60.09 Sternotomy for Hemorrhage Control in Trauma

Posted on January 31, 2019

L. Al-Khouja1, A. Grigorian1, S. Schubl1, K. Galvin1, A. Kong1, M. Lekawa1, T. Chin1, J. Nahmias1  1University Of California – Irvine,Department Of Trauma And Critical Care Surgery,Orange, CA, USA

Introduction: Thoracic trauma accounts for 20-25% of trauma deaths. Little is known about the injuries, mechanisms, and outcomes in trauma patients undergoing sternotomy for hemorrhage control. The purpose of this study is to perform a descriptive analysis of trauma patients undergoing sternotomy for hemorrhage control and identify which thoracic injuries and other factors are predictors of mortality. We hypothesize blunt trauma is associated with higher risk of death compared to penetrating trauma within this population.

Methods: The Trauma Quality Improvement Program (2010-2016) database was queried for patients undergoing sternotomy for hemorrhage control within 24-hours of admission. Patients with blunt and penetrating trauma were compared using chi-square and Mann-Whitney U test. A multivariable logistic regression model was used to determine risk of mortality.

Results: Of the 584 patients undergoing sternotomy for hemorrhage control, 322 (55.1%) were involved in penetrating trauma, 69 (11.8%) in blunt trauma and 193 (33.3%) involved in “other/unknown” mechanism. The median injury severity score (ISS) was 25.0 and the most common known mechanism was a stab wound (49.9%) followed by gunshot wound (19.2%). The overall time to hemorrhage control was 52.8 minutes and was longer in those with blunt compared to penetrating trauma (84.6 vs. 49.8 minutes, p <0.001). Open-cardiac injury (26.4%) and hemothorax (23.5%) were the most common thoracic injuries. The most common procedures involved cardiac repair (44.7%) followed by lung repair (15.8%). The overall mortality rate was 18.2% and was higher in patients with blunt compared to penetrating trauma (29.0% vs. 12.7%, p<0.001). However, after adjusting for covariates there was no difference in risk of mortality between blunt and penetrating trauma (p=0.49). We did not find any independent predictors of mortality in patients with blunt trauma. The strongest independent predictor of mortality in patients with penetrating trauma was ISS≥25 (OR=6.24, CI=2.04-19.11, p=0.001).

Conclusion: Trauma patients who undergo sternotomy for bleeding often achieve hemorrhage control in less than one hour. Nearly half the patients present after a stab wound and require cardiac repair. Trauma patients requiring sternotomy for hemorrhage control after blunt trauma had a higher mortality rate, compared to those involved in penetrating trauma. However, after adjusting for known predictors of mortality in trauma, there was no difference in risk of mortality despite nearly double the time to hemorrhage control in patients presenting after blunt trauma.

60.08 Use of Indirect Calorimetry in a Cohort of Patients with Enterocutaneous Fistula

Posted on January 31, 2019

S. Bou Zein Eddine1, A. Kamien1, Z. Yin1, C. Trevino1, R. Fritzshall1, A. Stachowiak1, A. Szabo1, J. S. Juern1, J. R. Peschman1, M. A. De Moya1, P. A. Codner1  1Medical College Of Wisconsin,Surgery/ Trauma And Acute Care Surgery,Milwaukee, WI, USA

Introduction:
Nutrition is key in Enterocutaneous Fistula (ECF) management.  Predictive Equations (PE) and Indirect Calorimetry (IC) are used to calculate nutritional needs but do not correlate well to each other. We hypothesize that even though IC is the gold standard, it is difficult to implement logistically and has a weak concordance with PE.

Methods:
The study included a retrospective cohort of all patients aged ≥18 with ECF admitted between January 2011 and April 2016 and a prospective cohort admitted between April 2016 and February 2017. Demographics, comorbidities, surgeries, and fistula output were determined.  A dietitian determined nutritional therapy and caloric requirements at initial consult, hospital discharge, and outpatient using the Mifflin St Jeor (MSJ) equation. The baseline Resting Energy Expenditure from PE (REEPE) and REE from IC (REEIC) was evaluated for both cohorts, the difference was calculated, and the concordance were plotted in a Bland-Altman plot.

