- Email: [email protected]
Factors that predict the need for early surgeon presence in the setting of pediatric trauma
YJPSU-59224; No of Pages 4 Journal of Pediatric Surgery xxx (xxxx) xxx
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Factors that predict the need for early surgeon presence in the setting of pediatric trauma☆ Paul McGaha II ⁎, Tabitha Garwe, Kenneth Stewart, Zoona Sarwar, Justin Robbins, Jeremy Johnson, Robert W Letton University of Oklahoma Health Sciences Center, Oklahoma City, OK
a r t i c l e
i n f o
Article history: Received 13 February 2019 Received in revised form 9 May 2019 Accepted 11 May 2019 Available online xxxx Key words: Pediatric trauma Surgeon presence Injury severity score Trauma activation
a b s t r a c t Introduction: Evidence based variables predicting the need for surgeon presence (NSP) on arrival of an injured child are limited. We sought to identify prehospital factors that best correlate with NSP and highest level of activation in pediatric trauma. A secondary analysis was also performed to determine whether injury severity score (ISS) was predictive of NSP in pediatric trauma. Methods: This was a retrospective, single institution study of injured patients age ≤ 16 years delivered from scene to our Pediatric Level I trauma center between January 2016 and June 2017. 526 patients had complete data available for analysis. NSP was previously described as the presence of any of these factors: intubation, transfusion, emergent operation with the trauma team/craniotomy with the neurosurgery team, vasopressors, interventional radiology, spinal cord Injury, chest tube, emergency department thoracotomy, intracranial pressure monitor, pericardiocentesis, or death in the trauma bay. Multivariable analysis was performed with covariates of interest including scene and ED arrival vitals and interventions. Results: Independent predictors of NSP and highest level of activation were GCS of ≤12 (OR 22.3), penetrating trauma (OR 5.4), and hypotension (age adjusted) (OR 10.2). We also found that ISS ≥ 16 was a poor indicator of NSP with a sensitivity of only 61%. Conclusion: A validated model based on these variables may be useful in predicting NSP and highest level of activation prior to arrival of pediatric trauma patients. NSP may augment assessment of over and undertriage in pediatric trauma patients as compared to the ISS/Cribari system alone. Level of evidence Level III, retrospective cohort study © 2019 Elsevier Inc. All rights reserved.
Unintentional injury remains the leading cause of death in the pediatric population. According to the CDC, approximately 13,000 children 0 to 19 years of age die each year in the United States from an unintentional injury [1]. An estimated 9.2 million nonfatal unintentional injuries occur in the pediatric population annually. Many of these injuries will require delivery or transfer to a designated or verified trauma center and may lead to a trauma team activation. Currently, the guidelines
Abbreviations: NSP, Need for surgeon presence; ISS, injury severity score. ☆ Conflicts of Interest: The authors declare there are no conflicts of interest to report Presented: 5th Annual Pediatric Trauma Society Meeting in Houston, TX, November 8–10, 2018. Funding: Not-applicable. ⁎ Corresponding author at: 800 Stanton L Young BLVD, Suite 9000, Oklahoma City, OK, 73104. E-mail addresses: [email protected] (P. McGaha), [email protected] (T. Garwe), [email protected] (K. Stewart), [email protected] (Z. Sarwar), [email protected] (J. Robbins), [email protected] (J. Johnson), [email protected] (R.W. Letton).
for trauma activation assessment in pediatric trauma patients are mostly weighted on adult criteria and their appropriateness is based on the injury severity score (ISS) [2,3]. ISS was first reported as a scoring method in 1974 as a way to define severity of injury by scoring body regions and giving the injury certain weights [4]. The original study included only patients injured in motor vehicle collisions and did not account for age of the patient. ISS provides an overall score calculated from assigned Abbreviated Injury Scale (AIS) scores of six body regions (head and neck, face, chest, abdomen, extremity, and external injuries). This is calculated based on three most severely injured body regions using only the highest AIS score in each body region [4–6]. The ISS scoring system led to the development of the Cribari matrix which has been the standard tool for assessing undertriage and overtriage rates within a trauma system or for a trauma center. This matrix has been used by many trauma centers to assess the accuracy of their trauma activation. In this matrix an ISS of ≥16 is considered severe injury and is used as a threshold to define over and undertriage rates. The undertriage rate is calculated as the percentage of severely injured
https://doi.org/10.1016/j.jpedsurg.2019.05.010 0022-3468/© 2019 Elsevier Inc. All rights reserved.
Please cite this article as: P. McGaha, T. Garwe, K. Stewart, et al., Factors that predict the need for early surgeon presence in the setting of pediatric trauma, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.05.010
2
P. McGaha II et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
patients who did not receive a trauma activation and the overtriage rate is calculated as the percentage of patients who did receive an activation. The 2014 American College of Surgeons Committee on Trauma Optimal Resource Department defined the acceptable overtriage rate range as 25%–35% and the acceptable undertriage rate as 5% or less [7]. There are a few reasons why the use of ISS to assess over and undertriage in pediatric patients may be limited. First, in calculating the ISS, equal weights are given to the six body regions and the calculation does not account for multiple injuries to the same body region limiting its use in predicting trauma resource requirements. Second, ISS does not take into account the physiologic derangement of the patient. Finally, the Cribari methodology has only been validated in adult trauma populations and so it's unclear whether the Cribari ISS threshold is the optimal threshold for assessing triage accuracy in the pediatric population. However, since the introduction of the Cribari matrix the standard for evaluating over and undertriage has been based on this methodology for both pediatric and adult trauma. There have been previous studies evaluating triage criteria in pediatric trauma with particular attention paid to simplifying the complexities of the current standard of over and undertriage and relating the frequency of ‘high-level’ resources utilized by patients with certain criteria [7,8]. However, these studies did not perform multivariable analyses to identify independent criteria and are still based on the ISS system of over and undertriage, which is primarily based on criteria validated in the adult population [4,9,10]. In 2015, Lerner et al. suggested a series of factors defining the need for surgeon presence (NSP) in pediatric trauma. NSP refers to the specific trauma team including the trauma surgeon. There are other surgical subspecialties such as orthopedic surgery frequently needed in pediatric trauma care, but NSP relates specifically to that of the trauma team and trauma surgeon. NSP has been defined by the following factors: intubation, transfusion, operating room for hemorrhage control/ craniotomy, vasopressor requirement, interventional radiology, spinal cord injury, tube thoracostomy, emergency department thoracotomy, cesarean delivery, intracranial pressure monitor, pericardiocentesis, or death in the trauma bay [1]. The suggested criteria were based on expert opinion using a modified Delphi Survey technique. Despite previous work regarding pediatric trauma triage criteria, questions remain regarding accuracy and currently, no distinct, validated, universally applicable/accepted prehospital criteria for predicting which pediatric trauma patients would benefit from the highest level of trauma activation requiring NSP [9]. The purpose of our study was to identify easily recognizable, objective prehospital factors in pediatric trauma patients independently predictive of NSP. 1. Methods This was a retrospective, prognostic research study of pediatric patients age ≤ 16 years transported directly from the scene to our Level I trauma center between January 2016 and June 2017. A total of 526 patients were included in our study. We collected the data via chart review of prehospital emergency personnel trauma run-sheets in combination with the documentation obtained in the trauma bay. The prognostic outcome of interest was NSP, defined as having any one of the following: intubation (whether in the field or immediately upon arrival in the trauma bay), transfusion within four hours of arrival, emergent operation with the trauma team/craniotomy with the neurosurgery team, vasopressor requirement, interventional radiology, spinal cord injury, tube thoracostomy placed, emergency department thoracotomy, urgent need for intracranial pressure monitor, pericardiocentesis, or death in the trauma bay. Demographic and clinical variables considered as predictors for the prognostic model included age, gender, mechanism of injury and prehospital vital signs. Prehospital vitals included: heart rate (HR), systolic blood pressure (SBP), respiratory rate (RR), and Glasgow Coma Scale (GCS). If a patient did not have prehospital vitals recorded, initial arrival vitals were substituted in the place of the missing vital.
