The American Journal of Surgery (2015) -, -–-
A simplified trauma triage system safely reduces overtriage and improves provider satisfaction: a prospective study Robert R. Shawhan, M.D.a, Derek P. McVay, D.O.a, Linda Casey, R.N.a, Tara Spears, R.N.a, Scott R. Steele, M.D.a, Matthew J. Martin, M.D., F.A.C.S.a,b,* a
Department of Surgery, Madigan Army Medical Center, 9040-A Fitzsimmons Drive, Tacoma, WA 98431, USA; bDepartment of Surgery, Legacy Emanuel Medical Center, Portland, OR, USA KEYWORDS: Trauma triage; Trauma system; Undertriage; Overtriage
Abstract BACKGROUND: Standard triage systems result in high rates of overtriage to achieve acceptably low undertriage. We previously validated optimal triage variables and used these to implement a new simplified triage system (NEW) at our hospital. METHODS: All trauma entries from May 2010 to Feb 2013 were prospectively reviewed. Calculation of the undertriage and overtriage rates was based on the need for any urgent or life-saving intervention. RESULTS: We identified 704 trauma patients. Level 1 activations were reduced from 32% (OLD) to 19% in the NEW system (P , .05). Overtriage was reduced from 79% (OLD) to 44% in the NEW system (P , .01). The undertriage rate was 1.6% in the NEW system, compared with 1.2% in the OLD system (P 5 nonsignificant). Of all patients, 14% (63) required a life-saving intervention. There were no deaths among undertriaged patients. CONCLUSION: The NEW simplified triage system significantly reduced the rate of overtriage, while safely maintaining a low undertriage rate. Published by Elsevier Inc.
The opinions and assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of Defense. The authors declare no conflicts of interest. Presented as a podium presentation at the North Pacific Surgical Association Annual Meeting, November 14–15, 2014, Seattle, Washington. * Corresponding author. Tel.: 11-253-968-2361; fax: 11-253-968-5900. E-mail address: [email protected] Manuscript received October 15, 2014; revised manuscript January 5, 2015
0002-9610/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.amjsurg.2015.01.008
The appropriate and accurate triage of the acutely injured patient is a cornerstone of modern trauma care and effective trauma systems. Multiple studies have demonstrated the positive impact of appropriate triage on morbidity and mortality.1–3 Although the perfect triage system would always match patient needs to the available resources, current systems inherently result in various degrees of overtriage, undertriage, and mistriage. Among these, the most widely discussed and feared is undertriage, which can result in patient morbidity and mortality because of delayed or missed interventions.2,4 For this reason, most triage systems have focused on minimizing undertriage
2
The American Journal of Surgery, Vol -, No -, - 2015
rates as much as possible without regard to the impact on overtriage or mistriage. The result has been an acceptance of overtriage rates well above 50% as a ‘‘necessary evil’’ of avoiding undertriage and missed significant injuries.5,6 Although the negative impacts of undertriage have been well described, there has been much less investigation of the potential negative impacts or burden of high overtriage rates. The most obvious of these is a waste of resources, including personnel, time, and equipment, on a patient with minimal or no injuries. Additional demonstrated or hypothesized effects include staff and resident frustration, provider and hospital unwillingness to participate in trauma coverage, and the exposure of patients to unnecessary or excessive tests and interventions. The degree of impact of overtriage may also be highly variable depending on the particular system and the level of available resources, and would be expected to have more of a negative impact at centers with fewer resources (Level 2 or 3) compared with a Level 1 trauma center. The majority of the data currently used to design triage systems comes from experiences at Level 1 centers, and may not take into account the significant differences in resources and infrastructure at Level 2 or 3 centers.7 Our Level 2 trauma center had previously used the Pierce Country emergency medical system (EMS) Trauma Triage System (Fig. 1), a 3-tiered graded system of trauma activation that considered physiologic, demographic, and mechanistic variables to determine the level of trauma team activation. We found this system to be both confusing and poorly predictive of the need for trauma team
resources, mainly because of the inclusion of a number of variables that had little to no independent predictive ability to identify injury or need for intervention. On retrospective analysis, we found that this system was associated with a 79% overtriage rate and a 14% mistriage rate. In addition, we were able to identify the important triage variables that independently predicted the presence of injuries and need for urgent life-saving interventions (LSIs).8 Based on this analysis, our Trauma Committee developed a new simplified system for in-hospital trauma triage and trauma team activation that was implemented in April 2010. The purpose of this study is to prospectively analyze the safety, efficacy, and surgeon satisfaction with the newly introduced triage system.
Figure 1
Patients and Methods The study was approved by the Madigan Army Medical Center Institutional Review Board. Our hospital is a Level II trauma center located in Pierce County, Washington. Based on the prior studies at our institution,8,9 we changed our activation criteria from the 3-tiered Pierce County EMS system to a more simplified 2-tiered system beginning in April 2010 (Fig. 2). We established a prospective database to closely track all patients triaged using the newly implemented system. Data collected included the level of trauma activation, reason for activation, prehospital vital signs, emergency department (ED) vital signs, demographics, injuries, and all ED or operative interventions. The
Pierce County pre-hospital trauma triage procedures. Used before implementation of the NEW triage system.
R.R. Shawhan et al.
A simplified trauma triage system
Level 1 Trauma Activation Criteria Hypotension (SBP 120 (or age appropriate tachycardia for pediatrics) Mangled extremity or distal amputations Age > 65 + mechanism (excludes ground level falls) Neurologic deficit (paralysis, suspected spinal cord injury) Burns > 20% BSA or inhalation Multiple long bone fractures or mangled extremity Flail chest Peritonitis on abdominal exam Pregnancy Trauma Consultation All other traumatic mechanisms GCS 15 and normal vital signs
Figure 2 NEW 2-tiered triage system implemented in our hospital starting in April 2010.
appropriateness of triage, need for urgent intervention, and over triage, under triage, and mistriage rates were assessed and compared with those from the previous triage system.8 Our inclusion criteria were all trauma patients older than 16 years of age who presented to our trauma center during the specified time period. Detailed demographics and injury data were only collected on all patients with a formal trauma team activation (Level 1 or Level 2). Patients who did not receive a trauma team activation but were triaged as a ‘‘Trauma Consult’’ were analyzed for any incident of undertriage or mistriage and were included in the denominator of all triage calculations. To allow for easy comparison with our previous study results, we used the same definitions to define appropriate and inappropriate triage rates. Appropriate triage was defined as full trauma team activation (Level 1) in a patient who required some urgent intervention or LSI. Urgent interventions were defined as follows: intubation, emergent or surgical airway, tube thoracostomy or needle thoracostomy, emergency room thoracotomy, packed red blood cell transfusion, placement of a central venous line, or emergent transfer to the operating room (OR). Overtriage was defined as any Level 1 trauma (full activation), which did not necessitate an urgent intervention either in the ED or OR. Undertriage was defined as any patient requiring an urgent intervention that was not triaged as a Level 1 trauma activation. Mistriage was defined as any patient assigned to an incorrect level of trauma activation based on their prehospital data. Full trauma team activation (Level 1) consists of all available surgical resources mobilized to the ED: staff surgeon on call, chief surgical resident, junior surgical resident, and intern surgery resident on call. Additionally, the ED staff and junior resident, anesthesia provider on call, respiratory therapy, radiology technician,
3 radiologist and blood bank representative were also required as part of all trauma activations despite the categorization. Modified trauma activation (Level 2) consists of the full trauma response with the exception of the staff surgeon and blood bank personnel. For trauma patients not meeting criteria for a Level 1 or 2 trauma activation, the initial evaluation was performed by ED personnel with subsequent consultation of the trauma service and were defined as trauma consults. The primary endpoints were the overtriage, undertriage, and mistriage rates, as defined above. A detailed analysis was performed of the undertriaged patient cohort, including the primary reason for undertriage and the development of any subsequent morbidity or mortality. If there were any adverse events among the undertriaged patients, then they were formally presented and discussed by the trauma committee and a determination was made regarding the contribution of undertriage to the adverse outcome. In addition, an anonymous online survey was distributed to all staff surgeons participating in trauma coverage at our institution at 15 months after implementation of the NEW triage system. The survey instrument assessed the staff surgeon’s perception of the safety, efficiency, and overall satisfaction with the NEW triage system compared with the previous system. Statistical analysis was performed using PASW 19.0 (IBM, Chicago, IL). Descriptive statistics were used to describe the study population and continuous variables are shown as mean 6 standard deviation. Dichotomous variables are presented as percentages. For all statistical analysis, P value less than .05 was considered statistically significant. Chi-square test was used to analyze categorical variables and t-tests were used for the comparison of continuous variables.