Results:
A total of 33 patients were included. In the prospective arm (n=12) mean age was 54.3 (±18.9) and 66.7% were male. In the retrospective arm (n=21), mean age was 53.7 (±18.3) and 66.7% were male. The median equation difference of REEPE from REEIC, was 71.0 kcal/24 h (IQR, -203.0 – 173.5) in the prospective arm and 123.0 kcal/24 h (IQR, -97.0 – 177.0) in the retrospective arm. The concordance correlation coefficient between REEPE and REEIC was 0.541 (95%CI, 0.213, 0.785). The Bland-Altman plot for the concordance between the REEIC and REEPE (Figure 1) had wide limits of agreement.

Conclusion:
Assessing accurate nutritional requirements remains an increasingly challenging clinical problem. There’s a weak to moderate agreement between REEPE and REEIC and the standardization for IC measurements is logistically difficult to perform.
 

60.07 Is Hospital-Associated Venous Thromboembolism Preventable in Trauma Patients?

Posted on January 31, 2019

K. L. Florecki1, B. Lau1, O. Owodunni1, J. Sakran1, M. Streiff1, E. Haut1  1The Johns Hopkins University School Of Medicine,Baltimore, MD, USA

Introduction: Venous thromboembolism is a common complication associated with significant morbidity and mortality in the trauma population. We strived  to characterize the true preventability of venous thromboembolism (VTE) in the trauma patient population and the patients who receive “defect-free care” prophylaxis for VTE. 

Methods:  Retrospective review of trauma patients with hospital-acquired VTE (including deep vein thrombosis [DVT] and/or pulmonary embolism [PE]) identified at The Johns Hopkins Hospital from January 2008 – June 2016.  Data on patient risk assessment for VTE, prescription of risk-appropriate VTE prophylaxis and pharmacologic VTE prophylaxis medication administration were extracted from the electronic health record system. Patients were dichotomized into prophylaxis failure vs. defect-free care, defined as receiving all doses of risk-appropriate VTE prophylaxis recommended by our validated, mandatory computerized clinical decision support tool. Prophylaxis failure was divided into prescription failures or dose-administration failures (missed doses).

Results: 92 trauma patients had hospital-acquired VTE. All 92 (100%) were assessed using the VTE risk assessment mandatory computerized clinical decision support tool, however only 75 (81.5%) were prescribed risk-appropriate prophylaxis. Of the 92 patients, 33 (35.9%) received defect free-care. Of the 59 (64.1%) who received suboptimal care, 17 (28.8%) were not prescribed risk-appropriate prophylaxis and 42 (71.2%) missed at least one dose of pharmacologic VTE prophylaxis. 

Conclusion: Our study identifies the need to reevaluate outcome and process measures for VTE prevention after trauma. 33 VTE events occurred in patients who received best-practice defect-free care, showing not all VTE are truly preventable, and should not be included in outcome measures as “potentially preventable events.”  Our findings also identify specific targets and emphasize the importance of ongoing efforts to improve prescription and administration of risk appropriate VTE prophylaxis in trauma centers. 

 

60.06 Trauma Patients with Mangled Lower Extremities Experience a Higher Incidence of VTE

Posted on January 31, 2019

D. Freitas1, E. Warnack1, M. Kim1, C. DiMaggio1, S. Frangos1, M. Klein1, C. Berry1, M. Bukur1  1New York University School Of Medicine,New York, NY, USA

Introduction:
The mangled lower extremity (MLE) is a limb sustaining complex structural injuries (i.e. soft tissue, bone, nerves, and vessels) resulting from a high- energy mechanism. We hypothesized that trauma patients who present with MLE experience a higher rate of deep venous thrombosis (DVT) and pulmonary embolism (PE) when matched with trauma patients of similar injury burden without MLE.