Testing differences in prehospital and arrival vitals for SBP and HR was done using paired t-tests. Both were nonsignificant with p-values of 0.11 and 0.16. Weighted kappa for GCS was .77 indicating substantial agreement of prehospital and arrival GCS. Arrival vitals obtained included: arrival HR, arrival SBP, arrival RR, and arrival GCS. Moderate or severe traumatic brain injury, was defined as a GCS score ≤ 12 [11]. Similarly, hypotension was stratified by age with a systolic blood pressure at age b 1 ≤ 60 mmHg; ages1–2 ≤ 70 mmHg; ages 3–5 ≤ 75 mmHg; ages 6–12 ≤ 80 mmHg; ages 13–16 ≤ 90 mmHg [7]. Tachycardia was stratified by age with 0 to b2 years old N190 beats per minute (BPM), 2 to 10 years old N 140 BPM, and N 10 years old N100 BPM [12]. High RR was defined as a RR of N20. Shock Index Pediatric-Adjusted (SIPA) was calculated by dividing heart rate with systolic blood pressure and thresholds for unfavorable scores adjusted based on specific pediatric age groups [13]. SIPA has been define in previous literature as heart rate/SBP, and is age adjusted with maximum normal values. For age 4–6 the maximum normal value is 1.22. For age 6–12 the maximum normal value is 1. For age greater than 12 the maximum normal value is 0.9 [13]. In our analysis, combined SIPA and GCS were calculated as follow: SIPA negative and mild GCS ≥ 13, SIPA positive and mild GCS ≥ 13, SIPA negative and moderate to severe GCS ≤ 12, and SIPA positive and GCS ≤ 12. 1.1. Statistical analyses Means and proportions were used to summarize the data by NSP status. Unadjusted comparisons between the two NSP groups were performed using the independent Student's t-test or Mann–Whitney Wilcoxon test for continuous variables, and for categorical variables, the chi-square or Fisher's exact tests were used. Multivariable logistic regression was used to identify independent predictors of NSP. Variables significant on univariate analysis were considered for multivariable modeling as well as variables determined a priori to be key potential predictors or modifiers e.g. age. Heart rate and systolic blood pressure were considered both as individual variables as well as using the combined SIPA variable in the multivariable model. In terms of multivariable model assessment, the C-statistic was used to assess how well the model discriminated between the two NSP groups, and the Hosmer–Lemeshow goodness-of-fit test assessed the final model's overall adequacy of fit with the data. Secondary analyses assessing the validity of different ISS thresholds in differentiating between NSP and non-NSP patients were also performed. Using NSP as the reference standard, we calculated sensitivity, specificity, positive and negative predictive values for two ISS thresholds. We considered an ISS ≥ 16 threshold based on the accepted traditional definition of severe injury as well an ISS threshold of 12 or greater, an optimal cutoff point which maximized sensitivity and specificity in differentiating NSP status in our data. An alpha of 0.05 was used to determine all statistical significance. All analyses were performed using SAS software version 9.4 (SAS 9.4, SAS Institute, Cary, NC). 2. Results Table 1 summarizes demographic and clinical characteristics by NSP status for the 526 patients. No significant differences (p N 0.05) were noted in the distribution of age, gender, and race. However, patients with NSP were significantly (p b 0.05) more likely to have a penetrating injury, lower GCS scores, hypotension, tachycardia, positive SIPA scores (indicating pediatric shock), and higher respiratory rate. Covariates initially considered for the multivariable model included age, mechanism of injury, prehospital vital signs including the SIPA score variables. Upon multivariable analysis, three significant independent predictors were retained in the final model (Table 2). These included TBI, age-adjusted hypotension, and penetrating trauma. In terms of the independent predictors of NSP identified in the final model, the odds of NSP in patients with a traumatic brain injury (GCS
Please cite this article as: P. McGaha, T. Garwe, K. Stewart, et al., Factors that predict the need for early surgeon presence in the setting of pediatric trauma, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.05.010
P. McGaha II et al. / Journal of Pediatric Surgery xxx (xxxx) xxx Table 1 Comparison of patient demographic and clinical characteristics by NSP. Variable
Surgeon Needed n = 72
Age, Mean (±SD) Age groups
10.6 (5.3) ≤6 yrs. 17 (23.6) 7–12 yrs. 19 (26.4) 13–16 yrs. 36 (50.0)
Surgeon not Needed n = 454 9.5 (5.1)
Table 3 The Standard Cribari matrix, NSP matrix, and combined NSP/ISS matrix.a
0.0905 0.1795
a
0.0949 Male 48 (66.7) Female 24 (33.3)
254 (56.2) 198 (43.8)
Race, n (%) ** 1 missing White Black American Indian Asian Other Mechanism of Injury, n (%) **1 missing Traffic-related Gun Shot Wound/Stabbing Falls Other Injury Type, n (%) Blunt Penetrating Burn Penetrating Trauma Traumatic Brain Injury b13 GCS ≥13 GCS c Systolic Blood Pressure (Age Adjusted) **1 missing Hypotension Normal Heart Rate, Mean (±SD) c Tachycardia, n (%) c Respiratory Rate (ED RR and age group), n (%) Normal Abnormal RR SIPA, n (%) SIPA and GCS, n (%) SIPA (−) and GCS as mild SIPA (+) and GCS is Mild SIPA (−) and GCS is moderate or severe SIPA (+) and GCS as moderate or severe
ISS 1–14
p-value
141 (31.2) 136 (30.1) 175 (38.7)
Gender, n (%)
Full Trauma Activation 1 Limited or No Activation 2–3 Undertriage Rate Overtriage Rate
272 (60.0) 92 (20.3) 16 (3.6) 4 (0.9) 69 (15.2) b.0001
40 (56.4) 15 (21.1) 4 (5.6) 12 (16.9)
255 (56.2) 21 (4.6) 96 (21.2) 82 (18.0)
53 (73.6) 17 (23.6) 2 (2.8) 17 (22.4)
412 (90.8) 33 (7.3) 9 (2.0) 33 (7.4)
41 (56.9) 35 (43.1)
23(5.1) 431 (94.9)
b.0001
b.0001 b.0001 b.0001b
19 (26.4) 53 (73.6) 91.7 (41.4) 22 (30.6)
7 (1.6) 444 (98.5) 101.4 (25.1) 82 (18.1)
37 (51.4) 35 (48.6) 24 (33.3)
314 (69.2) 140 (30.8) 84 (18.5)
22 (30.6) 9 (12.5) 26 (36.1) 15 (20.8)
356 (78.4) 75 (16.5) 14 (3.1) 9 (2.0)
0.0583a 0.0134 0.0029
0.0038 b.0001
a
p-value for Student's T-Test. p-value for Fishers exact test. c Abnormal RR [Low RR (b 3 years old RR b24, 3 to 5 years old RR b22, 6 to 10 years old RR b18, and N 10 years old RR b12). Tachypnea (b 3 years old RR N40, 3 to 5 years old RR N34, 6 to 10 years old RR N30, and N 10 years old RR N20)]. Tachycardia was stratified by age with 0 to b2 years old N190 (BPM), 2 to 10 years old N140 BPM, and N 10 years old N100 BPM. Systolic Blood Pressure was stratified by age b 1 ≤ 60; ages1–2 ≤ 70; ages 3–5 ≤ 75; ages 6–12 ≤ 80; ages 13–16 ≤ 90. b
