Comparative Performance of the BaxtTrauma Triage Rule CHARLES L. EMERMAN, JOHN KUBINCANEKt


New methods of deciding which patients require trauma center transport continue to be devised. Bad recently published a Trauma Triage Rule (TTR) using anatomic injury, blood pressure, and elements of the Glascow Coma scale which can be used to identity adult major trauma patients. The purpose of this study was to compare the TTR against three previously published trauma triage instruments; the Triage-Revised Trauma Score, the Prehospital Index, and the CRAMS scale. We applied these rules to a data base of trauma patients transported by the Cleveland, Ohlo Emergency Medical Services System. All four Instruments identified adult trauma patients who elther died or required emergent opemtlons with sensitlvities of at least .85. The specificity of the TTR exceeded that of the CRAMS. We conclude that the TTR is an effective means of identifying patients who either die or require emergent operation. (Am J Emerg Med 1992; l&294-297. Copyright 0 1992 by W.B. Saunders Company)

A number of scoring systems have been devised to identify patients who require transport to regional trauma centers. The ideal scoring system would limit overtriage, which has the potential to adversely impact emergency medical services availability, while at the same time limiting undertriage which has the potential to adversely affect patient care. These scoring systems have included checklists or triage rules based on mechanism of injury, anatomic injury, and changes in physiologic variables. In addition, some rules such as the American College of Surgeons guidelines or Kanes Triage Checklist ‘.’ use combinations of these factors. Recently a scoring system was proposed by Baxt which detines adult major trauma victims as patients with a systolic blood pressure less than 85 mm Hg, a motor component of the Glascow Coma scale less than 5, or penetrating injuries of the head, neck, or trunk.3 In the original study group, this rule had a sensitivity of .92 and a specificity of .92 in identifying major trauma victims. The Trauma Triage Rule (‘ITR) contains elements in common with the Prehospital Index (PHI) and the CRAMS scale (Table 1) in that they use both anatomic injury and physioThe Triage-Revised Trauma score logic variables.4*5 (T-RTS) does not include any assessment of anatomic injury in assigning trauma severity.6 The TTR has not been asFrom the *Department of Emergency Medicine, Department of Surgery, MetroHealth Medical Center, Case Western Reserve University, Cleveland, OH; and *City of Cleveland Emergency Medical Services, Cleveland, OH. Manuscript received July 29, 1991; revision accepted January 27, 1992. Address reprint requests to Dr Emerman, Department of Emeraencv Medicine, Sl-203. MetroHealth Medical Center, 2500 MetriHealih Rd, Cleveland, CH 44109. Key Words: Trauma, triage, prehospital care, emergency medical services. Copyright 0 1992 by W.B. Saunders Company 0735-6757/92/l 004-0005$5.00/O 294

sessed in a system other than the one in which it was devised and, further, no comparisons with the other trauma triage tools has been reported. The purpose of this study, therefore, was to compare the TTR against the other traumatriage instruments (PHI, CRAMS, and T-RTS) in a cohort of adult trauma patients from our community. In particular, we wished to assess these four scoring systems as applied to a cohort of patients as seen from the viewpoint of the prehospita1 care provider. We purposely did not limit the study to severe trauma or to patients transported only to trauma centers, recognizing that this would limit our ability to gather detailed’data such as the Injury Severity Score (ISS), duration of hospital stay, results of operative treatment, or residual disability and would include a large number of patients not normally seen by trauma centers. This approach was chosen to assess applicability to EMS systems rather than as a tool for use in the trauma center. METHODS

We have previously accumulated a database of trauma patients transported by the City of Cleveland EMS System.’ The EMS system, at the time, was composed of eight advanced life support and seven basic life support squads, covering a population of 560,000. The data was collected from August 1989 through January 1990. Both paramedic and emergency medical technician (EMT)-A personnel participated in this study. All adult (age 15 years or greater) trauma patients transported by either paramedic or basic squad to any Cuyahoga County hospital were included in the study. Patients transported by the Cleveland Fire Department or by private ambulances or vehicles were not included in the study. Patient information was collected through the use of a data sheet completed by the EMTs. This included information on demographics, mechanism of injury, vital signs, physical examination, and information adequate to calculate a T-RTS, PHI, CRAMS Scale, TTR, and the Glascow Coma scale. Information not included on the data sheets was obtained from the EMS run reports. For the purposes of our study, the CRAMS scale was calculated using the modification suggested by Clemmer and colleagues, with a respiratory rate greater than 35 as added criteria.* Our EMTs noted chest or abdominal injury on a three-point scale as none, mildmoderate, or severe. This was applied to the abdominal portion of the CRAMS scale as corresponding to nontender abdomen/thorax, tender abdomen/thorax, and rigid abdomen/ flail chest, respectively. The T-RTS, PHI, and ‘ITR were used as proposed without modification. All EMS personnel attended inservices immediately prior to initiation of this study to acquaint them with the study instrument. Information was gathered from both the study sheet and the EMS