Methods:
Data were abstracted from the Trauma Quality Improvement Program (TQIP) database from 2013- 2016. Characteristics of blunt trauma patients who presented with MLE versus all other patients were compared using Chi-squared, Student’s t-test, or Mann-U Whitney test where appropriate. Propensity score matching, using a 2:1 match of controls versus patients with MLE, was used to compare the occurrence of DVT and PE.

Results:
A total of 1,060 patients presented with MLE from 2013 through 2016. Patients with MLE were younger (median age 42 IQR [28,54] vs. 54 [32,72], p < .001) and were more likely to need immediate operative intervention when compared to other trauma patients (60.5% vs. 11%, p < .001). In controlled models, while patients with MLE had a higher odds of receiving VTE prophylaxis (86% vs. 58.8%, AOR = 3.23, p < .001) and filter placement (6% vs. 1.5%, AOR = 2.81; p < .001), they were over two times more likely to experience DVT (5.4% vs. 1.4%, AOR 2.5, p < .001) or PE (2.3% vs. 0.6%, AOR 2.07, p = .155), although the latter was not statistically significant. In a propensity score model, MLE conferred a significant absolute risk increase of 3.6% for DVT or PE with an odds ratio of 2.43 (7% MLE vs. 3.3% no MLE). Mortality was not significantly lower in MLE patients versus other trauma patients (4.5% vs. 5.5%, p = .157).

Conclusion:
Trauma patients with MLE have a 2.4 times higher odds of having a DVT or PE with an absolute risk increase of 3.6%. Early aggressive prophylactic measures are warranted in this high-risk population.
 

60.05 Nutritional Support Disparities in Trauma ICUs in the US: An Assessment of the AAST TRIPP Database

Posted on January 31, 2019

L. TUNG1, R. Dumas1, D. N. Holena1, M. J. Seamon1, L. J. Kaplan1, A. B. Newcomb2, C. P. Michetti2, N. D. Martin1  1University Of Pennsylvania,Philadelphia, PA, USA 2Inova Fairfax Hospital,Falls Church, VA, USA

Introduction:  Adequate nutritional support during critical illness has been shown to improve outcomes. Timely initiation of any nutrition can be variable by patient population. We hypothesize that all patient types in a trauma-designated ICU would receive a similar onset of nutritional support, and that pressors would not play a role in nutrition.

Methods:  The TRIPP database is a 2-day, multicenter prevalence study of all patients present in a trauma-designated intensive care unit (ICU) on 11/2/17 and 4/10/2018. Patients were stratified as either trauma, general surgery, or medical ICU patients. The presence of any nutritional support was evaluated based on the current ICU length of stay (LOS). Differences in nutritional support over the first 7 days were determined by t-test. Differences in either enteral feeds or parenteral nutrition above or below a total pressor equivalence of norepinephrine (NE) of 0.11mcg/kg/min was determined by chi-square.

Results: 1216 patients from 49 trauma-designated ICUs contributed data. 562 (46.2%) patients were traumatically injured, 458 (37.7%) were general surgery, and 196 (16.1%) were medicine patients. On ICU day 1, 45.7%, 51.1%, 45.5% of trauma, general surgery, and medical ICU patients were receiving nutritional support, respectively. By ICU day 7, this increased to 89.3%, 78.9%, and 80.0%, respectively, with a trend toward statistical significance. There was also no significant difference between the number of patients who were receiving nutrition on NE ≥0.11 mcg/kg/min compared to those on NE ≤0.11mcg/kg/min (p=0.5).

Conclusion: After 7 days of critical illness, a significant percentage of Trauma ICU patients were not receiving nutritional support; further, disparities between ICU populations within trauma-designated ICUs also seemed to exist.  Vasopressors however, were not associated with decreased nutritional support.  Additional research into the etiologies of these findings is warranted.

