A. Rozenberg et al.
transfer all severely injured patients to a Level I trauma center (1–4). To create a triage system that is safe and more efficient, identifying outcomes and common injuries of transfer patients are essential. The objective of this study is to assess whether there is a difference in outcomes of patients transferred to Level I centers compared with direct admits and to compare their sociodemographic characteristics, severity of injuries, hospital mortality, and hospital resource utilization. A secondary goal was to identify leading admitting diagnoses for trauma transfers. METHODS Data Source The Nationwide Inpatient Sample (NIS) files for the years 2003 to 2008 were used for this study (5). The NIS is the largest all-payer hospital discharge database developed as part of the Healthcare Cost and Utilization Project, sponsored by the Agency for Health Care Research and Quality. This publicly available database contains data from approximately 8,000,000 yearly hospital stays from more than 1000 hospitals and represents a 20% sample of hospitals in the United States. The NIS sampling and weighting strategies provide researchers with the opportunity to obtain national estimates with variances for the variables of interest. Study Population Patients who were involved in motor vehicle accidents were identified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD9-CM) E-codes E810–E819. These codes represent supplementary classification of external causes of injury and poisoning used to classify environmental events, circumstances, and conditions as the cause of the injury, poisoning, and other adverse effects. These codes are intended to be used in addition to a code from one of the main chapters of ICD-9-CM, indicating the nature of the condition. Only patients that were admitted to Level I trauma centers were selected for this analysis. Among all hospitals in the database, we identified Level I trauma centers as designated by the American Trauma Society and Level I centers accredited by the American College of Surgeons. Level I direct admissions vs. transfer admissions groups were compared using age, gender, race, insurance status, length of hospital stay, total hospital charges, mortality rate, risk of mortality subclasses, and severity of illness subclasses. To define injury severity and mortality risk, the All-Patient Refined DiagnosisRelated Groups (APR-DRGs) developed by 3M Health
Information Systems were used, which are included in the NIS database (6). The APR-DRGs expand the basic DRG structure by adding four subclasses to each DRG. The four sub-classes address patient differences relating to severity of illness and risk of mortality. The four severity of illness subclasses and the four risk of mortality subclasses are numbered sequentially from 1 to 4, indicating minor, moderate, major, or extreme severity of illness or risk of mortality. Statistical Analysis We used SAS 9.2 software (SAS Institute, Cary, NC) for the analysis of the database and all statistics. For categorical variables, we utilized chi-squared analysis with calculated odds ratio and 95% confidence intervals (CI). For continuous variables, a t-test or Wilcoxon rank-sum test (with normal and nonnormal distribution, respectively) were used. A p value of <0.05 was considered statistically significant. RESULTS From the NIS, 343,868 patients were hospitalized after sustaining vehicular trauma (national estimate of 1,719,340 patients). Out of all hospitalized patients, 100,297 (29.2%, national estimate of 501,485 patients) were admitted to Level I trauma centers. Among them, 94,606 patients (94.3%, national estimate of 473,030 patients) were directly admitted to a Level I center and 5691 patients (5.7%, national estimate of 28,455 patients) were transferred from other facilities. We analyzed and compared age, gender, race, and insurance status of patients in both groups, as shown in Table 1. Although transferred patients were slightly older than directly admitted patients (38.4 6 20.6 years vs. 37.7 6 19.7 years, respectively; p = 0.017), comparison of these groups by decade did not demonstrate statistically significant differences in age. Among patients who were directly admitted to the hospital, 63.6% were male and 36.4% were female. In the group of transferred patients, there were 62.8% males and 37.2% females (p = 0.236). Caucasians were 2.62 times more likely to be transferred, compared with African-Americans (95% CI 2.32–2.97), and 3.71 times more likely than Hispanic patients (95% CI 3.25–4.23). African-American patients were 1.41 times more likely to be transferred, compared with Hispanic patients (95% CI 1.19–1.68). In the transfer group, the lead admitting diagnoses were pelvic or acetabular fractures (11.5% of transfers), followed by femoral fractures (7.8%), intracranial hemorrhage (7.4%), and tibial or fibular fractures (6.7%). Table 2 lists the lead admitting diagnoses in the transfer group by decreasing order of frequency.
Outcomes after Motor Vehicle Trauma
3
Table 1. Demographic Characteristics of Patient Transferred and Admitted to Level I Trauma Centers Transfers
Direct Admits
p Value
38.4 6 20.6 36.4 (23.3–54.6)
37.7 6 19.7 36.1 (22.9–53.8)
0.017 NS
3576 (62.8%) 2115 (37.2%)
60,429 (63.6%) 34,177 (36.4%)
NS NS
3194 (56.1%) 287 (5.0%) 249 (4.4%)
42,238 (44.6%) 9954 (10.5%) 12,211 (12.9%)
<0.0001 <0.0001 <0.0001
339 (5.9%) 717 (13.5%) 3376 (59.3%) 672 (11.8%)
5013 (5.3%) 12,123 (12.8%) 57,004 (60.3%) 11,226 (11.9%)
NS NS NS NS
Demographics Age Average Median (interquartile range) Sex Male Female Race Caucasian African-American Hispanic Insurance Medicare Medicaid Private insurance Self-pay
To compare the severity of patients’ conditions we used two measures that were available in the NIS database. These measures were the severity of illness that evaluated loss of function, and the risk of mortality that evaluated the likelihood of dying. According to the magnitude of these measures, they were categorized into the following four subclasses: minor, moderate, major, and extreme. For example, a patient with a pelvic fracture as the only diagnosis is considered to have a high severity of illness but a lower risk of mortality. If additional diagnoses are present along with the pelvic fracture, patient severity of illness and risk of mortality may increase. If hypotension is present along with the pelvic fracture, the patient is considered to have an extreme severity of illness and a major risk of mortality. Because severity of illness and risk of mortality are distinct attributes, separate subclasses are assigned to each category.