Scoring of Four Trauma Triage Rules

Prehospital Index’/Score 0 1 2 3 4 5

CRAMS Scalet Score 2 1 0

Respirations (breaths/min)

Blood Pressure (mm Hg)

Pulse (beats/min)

> 100 86-l 00 75-85 o-74

51-119 -

Normal -


> 120 -

Shallow/labored -

Confused/combative -

< 50

< lO/min or intubation

No intelligible

Normal capillary refill orBP>lOO Delayed capillary refill or85 29 6-9 l-5 0

Baxt Trauma Triage Rules

Glascow Coma Scale

Blood Pressure (mm Hg)

Motor component < 5

Systolic < 85

ABBREVIATION:BP, blood pressure. Add scores for each component. Score t Add scores for each component. Total $ Add scores for each component. Total 5 Major trauma identified if any of above l

Injury Penetrating head, neck, or trunk wound

> 3 indicates major trauma. score < 9 indicates major trauma. score < 12 indicates major trauma. three components are present.

run report, both of which were completed prospectively. The EMTs did not calculate any of the scores on either the study sheet or the EMS run sheet. Patient outcome was verified by review of the Cuyahoga County coroner records and by review of information contained in the trauma logs maintained by five of the major teaching hospitals in Cleveland. At the time of the study, these trauma logs did not include ISS scores or detailed data on inpatient course. For the purposes of this study a major trauma victim was defined as a patient who either died or required emergent general surgical (laparotomy, thoracotomy, vascular) or neurosurgical (craniotomy) operation within 2 hours of arrival in the emergency department. This approach was chosen since it identifies patients in whom any delay in definitive operative care may be harmful. It is also similar to the approach used by Gormican4 and Koehler et al.’ A T-RTS value below 12, CRAMS scale less than 9, and a PHI greater than 3 were used as threshold values indicating an identification by those rules of a major trauma patient. The T-RTS scale, CRAMS scale, and PHI were compared

against the TTR using McNemar’s test. A P-value < .OS was taken to indicate statistical significance. Data are reported as the mean & standard deviation. RESULTS One thousand two hundred seventy-five trauma patients were studied during a 6-month period. Complete data adequate to calculate the four scores were available for 1,027 (81%) of the patients. Fifty-seven of the 248 incomplete charts (23%) were missing data to calculate the Glascow Coma scale, 39 (16%) were missing pulse, 102 (41%) were missing respiratory rates, 46 (19%) were missing blood pressure and capillary refill, while 24 (10%) were missing mechanism of injury. The demographics of the study patients are given in Table 2. There were 659 males and 368 females, with an average age of 34.4 * 17.2 years. Three hundred forty-six (34%) of the 1,027 patients included were between 20 and 30 years of age. Most of the injuries (86%) were the result of blunt trauma; however, 17 (46%) of the deaths were as a result of penetrating injury. As seen in Table 3, all of the rules pre-








Sex M/F Age (~1 Penetrating injury Advanced life support squad Severe head injury* Severe abdominal injury’ Severe chest injury’ As clinically


6291361 35.0 2 17.4 2% 65% 4% 1% 1%

mortality with a sensitivity of 1.0. The specificity of the TTR (.90) was similar to that of the PHI (.88) and the T-RTS (.88), although it was significantly better than the CRAMS scale (.83; P < .OOl). The data were further analyzed by using the information on the need for emergent general (laparotomy, thoracotomy, vascular repair) or neurosurgical operation (craniotomy) available on those patients taken to the five teaching hospitals. Six hundred thirteen patients, including 205 females and 408 males with an average age of 33.4 * 15.7 years, were included in this group. Of the patients 247 (40%) were between the ages of 20 and 30. Thirty-eight patients (6.2%) required emergent operations, while there were 29 deaths in this group of patients, 14 (48%) of which were from penetrating trauma. At the recommended threshold, all four triage instruments had a sensitivity of at least .85 (Table 4). The higher sensitivity of the TTR (.94) over the T-RTS (.85) did not reach statistical significance (P = .07). The specificity of the TTR (.89), PHI (.88), and T-RTS (.88) were better than that of the CRAMS scale (.80). The overall accuracy of the TTR (.89) exceeded that of the CRAMS scale (.81). Three patients were incorrectly categorized by the TTR as having minor trauma who subsequently required emergent operation. One patient had a head injury with intermittent confusion following a motor vehicle accident injury, while the other two patients had significant blunt chest injury without hypotension. None of these three patients were correctly categorized by the T-RTS, while the PHI and the CRAMS scale each correctly categorized one patient.