 

60.04 Comparison Of Urban Off-road Vehicle And Motorcycle Injuries At A Level 1 Trauma Center

Posted on January 31, 2019

C. A. Butts1, R. Gonzalez2, J. P. Gaughan2, J. San Roman2, S. Ross2, J. Porter2, J. P. Hazelton2  1Rutgers-Robert Wood Johnson University Hospital,Acute Care Surgery,New Brunswick, NJ, USA 2Cooper University Hospital,Trauma, Surgical Critical Care, & Acute Care Surgery,Camden, NJ, USA

Introduction:
Recently, there has been an increase in the use of dirt bikes (DB) and all-terrain vehicles (ATV) in urban environments. Our previous work found that crashes involving urban off-road vehicles (UORV), defined as any ATV or DB accident which occurred on paved inner city/suburban/major roadways, resulted in different injury patterns than crashes which occurred in rural environments.  The aim of this study is to compare the injury patterns of patients involved in crashes while riding urban off-road vehicles with motorcycles (MC).

Methods:

A retrospective review (2005-2016) of patients who presented to our urban Level I trauma center as a result of any MC or UORV crash was performed. Patients who presented more than 48 hours from time of accident were excluded. A p<0.05 was considered significant.  

Results:

1556 patients were identified to have an MC or UORV crash resulting in injury [MC: n=1324 (85%); UORV: n=232 (15%)]. Patients involved in UORV accidents were younger (26.2 vs 39.6; p <0.05), less likely to be helmeted (39.6% vs 90.2%; p<0.05), but required fewer emergent trauma bay procedures (28.4% vs. 36.7%; p<0.05) and fewer operative interventions (45.9% vs 54.2%; p <0.05).  Both group of patients had a similar ISS (12.2 vs 12.6; p=0.54) and presenting GCS (13.8 vs 13.5; p=0.46). Interestingly, UORV patients had a lower mortality (0.9% vs 4.7%; p <0.05) when compared to MC crash patients despite injury patterns being similar between both groups.

Conclusion:
Our data demonstrates that UORV riders are younger and less likely to be helmeted, but have a lower mortality rate following a crash, despite having similar injury patterns to motorcyclists.  Even though our data suggests a lower mortality in this patient population, we feel that the fact that fewer than 50% of UORV patients were wearing helmets is unacceptable in this younger patient population.  This data may provide a framework for legislative intervention for helmet laws to be more strictly enforced in this population and facilitate a more aggressive community outreach focusing on increased helmet compliance.

60.03 CSF Cultures in Traumatic Brain Injury: Is It Worth It? A Two-Center Study

Posted on January 31, 2019

N. K. Dhillon1, S. Sahi2, G. Barmparas1, N. T. Linaval1, T. Lin1, S. Lahiri1, C. V. Brown2, E. J. Ley1  1Cedars-Sinai Medical Center,Los Angeles, CA, USA 2Dell Medical School, University of Texas at Austin,Austin, TEXAS, USA

Introduction:  Patients with traumatic brain injury (TBI) frequently develop leukocytosis, fever, and tachycardia which may lead to extensive medical investigations to rule out an infectious process. Cerebral spinal fluid (CSF) is often acquired during this work-up, however the utility of this practice has not been previously studied. We hypothesized that CSF cultures would unlikely yield positive results in patients with TBI.

Methods:  A retrospective review was conducted of all TBI patients admitted to two Level 1 trauma center at an urban, academic medical center from 01/2009 to 12/2016. Data collected included patient demographics, presenting GCS, injury profile, injury severity scores (ISS), regional abbreviated injury scale (AIS), hospital and ICU length of stay (LOS), ventilator days, and culture results. For purposes of the analysis, CSF cultures with Staphylcoccus epidermidis, Staphylococcus aureus, or Candida were considered contaminated and deemed negative.