≤12) were 22 times those of patients having GCS ≥13 (OR: 22.3, 95% CI: 11.1–44.7). Similarly the odds of NSP in a hypotensive (age adjusted) patient were 10.2 times (OR 10.2, 95% CI: 3.2–32.4) those of normotensive patients. Finally, the odds of NSP for patients presenting with gunshot injuries, stabbing and other penetrating injuries were 5.4 times (OR Table 2 Independent predictors of need for surgeon presence in pediatric trauma patients.a Effect
OR (95% CI)
p-value
Traumatic Brain Injury (GCS b 13) Hypotension (aAge Adjusted) Penetrating Trauma
22.3 (11.1–44.7) 10.2 (3.2–32.4) 5.4 (2.4–12.3)
b.0001 b.0001 0.0001
Hosmer and Lemeshow Goodness-of-Fit Test p value = 0.65. AUC (C-statistic) = 0.83. a Age-adjusted hypotension defined as follows based on systolic blood pressure (SBP): age b 1 = SBP ≤ 60; ages 1–2 = SBP ≤ 70; ages 3–5 = SBP ≤ 75; ages 6–12 = SBP ≤ 80; ages 13–16 = SBP ≤90.
49 400 7.0% 55.1% No NSP
Full Trauma Activation 1 Limited or No Activation 2–3 Undertriage Rate Overtriage Rate
37 417 4.6% 41.6% No NSP or ISS b 15
Full Trauma Activation 1 Limited or No Activation 2–3 Undertriage Rate Overtriage Rate
31 388 9.3% 34.8%
0.7116 41 (56.9) 16 (22.2) 1 (1.4) 1 (1.4) 13 (18.1)
3
a
ISS 16–75
Total
40 30
89 430
Ideally ≤5% Ideally ≤50% NSP Total 52 20
89 437
Ideally ≤5% Ideally ≤50% NSP or ISS ≥ 16 Total 58 89 40 428 Ideally ≤5% Ideally ≤50%
Missing ISS for 7 patients.
5.4, 95% CI: 2.-12.3) those of patients with blunt injuries. The sensitivity of ISS in predicting NSP is only 61%. We then performed an analysis using the Standard Cribari Matrix, our created NSP matrix, and combined NSP/ISS Matrix. The NSP/ISS combined matrix consists of patient with any 1 NSP variable or and ISS of ≥16.Table 3 shows that the standard Cribari matrix of over and undertriage for our data set demonstrates undertriage rate of 7.0% and an overtriage rate of 55.1%. The NSP matrix demonstrates an undertriage rate of 4.6% and an overtriage rate of 41.6%. Finally, the combined NSP/ISS matrix demonstrated an undertriage rate of 9.3% and an overtriage rate of 34.8%. 3. Discussion Recently, it has been suggested that surgeon presence may not be needed for lower level pediatric trauma resuscitations [14]. Unnecessary highest level activation can lead to a stretching of resources which may drive up hospital costs as well as decrease the quality of patient care. Seriously injured patients resuscitated by an organized trauma team with surgeon presence allows for faster time to CT scan, shorter emergency department times, and shorter time to transport to the operating room [15]. Therefore, identifying pediatric trauma patients with an immediate need for a surgeon's presence upon arrival is critical, and makes the proper evaluation of over and undertriage paramount in the care of the pediatric trauma patient. The goal of our study was to analyze our prehospital data and develop a model which reveals the factors most predictive of NSP and highest level of activation. Development of a prehospital model predictive of NSP may allow improved accuracy of assessment of over and undertriage. Our study included collection of initial prehospital values which allowed us to assess their association for NSP independently of one another. We analyzed a host of factors in a multivariable fashion with moderate to severe traumatic brain injury, SBP, and penetrating trauma being the factors most significantly associated with NSP and highest level of activation. It is interesting the 3 factors identified here were in similarly represented among the 8 factors proposed by Falcone et al. in 2012, however multivariable analyses of those factors was not presented. It is possible multivariable adjustment would have arrived at a reduced model even more similar to that presented here. The Cribari Matrix, which has been widely adopted to determine appropriateness of triage, uses ISS in assessing accuracy of over and undertriage. Severe injury is defined as an ISS ≥ 16 as it was found to predict a 10% mortality in adult trauma patients [16]. Several potential limitations exist using this system. ISS does not take into account physiologic derangement or clinical judgment of the treating physician. The ISS score weighs injuries equally (i.e. head and extremity injuries) when there has been a significant amount of evidence that this is not entirely
Please cite this article as: P. McGaha, T. Garwe, K. Stewart, et al., Factors that predict the need for early surgeon presence in the setting of pediatric trauma, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.05.010
4
P. McGaha II et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
accurate in the pediatric trauma population. A pediatric patient with an intracranial hemorrhage may have an ISS b 16 but may need significant attention, while a patient with long bone fractures may have an ISS N 14 and be much more stable clinically [10]. Since traumatic brain injury is largely responsible for morbidity and mortality in the pediatric trauma population, ISS based systems can lead to significant undertriage [17]. Currently, ISS is not sufficient to in adequately provide an assessment of over and undertriage in the pediatric trauma patient. Our data suggest the use of only ISS dependent triage matrices may markedly underestimate the rate of undertriage. As shown in Table 3, combining ISS and NSP shows a 4% increase in our undertriage rate (undertriage rate of 9.3% when ISS/NSP is combined). This suggests a substantial portion of patients are undertriaged and not accounted for in the currently utilized systems. In reviewing the undertriaged patients, there were multiple patients who were undertriaged by the Cribari matrix. Many of these patient were intubated secondary to decreased GCS. This is a flaw in the current standard for over/undertriage as many times head injuries have a low scoring ISS in spite of their severity. There were one patient who needed an emergent decompressive craniotomy, one patient who had a suspected spinal cord injury, and multiple patients who needed emergent operative intervention for neck lacerations. In our study, ISS demonstrated only a 61% sensitivity for NSP. This also suggests ISS-based over and undertriage measures may not be predictive of true injury severity, NSP, and highest level of activation. In addition, the inability to calculate the ISS until all injuries have been identified, limits the utility of ISS in assessing concurrent over and undertriage accuracy in the pediatric trauma patient. NSP may allow for a more concurrent and complete assessment of over and undertriage in the pediatric trauma patient than the ISS alone based Cribari Matrix. These factors suggest that adding a NSP matrix which is related to physiologic derangements of trauma patients may augment the ISS system. Shock Index Pediatric Age-adjusted (SIPA) has