Results During the study period, 460 patients met our inclusion criteria. A full (Level 1) trauma team activation was initiated for 89 patients (19%), with the remainder receiving a modified trauma response (Level 2 or trauma consult). Table 1 shows the demographics of all Level 1 and Level 2 trauma patients. Most trauma patients were male. Compared with Level 2 trauma patients, Level 1 trauma patients had a lower field and ED Glasgow coma score (GCS), were more likely to require intensive care unit (ICU) admission, and had a higher mean injury severity score. Emergent intervention was required in 14% (65) of all trauma patients, and was more frequently required in Level 1 trauma activations (56%) versus a Level 2 or trauma consult (4%). Table 2 shows the breakdown of urgent interventions performed by the level of triage and trauma team activation. In comparing our NEW triage criteria with the OLD system, the overtriage rate for full trauma team activation (Level 1) was significantly decreased from 79% to 44% (P , .001, Fig. 3). The overall undertriage rate
The American Journal of Surgery, Vol -, No -, - 2015
4 Table 1 Demographics and triage data of the study population Variable Age Male sex Blunt mechanism Field GCS Field heart rate Field systolic blood pressure ED GCS ED heart rate ED systolic blood pressure ED hematocrit Alcohol involved Admission ICU admission Injury severity score
Level 1 (n 5 89)
Level 2 (n 5 146)
37.1 6 18.2 61 (68%) 72 (81%) 11.1 6 4.5 95.3 6 28.0 131.0 6 34.9
40.4 6 21.8 86 (58%) 134 (92%) 14.4 6 1.2 98.1 6 18.6 135 6 28.3
10.8 6 5.1 90.3 6 22.0 134.8 6 31.4
14.7 6 .78 91.5 6 19.5 144.1 6 24.9
38.7 6 9.5 26 (29%) 78 (88%) 56 (63%) 13.7 6 12.2
30.4 6 9.8 27 (18%) 116 (79%) 26 (18%) 5.7 6 6.9
Values expressed as means and standard deviation. ED 5 emergency department; GCS 5 Glasgow coma score; ICU 5 intensive care unit.
demonstrated no statistically significant change (1.6% vs 1.2%, P 5 .66; Fig. 3). Finally, mistriage rates were significantly decreased at 6% compared with 14% (P , .05; Fig. 3) using the OLD triage system.9 In the patients who were undertriaged, all were immediately evaluated by the ED staff, and then seen as ‘‘trauma consults’’ by the trauma service. They were all hemodynamically stable on presentation. Of those, 1 patient needed a tube thoracostomy for a small and asymptomatic pneumothorax and 2 patients required intubation for combativeness because of intoxication but had no significant injuries. Finally, 3 patients were transferred directly to the OR from the ED, 2 for orthopedic injuries and 1 for a craniotomy after a small amount of pneumocephalus was found on trauma workup. There were no deaths among the undertriaged patients, and all were reviewed by the trauma committee and found to have no delays in receiving
Table 2
appropriate diagnostic studies or care. The trauma activation criteria correctly identified 100% of patients needing an emergent surgical airway, 85% of intubations, and 82% of those needing tube thoracostomy. An online anonymous survey was distributed to 11 staff trauma surgeons who had provided trauma coverage under both the old triage system and using the new simplified triage system. Nine completed surveys were returned, for a response rate of 82%. The respondents unanimously preferred the NEW triage system and felt that it more accurately triaged trauma patients, reduced overtriage rates, and was a better use of hospital and staff resources compared with the previous system. All respondents (100%) preferred the NEW system to the OLD system, and most importantly all (100%) felt that the NEW triage system was safe and adequately captures all seriously injured patients.
Comments Optimal care of the traumatically injured patient relies on a rapid and prioritized approach to identifying anatomic and physiologic derangements that require urgent intervention. This process begins with prehospital identification and triage of the injured patient to an appropriately staffed and resourced trauma center that is able to care for their injuries. This field triage is then followed by a separate and individualized in-hospital triage process to activate the appropriate level of personnel and resources to provide optimal trauma care. As the exact nature and extent of injuries are unknown before the patient’s arrival, both these triage processes involve estimations based on limited available information. Multiple variables and algorithms have been proposed to help guide prehospital and ED personnel in making decisions about entering a patient into a trauma system and triggering trauma team activation. Although some of these systems have been studied and validated, the majority are anecdotal or of unproven predictive ability. In addition, the complexity of many of these proposed schemes may limit their application and reliability, and result in a high rate of inappropriate or erroneous triage categorizations (mistriage).
Urgent interventions by level of triage and trauma team activation
Intervention
Overall (n 5 460)
Level 1 (n 5 89)
Level 2 (n 5 146)
Trauma consult (n 5 225)
Any intervention* Intubation Surgical airway Chest tube Central line Blood products CPR Urgent surgery
65 33 1 17 13 18 5 15
50 28 1 14 11 12 5 12
8 1 0 2 2 5 0 1
7 4 0 1 0 1 0 2
(14%) (9%) (.2%) (4%) (3%) (4%) (1%) (3%)
(56%) (31%) (1%) (16%) (12%) (13%) (6%) (13%)
(5%) (.7%) (1%) (1%) (3%) (.7%)
(3%) (1.8%) (.4%) (.4%) (.9%)
CPR 5 cardiopulmonary resuscitation. *Totals not equal to the sum of individual interventions as individual patients may have required multiple interventions.
R.R. Shawhan et al.
A simplified trauma triage system
Figure 3 Comparison of undertriage, overtriage, and mistriage rates in the OLD and NEW system. Values expressed as percentages. *Statistically significant trend.