Proportions of patients with minor and moderate loss of function among the transferred patients (18.0% and 35.4%, respectively) were smaller than among directly admitted patients (28.5%, p < 0.0002 and 37.5%, p = 0.002), whereas proportions of those with moderate and extreme loss of functions were significantly greater (30.4% vs. 22.3% and 16.3% vs. 11.7%, p < 0.0002 for both). Table 3 shows the comparison of the severity of illness between the two groups. Similarly, the subclass with a minor likelihood of dying in the transferred group (59.2%) was smaller than in the directly admitted group (68.8%, p < 0.0002), whereas subclasses with higher likelihood of dying were significantly greater (moderate: 18.9% vs. 15.3%; major: 13.5% vs. 9.6%; extreme: 8.4% vs. 6.3%; p < 0.0002 for all comparisons). Table 4 shows the comparison of mortality risk between the two groups. Nevertheless, hospital mortality rates among transfers (3.6%)
Table 2. Most Common Admitting Diagnosis in Transferred Patients Admitting Diagnosis
Number of Patients
National Estimate of Patients
Percent of Total Transferred
Pelvic or acetabular fracture Femoral fracture Intracranial hemorrhage Tibial or fibular fracture Cervical spine fracture(s) Lumbar spine fracture(s) Thoracic spine fracture(s) Pneumothorax Head injury with coma Skull base fracture Splenic laceration Ankle fracture Lung contusion Liver injury or laceration Thoracic aorta injury Maxillofacial fracture Sacrum or coccyx fracture Total
653 443 423 384 362 303 291 210 159 126 102 99 98 92 69 50 47 3911
3265 2215 2115 1920 1810 1515 1455 1050 795 630 510 495 490 460 345 250 235 19,555
11.5% 7.8% 7.4% 6.7% 6.3% 5.3% 5.1% 3.7% 2.8% 2.2% 1.8% 1.7% 1.7% 1.6% 1.2% 0.9% 0.8% 68.5%
4
A. Rozenberg et al.
Table 3. Comparison of Severity of Illness that Evaluates Loss of Function (using All Patient Refined Diagnosis-Related Groups) Between Transferred and Directly Admitted Patients Severity of Illness
Transfers
Direct Admits
p Value
I – Minor loss of function II – Moderate loss of function III – Major loss of function IV – Extreme loss of function
18.0% 35.4% 30.4% 16.3%
28.5% 37.5% 22.3% 11.7%
<0.0002 0.002 <0.0002 <0.0002
and direct admits (3.6%) were not statistically significant (p = 0.95). We compared hospital resource utilization measures such as length of stay and total hospital charges between transferred and directly admitted patients. Because both these measures are highly skewed to the right, we calculated their medians for comparisons. The median length of stay for transferred patients (5 days) was significantly greater than for directly admitted (3 days, p < 0.0001). Accordingly, median for total hospital charges for transfers ($35,287) was also greater than for directly admitted patients ($28,171; p < 0.0001). DISCUSSION Motor vehicle trauma represents one of the largest contributors to trauma center admissions, particularly to Level I centers (7). Trauma transfers seem to be strongly weighted toward blunt trauma, including head injury and orthopedic injury, which are commonly seen in vehicular trauma (4,8). This analysis has demonstrated that motor vehicle trauma patients represent a large number of Level I trauma admissions nationally, and nearly 6% of these admissions are trauma transfers. This group of patients subsequently had high hospital utilization with increased length of stay and total hospital charges. Despite higher adjusted risk severity and adjusted risk of mortality, trauma transfers had no difference in overall mortality, as compared with direct admissions. Although protocols are known to be widely variable nationally, Emergency Medical Services professionals and paramedics who transport patients from motor Table 4. Comparison of Risk of Mortality (using All Patient Refined Diagnosis Related Groups) Between Transferred and Directly Admitted Patients Risk of Mortality
Transfers
Direct Admits
p Value
I – Minor risk of mortality II – Moderate risk of mortality III – Major risk of mortality IV – Extreme risk of mortality
59.2% 18.9% 13.5% 8.4%
68.8% 15.3% 9.6% 6.3%
<0.0002 <0.0002 <0.0002 <0.0002
vehicle accident scenes in the field use data-driven triage algorithms to determine if an injured patient should go to a trauma center (9,10). Prior studies have demonstrated that a degree of under-triage exists in the field, and this could be a contributing factor for the need of trauma patients to be subsequently transferred to higher-level centers (11). Most Level I admissions were direct in our analysis as well as others, yet still a small proportion of these admissions related to transfers (12). Reasons for transfers vary, but high injury severity, need for specialized surgical services, and experienced intensive care unit staffing are cited (13). Trauma transfer patients in our study had primary ICD-9-CM code diagnoses that were strongly weighted toward neurologic injuries and orthopedic injuries of the pelvis, femur, and spine. Similar studies have demonstrated that orthopedic injuries are common in transfers, citing the need for specialized surgical services, which are often unavailable at the transferring facility (14). A similar pattern emerges with respect to head injury, as neurosurgery services are not always available at non-Level I centers (15,16). In accordance with the literature, our analysis found that Caucasians were two to three times as likely to be transferred, compared with African-Americans or Hispanics (17–19). Additionally, African-Americans were approximately 1.5 times as likely to be transferred as Hispanics. For patients that are involved in motor vehicle accidents in urban areas, the Level I trauma center may be the closest hospital, which may provide some of the explanation for the disparities in transfers. For example, if minority patients are more likely to live and drive in urban areas served by Level I trauma centers, then they will be more likely to be directly admitted to these centers—and if nonminorities are more likely to live outside of urban areas, they may be more likely to be taken initially to other hospitals and then require eventual transfer to a Level I trauma center. Using NIS-provided All-Patient Refined DRGs data points, we found that patients in the transfer group had greater severity of illness and risk of mortality, without actual mortality rate increase (6). Explanations for the difference in severity of illness has been discussed in the literature, and includes delay of care in transfer, worsening or exacerbation of injuries during transfer, and selection of the sickest patients who require more advanced care (20–24). Of note, even though patients who were transferred had significantly higher adjusted risk of mortality and severity of illness scores, their mortality was not increased in our study, suggesting that transfer to a Level I center may be of benefit in this group. We have demonstrated that on a national level, Level I vehicular trauma transfers do not have an increase in mortality when compared with directly admitted patients. Patients who were transferred had significantly higher