been previously shown to predict severe injury based on ISS [13,18]. Although the use of SIPA was somewhat predictive of NSP in univariable analysis, it was not when subjected to multivariable modeling. This is likely because changes in GCS independent of SIPA were so strong in predicting NSP that SIPA ultimately fell out of the final model. Identifying prehospital factors that predict NSP may allow for a more appropriate trauma activation of pediatric trauma patients. In addition, accurate assessment of over and undertriage is essential to allow for optimal resource utilization by the medical facility and ultimately benefits the patient, physicians, and hospital. Our data suggest that assessing the appropriateness of triage based on ISS alone may not accurately reflect the rate of undertriage in the pediatric trauma patient. Use of the model itself may aid in field evaluation of pediatric trauma patients. Having a patient with any one of the following prehospital characteristics would benefit from the highest level of activation: GCS ≤12, a low SBP as defined by the model, and the victim of penetrating trauma. 3.1. Limitations The first limitation of this study is that it was a retrospective study and at a single institution. This limits the extrapolation of this model currently; however, we are planning a multi-institutional study in order to externally validate this data. Additionally, owing to the difficulty in tracking the presence of subjective, mechanistic activation criteria in the prehospital reports, only objective data were examined. Perhaps other mechanistic criteria impact NSP and highest level of activation as well [19]. We also plan to validate the model for prehospital predictive factors involved in predicting NSP and highest level of activation through a prospective study at our institution, which is currently ongoing. This will ensure that almost all prehospital vital signs and NSP factors are captured, as well as mechanistic variables, thereby increasing the validity of the study. In our prospective study, we are also collecting data on
transfers and the NSP; as this study only included patients that came to our facility directly from the scene of injury. Finally, like ISS, NSP is somewhat retrospective in nature, and limited with respect to certain types of injuries like eye, oral–maxillofacial injuries, tendon/nerve injury in hands/extremities, laceration/incisive wounds in conspicuous places. Also, it may miss nonaccidental trauma patients where specialist intervention may be needed after resuscitation and admission to the hospital. Furthermore, NSP does not necessarily mean an operation is needed by a trauma surgeon, rather, it is a marker for pediatric trauma patients requiring the highest level of care, even if the end results is nonoperative. 4. Conclusion A validated model of factors predictive of NSP and highest level of activation in pediatric trauma patients has yet to be developed. NSP can be predicted based on the following factors: GCS ≤ 12, penetrating trauma, and age adjusted hypotension. Such a prognostic prehospital model may allow for an improved assessment of pediatric trauma activation criteria. Our data also suggest that a triage matrix based on NSP may augment assessment of over and undertriage in the pediatric trauma patient as compared to the ISS/Cribari system alone. References [1] Lerner EB, Drendel AL, Falcone Jr RA, et al. A consensus-based criterion standard definition for pediatric patients who needed the highest-level trauma team activation. J Trauma Acute Care Surg 2015;78(3):634–8. [2] Brown JB, Gestring ML, Leeper CM, et al. The value of the injury severity score in pediatric trauma: time for a new definition of severe injury? J Trauma Acute Care Surg 2017;82(6):995–1001. [3] Rogers AT, Gross BW, Cook AD, et al. Outcome differences in adolescent blunt severe polytrauma patients managed at pediatric versus adult trauma centers. J Trauma Acute Care Surg 2017;83(6):1082–7. [4] Baker SP, O'Neill B, Haddon Jr W, et al. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 1974;14(3):187–96. [5] Copes WS, Champion HR, Sacco WJ, et al. The injury severity score revisited. J Trauma 1988;28(1):69–77. [6] Beverland DE, Rutherford WH. An assessment of the validity of the injury severity score when applied to gunshot wounds. Injury 1983;15(1):19–22. [7] Falcone Jr RA, Haas L, King E, et al. A multicenter prospective analysis of pediatric trauma activation criteria routinely used in addition to the six criteria of the American College of Surgeons. J Trauma Acute Care Surg 2012;73(2):377–84 [discussion 84]. [8] Chen LE, Snyder AK, Minkes RK, et al. Trauma stat and trauma minor: are we making the call appropriately? Pediatr Emerg Care 2004;20(7):421–5. [9] van der Sluijs R, van Rein EAJ, Wijnand JGJ, et al. Accuracy of pediatric trauma field triage: a systematic review. JAMA Surg 2018;153(7):671–6. [10] Roden-Foreman JW, Rapier NR, Yelverton L, et al. Avoiding Cribari gridlock: the standardized triage assessment tool improves the accuracy of the Cribari matrix method in identifying potential overtriage and undertriage. J Trauma Acute Care Surg 2018; 84(5):718–26. [11] ATLS. 10th edition of the Advanced Trauma Life Support® (ATLS®) student course manual. Chicago. IL: American College of Surgeons; 2018. [12] Berg MD, Schexnayder SM, Chameides L, et al. Part 13: pediatric basic life support: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;122(18 suppl 3):S862–75. [13] Acker SN, Ross JT, Partrick DA, et al. Pediatric specific shock index accurately identifies severely injured children. J Pediatr Surg 2015;50(2):331–4. [14] Boomer LA, Nielsen JW, Lowell W, et al. Managing moderately injured pediatric patients without immediate surgeon presence: 10 years later. J Pediatr Surg 2015;50(1):182–5. [15] Vernon DD, Furnival RA, Hansen KW, et al. Effect of a pediatric trauma response team on emergency department treatment time and mortality of pediatric trauma victims. Pediatrics 1999;103(1):20–4. [16] Boyd CR, Tolson MA, Copes WS. Evaluating trauma care: the TRISS method. Trauma Score and the Injury Severity Score J Trauma 1987;27(4):370–8. [17] DiBrito SR, Cerullo M, Goldstein SD, et al. Reliability of Glasgow Coma Score in pediatric trauma patients. J Pediatr Surg 2018;53(9):1789–94. [18] Nordin A, Coleman A, Shi J, et al. Validation of the age-adjusted shock index using pediatric trauma quality improvement program data. J Pediatr Surg 2017 [Oct 12. pii: S0022-3468(17)30645-0]. [19] Recicar J, Barczyk A, Duzinski S, et al. Does restraint status in motor vehicle crash with rollover predict the need for trauma team presence on arrival? An ATOMAC study. J Pediatr Surg 2016;51(2):319–22.