The priorities for triage of the injured patient have been well described, and are outlined by the American College of Surgeons guidelines for establishing trauma systems and centers.10 The most important and emphasized goal of these systems is to minimize undertriage, which occurs when there is a failure to recognize or appreciate the presence of significant injury or derangements that require trauma center care. Undertriage has been well studied and shown to contribute to potentially preventable patient morbidity or even mortality.11 All trauma triage systems that aim to minimize undertriage will uniformly result in some degree of overtriage of uninjured or minimally injured patients. This is a well understood and accepted consequence of maximizing patient safety. The American College of Surgeons guidelines give an acceptable overtriage rate of up to 50% for prehospital triage, and provide no guidance on the acceptable or optimal overtriage rate for in-hospital triage and trauma team activation.10 Despite the universal presence of a high degree of overtriage at most trauma centers, there has been relatively little published data analyzing the reasons for overtriage and the potential impacts at the patient, hospital, or system level. Uleberg et al6 analyzed a series of 809 trauma patients and demonstrated an overtriage rate of 78% at their institution, which is comparable with the 79% overtriage rate we documented with our previous triage system. They examined the reasons for overtriage and identified the inclusion of variables related to mechanism of injury in their triage system as the main cause. Other series have found that poorly designed and complex triage systems, lack of provider education, and failure to comply with triage protocols are also significant contributors to overtriage.12,13 A 9-year study of triage in the Orange County EMS system demonstrated an increase in the overtriage rate from 43% to 74% over the study period, with growth in the trauma population exceeding the overall
5 population growth.14 Despite these concerning numbers, the short- and long-term impacts of overtriage remain poorly studied and often under-appreciated in comparison with undertriage. The overtriage of uninjured or minimally injured patients can result in significant wasted costs and resources, inappropriate and unnecessary patient interventions and imaging, unnecessary disruption of ED and OR work flow, and can contribute to significant provider-level dissatisfaction with the trauma system and providing trauma coverage.5,6,15 In the current era of crisis regarding healthcare costs, shortages of personnel willing to provide trauma coverage, and declining resident interest in careers in trauma, the problem of overtriage should warrant closer inspection. It is also important to note that both undertriage and overtriage rates will vary widely between studies based on the definitions that are chosen to define what an appropriate or inappropriate triage decision is. We used a relatively strict definition of appropriate triage based on the need for some urgent intervention or LSI performed during the initial evaluation or shortly after in the OR. Other studies that have used much looser criteria, such as the presence of any major traumatic injury, the need for hospital or ICU admission, or some cutoff injury severity score, will naturally have a lower overtriage rate compared with our data. Although our choice of a stricter set of criteria for appropriate triage did result in a high baseline overtriage rate for the OLD system, this rate is comparable with several other studies at Level 1 trauma centers. We did demonstrate a significant reduction in overtriage using the same set of criteria for both the OLD and NEW systems, but further comparative studies to other triage systems are required to determine if our system has any advantage over well-designed triage criteria at other centers. To minimize the chance of an undertriage event, most trauma systems have adopted extensive lists of variables related to the mechanism of injury and patient demographics, which often have little to no scientific validation. An unintended result of inclusion of many of these variables is an extremely high rate of overtriage with no significant impact on undertriage. In this situation, the addition of these variables only serves to magnify all the adverse effects of overtriage, with no impact on improving patient safety or outcomes. Multiple authors have examined the independent validity of triage variables and concluded that many can be safely eliminated from triage systems. Dehli et al16 examined triage variables associated with a 71% overtriage incidence in their trauma system, and found that mechanism of injury variables made the largest contribution to overtriage. Another study of prehospital triage demonstrated that adding variables such as ‘‘high velocity impact’’ and ‘‘ emergency medical technician (EMT) judgment’’ significantly increased overtriage without improving triage sensitivity for detecting severe injury.17 Multiple additional studies have demonstrated that overtriage can be significantly and safely reduced by excluding select variables from both prehospital and in-hospital triage guidelines.18–20 A common
6
The American Journal of Surgery, Vol -, No -, - 2015
finding among these studies has been the poor predictive value of mechanistic variables, which appears to have no additional value when used in conjunction with anatomic and physiologic variables. We came to similar conclusions when we performed a detailed multivariate analysis of the triage variables used in the Pierce County triage system (Fig. 1), and concluded that variables such as ‘‘high speed,’’ ‘‘significant vehicle intrusion,’’ ‘‘prolonged extrication,’’ and so on had no additional value to the more reliable anatomic and physiologic triage variables.21 We found that 4 variables were highly predictive of the need for urgent interventions, hypotension, altered respirations, GCS less than 13, and penetrating truncal injury, and used these as the core variables for our simplified triage system for a Level 1 trauma activation. The strongest individual predictors of the need for urgent intervention in this study were (1) penetrating truncal wound, (2) hypotension, and (3) altered respirations or intubation, in agreement with our previous retrospective data. The other triage criteria were less predictive, or several were added because of requirements by the state of Washington. We also added several criteria that are unable to be validated because of their rarity, but pass the common sense test of having a high probability of needing urgent LSIs (such as extremity amputation). The true incidence of overtriage and undertriage will also vary highly depending on the setting and the definitions used. The majority of published data regarding triage systems comes from highly resourced Level 1 trauma centers, and may not be applicable to Level 2 or lower centers such as ours. Another major factor that must be appreciated is that the goals of prehospital triage (whether to transport to a trauma center) and in-hospital triage (activating the appropriate resources) systems are markedly different and should be separately designed, implemented, and analyzed. Our previous system of triage, and that used by other trauma centers in our area, essentially uses a triage system designed for prehospital use to also make in-hospital triage decisions about the level of trauma team activation. We found that while this triage system (Fig. 1) was appropriate for prehospital determination regarding the need for transport to a trauma center, it was not adequately designed to reasonably predict in-hospital injury acuity or resource needs. In addition, the complexity of the 3-tiered system and high number of variables to consider resulted in a constant rate of mistriage in the 10% to 15% range. This study has several significant limitations that should be noted. We used the requirement for urgent interventions as the measure of overtriage or undertriage, and this may not adequately account for all patients who could benefit from a full trauma team activation. We also specifically focused on urgent interventions in the ED or OR, and not other outcome measures that may be important such as the need for ICU admission or procedures. This experience is from a Level 2 trauma center with a significantly lower trauma volume and a relatively lower incidence of severely injured patients compared with larger Level 1 trauma
centers. Although we demonstrated no change in the undertriage rate with the NEW system, the low incidence of these events may require significantly larger sample sizes to definitively demonstrate the safety of our more restrictive triage criteria. This difference must also be taken into account when considering the implications of overtriage and undertriage, which we believe has greater impact at facilities with less personnel and resources compared with larger dedicated trauma centers. This study also represents the initial experience with a new triage system, and further collection and analysis of the data may identify additional factors or variables that should be added or deleted to maximize the safety and efficiency of this system. Finally, although the survey results overwhelmingly favored the NEW triage system, this represents opinion data and was only collected from staff trauma surgeons. We did not perform an analysis of the impact or impressions of the triage system from the perspective of other key trauma team personnel such as residents, nurses, medics, and ED staff members. In addition, we did not analyze potential impacts on other resource utilization measures such as time to evaluation, time to disposition decisions, and time to admission or discharge from the ED. We recognize that these are significant and often underestimated factors that should be considered and analyzed when making any major changes to trauma triage and activation criteria.
Conclusions In conclusion, this study serves to prospectively validate the safety and efficacy of our 2-tiered simplified triage system for in-hospital activation of the trauma team at our Level 2 trauma center. The overtriage rate for full trauma team activation demonstrated a significant reduction from 79% to 44% for patients who required any type of urgent intervention in the ED or OR. Most importantly, there was no significant increase in the incidence of undertriage events and detailed analysis of the undertriaged patients demonstrated no significant morbidity or mortality associated with a delay in care. The improvements we noted in reducing the overtriage rate were mainly attributable to the elimination of poorly predictive mechanism of injury variables, as well as eliminating isolated mild confusion or intoxication (GCS of 14) as a trigger for a Level 1 trauma activation. These improvements to the trauma system were made only after previous retrospective analysis of our trauma population. Also of significant importance, particularly for a Level 2 trauma center, was the significant reduction in hospital and staff resources and the improved satisfaction of the participating staff trauma surgeons. Further prospective data collection and potentially even controlled trials would be of benefit to optimize in-hospital trauma triage systems and maximize both patient safety and the utilization of critical trauma and emergency resources.
R.R. Shawhan et al.