Outcomes after Motor Vehicle Trauma
adjusted risk of mortality and severity of illness scores; however, their mortality was not increased in our study, suggesting that transfer to a Level I center may be of benefit to patients in this group. Similarly, multiple smaller studies have shown no difference in outcomes between directly admitted trauma patients and those that were transferred after initial stabilization at an outside facility (25–27). Additionally, our analysis demonstrates that Caucasians are two to three times as likely to be transferred as African-Americans and three to four times as likely as Hispanics. Although the reasons are unclear, the findings could be explained by the greater number of Level I trauma centers in urban areas with larger minority populations. Limitations We acknowledge several limitations of this retrospective study. The administrative data were originally intended for billing purposes and carry the innate limitation of billing data. The potential for inclusion bias is based on limited coding schemes. Additionally, due to the large number of hospitals reporting the data, and the variability of the individual coders entering the data, there is a potential for coding errors. Given the possible biases, it is often found that coding of multiple injuries may be underreported or under-coded (28). As well, the NIS is limited by coding and cannot offer the specific clinical detail of a single-center series or trial. It is also important to note that when comparing the two groups, patients too unstable to transfer are less likely to be transferred, and stable patients needing specialty surgeries are more likely to be transferred. Additionally, the NIS database does not provide information about the distance of transfer between the initial hospital and receiving Level I center, which may help explain the decision to not transport certain patients. Though our analysis did not control for injury severity with a specific trauma marker, future studies should be stratified according to specific trauma injury severity markers such as Trauma Mortality Prediction Model (29). CONCLUSION A significant proportion of transferred patients experienced traumatic injuries of the brain, pelvis, femur, and spine. The transferred patients had higher severity of illness and risk of mortality, but had no significant difference in mortality compared with direct admits. Future programs should examine if the commonly transferred injuries benefit from direct admission to Level I centers vs. transfer, focusing on mortality and efficient resource utilization.
5
REFERENCES 1. Durham R, Pracht E, Orban B, Lottenburg L, Tepas J, Flint L. Evaluation of a mature trauma system. Ann Surg 2006;243:775–783. discussion 783–5. 2. 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;63:965–971. 3. Nirula R, Maier R, Moore E, Sperry J, Gentilello L. Scoop and run to the trauma center or stay and play at the local hospital: hospital transfer’s effect on mortality. J Trauma 2010;69:595–599. discussion 599–601. 4. 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;61:1380–1386. discussion 1386–8. 5. HCUP Databases. Healthcare Cost and Utilization Project (HCUP). 2003-2008. Agency for Healthcare Research and Quality, Rockville, MD. Available at: www.hcup-us.ahrq.gov/databases.jsp. 6. Averill R, Goldfield N, Hughes JS, et al. 3M APR DRG Classification System. 31st ed. Wallingford, CT: Agency for Healthcare Research and Quality; 2013. Available at: http://www.hcupus. ahrq.gov/db/nation/nis/grp031_aprdrg_meth_ovrview.pdf. Accessed November 15, 2016. 7. Ovadia P, Szewczyk D, Walker K, Abdullah F, Schmidt-Gillespie S, Rabinovici R. Admission patterns of an urban level I trauma center. Am J Med Qual 2000;15:9–15. 8. Mann NC, Hedges JR, Mullins RJ, Helfand M, Worrall W, Zechnich AD. Rural hospital transfer patterns before and after implementation of a statewide trauma system. OHSU Rural Trauma Research Group. Acad Emerg Med 1997;4:764–-771. 9. 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(RR-1):1–35. 10. Lerner EB, Shah MN, Swor RA, et al. Comparison of the 1999 and 2006 trauma triage guidelines: where do patients go? Prehosp Emerg Care 2011;15:12–17. 11. Haas B, Gomez D, Zagorski B, Stukel TA, Rubenfeld GD, Nathens AB. Survival of the fittest: the hidden cost of undertriage of major trauma. J Am Coll Surg 2010;211:804–811. 12. Sampalis JS, Denis R, Frechette P, Brown R, Fleiszer D, Mulder D. 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; 43:288–295. discussion 295–6. 13. Ball CG, Sutherland FR, Dixon E, et al. Surgical trauma referrals from rural level III hospitals: should our community colleagues be doing more, or less? J Trauma 2009;67:180–184. 14. 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;24:331–335. 15. Gray A, Bush S, Whiteley S. Secondary transport of the critically ill and injured adult. Emerg Med J 2004;21:281–285. 16. Hartl R, Gerber LM, Iacono L, Ni Q, Lyons K, Ghajar J. Direct transport within an organized state trauma system reduces mortality in patients with severe traumatic brain injury. J Trauma 2006;60: 1250–1256. discussion 1256. 17. Haider AH, Chang DC, Efron DT, Haut ER, Crandall M, Cornwell EE 3rd. Race and insurance status as risk factors for trauma mortality. Arch Surg 2008;143:945–949. 18. 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: 336–339. 19. Alban RF, Berry C, Ley E, et al. Does health care insurance affect outcomes after traumatic brain injury? Analysis of the National Trauma Databank. Am Surg 2010;76:1108–1111.