Please cite this article as: P. McGaha, T. Garwe, K. Stewart, et al., Factors that predict the need for early surgeon presence in the setting of pediatric trauma, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.05.010
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Factors that predict the need for early surgeon presence in the setting of pediatric trauma☆ Paul McGaha II ⁎, Tabitha Garwe, Kenneth Stewart, Zoona Sarwar, Justin Robbins, Jeremy Johnson, Robert W Letton University of Oklahoma Health Sciences Center, Oklahoma City, OK
a r t i c l e
i n f o
Article history: Received 13 February 2019 Received in revised form 9 May 2019 Accepted 11 May 2019 Available online xxxx Key words: Pediatric trauma Surgeon presence Injury severity score Trauma activation
a b s t r a c t Introduction: Evidence based variables predicting the need for surgeon presence (NSP) on arrival of an injured child are limited. We sought to identify prehospital factors that best correlate with NSP and highest level of activation in pediatric trauma. A secondary analysis was also performed to determine whether injury severity score (ISS) was predictive of NSP in pediatric trauma. Methods: This was a retrospective, single institution study of injured patients age ≤ 16 years delivered from scene to our Pediatric Level I trauma center between January 2016 and June 2017. 526 patients had complete data available for analysis. NSP was previously described as the presence of any of these factors: intubation, transfusion, emergent operation with the trauma team/craniotomy with the neurosurgery team, vasopressors, interventional radiology, spinal cord Injury, chest tube, emergency department thoracotomy, intracranial pressure monitor, pericardiocentesis, or death in the trauma bay. Multivariable analysis was performed with covariates of interest including scene and ED arrival vitals and interventions. Results: Independent predictors of NSP and highest level of activation were GCS of ≤12 (OR 22.3), penetrating trauma (OR 5.4), and hypotension (age adjusted) (OR 10.2). We also found that ISS ≥ 16 was a poor indicator of NSP with a sensitivity of only 61%. Conclusion: A validated model based on these variables may be useful in predicting NSP and highest level of activation prior to arrival of pediatric trauma patients. NSP may augment assessment of over and undertriage in pediatric trauma patients as compared to the ISS/Cribari system alone. Level of evidence Level III, retrospective cohort study © 2019 Elsevier Inc. All rights reserved.
Unintentional injury remains the leading cause of death in the pediatric population. According to the CDC, approximately 13,000 children 0 to 19 years of age die each year in the United States from an unintentional injury [1]. An estimated 9.2 million nonfatal unintentional injuries occur in the pediatric population annually. Many of these injuries will require delivery or transfer to a designated or verified trauma center and may lead to a trauma team activation. Currently, the guidelines
Abbreviations: NSP, Need for surgeon presence; ISS, injury severity score. ☆ Conflicts of Interest: The authors declare there are no conflicts of interest to report Presented: 5th Annual Pediatric Trauma Society Meeting in Houston, TX, November 8–10, 2018. Funding: Not-applicable. ⁎ Corresponding author at: 800 Stanton L Young BLVD, Suite 9000, Oklahoma City, OK, 73104. E-mail addresses: [email protected] (P. McGaha), [email protected] (T. Garwe), [email protected] (K. Stewart), [email protected] (Z. Sarwar), [email protected] (J. Robbins), [email protected] (J. Johnson), [email protected] (R.W. Letton).
for trauma activation assessment in pediatric trauma patients are mostly weighted on adult criteria and their appropriateness is based on the injury severity score (ISS) [2,3]. ISS was first reported as a scoring method in 1974 as a way to define severity of injury by scoring body regions and giving the injury certain weights [4]. The original study included only patients injured in motor vehicle collisions and did not account for age of the patient. ISS provides an overall score calculated from assigned Abbreviated Injury Scale (AIS) scores of six body regions (head and neck, face, chest, abdomen, extremity, and external injuries). This is calculated based on three most severely injured body regions using only the highest AIS score in each body region [4–6]. The ISS scoring system led to the development of the Cribari matrix which has been the standard tool for assessing undertriage and overtriage rates within a trauma system or for a trauma center. This matrix has been used by many trauma centers to assess the accuracy of their trauma activation. In this matrix an ISS of ≥16 is considered severe injury and is used as a threshold to define over and undertriage rates. The undertriage rate is calculated as the percentage of severely injured
https://doi.org/10.1016/j.jpedsurg.2019.05.010 0022-3468/© 2019 Elsevier Inc. All rights reserved.
Please cite this article as: P. McGaha, T. Garwe, K. Stewart, et al., Factors that predict the need for early surgeon presence in the setting of pediatric trauma, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.05.010
2
P. McGaha II et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
patients who did not receive a trauma activation and the overtriage rate is calculated as the percentage of patients who did receive an activation. The 2014 American College of Surgeons Committee on Trauma Optimal Resource Department defined the acceptable overtriage rate range as 25%–35% and the acceptable undertriage rate as 5% or less [7]. There are a few reasons why the use of ISS to assess over and undertriage in pediatric patients may be limited. First, in calculating the ISS, equal weights are given to the six body regions and the calculation does not account for multiple injuries to the same body region limiting its use in predicting trauma resource requirements. Second, ISS does not take into account the physiologic derangement of the patient. Finally, the Cribari methodology has only been validated in adult trauma populations and so it's unclear whether the Cribari ISS threshold is the optimal threshold for assessing triage accuracy in the pediatric population. However, since the introduction of the Cribari matrix the standard for evaluating over and undertriage has been based on this methodology for both pediatric and adult trauma. There have been previous studies evaluating triage criteria in pediatric trauma with particular attention paid to simplifying the complexities of the current standard of over and undertriage and relating the frequency of ‘high-level’ resources utilized by patients with certain criteria [7,8]. However, these studies did not perform multivariable analyses to identify independent criteria and are still based on the ISS system of over and undertriage, which is primarily based on criteria validated in the adult population [4,9,10]. In 2015, Lerner et al. suggested a series of factors defining the need for surgeon presence (NSP) in pediatric trauma. NSP refers to the specific trauma team including the trauma surgeon. There are other surgical subspecialties such as orthopedic surgery frequently needed in pediatric trauma care, but NSP relates specifically to that of the trauma team and trauma surgeon. NSP has been defined by the following factors: intubation, transfusion, operating room for hemorrhage control/ craniotomy, vasopressor requirement, interventional radiology, spinal cord injury, tube thoracostomy, emergency department thoracotomy, cesarean delivery, intracranial pressure monitor, pericardiocentesis, or death in the trauma bay [1]. The suggested criteria were based on expert opinion using a modified Delphi Survey technique. Despite previous work regarding pediatric trauma triage criteria, questions remain regarding accuracy and currently, no distinct, validated, universally applicable/accepted prehospital criteria for predicting which pediatric trauma patients would benefit from the highest level of trauma activation requiring NSP [9]. The purpose of our study was to identify easily recognizable, objective prehospital factors in pediatric trauma patients independently predictive of NSP. 1. Methods This was a retrospective, prognostic research study of pediatric patients age ≤ 16 years transported directly from the scene to our Level I trauma center between January 2016 and June 2017. A total of 526 patients were included in our study. We collected the data via chart review of prehospital emergency personnel trauma run-sheets in combination with the documentation obtained in the trauma bay. The prognostic outcome of interest was NSP, defined as having any one of the following: intubation (whether in the field or immediately upon arrival in the trauma bay), transfusion within four hours of arrival, emergent operation with the trauma team/craniotomy with the neurosurgery team, vasopressor requirement, interventional radiology, spinal cord injury, tube thoracostomy placed, emergency department thoracotomy, urgent need for intracranial pressure monitor, pericardiocentesis, or death in the trauma bay. Demographic and clinical variables considered as predictors for the prognostic model included age, gender, mechanism of injury and prehospital vital signs. Prehospital vitals included: heart rate (HR), systolic blood pressure (SBP), respiratory rate (RR), and Glasgow Coma Scale (GCS). If a patient did not have prehospital vitals recorded, initial arrival vitals were substituted in the place of the missing vital.