A simplified trauma triage system
References 1. Davis T, Dinh M, Roncal S, et al. Prospective evaluation of a twotiered trauma activation protocol in an Australian major trauma referral hospital. Injury 2010;41:470–4. 2. Mackersie RC. History of trauma field triage development and the American College of Surgeons criteria. Prehosp Emerg Care 2006; 10:287–94. 3. Nasr A, Mikrogianakis A, McDowall D, et al. External validation and modification of a pediatric trauma triage tool. J Trauma 2007;62:606–9. 4. Raynaud L, Borne M, Coste S, et al. Triage protocol: both undertriage and overtriage need to be evaluated. J Trauma 2010;69:998. 5. Ciesla DJ, Sava JA, Street JH, et al. Secondary overtriage: a consequence of an immature trauma system. J Am Coll Surg 2008;206:131–7. 6. Uleberg O, Vinjevoll OP, Eriksson U, et al. Overtriage in traumad what are the causes? Acta Anaesthesiol Scand 2007;51:1178–83. 7. Shapiro MJ, McCormack JE, Jen J. Let the surgeon sleep: trauma team activation for severe hypotension. J Trauma 2008;65:1245–50; discussion, 1250–2. 8. Lehmann R, Brounts L, Lesperance K, et al. A simplified set of trauma triage criteria to safely reduce overtriage: a prospective study. Arch Surg 2009;144:853–8. 9. Lehmann RK, Arthurs ZM, Cuadrado DG, et al. Trauma team activation: simplified criteria safely reduces overtriage. Am J Surg 2007;193:630–5. 10. Sasser SM, Hunt RC, Faul M, et al. Guidelines for field triage of injured patients: recommendations of the National Expert Panel on Field Triage, 2011. MMWR Recomm Rep 2012;61:1–20. 11. Nakahara S, Matsuoka T, Ueno M, et al. Predictive factors for undertriage among severe blunt trauma patients: what enables them to slip through an established trauma triage protocol? J Trauma 2010;68:1044–51. 12. Cook CH, Muscarella P, Praba AC, et al. Reducing overtriage without compromising outcomes in trauma patients. Arch Surg 2001;136:752–6. 13. Knudson P, Frecceri CA, DeLateur SA. Improving the field triage of major trauma victims. J Trauma 1988;28:602–6. 14. O’Rourke B, Bade RH, Drezner T. Trauma triage: a nine-year experience. Ann Emerg Med 1992;21:680–7. 15. Hoff WS, Tinkoff GH, Lucke JF, et al. Impact of minimal injuries on a level I trauma center. J Trauma 1992;33:408–12. 16. Dehli T, Fredriksen K, Osbakk SA, et al. Evaluation of a university hospital trauma team activation protocol. Scand J Trauma Resusc Emerg Med 2011;19:18. 17. Lavoie A, Emond M, Moore L, et al. Evaluation of the prehospital index, presence of high-velocity impact and judgment of emergency medical technicians as criteria for trauma triage. CJEM 2010;12:111–8. 18. Isenberg D, Cone DC, Vaca FE. Motor vehicle intrusion alone does not predict trauma center admission or use of trauma center resources. Prehosp Emerg Care 2011;15:203–7. 19. Kouzminova N, Shatney C, Palm E, et al. The efficacy of a two-tiered trauma activation system at a level I trauma center. J Trauma 2009;67: 829–33. 20. Shatney CH, Sensaki K. Trauma team activation for ‘‘mechanism of injury’’ blunt trauma victims: time for a change? J Trauma 1994;37: 275–82. 21. Lehmann R, Beekley A, Casey L, et al. The impact of advanced age on trauma triage decisions and outcomes: a statewide analysis. Am J Surg 2009;197:571–4.
Discussion William Long, M.D.: Our colleagues from Madigan Army Medical Center (MAMC), a Tacoma, Washington State Level 2 trauma center developed a 2 tiered in-hospital trauma triage and trauma team activation triage scheme to replace the previous three tiered in-hospital trauma activation scheme that was in place prior to April
7 2010. They gathered data on the pre-hospital trauma triage criteria used in their community (Pierce County), compared this with in-hospital trauma triage and trauma team activation for the previous 3 tiered system and the introduced 2 tiered system. They measured over and under triage, mis-triage and surgeons’ satisfaction with the 2 trauma activation systems. In this paper, we have similar triage terminology, but different types of triage conducted for two reasons: 1. Selection of scene trauma patients to transport to a trauma center by EMS. 2. Selection of level of hospital resources to assess and treat the trauma patient brought to trauma center. Pre-hospital Trauma triage in general is complicated by multiple triage guidelines, influenced by National guidelines (ACSCOT, CDC), State guidelines, County guidelines, EMS politics, and regional hospital politics involving non trauma centers, and trauma center internal politics. A review of civilian trauma system development and the role of pre-hospital trauma triage criteria will characterize the complexity of this subject. As the authors pointed out, the ideal urban trauma system would have pre-hospital trauma triage criteria that misses no seriously injured patient (0% undertriage) and has no overtriage (minor trauma patients being transported by EMS to a trauma center). To understand the magnitude of this problem, in 1985, Multnomah County, Oregon did a study on 19,000 patients entering the EMS system with a trauma diagnosis. Only 5,000 patients retrospectively needed trauma center care at that time. Hospital administrators and emergency physicians at non trauma centers were concerned about the economic consequences of over-triage, and the effect on their Emergency Department revenues. Over 26 years have passed since Oregon began a trauma system. While patient numbers for mandatory trauma system entry criteria have not changed significantly, the number of patients entered via EMT discretionary trauma system entry have increased! Why has this occurred? Surgical specialists at non trauma centers now have less interest in the minor trauma patients, and ask their emergency physicians to transfer these cases to the trauma centers. The emergency physicians, having no surgical specialists to admit these patients to, then call the ambulance crews to transport the minor trauma patients to the Trauma centers. Some Portland area EMTs are now bypassing the non trauma centers to take patients with minor trauma to trauma centers. Time and research has shown that an ideal trauma pre-hospital triage criteria has not been achieved, nor likely to be achieved. Internal trauma triage for trauma centers is another matter. The surgeons at the Level 1 & 2 trauma centers now are assessing and treating an increased number of minor trauma patients, many of whom are chronically mentally ill, alcohol and/or drug intoxicated, with few social and financial resources. Some trauma specialists, namely
The American Journal of Surgery, Vol -, No -, - 2015
8 orthopedic surgeons and neurosurgeons, were and are concerned about too many patients with minor injuries in their specialty, coming to the Trauma Center, and occupying their time and hospital resources to the disadvantage of elective patients and surgeries. The increased number of minor trauma patients coming to trauma centers leads to internal trauma triage schemes like the Madigan authors propose to allow the trauma team members and the surgical specialists be more efficient. Our MAMC colleagues have shown their two tiered in-hospital triage criteria system uses their trauma center and hospital resources more appropriately and efficiently. Internal or In-hospital triage helps sort the arriving trauma patients to those who get priority resources versus those that can wait for their diagnostic workup and care, perhaps by non-trauma personnel. MAMC reports that there were no mortalities in their level 2 in-hospital trauma triage, and this is laudable. This raises the following questions: 1. Were there any significant complications for the level 2 patients and under-triaged patients from delays in hospital resources? 2. How available are the non-general surgeon specialists (Neurosurgeons, Orthopedic surgeons Thoracic and OMFS surgeons) to do immediate surgery and mop-
3. 4.
5.
6.
7.
8.
up surgery on trauma system patients at MAMC, compared to other trauma centers? How many times is MAMC on divert for lack of resources, providers and beds? Is the MAMC trauma center transferring civilian trauma to other trauma centers to lighten their trauma load of patients and make their resources more available to the military personnel and their families? Has MAMC the opportunity to upgrade a trauma patient from a level 2 in-hospital resource to a level 1, and are the response times for the surgical specialists any different? What happens when an under-triage patient with a fractured femur and a pulse deficit shows up? The Internal triage criteria doesn’t mention this? What happens with a conscious patient, not older than 65, who is taking Prodaxa, hits his or her head? Where does this fit into their in-hospital triage scheme? With regard to surgeon satisfaction on the two levels of internal triage for mobilization of resources, how realistic is it to expect anonymity of the surgeon responders in a military hospital, where everything is under close surveillance by the trauma medical director? Is there a kernel of truth in this data?