Results: There were 148 patients who had CSF cultures obtained with a median age of 40 years and 77.7% being male. The majority of patients presented following blunt trauma with median GCS of 6, head AIS of 4, and ISS of 26. These patients had prolonged median ICU and hospital stay at 14 and 22 days, respectively. Seven (4.7%) CSF cultures demonstrated growth. Four (2.7%) were deemed to be contaminants, with two growing Staphylcoccus epidermidis only, one with both Staphylcoccus epidermidis and Staphylococcus aureus, and one with Candida. Three cultures (2.0%) were positive; two had Enterobacter Cloacae and one Klebsiella Pneumoniae. Of note, all three patients with positive cultures also had instrumentation with either an external ventricular drain or a lumbar drain.

Conclusion: Obtaining CSF cultures in TBI patients without instrumentation is of extremely low yield. Other sources of infectious etiologies should be considered in this patient population.

 

60.02 Time to Prothrombin Complex Conctretrate Administration Effect on Intracranial Hemorrhage Outcomes.

Posted on January 31, 2019

L. S. Kuzomunhu1, M. M. Fleming2, R. R. Jean2, K. Y. Pei2  1Yale University School Of Medicine,New Haven, CT, USA 2Yale University School Of Medicine,Department Of Surgery,New Haven, CT, USA

Introduction: Prothrombin complex concentrate (PCC) is indicated in patients sustaining traumatic intracranial hemorrhage (ICH) while taking warfarin, but the optimal timing is unknown. This study evaluates the effect of timing to PCC administration on outcomes including intracranial hemorrhage expansion and length of hospital stay.

Methods: We retrospectively reviewed patients presenting with ICH who received PCC at our tertiary care hospital between March 2013 to November 2017. Only patients on warfarin and with repeat computed tomography of head were included. Time to PCC was defined as the time from order entry to patient administration as documented contemporaneously in the electronic health record. Time was stratified as early (≤30 minutes) or late (>30 minutes). Multivariable logistic regression with stepwise selection was used to predict ICH expansion between initial and repeat cross sectional imaging. Linear regression identified predictors for increased hospital and intensive care unit length of stay.

Results: In total, 127 patients with ICH on warfarin were included for analysis. Mean time to PCC administration was 82.3 ± 43.7 minutes. The majority of patients who demonstrated expansion of ICH (31.5% of patients) had admission motor Glasgow Coma Score (GCS) less than 6 (p<0.05), a higher Charlson Comorbidity Index (CCI) (p<0.05) and higher inpatient mortality (p<0.01). On multivariable analysis, only admission motor GCS<6 was independently associated with ICH expansion (OR 3.016, 95% CI 1.158-7.858). Time to PCC (early versus late), admission INR and anticoagulation indication were not associated with ICH expansion. On linear regression admission motor GCS<6 was associated with increased length of ICU stay (β=8.261, SE 2.070, p=0.0001); however no other patient characteristics or PCC administration time was associated with hospital length of stay.

Conclusion: Timing to PCC administration was not associated with ICH expansion or mortality after traumatic ICH. Further multi-institutional studies are needed to evaluate clinical and process measures to streamline PCC administration.

60.01 Axillary Vessel Injuries: What have we Learned From an Analysis of the National Trauma Data Bank

Posted on January 31, 2019

M. A. Dale1, A. Person1, G. Mendoza1, S. Brown1, D. Keleny1, D. Rigg1, J. Dabestani1, D. Agrawal1, J. Asensio1  1Creighton University Medical Center,Department Of Surgery,Omaha, NE, USA

Introduction: Axillary vessel injuries remain uncommon even in busy urban trauma centers. Most trauma surgeon possess limited experience with these injuries. The proximity of the surrounding anatomic structures account for a large number of associated injuries and mobidity associated with such injuries.

Methods: The National Trauma Data bank was queried for pre-hospital admission data for axillary vessel injuries. Data extracted included demographics, physiologic conditioins, surgical interventions and Functional Independent Measure Score (FIM). Univariate and stepwise logistic regression analysis were used.