Testing differences in prehospital and arrival vitals for SBP and HR was done using paired t-tests. Both were nonsignificant with p-values of 0.11 and 0.16. Weighted kappa for GCS was .77 indicating substantial agreement of prehospital and arrival GCS. Arrival vitals obtained included: arrival HR, arrival SBP, arrival RR, and arrival GCS. Moderate or severe traumatic brain injury, was defined as a GCS score ≤ 12 [11]. Similarly, hypotension was stratified by age with a systolic blood pressure at age b 1 ≤ 60 mmHg; ages1–2 ≤ 70 mmHg; ages 3–5 ≤ 75 mmHg; ages 6–12 ≤ 80 mmHg; ages 13–16 ≤ 90 mmHg [7]. Tachycardia was stratified by age with 0 to b2 years old N190 beats per minute (BPM), 2 to 10 years old N 140 BPM, and N 10 years old N100 BPM [12]. High RR was defined as a RR of N20. Shock Index Pediatric-Adjusted (SIPA) was calculated by dividing heart rate with systolic blood pressure and thresholds for unfavorable scores adjusted based on specific pediatric age groups [13]. SIPA has been define in previous literature as heart rate/SBP, and is age adjusted with maximum normal values. For age 4–6 the maximum normal value is 1.22. For age 6–12 the maximum normal value is 1. For age greater than 12 the maximum normal value is 0.9 [13]. In our analysis, combined SIPA and GCS were calculated as follow: SIPA negative and mild GCS ≥ 13, SIPA positive and mild GCS ≥ 13, SIPA negative and moderate to severe GCS ≤ 12, and SIPA positive and GCS ≤ 12. 1.1. Statistical analyses Means and proportions were used to summarize the data by NSP status. Unadjusted comparisons between the two NSP groups were performed using the independent Student's t-test or Mann–Whitney Wilcoxon test for continuous variables, and for categorical variables, the chi-square or Fisher's exact tests were used. Multivariable logistic regression was used to identify independent predictors of NSP. Variables significant on univariate analysis were considered for multivariable modeling as well as variables determined a priori to be key potential predictors or modifiers e.g. age. Heart rate and systolic blood pressure were considered both as individual variables as well as using the combined SIPA variable in the multivariable model. In terms of multivariable model assessment, the C-statistic was used to assess how well the model discriminated between the two NSP groups, and the Hosmer–Lemeshow goodness-of-fit test assessed the final model's overall adequacy of fit with the data. Secondary analyses assessing the validity of different ISS thresholds in differentiating between NSP and non-NSP patients were also performed. Using NSP as the reference standard, we calculated sensitivity, specificity, positive and negative predictive values for two ISS thresholds. We considered an ISS ≥ 16 threshold based on the accepted traditional definition of severe injury as well an ISS threshold of 12 or greater, an optimal cutoff point which maximized sensitivity and specificity in differentiating NSP status in our data. An alpha of 0.05 was used to determine all statistical significance. All analyses were performed using SAS software version 9.4 (SAS 9.4, SAS Institute, Cary, NC). 2. Results Table 1 summarizes demographic and clinical characteristics by NSP status for the 526 patients. No significant differences (p N 0.05) were noted in the distribution of age, gender, and race. However, patients with NSP were significantly (p b 0.05) more likely to have a penetrating injury, lower GCS scores, hypotension, tachycardia, positive SIPA scores (indicating pediatric shock), and higher respiratory rate. Covariates initially considered for the multivariable model included age, mechanism of injury, prehospital vital signs including the SIPA score variables. Upon multivariable analysis, three significant independent predictors were retained in the final model (Table 2). These included TBI, age-adjusted hypotension, and penetrating trauma. In terms of the independent predictors of NSP identified in the final model, the odds of NSP in patients with a traumatic brain injury (GCS
Please cite this article as: P. McGaha, T. Garwe, K. Stewart, et al., Factors that predict the need for early surgeon presence in the setting of pediatric trauma, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.05.010
P. McGaha II et al. / Journal of Pediatric Surgery xxx (xxxx) xxx Table 1 Comparison of patient demographic and clinical characteristics by NSP. Variable
Surgeon Needed n = 72
Age, Mean (±SD) Age groups
10.6 (5.3) ≤6 yrs. 17 (23.6) 7–12 yrs. 19 (26.4) 13–16 yrs. 36 (50.0)
Surgeon not Needed n = 454 9.5 (5.1)
Table 3 The Standard Cribari matrix, NSP matrix, and combined NSP/ISS matrix.a
0.0905 0.1795
a
0.0949 Male 48 (66.7) Female 24 (33.3)
254 (56.2) 198 (43.8)
Race, n (%) ** 1 missing White Black American Indian Asian Other Mechanism of Injury, n (%) **1 missing Traffic-related Gun Shot Wound/Stabbing Falls Other Injury Type, n (%) Blunt Penetrating Burn Penetrating Trauma Traumatic Brain Injury b13 GCS ≥13 GCS c Systolic Blood Pressure (Age Adjusted) **1 missing Hypotension Normal Heart Rate, Mean (±SD) c Tachycardia, n (%) c Respiratory Rate (ED RR and age group), n (%) Normal Abnormal RR SIPA, n (%) SIPA and GCS, n (%) SIPA (−) and GCS as mild SIPA (+) and GCS is Mild SIPA (−) and GCS is moderate or severe SIPA (+) and GCS as moderate or severe
ISS 1–14
p-value
141 (31.2) 136 (30.1) 175 (38.7)
Gender, n (%)
Full Trauma Activation 1 Limited or No Activation 2–3 Undertriage Rate Overtriage Rate
272 (60.0) 92 (20.3) 16 (3.6) 4 (0.9) 69 (15.2) b.0001
40 (56.4) 15 (21.1) 4 (5.6) 12 (16.9)
255 (56.2) 21 (4.6) 96 (21.2) 82 (18.0)
53 (73.6) 17 (23.6) 2 (2.8) 17 (22.4)
412 (90.8) 33 (7.3) 9 (2.0) 33 (7.4)
41 (56.9) 35 (43.1)
23(5.1) 431 (94.9)
b.0001
b.0001 b.0001 b.0001b
19 (26.4) 53 (73.6) 91.7 (41.4) 22 (30.6)
7 (1.6) 444 (98.5) 101.4 (25.1) 82 (18.1)
37 (51.4) 35 (48.6) 24 (33.3)
314 (69.2) 140 (30.8) 84 (18.5)
22 (30.6) 9 (12.5) 26 (36.1) 15 (20.8)
356 (78.4) 75 (16.5) 14 (3.1) 9 (2.0)
0.0583a 0.0134 0.0029
0.0038 b.0001
a
p-value for Student's T-Test. p-value for Fishers exact test. c Abnormal RR [Low RR (b 3 years old RR b24, 3 to 5 years old RR b22, 6 to 10 years old RR b18, and N 10 years old RR b12). Tachypnea (b 3 years old RR N40, 3 to 5 years old RR N34, 6 to 10 years old RR N30, and N 10 years old RR N20)]. Tachycardia was stratified by age with 0 to b2 years old N190 (BPM), 2 to 10 years old N140 BPM, and N 10 years old N100 BPM. Systolic Blood Pressure was stratified by age b 1 ≤ 60; ages1–2 ≤ 70; ages 3–5 ≤ 75; ages 6–12 ≤ 80; ages 13–16 ≤ 90. b