A simplified trauma triage system safely reduces overtriage and improves provider satisfaction: a prospective study Robert R. Shawhan, M.D.a, Derek P. McVay, D.O.a, Linda Casey, R.N.a, Tara Spears, R.N.a, Scott R. Steele, M.D.a, Matthew J. Martin, M.D., F.A.C.S.a,b,* a
Department of Surgery, Madigan Army Medical Center, 9040-A Fitzsimmons Drive, Tacoma, WA 98431, USA; bDepartment of Surgery, Legacy Emanuel Medical Center, Portland, OR, USA KEYWORDS: Trauma triage; Trauma system; Undertriage; Overtriage
Abstract BACKGROUND: Standard triage systems result in high rates of overtriage to achieve acceptably low undertriage. We previously validated optimal triage variables and used these to implement a new simplified triage system (NEW) at our hospital. METHODS: All trauma entries from May 2010 to Feb 2013 were prospectively reviewed. Calculation of the undertriage and overtriage rates was based on the need for any urgent or life-saving intervention. RESULTS: We identified 704 trauma patients. Level 1 activations were reduced from 32% (OLD) to 19% in the NEW system (P , .05). Overtriage was reduced from 79% (OLD) to 44% in the NEW system (P , .01). The undertriage rate was 1.6% in the NEW system, compared with 1.2% in the OLD system (P 5 nonsignificant). Of all patients, 14% (63) required a life-saving intervention. There were no deaths among undertriaged patients. CONCLUSION: The NEW simplified triage system significantly reduced the rate of overtriage, while safely maintaining a low undertriage rate. Published by Elsevier Inc.
The opinions and assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of Defense. The authors declare no conflicts of interest. Presented as a podium presentation at the North Pacific Surgical Association Annual Meeting, November 14–15, 2014, Seattle, Washington. * Corresponding author. Tel.: 11-253-968-2361; fax: 11-253-968-5900. E-mail address: [email protected] Manuscript received October 15, 2014; revised manuscript January 5, 2015
0002-9610/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.amjsurg.2015.01.008
The appropriate and accurate triage of the acutely injured patient is a cornerstone of modern trauma care and effective trauma systems. Multiple studies have demonstrated the positive impact of appropriate triage on morbidity and mortality.1–3 Although the perfect triage system would always match patient needs to the available resources, current systems inherently result in various degrees of overtriage, undertriage, and mistriage. Among these, the most widely discussed and feared is undertriage, which can result in patient morbidity and mortality because of delayed or missed interventions.2,4 For this reason, most triage systems have focused on minimizing undertriage
2
The American Journal of Surgery, Vol -, No -, - 2015
rates as much as possible without regard to the impact on overtriage or mistriage. The result has been an acceptance of overtriage rates well above 50% as a ‘‘necessary evil’’ of avoiding undertriage and missed significant injuries.5,6 Although the negative impacts of undertriage have been well described, there has been much less investigation of the potential negative impacts or burden of high overtriage rates. The most obvious of these is a waste of resources, including personnel, time, and equipment, on a patient with minimal or no injuries. Additional demonstrated or hypothesized effects include staff and resident frustration, provider and hospital unwillingness to participate in trauma coverage, and the exposure of patients to unnecessary or excessive tests and interventions. The degree of impact of overtriage may also be highly variable depending on the particular system and the level of available resources, and would be expected to have more of a negative impact at centers with fewer resources (Level 2 or 3) compared with a Level 1 trauma center. The majority of the data currently used to design triage systems comes from experiences at Level 1 centers, and may not take into account the significant differences in resources and infrastructure at Level 2 or 3 centers.7 Our Level 2 trauma center had previously used the Pierce Country emergency medical system (EMS) Trauma Triage System (Fig. 1), a 3-tiered graded system of trauma activation that considered physiologic, demographic, and mechanistic variables to determine the level of trauma team activation. We found this system to be both confusing and poorly predictive of the need for trauma team
resources, mainly because of the inclusion of a number of variables that had little to no independent predictive ability to identify injury or need for intervention. On retrospective analysis, we found that this system was associated with a 79% overtriage rate and a 14% mistriage rate. In addition, we were able to identify the important triage variables that independently predicted the presence of injuries and need for urgent life-saving interventions (LSIs).8 Based on this analysis, our Trauma Committee developed a new simplified system for in-hospital trauma triage and trauma team activation that was implemented in April 2010. The purpose of this study is to prospectively analyze the safety, efficacy, and surgeon satisfaction with the newly introduced triage system.
Figure 1
Patients and Methods The study was approved by the Madigan Army Medical Center Institutional Review Board. Our hospital is a Level II trauma center located in Pierce County, Washington. Based on the prior studies at our institution,8,9 we changed our activation criteria from the 3-tiered Pierce County EMS system to a more simplified 2-tiered system beginning in April 2010 (Fig. 2). We established a prospective database to closely track all patients triaged using the newly implemented system. Data collected included the level of trauma activation, reason for activation, prehospital vital signs, emergency department (ED) vital signs, demographics, injuries, and all ED or operative interventions. The
Pierce County pre-hospital trauma triage procedures. Used before implementation of the NEW triage system.
R.R. Shawhan et al.
A simplified trauma triage system
Level 1 Trauma Activation Criteria Hypotension (SBP 120 (or age appropriate tachycardia for pediatrics) Mangled extremity or distal amputations Age > 65 + mechanism (excludes ground level falls) Neurologic deficit (paralysis, suspected spinal cord injury) Burns > 20% BSA or inhalation Multiple long bone fractures or mangled extremity Flail chest Peritonitis on abdominal exam Pregnancy Trauma Consultation All other traumatic mechanisms GCS 15 and normal vital signs
Figure 2 NEW 2-tiered triage system implemented in our hospital starting in April 2010.
appropriateness of triage, need for urgent intervention, and over triage, under triage, and mistriage rates were assessed and compared with those from the previous triage system.8 Our inclusion criteria were all trauma patients older than 16 years of age who presented to our trauma center during the specified time period. Detailed demographics and injury data were only collected on all patients with a formal trauma team activation (Level 1 or Level 2). Patients who did not receive a trauma team activation but were triaged as a ‘‘Trauma Consult’’ were analyzed for any incident of undertriage or mistriage and were included in the denominator of all triage calculations. To allow for easy comparison with our previous study results, we used the same definitions to define appropriate and inappropriate triage rates. Appropriate triage was defined as full trauma team activation (Level 1) in a patient who required some urgent intervention or LSI. Urgent interventions were defined as follows: intubation, emergent or surgical airway, tube thoracostomy or needle thoracostomy, emergency room thoracotomy, packed red blood cell transfusion, placement of a central venous line, or emergent transfer to the operating room (OR). Overtriage was defined as any Level 1 trauma (full activation), which did not necessitate an urgent intervention either in the ED or OR. Undertriage was defined as any patient requiring an urgent intervention that was not triaged as a Level 1 trauma activation. Mistriage was defined as any patient assigned to an incorrect level of trauma activation based on their prehospital data. Full trauma team activation (Level 1) consists of all available surgical resources mobilized to the ED: staff surgeon on call, chief surgical resident, junior surgical resident, and intern surgery resident on call. Additionally, the ED staff and junior resident, anesthesia provider on call, respiratory therapy, radiology technician,
3 radiologist and blood bank representative were also required as part of all trauma activations despite the categorization. Modified trauma activation (Level 2) consists of the full trauma response with the exception of the staff surgeon and blood bank personnel. For trauma patients not meeting criteria for a Level 1 or 2 trauma activation, the initial evaluation was performed by ED personnel with subsequent consultation of the trauma service and were defined as trauma consults. The primary endpoints were the overtriage, undertriage, and mistriage rates, as defined above. A detailed analysis was performed of the undertriaged patient cohort, including the primary reason for undertriage and the development of any subsequent morbidity or mortality. If there were any adverse events among the undertriaged patients, then they were formally presented and discussed by the trauma committee and a determination was made regarding the contribution of undertriage to the adverse outcome. In addition, an anonymous online survey was distributed to all staff surgeons participating in trauma coverage at our institution at 15 months after implementation of the NEW triage system. The survey instrument assessed the staff surgeon’s perception of the safety, efficiency, and overall satisfaction with the NEW triage system compared with the previous system. Statistical analysis was performed using PASW 19.0 (IBM, Chicago, IL). Descriptive statistics were used to describe the study population and continuous variables are shown as mean 6 standard deviation. Dichotomous variables are presented as percentages. For all statistical analysis, P value less than .05 was considered statistically significant. Chi-square test was used to analyze categorical variables and t-tests were used for the comparison of continuous variables.