6 20. Harrington DT, Connolly M, Biffl WL, Majercik SD, Cioffi WG. Transfer times to definitive care facilities are too long: a consequence of an immature trauma system. Ann Surg 2005;241:961– 966. discussion 966–8. 21. Meisler R, Thomsen AB, Abildstrom H, et al. Triage and mortality in 2875 consecutive trauma patients. Acta Anaesthesiol Scand 2010;54:218–223. 22. Utter GH, Victorino GP, Wisner DH. Interhospital transfer occurs more slowly for elderly acute trauma patients. J Emerg Med 2008;35:415–420. 23. Young JS, Bassam D, Cephas GA, Brady WJ, Butler K, Pomphrey M. Interhospital versus direct scene transfer of major trauma patients in a rural trauma system. Am Surg 1998;64:88– 91. discussion 91–2. 24. Dunn LT. Secondary insults during the interhospital transfer of head-injured patients: an audit of transfers in the Mersey Region. Injury 1997;28:427–431.
A. Rozenberg et al. 25. 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;69:1362–1366. 26. Nathens AB, Maier RV, Brundage SI, Jurkovich GJ, Grossman DC. The effect of interfacility transfer on outcome in an urban trauma system. J Trauma 2003;55:444–449. 27. Rivara FP, Koepsell TD, Wang J, Nathens A, Jurkovich GA, Mackenzie EJ. Outcomes of trauma patients after transfer to a level I trauma center. J Trauma 2008;64:1594–1599. 28. Hemmila MR, Jakubus JL, Wahl WL, et al. Detecting the blind spot: complications in the trauma registry and trauma quality improvement. Surgery 2007;142:439–448. discussion 448–9. 29. Glance LG, Osler TM, Mukamel DB, Meredith W, Wagner J, Dick AW. TMPM-ICD9: a trauma mortality prediction model based on ICD-9-CM codes. Ann Surg 2009; 249:1032–1039.
Outcomes after Motor Vehicle Trauma
ARTICLE SUMMARY 1. Why is this topic important? Despite the fact that triage systems exist and continue to be developed, trauma patients are still transferred regularly from other hospital facilities to high-level centers for a variety of reasons, including injury severity and availability of specialized services. Additionally, hospital-toLevel-I-center transfers seem to be increasing in number; however, few national studies exist in identifying patients that are transferred to Level I centers. 2. What does this study attempt to show? We have attempted to demonstrate that on a national level, Level I vehicular trauma transfers do not have an increase in mortality when compared with directly admitted patients, even though transferred patients had greater severity of illness. As expected, the data confirm that transferred patients have increased hospital charges and increased health care utilization. 3. What are the key findings? Despite transfer patients having higher adjusted riskseverity scores and higher adjusted risk of mortality, there was no statistical difference in mortality between these groups. Additionally, hospitalizations due to transfer to a Level I trauma center had higher hospital resource utilization, incurring greater length of stay and total hospital charges. 4. How is patient care impacted? Patients who were transferred had significantly higher adjusted risk of mortality and severity of illness scores; however, their mortality was not increased in our study, suggesting that transfer to a Level I center may be of benefit to certain patients in this group. We believe our data provide the groundwork for further studies to examine if specific patients or injuries may benefit from direct admission to Level I trauma centers.
7
http://dx.doi.org/10.1016/j.jemermed.2017.04.001
Original Contributions
OUTCOMES AFTER MOTOR VEHICLE TRAUMA: TRANSFERS TO LEVEL I TRAUMA CENTERS COMPARED WITH DIRECT ADMISSIONS Aleksandr Rozenberg, MD,* Timothy Danish, MD,† Viktor Y. Dombrovskiy, MD, PHD,‡ and Todd R. Vogel, MD, MPH§ *Department of Radiology, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania, †Department of Surgery, Division of Thoracic Surgery, Vanderbilt University Medical Center, Nashville, Tennessee, ‡Department of Surgery, Rutgers-Robert Wood Johnson Medical School, New Brunswick, New Jersey, and §Department of Surgery, Division of Vascular Surgery, University of Missouri School of Medicine, Columbia, Missouri Reprint Address: Aleksandr Rozenberg, MD, Department of Radiology, Thomas Jefferson University Hospital, 1087 Main Building, 132 S. 10th Street, Philadelphia, PA 19107
, Abstract—Background: The multilevel designation system given to U.S. trauma centers has proven useful in providing injury-level-appropriate care and guiding field triage. Despite the system, patients are often transferred to Level I trauma centers for higher-level care/specialized services. Objectives: The objective of this study is to assess whether there is a difference in outcomes of patients transferred to Level I centers compared with direct admissions. Methods: The Nationwide Inpatient Sample was queried to identify patients involved in motor vehicle accidents, using International Classification of Diseases, Ninth Revision, Clinical Modification E-codes. Patients that were admitted to Level I trauma centers were identified using American College of Surgeons or American Trauma Society designations. Results: There were 343,868 patients that met inclusion criteria. Of these patients, 29.2% (100,297) were admitted to Level I trauma centers, 5.7% (5691) of which were identified as trauma transfers. The lead admitting diagnosis for transfers was pelvic fracture (11.5%). Caucasians were 2.62 times as likely to be transferred as African-Americans (confidence interval 2.32–2.97), and 3.71 times as likely as Hispanics (confidence interval 3.25–4.23). Despite transfer patients having higher adjusted severity scores and higher adjusted risk of mortality, there were no differences in mortality (p = 0.95). Conclusions: Nationally, trauma transfers do not have an increase in mortality when compared with directly admitted patients, despite a higher adjusted severity of illness and
higher adjusted risk of mortality. All rights reserved.