DISCUSSION The TTR represents the latest of a number of trauma triage instruments that have been developed. In this study we have compared it’s performance to three other triage rules. The T-RTS is the successor to several previous trauma scores that use physiologic variables as measured of injury severity. The T-RTS is significantly different from the PHI, the TABLE 3. Performance Mortality Only



3017 33.1 k 13.3 12% 57% 33% 24% 46%

< .05 NS < ,001 NS < ,001 < .OOl c.001

by EMS personnel.


of the Four Triage

Rules in Predicting

CRAMS scale, and the TTR in that it does not use any measures of anatomic injury. The American College of Surgeons guidelines do suggest, however, using the T-RTS in conjunction with anatomic injury and mechanism of injury.’ In the original study, a T-RTS less than 12 identified all but .005% of patients dying from trauma. There have been relatively few studies which have validated the use of the T-RTS in prehospital trauma triage. Eichelburger et al found that the T-RTS, modified for pediatric patients, identified pediatric trauma deaths with a sensitivity of .73 and specificity of .74.9 Baxt et al found that while the T-RTS identified mortality with a sensitivity and specificity greater than .85, it failed to identify survivors with major trauma, as defined by an ISS greater than 14 with a simultaneous sensitivity and specificity of greater than .70.” In a previous study, we demonstrated that the CRAMS scale, PHI, and EMT judgment were superior to the T-RTS in predicting patients who either died or required emergent operation.’ The CRAMS scale has been studied by several investigators since its development by Gormican.4 Ornato et al reviewed 5,130 trauma runs and found that the CRAMS scale identified only 20% of patients who required emergent operation. ’ ’ Hedges and colleagues studied 130 patients transported by a rural EMS system and found that the CRAMS scale identified major trauma victims defined as patients who either died, required emergent operation, or intensive care unit admission with a sensitivity of .85 and a specificity of .54.i2 As was the case with the T-RTS, Baxt et al found that the CRAMS scale performed poorly in identifying patients with an ISS greater than 14. The PHI was validated against a population of 3,581 patients in a multicenter study. I3 Using a PHI greater than 3 as the indicator of major trauma, Koehler et al found that the rule performed with a sensitivity of .93 and a specificity of .93. In the same study described above, Hedges and colleagues found that the PHI performed with a sensitivity of .73 and a specificity of .75.8 In our study, however, we have TABLE 4. Performance of the Four Trauma Triage Predicting Death or Need for Emergent Operation TTR

l-w Sensitivity Specificity Predictive value Predictive value Overall accuracy

1 .o

(+) (-)


PHI 1.0






.27 .87 .90

.18 .80 .83

.24 .85 .89

.24 .85 .89

P < ,001 compared with the Trauma Triage Rule.

Rules In




Et .31 .99 ,817

.94 .88 .42 99 .88

.85’ .88 .39 .98 .87

T-RTS Sensitivity Specificity Predictive value (+) Predictive value (-) Overall accuracy

.94 .89 .45 99 .89

P = .07 compared with Trauma Triage Rule. t P < .Ol compared with Trauma Triage Rule.