Results During the study period, 460 patients met our inclusion criteria. A full (Level 1) trauma team activation was initiated for 89 patients (19%), with the remainder receiving a modified trauma response (Level 2 or trauma consult). Table 1 shows the demographics of all Level 1 and Level 2 trauma patients. Most trauma patients were male. Compared with Level 2 trauma patients, Level 1 trauma patients had a lower field and ED Glasgow coma score (GCS), were more likely to require intensive care unit (ICU) admission, and had a higher mean injury severity score. Emergent intervention was required in 14% (65) of all trauma patients, and was more frequently required in Level 1 trauma activations (56%) versus a Level 2 or trauma consult (4%). Table 2 shows the breakdown of urgent interventions performed by the level of triage and trauma team activation. In comparing our NEW triage criteria with the OLD system, the overtriage rate for full trauma team activation (Level 1) was significantly decreased from 79% to 44% (P , .001, Fig. 3). The overall undertriage rate
The American Journal of Surgery, Vol -, No -, - 2015
4 Table 1 Demographics and triage data of the study population Variable Age Male sex Blunt mechanism Field GCS Field heart rate Field systolic blood pressure ED GCS ED heart rate ED systolic blood pressure ED hematocrit Alcohol involved Admission ICU admission Injury severity score
Level 1 (n 5 89)
Level 2 (n 5 146)
37.1 6 18.2 61 (68%) 72 (81%) 11.1 6 4.5 95.3 6 28.0 131.0 6 34.9
40.4 6 21.8 86 (58%) 134 (92%) 14.4 6 1.2 98.1 6 18.6 135 6 28.3
10.8 6 5.1 90.3 6 22.0 134.8 6 31.4
14.7 6 .78 91.5 6 19.5 144.1 6 24.9
38.7 6 9.5 26 (29%) 78 (88%) 56 (63%) 13.7 6 12.2
30.4 6 9.8 27 (18%) 116 (79%) 26 (18%) 5.7 6 6.9
Values expressed as means and standard deviation. ED 5 emergency department; GCS 5 Glasgow coma score; ICU 5 intensive care unit.
demonstrated no statistically significant change (1.6% vs 1.2%, P 5 .66; Fig. 3). Finally, mistriage rates were significantly decreased at 6% compared with 14% (P , .05; Fig. 3) using the OLD triage system.9 In the patients who were undertriaged, all were immediately evaluated by the ED staff, and then seen as ‘‘trauma consults’’ by the trauma service. They were all hemodynamically stable on presentation. Of those, 1 patient needed a tube thoracostomy for a small and asymptomatic pneumothorax and 2 patients required intubation for combativeness because of intoxication but had no significant injuries. Finally, 3 patients were transferred directly to the OR from the ED, 2 for orthopedic injuries and 1 for a craniotomy after a small amount of pneumocephalus was found on trauma workup. There were no deaths among the undertriaged patients, and all were reviewed by the trauma committee and found to have no delays in receiving
Table 2
appropriate diagnostic studies or care. The trauma activation criteria correctly identified 100% of patients needing an emergent surgical airway, 85% of intubations, and 82% of those needing tube thoracostomy. An online anonymous survey was distributed to 11 staff trauma surgeons who had provided trauma coverage under both the old triage system and using the new simplified triage system. Nine completed surveys were returned, for a response rate of 82%. The respondents unanimously preferred the NEW triage system and felt that it more accurately triaged trauma patients, reduced overtriage rates, and was a better use of hospital and staff resources compared with the previous system. All respondents (100%) preferred the NEW system to the OLD system, and most importantly all (100%) felt that the NEW triage system was safe and adequately captures all seriously injured patients.
Comments Optimal care of the traumatically injured patient relies on a rapid and prioritized approach to identifying anatomic and physiologic derangements that require urgent intervention. This process begins with prehospital identification and triage of the injured patient to an appropriately staffed and resourced trauma center that is able to care for their injuries. This field triage is then followed by a separate and individualized in-hospital triage process to activate the appropriate level of personnel and resources to provide optimal trauma care. As the exact nature and extent of injuries are unknown before the patient’s arrival, both these triage processes involve estimations based on limited available information. Multiple variables and algorithms have been proposed to help guide prehospital and ED personnel in making decisions about entering a patient into a trauma system and triggering trauma team activation. Although some of these systems have been studied and validated, the majority are anecdotal or of unproven predictive ability. In addition, the complexity of many of these proposed schemes may limit their application and reliability, and result in a high rate of inappropriate or erroneous triage categorizations (mistriage).
Urgent interventions by level of triage and trauma team activation
Intervention
Overall (n 5 460)
Level 1 (n 5 89)
Level 2 (n 5 146)
Trauma consult (n 5 225)
Any intervention* Intubation Surgical airway Chest tube Central line Blood products CPR Urgent surgery
65 33 1 17 13 18 5 15
50 28 1 14 11 12 5 12
8 1 0 2 2 5 0 1
7 4 0 1 0 1 0 2
(14%) (9%) (.2%) (4%) (3%) (4%) (1%) (3%)
(56%) (31%) (1%) (16%) (12%) (13%) (6%) (13%)
(5%) (.7%) (1%) (1%) (3%) (.7%)
(3%) (1.8%) (.4%) (.4%) (.9%)
CPR 5 cardiopulmonary resuscitation. *Totals not equal to the sum of individual interventions as individual patients may have required multiple interventions.
R.R. Shawhan et al.
A simplified trauma triage system
Figure 3 Comparison of undertriage, overtriage, and mistriage rates in the OLD and NEW system. Values expressed as percentages. *Statistically significant trend.