Ó 2017 Elsevier Inc.
, Keywords—motor vehicle accidents; Level I trauma center; Trauma transfer; Direct admission; Outcomes
INTRODUCTION The evolution of trauma systems has created a field of study evaluating their organization and potential benefit to patients, society, and the health care delivery system. Such systems should theoretically offer trauma-centertriaged patients a distinct survival advantage, as well as being cost effective (1). Studies of trauma systems that have a multilevel designation system as well as a field triage algorithm suggest that certain groups of severely injured patients have decreased mortality when transferred to a Level I trauma center (2,3). Despite the fact that triage systems exist and continue to be developed, trauma patients are still transferred regularly from other hospital facilities to high-level centers for a variety of reasons, including injury severity and availability of specialized services (4). Additionally, Level I center transfers seem to be increasing in number; however, current research does not definitively support the decision to
RECEIVED: 11 January 2017; FINAL SUBMISSION RECEIVED: 2 April 2017; ACCEPTED: 5 April 2017 1
2
A. Rozenberg et al.
transfer all severely injured patients to a Level I trauma center (1–4). To create a triage system that is safe and more efficient, identifying outcomes and common injuries of transfer patients are essential. The objective of this study is to assess whether there is a difference in outcomes of patients transferred to Level I centers compared with direct admits and to compare their sociodemographic characteristics, severity of injuries, hospital mortality, and hospital resource utilization. A secondary goal was to identify leading admitting diagnoses for trauma transfers. METHODS Data Source The Nationwide Inpatient Sample (NIS) files for the years 2003 to 2008 were used for this study (5). The NIS is the largest all-payer hospital discharge database developed as part of the Healthcare Cost and Utilization Project, sponsored by the Agency for Health Care Research and Quality. This publicly available database contains data from approximately 8,000,000 yearly hospital stays from more than 1000 hospitals and represents a 20% sample of hospitals in the United States. The NIS sampling and weighting strategies provide researchers with the opportunity to obtain national estimates with variances for the variables of interest. Study Population Patients who were involved in motor vehicle accidents were identified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD9-CM) E-codes E810–E819. These codes represent supplementary classification of external causes of injury and poisoning used to classify environmental events, circumstances, and conditions as the cause of the injury, poisoning, and other adverse effects. These codes are intended to be used in addition to a code from one of the main chapters of ICD-9-CM, indicating the nature of the condition. Only patients that were admitted to Level I trauma centers were selected for this analysis. Among all hospitals in the database, we identified Level I trauma centers as designated by the American Trauma Society and Level I centers accredited by the American College of Surgeons. Level I direct admissions vs. transfer admissions groups were compared using age, gender, race, insurance status, length of hospital stay, total hospital charges, mortality rate, risk of mortality subclasses, and severity of illness subclasses. To define injury severity and mortality risk, the All-Patient Refined DiagnosisRelated Groups (APR-DRGs) developed by 3M Health
Information Systems were used, which are included in the NIS database (6). The APR-DRGs expand the basic DRG structure by adding four subclasses to each DRG. The four sub-classes address patient differences relating to severity of illness and risk of mortality. The four severity of illness subclasses and the four risk of mortality subclasses are numbered sequentially from 1 to 4, indicating minor, moderate, major, or extreme severity of illness or risk of mortality. Statistical Analysis We used SAS 9.2 software (SAS Institute, Cary, NC) for the analysis of the database and all statistics. For categorical variables, we utilized chi-squared analysis with calculated odds ratio and 95% confidence intervals (CI). For continuous variables, a t-test or Wilcoxon rank-sum test (with normal and nonnormal distribution, respectively) were used. A p value of <0.05 was considered statistically significant. RESULTS From the NIS, 343,868 patients were hospitalized after sustaining vehicular trauma (national estimate of 1,719,340 patients). Out of all hospitalized patients, 100,297 (29.2%, national estimate of 501,485 patients) were admitted to Level I trauma centers. Among them, 94,606 patients (94.3%, national estimate of 473,030 patients) were directly admitted to a Level I center and 5691 patients (5.7%, national estimate of 28,455 patients) were transferred from other facilities. We analyzed and compared age, gender, race, and insurance status of patients in both groups, as shown in Table 1. Although transferred patients were slightly older than directly admitted patients (38.4 6 20.6 years vs. 37.7 6 19.7 years, respectively; p = 0.017), comparison of these groups by decade did not demonstrate statistically significant differences in age. Among patients who were directly admitted to the hospital, 63.6% were male and 36.4% were female. In the group of transferred patients, there were 62.8% males and 37.2% females (p = 0.236). Caucasians were 2.62 times more likely to be transferred, compared with African-Americans (95% CI 2.32–2.97), and 3.71 times more likely than Hispanic patients (95% CI 3.25–4.23). African-American patients were 1.41 times more likely to be transferred, compared with Hispanic patients (95% CI 1.19–1.68). In the transfer group, the lead admitting diagnoses were pelvic or acetabular fractures (11.5% of transfers), followed by femoral fractures (7.8%), intracranial hemorrhage (7.4%), and tibial or fibular fractures (6.7%). Table 2 lists the lead admitting diagnoses in the transfer group by decreasing order of frequency.