≤12) were 22 times those of patients having GCS ≥13 (OR: 22.3, 95% CI: 11.1–44.7). Similarly the odds of NSP in a hypotensive (age adjusted) patient were 10.2 times (OR 10.2, 95% CI: 3.2–32.4) those of normotensive patients. Finally, the odds of NSP for patients presenting with gunshot injuries, stabbing and other penetrating injuries were 5.4 times (OR Table 2 Independent predictors of need for surgeon presence in pediatric trauma patients.a Effect
OR (95% CI)
p-value
Traumatic Brain Injury (GCS b 13) Hypotension (aAge Adjusted) Penetrating Trauma
22.3 (11.1–44.7) 10.2 (3.2–32.4) 5.4 (2.4–12.3)
b.0001 b.0001 0.0001
Hosmer and Lemeshow Goodness-of-Fit Test p value = 0.65. AUC (C-statistic) = 0.83. a Age-adjusted hypotension defined as follows based on systolic blood pressure (SBP): age b 1 = SBP ≤ 60; ages 1–2 = SBP ≤ 70; ages 3–5 = SBP ≤ 75; ages 6–12 = SBP ≤ 80; ages 13–16 = SBP ≤90.
49 400 7.0% 55.1% No NSP
Full Trauma Activation 1 Limited or No Activation 2–3 Undertriage Rate Overtriage Rate
37 417 4.6% 41.6% No NSP or ISS b 15
Full Trauma Activation 1 Limited or No Activation 2–3 Undertriage Rate Overtriage Rate
31 388 9.3% 34.8%
0.7116 41 (56.9) 16 (22.2) 1 (1.4) 1 (1.4) 13 (18.1)
3
a
ISS 16–75
Total
40 30
89 430
Ideally ≤5% Ideally ≤50% NSP Total 52 20
89 437
Ideally ≤5% Ideally ≤50% NSP or ISS ≥ 16 Total 58 89 40 428 Ideally ≤5% Ideally ≤50%
Missing ISS for 7 patients.
5.4, 95% CI: 2.-12.3) those of patients with blunt injuries. The sensitivity of ISS in predicting NSP is only 61%. We then performed an analysis using the Standard Cribari Matrix, our created NSP matrix, and combined NSP/ISS Matrix. The NSP/ISS combined matrix consists of patient with any 1 NSP variable or and ISS of ≥16.Table 3 shows that the standard Cribari matrix of over and undertriage for our data set demonstrates undertriage rate of 7.0% and an overtriage rate of 55.1%. The NSP matrix demonstrates an undertriage rate of 4.6% and an overtriage rate of 41.6%. Finally, the combined NSP/ISS matrix demonstrated an undertriage rate of 9.3% and an overtriage rate of 34.8%. 3. Discussion Recently, it has been suggested that surgeon presence may not be needed for lower level pediatric trauma resuscitations [14]. Unnecessary highest level activation can lead to a stretching of resources which may drive up hospital costs as well as decrease the quality of patient care. Seriously injured patients resuscitated by an organized trauma team with surgeon presence allows for faster time to CT scan, shorter emergency department times, and shorter time to transport to the operating room [15]. Therefore, identifying pediatric trauma patients with an immediate need for a surgeon's presence upon arrival is critical, and makes the proper evaluation of over and undertriage paramount in the care of the pediatric trauma patient. The goal of our study was to analyze our prehospital data and develop a model which reveals the factors most predictive of NSP and highest level of activation. Development of a prehospital model predictive of NSP may allow improved accuracy of assessment of over and undertriage. Our study included collection of initial prehospital values which allowed us to assess their association for NSP independently of one another. We analyzed a host of factors in a multivariable fashion with moderate to severe traumatic brain injury, SBP, and penetrating trauma being the factors most significantly associated with NSP and highest level of activation. It is interesting the 3 factors identified here were in similarly represented among the 8 factors proposed by Falcone et al. in 2012, however multivariable analyses of those factors was not presented. It is possible multivariable adjustment would have arrived at a reduced model even more similar to that presented here. The Cribari Matrix, which has been widely adopted to determine appropriateness of triage, uses ISS in assessing accuracy of over and undertriage. Severe injury is defined as an ISS ≥ 16 as it was found to predict a 10% mortality in adult trauma patients [16]. Several potential limitations exist using this system. ISS does not take into account physiologic derangement or clinical judgment of the treating physician. The ISS score weighs injuries equally (i.e. head and extremity injuries) when there has been a significant amount of evidence that this is not entirely
Please cite this article as: P. McGaha, T. Garwe, K. Stewart, et al., Factors that predict the need for early surgeon presence in the setting of pediatric trauma, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.05.010
4
P. McGaha II et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
accurate in the pediatric trauma population. A pediatric patient with an intracranial hemorrhage may have an ISS b 16 but may need significant attention, while a patient with long bone fractures may have an ISS N 14 and be much more stable clinically [10]. Since traumatic brain injury is largely responsible for morbidity and mortality in the pediatric trauma population, ISS based systems can lead to significant undertriage [17]. Currently, ISS is not sufficient to in adequately provide an assessment of over and undertriage in the pediatric trauma patient. Our data suggest the use of only ISS dependent triage matrices may markedly underestimate the rate of undertriage. As shown in Table 3, combining ISS and NSP shows a 4% increase in our undertriage rate (undertriage rate of 9.3% when ISS/NSP is combined). This suggests a substantial portion of patients are undertriaged and not accounted for in the currently utilized systems. In reviewing the undertriaged patients, there were multiple patients who were undertriaged by the Cribari matrix. Many of these patient were intubated secondary to decreased GCS. This is a flaw in the current standard for over/undertriage as many times head injuries have a low scoring ISS in spite of their severity. There were one patient who needed an emergent decompressive craniotomy, one patient who had a suspected spinal cord injury, and multiple patients who needed emergent operative intervention for neck lacerations. In our study, ISS demonstrated only a 61% sensitivity for NSP. This also suggests ISS-based over and undertriage measures may not be predictive of true injury severity, NSP, and highest level of activation. In addition, the inability to calculate the ISS until all injuries have been identified, limits the utility of ISS in assessing concurrent over and undertriage accuracy in the pediatric trauma patient. NSP may allow for a more concurrent and complete assessment of over and undertriage in the pediatric trauma patient than the ISS alone based Cribari Matrix. These factors suggest that adding a NSP matrix which is related to physiologic derangements of trauma patients may augment the ISS system. Shock Index Pediatric Age-adjusted (SIPA) has been previously shown to predict severe injury based on ISS [13,18]. Although the use of SIPA was somewhat predictive of NSP in univariable analysis, it was not when subjected to multivariable modeling. This is likely because changes in GCS independent of SIPA were so strong in predicting NSP that SIPA ultimately fell out of the final model. Identifying prehospital factors that predict NSP may allow for a more appropriate trauma activation of pediatric trauma patients. In addition, accurate assessment of over and undertriage is essential to allow for optimal resource utilization by the medical facility and ultimately benefits the patient, physicians, and hospital. Our data suggest that assessing the appropriateness of triage based on ISS alone may not accurately reflect the rate of undertriage in the pediatric trauma patient. Use of the model itself may aid in field evaluation of pediatric trauma patients. Having a patient with any one of the following prehospital characteristics would benefit from the highest level of activation: GCS ≤12, a low SBP as defined by the model, and the victim of penetrating trauma. 