The priorities for triage of the injured patient have been well described, and are outlined by the American College of Surgeons guidelines for establishing trauma systems and centers.10 The most important and emphasized goal of these systems is to minimize undertriage, which occurs when there is a failure to recognize or appreciate the presence of significant injury or derangements that require trauma center care. Undertriage has been well studied and shown to contribute to potentially preventable patient morbidity or even mortality.11 All trauma triage systems that aim to minimize undertriage will uniformly result in some degree of overtriage of uninjured or minimally injured patients. This is a well understood and accepted consequence of maximizing patient safety. The American College of Surgeons guidelines give an acceptable overtriage rate of up to 50% for prehospital triage, and provide no guidance on the acceptable or optimal overtriage rate for in-hospital triage and trauma team activation.10 Despite the universal presence of a high degree of overtriage at most trauma centers, there has been relatively little published data analyzing the reasons for overtriage and the potential impacts at the patient, hospital, or system level. Uleberg et al6 analyzed a series of 809 trauma patients and demonstrated an overtriage rate of 78% at their institution, which is comparable with the 79% overtriage rate we documented with our previous triage system. They examined the reasons for overtriage and identified the inclusion of variables related to mechanism of injury in their triage system as the main cause. Other series have found that poorly designed and complex triage systems, lack of provider education, and failure to comply with triage protocols are also significant contributors to overtriage.12,13 A 9-year study of triage in the Orange County EMS system demonstrated an increase in the overtriage rate from 43% to 74% over the study period, with growth in the trauma population exceeding the overall
5 population growth.14 Despite these concerning numbers, the short- and long-term impacts of overtriage remain poorly studied and often under-appreciated in comparison with undertriage. The overtriage of uninjured or minimally injured patients can result in significant wasted costs and resources, inappropriate and unnecessary patient interventions and imaging, unnecessary disruption of ED and OR work flow, and can contribute to significant provider-level dissatisfaction with the trauma system and providing trauma coverage.5,6,15 In the current era of crisis regarding healthcare costs, shortages of personnel willing to provide trauma coverage, and declining resident interest in careers in trauma, the problem of overtriage should warrant closer inspection. It is also important to note that both undertriage and overtriage rates will vary widely between studies based on the definitions that are chosen to define what an appropriate or inappropriate triage decision is. We used a relatively strict definition of appropriate triage based on the need for some urgent intervention or LSI performed during the initial evaluation or shortly after in the OR. Other studies that have used much looser criteria, such as the presence of any major traumatic injury, the need for hospital or ICU admission, or some cutoff injury severity score, will naturally have a lower overtriage rate compared with our data. Although our choice of a stricter set of criteria for appropriate triage did result in a high baseline overtriage rate for the OLD system, this rate is comparable with several other studies at Level 1 trauma centers. We did demonstrate a significant reduction in overtriage using the same set of criteria for both the OLD and NEW systems, but further comparative studies to other triage systems are required to determine if our system has any advantage over well-designed triage criteria at other centers. To minimize the chance of an undertriage event, most trauma systems have adopted extensive lists of variables related to the mechanism of injury and patient demographics, which often have little to no scientific validation. An unintended result of inclusion of many of these variables is an extremely high rate of overtriage with no significant impact on undertriage. In this situation, the addition of these variables only serves to magnify all the adverse effects of overtriage, with no impact on improving patient safety or outcomes. Multiple authors have examined the independent validity of triage variables and concluded that many can be safely eliminated from triage systems. Dehli et al16 examined triage variables associated with a 71% overtriage incidence in their trauma system, and found that mechanism of injury variables made the largest contribution to overtriage. Another study of prehospital triage demonstrated that adding variables such as ‘‘high velocity impact’’ and ‘‘ emergency medical technician (EMT) judgment’’ significantly increased overtriage without improving triage sensitivity for detecting severe injury.17 Multiple additional studies have demonstrated that overtriage can be significantly and safely reduced by excluding select variables from both prehospital and in-hospital triage guidelines.18–20 A common
6
The American Journal of Surgery, Vol -, No -, - 2015
finding among these studies has been the poor predictive value of mechanistic variables, which appears to have no additional value when used in conjunction with anatomic and physiologic variables. We came to similar conclusions when we performed a detailed multivariate analysis of the triage variables used in the Pierce County triage system (Fig. 1), and concluded that variables such as ‘‘high speed,’’ ‘‘significant vehicle intrusion,’’ ‘‘prolonged extrication,’’ and so on had no additional value to the more reliable anatomic and physiologic triage variables.21 We found that 4 variables were highly predictive of the need for urgent interventions, hypotension, altered respirations, GCS less than 13, and penetrating truncal injury, and used these as the core variables for our simplified triage system for a Level 1 trauma activation. The strongest individual predictors of the need for urgent intervention in this study were (1) penetrating truncal wound, (2) hypotension, and (3) altered respirations or intubation, in agreement with our previous retrospective data. The other triage criteria were less predictive, or several were added because of requirements by the state of Washington. We also added several criteria that are unable to be validated because of their rarity, but pass the common sense test of having a high probability of needing urgent LSIs (such as extremity amputation). The true incidence of overtriage and undertriage will also vary highly depending on the setting and the definitions used. The majority of published data regarding triage systems comes from highly resourced Level 1 trauma centers, and may not be applicable to Level 2 or lower centers such as ours. Another major factor that must be appreciated is that the goals of prehospital triage (whether to transport to a trauma center) and in-hospital triage (activating the appropriate resources) systems are markedly different and should be separately designed, implemented, and analyzed. Our previous system of triage, and that used by other trauma centers in our area, essentially uses a triage system designed for prehospital use to also make in-hospital triage decisions about the level of trauma team activation. We found that while this triage system (Fig. 1) was appropriate for prehospital determination regarding the need for transport to a trauma center, it was not adequately designed to reasonably predict in-hospital injury acuity or resource needs. In addition, the complexity of the 3-tiered system and high number of variables to consider resulted in a constant rate of mistriage in the 10% to 15% range. This study has several significant limitations that should be noted. We used the requirement for urgent interventions as the measure of overtriage or undertriage, and this may not adequately account for all patients who could benefit from a full trauma team activation. We also specifically focused on urgent interventions in the ED or OR, and not other outcome measures that may be important such as the need for ICU admission or procedures. This experience is from a Level 2 trauma center with a significantly lower trauma volume and a relatively lower incidence of severely injured patients compared with larger Level 1 trauma
centers. Although we demonstrated no change in the undertriage rate with the NEW system, the low incidence of these events may require significantly larger sample sizes to definitively demonstrate the safety of our more restrictive triage criteria. This difference must also be taken into account when considering the implications of overtriage and undertriage, which we believe has greater impact at facilities with less personnel and resources compared with larger dedicated trauma centers. This study also represents the initial experience with a new triage system, and further collection and analysis of the data may identify additional factors or variables that should be added or deleted to maximize the safety and efficiency of this system. Finally, although the survey results overwhelmingly favored the NEW triage system, this represents opinion data and was only collected from staff trauma surgeons. We did not perform an analysis of the impact or impressions of the triage system from the perspective of other key trauma team personnel such as residents, nurses, medics, and ED staff members. In addition, we did not analyze potential impacts on other resource utilization measures such as time to evaluation, time to disposition decisions, and time to admission or discharge from the ED. We recognize that these are significant and often underestimated factors that should be considered and analyzed when making any major changes to trauma triage and activation criteria.
Conclusions In conclusion, this study serves to prospectively validate the safety and efficacy of our 2-tiered simplified triage system for in-hospital activation of the trauma team at our Level 2 trauma center. The overtriage rate for full trauma team activation demonstrated a significant reduction from 79% to 44% for patients who required any type of urgent intervention in the ED or OR. Most importantly, there was no significant increase in the incidence of undertriage events and detailed analysis of the undertriaged patients demonstrated no significant morbidity or mortality associated with a delay in care. The improvements we noted in reducing the overtriage rate were mainly attributable to the elimination of poorly predictive mechanism of injury variables, as well as eliminating isolated mild confusion or intoxication (GCS of 14) as a trigger for a Level 1 trauma activation. These improvements to the trauma system were made only after previous retrospective analysis of our trauma population. Also of significant importance, particularly for a Level 2 trauma center, was the significant reduction in hospital and staff resources and the improved satisfaction of the participating staff trauma surgeons. Further prospective data collection and potentially even controlled trials would be of benefit to optimize in-hospital trauma triage systems and maximize both patient safety and the utilization of critical trauma and emergency resources.
R.R. Shawhan et al.