Outcomes after Motor Vehicle Trauma
3
Table 1. Demographic Characteristics of Patient Transferred and Admitted to Level I Trauma Centers Transfers
Direct Admits
p Value
38.4 6 20.6 36.4 (23.3–54.6)
37.7 6 19.7 36.1 (22.9–53.8)
0.017 NS
3576 (62.8%) 2115 (37.2%)
60,429 (63.6%) 34,177 (36.4%)
NS NS
3194 (56.1%) 287 (5.0%) 249 (4.4%)
42,238 (44.6%) 9954 (10.5%) 12,211 (12.9%)
<0.0001 <0.0001 <0.0001
339 (5.9%) 717 (13.5%) 3376 (59.3%) 672 (11.8%)
5013 (5.3%) 12,123 (12.8%) 57,004 (60.3%) 11,226 (11.9%)
NS NS NS NS
Demographics Age Average Median (interquartile range) Sex Male Female Race Caucasian African-American Hispanic Insurance Medicare Medicaid Private insurance Self-pay
To compare the severity of patients’ conditions we used two measures that were available in the NIS database. These measures were the severity of illness that evaluated loss of function, and the risk of mortality that evaluated the likelihood of dying. According to the magnitude of these measures, they were categorized into the following four subclasses: minor, moderate, major, and extreme. For example, a patient with a pelvic fracture as the only diagnosis is considered to have a high severity of illness but a lower risk of mortality. If additional diagnoses are present along with the pelvic fracture, patient severity of illness and risk of mortality may increase. If hypotension is present along with the pelvic fracture, the patient is considered to have an extreme severity of illness and a major risk of mortality. Because severity of illness and risk of mortality are distinct attributes, separate subclasses are assigned to each category.
Proportions of patients with minor and moderate loss of function among the transferred patients (18.0% and 35.4%, respectively) were smaller than among directly admitted patients (28.5%, p < 0.0002 and 37.5%, p = 0.002), whereas proportions of those with moderate and extreme loss of functions were significantly greater (30.4% vs. 22.3% and 16.3% vs. 11.7%, p < 0.0002 for both). Table 3 shows the comparison of the severity of illness between the two groups. Similarly, the subclass with a minor likelihood of dying in the transferred group (59.2%) was smaller than in the directly admitted group (68.8%, p < 0.0002), whereas subclasses with higher likelihood of dying were significantly greater (moderate: 18.9% vs. 15.3%; major: 13.5% vs. 9.6%; extreme: 8.4% vs. 6.3%; p < 0.0002 for all comparisons). Table 4 shows the comparison of mortality risk between the two groups. Nevertheless, hospital mortality rates among transfers (3.6%)
Table 2. Most Common Admitting Diagnosis in Transferred Patients Admitting Diagnosis
Number of Patients
National Estimate of Patients
Percent of Total Transferred
Pelvic or acetabular fracture Femoral fracture Intracranial hemorrhage Tibial or fibular fracture Cervical spine fracture(s) Lumbar spine fracture(s) Thoracic spine fracture(s) Pneumothorax Head injury with coma Skull base fracture Splenic laceration Ankle fracture Lung contusion Liver injury or laceration Thoracic aorta injury Maxillofacial fracture Sacrum or coccyx fracture Total
653 443 423 384 362 303 291 210 159 126 102 99 98 92 69 50 47 3911
3265 2215 2115 1920 1810 1515 1455 1050 795 630 510 495 490 460 345 250 235 19,555
11.5% 7.8% 7.4% 6.7% 6.3% 5.3% 5.1% 3.7% 2.8% 2.2% 1.8% 1.7% 1.7% 1.6% 1.2% 0.9% 0.8% 68.5%