l l

4 n July 1992

10, Number

Demographics Survivors


W Volume



found that the PHI performed similarly to the TTR with a sensitivity of .95 and a specificity of .88, although we did not include intensive care unit admission as an indicator of major trauma, as did Hedges, due to lack of detailed outcome data. Of note, while these rules were very sensitive, over half of the patients identified by the TTR do not die or require emergent operation, while almost two thirds of patients identified by the CRAMS scale do not meet this definition of major trauma. Emergency medical services systems implementing trauma triage rules, therefore, must be willing to accept this degree of “overtriage.” One difficulty in comparing these rules is the difference in the definition of major trauma used by each study. Baxt and colleagues defined major trauma as a patient who died, required greater than 1 L of fluid resuscitation in the hospital to maintain a systolic blood pressure of 90 mm Hg, required intracranial pressure monitoring for a significantly positive computed tomography scan, or who had a nonorthopedic operative procedure with the first 48 hours that demonstrated a injury which would be life-threatening if not treated.3 Gormican used either death or direct admission to the operating room,4 while Koehler et al used death or emergent operation within 4 hours as a definition of major trauma.5 Because of the limitation of our trauma registries, we were not able to use fluid resuscitation, results of computed tomographic scans, or need for intracranial pressure monitoring in assessing our patient population. Additional variables that potentially could be used to assess whether the patient in fact had major trauma, including the ISS, duration of intensive care unit stay, duration of hospital admission, or residual deficits could not be assessed using our data base. While our study identifies patients requiring immediate access to operative surgical care, it does not identify patients who require the other benefits of trauma centers such as sophisticated emergency department resuscitation teams, specialized intensive care units, or other benefits of the regional trauma center. On the other hand, we feel this de& nition identities patients in whom the delay associated with mistriage of patients, with subsequent need for transfer, would be harmful. Only a small percentage of the patients included in our study met our definition of major trauma, as exemplified by a 3% mortality rate. This experience varies from other reports from populations taken only from major trauma centers, but instead reflects the population as seen from the untriaged viewpoint of an urban EMS system, consistent with our objective of studying these scoring systems from that vantage. Unfortunately, this leads to the weaknesses discussed above that preclude us from obtaining detailed outcome data on a large number of patients. As such, the results of this study should be viewed with regard to their applicability for prehospital care providers rather than as tools to be implemented towards the selected patient population seen in a trauma center. Finally, our EMS personnel did not calculate or use any of the scores during the period of study. All scores were calculated retrospectively, using the data available on the data sheets and run reports. Since we cannot be certain that both the data sheet and run reports were completed at the same time, and since these scores

were retrospectively calculated, this represents a potential source of bias in our study. Within the limitations of our study, as described above, we have found that the TTR performs well in identifying major trauma victims. In a previous study, we have demonstrated that EMT judgment performed as well as the PHI and the CRAMS scale and superior to the T-RTS. We cannot directly compare EMT judgment against the TTR, since in our original study the EMTs were not asked to make an absolute judgment on whether the patient should have been taken to a trauma center. It remains to be proven therefore whether implementation of the TTR would improve the triage of trauma patients to the appropriate hospital over that of independent EMT judgment. CONCLUSION We have verified that the Baxt TTR is an accurate method for the prehospital identification of major trauma victims who either die or require emergent general or neurosurgical operation. It performs similarly to the PHI and appears to be superior to the CRAMS scale in this function. Although it has not been directly compared against the EMT judgment, this validation suggests that it can be used as a means of identifying adult patients who should be transported to trauma centers. REFERENCES 1. Field Categorization of Trauma Patients in Resources for Optimal Care of the Injured Patient. Chicago, IL, American College of Surgeons, 1990 2. Kane G, Engelhardt R, Celentano J, et al: Empire development and evaluation of prehospital trauma triage instruments. J Trauma 1985;25:482-489 3. Baxt WG, Jones G, Fortlage D: The trauma triage rule: A new resource based approach to the prehospital identification of major trauma victims. Ann Emerg Med 1990;19:1401-1408 4. Gormican SP: CRAMS scale: Field triage of trauma victims. Ann Emerg Med 1982;11:132-135 5. Koehler JJ, Baer LJ, Malafa SA, et al: Prehospital index: A scoring system for field triage of trauma victims. Ann Emerg Med 1986;15:178-182 6. Champion HR, Sacco WJ, Copes WS, et al: A revision of the trauma score. J Trauma 1989;29:623-629 7. Emerman CL, Shade B, Kubincanek J: Comparison of EMT judgment and prehospital triage instruments. J Trauma 1991;31:1369-1375 8. Clemmer TP, Orme JF, Thomas F, et al: Prospective evaluation of the CRAMS scale for triaging major trauma. J Trauma 1985;25:188-191 9. Eichelberger MR, Gotschall CS, Sacco WJ, et al: A comparison of the trauma score, the revised trauma score, the pediatric trauma score. Ann Emerg Med 1989;18:1053-1058 10. Baxt WG, Berry CC, Epperson MD, et al: The failure of prehospital trauma prediction rules to classify trauma patients accurately. Ann Emerg Med 1989;18:1-8 11. Ornato J, Mlinek EJ, Crasen EJ, et al: Ineffectiveness of the trauma score and the CRAMS scale for triaging patients to trauma centers. Ann Emerg Med 1985;14:1061-1064 12. Hedges JR, Ferro S, Moore B, et al: Comparison of prehospital trauma triage instruments in a semirural population. J Emerg Med 1986;5:197-208 13. Koehler JJ, Malafa SA, Hillesend J, et al: A multicenter validation of the prehospital index. Ann Emerg Med 1987;16: 380-385

Comparative performance of the Baxt Trauma Triage Rule.

New methods of deciding which patients require trauma center transport continue to be devised. Baxt recently published a Trauma Triage Rule (TTR) usin...
568KB Sizes 0 Downloads 0 Views