A simplified trauma triage system
References 1. Davis T, Dinh M, Roncal S, et al. Prospective evaluation of a twotiered trauma activation protocol in an Australian major trauma referral hospital. Injury 2010;41:470–4. 2. Mackersie RC. History of trauma field triage development and the American College of Surgeons criteria. Prehosp Emerg Care 2006; 10:287–94. 3. Nasr A, Mikrogianakis A, McDowall D, et al. External validation and modification of a pediatric trauma triage tool. J Trauma 2007;62:606–9. 4. Raynaud L, Borne M, Coste S, et al. Triage protocol: both undertriage and overtriage need to be evaluated. J Trauma 2010;69:998. 5. Ciesla DJ, Sava JA, Street JH, et al. Secondary overtriage: a consequence of an immature trauma system. J Am Coll Surg 2008;206:131–7. 6. Uleberg O, Vinjevoll OP, Eriksson U, et al. Overtriage in traumad what are the causes? Acta Anaesthesiol Scand 2007;51:1178–83. 7. Shapiro MJ, McCormack JE, Jen J. Let the surgeon sleep: trauma team activation for severe hypotension. J Trauma 2008;65:1245–50; discussion, 1250–2. 8. Lehmann R, Brounts L, Lesperance K, et al. A simplified set of trauma triage criteria to safely reduce overtriage: a prospective study. Arch Surg 2009;144:853–8. 9. Lehmann RK, Arthurs ZM, Cuadrado DG, et al. Trauma team activation: simplified criteria safely reduces overtriage. Am J Surg 2007;193:630–5. 10. Sasser SM, Hunt RC, Faul M, et al. Guidelines for field triage of injured patients: recommendations of the National Expert Panel on Field Triage, 2011. MMWR Recomm Rep 2012;61:1–20. 11. Nakahara S, Matsuoka T, Ueno M, et al. Predictive factors for undertriage among severe blunt trauma patients: what enables them to slip through an established trauma triage protocol? J Trauma 2010;68:1044–51. 12. Cook CH, Muscarella P, Praba AC, et al. Reducing overtriage without compromising outcomes in trauma patients. Arch Surg 2001;136:752–6. 13. Knudson P, Frecceri CA, DeLateur SA. Improving the field triage of major trauma victims. J Trauma 1988;28:602–6. 14. O’Rourke B, Bade RH, Drezner T. Trauma triage: a nine-year experience. Ann Emerg Med 1992;21:680–7. 15. Hoff WS, Tinkoff GH, Lucke JF, et al. Impact of minimal injuries on a level I trauma center. J Trauma 1992;33:408–12. 16. Dehli T, Fredriksen K, Osbakk SA, et al. Evaluation of a university hospital trauma team activation protocol. Scand J Trauma Resusc Emerg Med 2011;19:18. 17. Lavoie A, Emond M, Moore L, et al. Evaluation of the prehospital index, presence of high-velocity impact and judgment of emergency medical technicians as criteria for trauma triage. CJEM 2010;12:111–8. 18. Isenberg D, Cone DC, Vaca FE. Motor vehicle intrusion alone does not predict trauma center admission or use of trauma center resources. Prehosp Emerg Care 2011;15:203–7. 19. Kouzminova N, Shatney C, Palm E, et al. The efficacy of a two-tiered trauma activation system at a level I trauma center. J Trauma 2009;67: 829–33. 20. Shatney CH, Sensaki K. Trauma team activation for ‘‘mechanism of injury’’ blunt trauma victims: time for a change? J Trauma 1994;37: 275–82. 21. Lehmann R, Beekley A, Casey L, et al. The impact of advanced age on trauma triage decisions and outcomes: a statewide analysis. Am J Surg 2009;197:571–4.
Discussion William Long, M.D.: Our colleagues from Madigan Army Medical Center (MAMC), a Tacoma, Washington State Level 2 trauma center developed a 2 tiered in-hospital trauma triage and trauma team activation triage scheme to replace the previous three tiered in-hospital trauma activation scheme that was in place prior to April
7 2010. They gathered data on the pre-hospital trauma triage criteria used in their community (Pierce County), compared this with in-hospital trauma triage and trauma team activation for the previous 3 tiered system and the introduced 2 tiered system. They measured over and under triage, mis-triage and surgeons’ satisfaction with the 2 trauma activation systems. In this paper, we have similar triage terminology, but different types of triage conducted for two reasons: 1. Selection of scene trauma patients to transport to a trauma center by EMS. 2. Selection of level of hospital resources to assess and treat the trauma patient brought to trauma center. Pre-hospital Trauma triage in general is complicated by multiple triage guidelines, influenced by National guidelines (ACSCOT, CDC), State guidelines, County guidelines, EMS politics, and regional hospital politics involving non trauma centers, and trauma center internal politics. A review of civilian trauma system development and the role of pre-hospital trauma triage criteria will characterize the complexity of this subject. As the authors pointed out, the ideal urban trauma system would have pre-hospital trauma triage criteria that misses no seriously injured patient (0% undertriage) and has no overtriage (minor trauma patients being transported by EMS to a trauma center). To understand the magnitude of this problem, in 1985, Multnomah County, Oregon did a study on 19,000 patients entering the EMS system with a trauma diagnosis. Only 5,000 patients retrospectively needed trauma center care at that time. Hospital administrators and emergency physicians at non trauma centers were concerned about the economic consequences of over-triage, and the effect on their Emergency Department revenues. Over 26 years have passed since Oregon began a trauma system. While patient numbers for mandatory trauma system entry criteria have not changed significantly, the number of patients entered via EMT discretionary trauma system entry have increased! Why has this occurred? Surgical specialists at non trauma centers now have less interest in the minor trauma patients, and ask their emergency physicians to transfer these cases to the trauma centers. The emergency physicians, having no surgical specialists to admit these patients to, then call the ambulance crews to transport the minor trauma patients to the Trauma centers. Some Portland area EMTs are now bypassing the non trauma centers to take patients with minor trauma to trauma centers. Time and research has shown that an ideal trauma pre-hospital triage criteria has not been achieved, nor likely to be achieved. Internal trauma triage for trauma centers is another matter. The surgeons at the Level 1 & 2 trauma centers now are assessing and treating an increased number of minor trauma patients, many of whom are chronically mentally ill, alcohol and/or drug intoxicated, with few social and financial resources. Some trauma specialists, namely
The American Journal of Surgery, Vol -, No -, - 2015
8 orthopedic surgeons and neurosurgeons, were and are concerned about too many patients with minor injuries in their specialty, coming to the Trauma Center, and occupying their time and hospital resources to the disadvantage of elective patients and surgeries. The increased number of minor trauma patients coming to trauma centers leads to internal trauma triage schemes like the Madigan authors propose to allow the trauma team members and the surgical specialists be more efficient. Our MAMC colleagues have shown their two tiered in-hospital triage criteria system uses their trauma center and hospital resources more appropriately and efficiently. Internal or In-hospital triage helps sort the arriving trauma patients to those who get priority resources versus those that can wait for their diagnostic workup and care, perhaps by non-trauma personnel. MAMC reports that there were no mortalities in their level 2 in-hospital trauma triage, and this is laudable. This raises the following questions: 1. Were there any significant complications for the level 2 patients and under-triaged patients from delays in hospital resources? 2. How available are the non-general surgeon specialists (Neurosurgeons, Orthopedic surgeons Thoracic and OMFS surgeons) to do immediate surgery and mop-
3. 4.
5.
6.
7.
8.
up surgery on trauma system patients at MAMC, compared to other trauma centers? How many times is MAMC on divert for lack of resources, providers and beds? Is the MAMC trauma center transferring civilian trauma to other trauma centers to lighten their trauma load of patients and make their resources more available to the military personnel and their families? Has MAMC the opportunity to upgrade a trauma patient from a level 2 in-hospital resource to a level 1, and are the response times for the surgical specialists any different? What happens when an under-triage patient with a fractured femur and a pulse deficit shows up? The Internal triage criteria doesn’t mention this? What happens with a conscious patient, not older than 65, who is taking Prodaxa, hits his or her head? Where does this fit into their in-hospital triage scheme? With regard to surgeon satisfaction on the two levels of internal triage for mobilization of resources, how realistic is it to expect anonymity of the surgeon responders in a military hospital, where everything is under close surveillance by the trauma medical director? Is there a kernel of truth in this data?