4
A. Rozenberg et al.
Table 3. Comparison of Severity of Illness that Evaluates Loss of Function (using All Patient Refined Diagnosis-Related Groups) Between Transferred and Directly Admitted Patients Severity of Illness
Transfers
Direct Admits
p Value
I – Minor loss of function II – Moderate loss of function III – Major loss of function IV – Extreme loss of function
18.0% 35.4% 30.4% 16.3%
28.5% 37.5% 22.3% 11.7%
<0.0002 0.002 <0.0002 <0.0002
and direct admits (3.6%) were not statistically significant (p = 0.95). We compared hospital resource utilization measures such as length of stay and total hospital charges between transferred and directly admitted patients. Because both these measures are highly skewed to the right, we calculated their medians for comparisons. The median length of stay for transferred patients (5 days) was significantly greater than for directly admitted (3 days, p < 0.0001). Accordingly, median for total hospital charges for transfers ($35,287) was also greater than for directly admitted patients ($28,171; p < 0.0001). DISCUSSION Motor vehicle trauma represents one of the largest contributors to trauma center admissions, particularly to Level I centers (7). Trauma transfers seem to be strongly weighted toward blunt trauma, including head injury and orthopedic injury, which are commonly seen in vehicular trauma (4,8). This analysis has demonstrated that motor vehicle trauma patients represent a large number of Level I trauma admissions nationally, and nearly 6% of these admissions are trauma transfers. This group of patients subsequently had high hospital utilization with increased length of stay and total hospital charges. Despite higher adjusted risk severity and adjusted risk of mortality, trauma transfers had no difference in overall mortality, as compared with direct admissions. Although protocols are known to be widely variable nationally, Emergency Medical Services professionals and paramedics who transport patients from motor Table 4. Comparison of Risk of Mortality (using All Patient Refined Diagnosis Related Groups) Between Transferred and Directly Admitted Patients Risk of Mortality
Transfers
Direct Admits
p Value
I – Minor risk of mortality II – Moderate risk of mortality III – Major risk of mortality IV – Extreme risk of mortality
59.2% 18.9% 13.5% 8.4%
68.8% 15.3% 9.6% 6.3%
<0.0002 <0.0002 <0.0002 <0.0002
vehicle accident scenes in the field use data-driven triage algorithms to determine if an injured patient should go to a trauma center (9,10). Prior studies have demonstrated that a degree of under-triage exists in the field, and this could be a contributing factor for the need of trauma patients to be subsequently transferred to higher-level centers (11). Most Level I admissions were direct in our analysis as well as others, yet still a small proportion of these admissions related to transfers (12). Reasons for transfers vary, but high injury severity, need for specialized surgical services, and experienced intensive care unit staffing are cited (13). Trauma transfer patients in our study had primary ICD-9-CM code diagnoses that were strongly weighted toward neurologic injuries and orthopedic injuries of the pelvis, femur, and spine. Similar studies have demonstrated that orthopedic injuries are common in transfers, citing the need for specialized surgical services, which are often unavailable at the transferring facility (14). A similar pattern emerges with respect to head injury, as neurosurgery services are not always available at non-Level I centers (15,16). In accordance with the literature, our analysis found that Caucasians were two to three times as likely to be transferred, compared with African-Americans or Hispanics (17–19). Additionally, African-Americans were approximately 1.5 times as likely to be transferred as Hispanics. For patients that are involved in motor vehicle accidents in urban areas, the Level I trauma center may be the closest hospital, which may provide some of the explanation for the disparities in transfers. For example, if minority patients are more likely to live and drive in urban areas served by Level I trauma centers, then they will be more likely to be directly admitted to these centers—and if nonminorities are more likely to live outside of urban areas, they may be more likely to be taken initially to other hospitals and then require eventual transfer to a Level I trauma center. Using NIS-provided All-Patient Refined DRGs data points, we found that patients in the transfer group had greater severity of illness and risk of mortality, without actual mortality rate increase (6). Explanations for the difference in severity of illness has been discussed in the literature, and includes delay of care in transfer, worsening or exacerbation of injuries during transfer, and selection of the sickest patients who require more advanced care (20–24). Of note, even though patients who were transferred had significantly higher adjusted risk of mortality and severity of illness scores, their mortality was not increased in our study, suggesting that transfer to a Level I center may be of benefit in this group. We have demonstrated that on a national level, Level I vehicular trauma transfers do not have an increase in mortality when compared with directly admitted patients. Patients who were transferred had significantly higher
Outcomes after Motor Vehicle Trauma
adjusted risk of mortality and severity of illness scores; however, their mortality was not increased in our study, suggesting that transfer to a Level I center may be of benefit to patients in this group. Similarly, multiple smaller studies have shown no difference in outcomes between directly admitted trauma patients and those that were transferred after initial stabilization at an outside facility (25–27). Additionally, our analysis demonstrates that Caucasians are two to three times as likely to be transferred as African-Americans and three to four times as likely as Hispanics. Although the reasons are unclear, the findings could be explained by the greater number of Level I trauma centers in urban areas with larger minority populations. Limitations We acknowledge several limitations of this retrospective study. The administrative data were originally intended for billing purposes and carry the innate limitation of billing data. The potential for inclusion bias is based on limited coding schemes. Additionally, due to the large number of hospitals reporting the data, and the variability of the individual coders entering the data, there is a potential for coding errors. Given the possible biases, it is often found that coding of multiple injuries may be underreported or under-coded (28). As well, the NIS is limited by coding and cannot offer the specific clinical detail of a single-center series or trial. It is also important to note that when comparing the two groups, patients too unstable to transfer are less likely to be transferred, and stable patients needing specialty surgeries are more likely to be transferred. Additionally, the NIS database does not provide information about the distance of transfer between the initial hospital and receiving Level I center, which may help explain the decision to not transport certain patients. Though our analysis did not control for injury severity with a specific trauma marker, future studies should be stratified according to specific trauma injury severity markers such as Trauma Mortality Prediction Model (29). CONCLUSION A significant proportion of transferred patients experienced traumatic injuries of the brain, pelvis, femur, and spine. The transferred patients had higher severity of illness and risk of mortality, but had no significant difference in mortality compared with direct admits. Future programs should examine if the commonly transferred injuries benefit from direct admission to Level I centers vs. transfer, focusing on mortality and efficient resource utilization.