3.1. Limitations The first limitation of this study is that it was a retrospective study and at a single institution. This limits the extrapolation of this model currently; however, we are planning a multi-institutional study in order to externally validate this data. Additionally, owing to the difficulty in tracking the presence of subjective, mechanistic activation criteria in the prehospital reports, only objective data were examined. Perhaps other mechanistic criteria impact NSP and highest level of activation as well [19]. We also plan to validate the model for prehospital predictive factors involved in predicting NSP and highest level of activation through a prospective study at our institution, which is currently ongoing. This will ensure that almost all prehospital vital signs and NSP factors are captured, as well as mechanistic variables, thereby increasing the validity of the study. In our prospective study, we are also collecting data on
transfers and the NSP; as this study only included patients that came to our facility directly from the scene of injury. Finally, like ISS, NSP is somewhat retrospective in nature, and limited with respect to certain types of injuries like eye, oral–maxillofacial injuries, tendon/nerve injury in hands/extremities, laceration/incisive wounds in conspicuous places. Also, it may miss nonaccidental trauma patients where specialist intervention may be needed after resuscitation and admission to the hospital. Furthermore, NSP does not necessarily mean an operation is needed by a trauma surgeon, rather, it is a marker for pediatric trauma patients requiring the highest level of care, even if the end results is nonoperative. 4. Conclusion A validated model of factors predictive of NSP and highest level of activation in pediatric trauma patients has yet to be developed. NSP can be predicted based on the following factors: GCS ≤ 12, penetrating trauma, and age adjusted hypotension. Such a prognostic prehospital model may allow for an improved assessment of pediatric trauma activation criteria. Our data also suggest that a triage matrix based on NSP may augment assessment of over and undertriage in the pediatric trauma patient as compared to the ISS/Cribari system alone. References [1] Lerner EB, Drendel AL, Falcone Jr RA, et al. A consensus-based criterion standard definition for pediatric patients who needed the highest-level trauma team activation. J Trauma Acute Care Surg 2015;78(3):634–8. [2] Brown JB, Gestring ML, Leeper CM, et al. The value of the injury severity score in pediatric trauma: time for a new definition of severe injury? J Trauma Acute Care Surg 2017;82(6):995–1001. [3] Rogers AT, Gross BW, Cook AD, et al. Outcome differences in adolescent blunt severe polytrauma patients managed at pediatric versus adult trauma centers. J Trauma Acute Care Surg 2017;83(6):1082–7. [4] Baker SP, O'Neill B, Haddon Jr W, et al. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 1974;14(3):187–96. [5] Copes WS, Champion HR, Sacco WJ, et al. The injury severity score revisited. J Trauma 1988;28(1):69–77. [6] Beverland DE, Rutherford WH. An assessment of the validity of the injury severity score when applied to gunshot wounds. Injury 1983;15(1):19–22. [7] Falcone Jr RA, Haas L, King E, et al. A multicenter prospective analysis of pediatric trauma activation criteria routinely used in addition to the six criteria of the American College of Surgeons. J Trauma Acute Care Surg 2012;73(2):377–84 [discussion 84]. [8] Chen LE, Snyder AK, Minkes RK, et al. Trauma stat and trauma minor: are we making the call appropriately? Pediatr Emerg Care 2004;20(7):421–5. [9] van der Sluijs R, van Rein EAJ, Wijnand JGJ, et al. Accuracy of pediatric trauma field triage: a systematic review. JAMA Surg 2018;153(7):671–6. [10] Roden-Foreman JW, Rapier NR, Yelverton L, et al. Avoiding Cribari gridlock: the standardized triage assessment tool improves the accuracy of the Cribari matrix method in identifying potential overtriage and undertriage. J Trauma Acute Care Surg 2018; 84(5):718–26. [11] ATLS. 10th edition of the Advanced Trauma Life Support® (ATLS®) student course manual. Chicago. IL: American College of Surgeons; 2018. [12] Berg MD, Schexnayder SM, Chameides L, et al. Part 13: pediatric basic life support: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;122(18 suppl 3):S862–75. [13] Acker SN, Ross JT, Partrick DA, et al. Pediatric specific shock index accurately identifies severely injured children. J Pediatr Surg 2015;50(2):331–4. [14] Boomer LA, Nielsen JW, Lowell W, et al. Managing moderately injured pediatric patients without immediate surgeon presence: 10 years later. J Pediatr Surg 2015;50(1):182–5. [15] Vernon DD, Furnival RA, Hansen KW, et al. Effect of a pediatric trauma response team on emergency department treatment time and mortality of pediatric trauma victims. Pediatrics 1999;103(1):20–4. [16] Boyd CR, Tolson MA, Copes WS. Evaluating trauma care: the TRISS method. Trauma Score and the Injury Severity Score J Trauma 1987;27(4):370–8. [17] DiBrito SR, Cerullo M, Goldstein SD, et al. Reliability of Glasgow Coma Score in pediatric trauma patients. J Pediatr Surg 2018;53(9):1789–94. [18] Nordin A, Coleman A, Shi J, et al. Validation of the age-adjusted shock index using pediatric trauma quality improvement program data. J Pediatr Surg 2017 [Oct 12. pii: S0022-3468(17)30645-0]. [19] Recicar J, Barczyk A, Duzinski S, et al. Does restraint status in motor vehicle crash with rollover predict the need for trauma team presence on arrival? An ATOMAC study. J Pediatr Surg 2016;51(2):319–22.
Please cite this article as: P. McGaha, T. Garwe, K. Stewart, et al., Factors that predict the need for early surgeon presence in the setting of pediatric trauma, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.05.010