5
REFERENCES 1. Durham R, Pracht E, Orban B, Lottenburg L, Tepas J, Flint L. Evaluation of a mature trauma system. Ann Surg 2006;243:775–783. discussion 783–5. 2. 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;63:965–971. 3. Nirula R, Maier R, Moore E, Sperry J, Gentilello L. Scoop and run to the trauma center or stay and play at the local hospital: hospital transfer’s effect on mortality. J Trauma 2010;69:595–599. discussion 599–601. 4. 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;61:1380–1386. discussion 1386–8. 5. HCUP Databases. Healthcare Cost and Utilization Project (HCUP). 2003-2008. Agency for Healthcare Research and Quality, Rockville, MD. Available at: www.hcup-us.ahrq.gov/databases.jsp. 6. Averill R, Goldfield N, Hughes JS, et al. 3M APR DRG Classification System. 31st ed. Wallingford, CT: Agency for Healthcare Research and Quality; 2013. Available at: http://www.hcupus. ahrq.gov/db/nation/nis/grp031_aprdrg_meth_ovrview.pdf. Accessed November 15, 2016. 7. Ovadia P, Szewczyk D, Walker K, Abdullah F, Schmidt-Gillespie S, Rabinovici R. Admission patterns of an urban level I trauma center. Am J Med Qual 2000;15:9–15. 8. Mann NC, Hedges JR, Mullins RJ, Helfand M, Worrall W, Zechnich AD. Rural hospital transfer patterns before and after implementation of a statewide trauma system. OHSU Rural Trauma Research Group. Acad Emerg Med 1997;4:764–-771. 9. 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(RR-1):1–35. 10. Lerner EB, Shah MN, Swor RA, et al. Comparison of the 1999 and 2006 trauma triage guidelines: where do patients go? Prehosp Emerg Care 2011;15:12–17. 11. Haas B, Gomez D, Zagorski B, Stukel TA, Rubenfeld GD, Nathens AB. Survival of the fittest: the hidden cost of undertriage of major trauma. J Am Coll Surg 2010;211:804–811. 12. Sampalis JS, Denis R, Frechette P, Brown R, Fleiszer D, Mulder D. 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; 43:288–295. discussion 295–6. 13. Ball CG, Sutherland FR, Dixon E, et al. Surgical trauma referrals from rural level III hospitals: should our community colleagues be doing more, or less? J Trauma 2009;67:180–184. 14. 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;24:331–335. 15. Gray A, Bush S, Whiteley S. Secondary transport of the critically ill and injured adult. Emerg Med J 2004;21:281–285. 16. Hartl R, Gerber LM, Iacono L, Ni Q, Lyons K, Ghajar J. Direct transport within an organized state trauma system reduces mortality in patients with severe traumatic brain injury. J Trauma 2006;60: 1250–1256. discussion 1256. 17. Haider AH, Chang DC, Efron DT, Haut ER, Crandall M, Cornwell EE 3rd. Race and insurance status as risk factors for trauma mortality. Arch Surg 2008;143:945–949. 18. 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: 336–339. 19. Alban RF, Berry C, Ley E, et al. Does health care insurance affect outcomes after traumatic brain injury? Analysis of the National Trauma Databank. Am Surg 2010;76:1108–1111.
6 20. Harrington DT, Connolly M, Biffl WL, Majercik SD, Cioffi WG. Transfer times to definitive care facilities are too long: a consequence of an immature trauma system. Ann Surg 2005;241:961– 966. discussion 966–8. 21. Meisler R, Thomsen AB, Abildstrom H, et al. Triage and mortality in 2875 consecutive trauma patients. Acta Anaesthesiol Scand 2010;54:218–223. 22. Utter GH, Victorino GP, Wisner DH. Interhospital transfer occurs more slowly for elderly acute trauma patients. J Emerg Med 2008;35:415–420. 23. Young JS, Bassam D, Cephas GA, Brady WJ, Butler K, Pomphrey M. Interhospital versus direct scene transfer of major trauma patients in a rural trauma system. Am Surg 1998;64:88– 91. discussion 91–2. 24. Dunn LT. Secondary insults during the interhospital transfer of head-injured patients: an audit of transfers in the Mersey Region. Injury 1997;28:427–431.
A. Rozenberg et al. 25. 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;69:1362–1366. 26. Nathens AB, Maier RV, Brundage SI, Jurkovich GJ, Grossman DC. The effect of interfacility transfer on outcome in an urban trauma system. J Trauma 2003;55:444–449. 27. Rivara FP, Koepsell TD, Wang J, Nathens A, Jurkovich GA, Mackenzie EJ. Outcomes of trauma patients after transfer to a level I trauma center. J Trauma 2008;64:1594–1599. 28. Hemmila MR, Jakubus JL, Wahl WL, et al. Detecting the blind spot: complications in the trauma registry and trauma quality improvement. Surgery 2007;142:439–448. discussion 448–9. 29. Glance LG, Osler TM, Mukamel DB, Meredith W, Wagner J, Dick AW. TMPM-ICD9: a trauma mortality prediction model based on ICD-9-CM codes. Ann Surg 2009; 249:1032–1039.
Outcomes after Motor Vehicle Trauma
ARTICLE SUMMARY 1. Why is this topic important? Despite the fact that triage systems exist and continue to be developed, trauma patients are still transferred regularly from other hospital facilities to high-level centers for a variety of reasons, including injury severity and availability of specialized services. Additionally, hospital-toLevel-I-center transfers seem to be increasing in number; however, few national studies exist in identifying patients that are transferred to Level I centers. 2. What does this study attempt to show? We have attempted to demonstrate that on a national level, Level I vehicular trauma transfers do not have an increase in mortality when compared with directly admitted patients, even though transferred patients had greater severity of illness. As expected, the data confirm that transferred patients have increased hospital charges and increased health care utilization. 3. What are the key findings? Despite transfer patients having higher adjusted riskseverity scores and higher adjusted risk of mortality, there was no statistical difference in mortality between these groups. Additionally, hospitalizations due to transfer to a Level I trauma center had higher hospital resource utilization, incurring greater length of stay and total hospital charges. 4. How is patient care impacted? Patients who were transferred had significantly higher adjusted risk of mortality and severity of illness scores; however, their mortality was not increased in our study, suggesting that transfer to a Level I center may be of benefit to certain patients in this group. We believe our data provide the groundwork for further studies to examine if specific patients or injuries may benefit from direct admission to Level I trauma centers.
7