Results: 590 cases were identified from 1,466,887 patients in NTDB from 2001-2005. Incidence = 0.004%. Mean age 33 ±  16, mean RTS 6.8 ±  2.3, mean Glasgow Coma Score 13 ±  4.4, mean ISS 14.9 ± 10.3. Mechanism of inury: penetrating n=329 (55.6%), blunt n=259 (43.9%), and non-specified n=2 (0.34%). Mean initial SBP 119.7 ±  37.7, mean in survivors 122.8 ± 24.8, mean in non-survivors 79.7 ±  56.1. Base deficit mean -3.7 ±  10.7, in survivors -3.3 ±  8.5, non-survivors -7.3 ± 19.2. Total number of injuries n=681, of these axillary artery n=455 (67%), axillary vein n=144 (21%), axillary nerve n=68 (9.98%), unspecified n=14 (2%). Sugical procedures for axillary injuries were documented in 238 patients, these included n=54 (15.4%) suture of the artery, n=38 (8.91%) vascular shunt or bypass. Complications n=2637 associated with axillary vessel injuries: pneumonia n=22 (16.05%), wound infection n=16 (11.67%), compartment syndrome n=14 (10.21%), ARDS n=12 (8.76%). Fuctional Independent Measures: dependent n=11 (1.86%), partial help required n=13 (2.2%), independent with device n=28 (4.75%), independent n=210 (35%), not applicable (7.3%), and not documented n=285 (48.3%). Survivors n=548 (92.88%), non-survivors n=42 (7.12%).

Conclusion: Axillary vessel injuries remain an uncommon occurence in the United States as denoted by their low incidence as reported in the NTDB and literature. Initial admitting systolic blood pressure has a higher correlation with morbidity based on mechanism of inury. However, when using stepwise logitistic models, ISS, TRISS, and LOS had the highest correlation with survival. Axillary vessel repairs require complex surgical interventions. Most injuries carry an associated high morbidity, good functional outcomes, and relative low mortality.

 

40.10 ISOLATED LOW-GRADE BLUNT SOLID ORGAN INJURY IN ADULT TRAUMA PATIENTS: TRANSFER TO A LEVEL I TRAUMA CENTER?

Posted on January 31, 2019

A. Azim1, K. Prabhakaran1, M. Khan1, G. Lombardo1, J. Con1, R. Latifi1  1Westchester Medical Center,Surgery,Valhalla, NEW YORK, USA

Background:The management of low grade (I-III) solid organ (spleen, liver, kidney) injury has evolved to a predominantly conservative management approach with few of these patients requiring procedural intervention. Nonetheless, transfer of these patients from non-trauma centers or lower level trauma centers to Level I trauma centers is common. The objective of this study was to determine the difference in outcomes between patients transferred to and managed at a higher-level trauma center and those managed at a lower level trauma center (Level II/III).

Methods: A two-year retrospective analysis of the National Trauma Data Bank was performed. All trauma patients greater than 18 years of age with isolated low grade solid organ injury were identified. Patients were identified using ICD-10 codes and the following data points were obtained: age, gender race, admission vitals, transfer status, trauma center (TC) designation, and procedures performed. Patients were divided between 2 groups; those directly admitted to Lower Level TC or those transferred to higher Level TC Outcome measures were intervention (angioembolization or surgery) and mortality. A Chi square analysis was performed.

Results: A total of 917865 trauma patients with blunt mechanism of injury were analyzed of which 25, 619 patients had solid organ injury. A total of 1256 patients with isolated low grade solid organ injury were identified. Mean age was 43 ± 25 years, 75.2% were male, and 79.7% were white. Of these, 53.5% were transferred to Level-I trauma center and 46.4% were admitted to lower Level TC. Overall mortality rate was 1.4% and need for intervention was 4.7%.  For low grade isolated splenic injuries, 47.7% were admitted to Low Level TC compared to 52.3% transferred to Level I TC. For low grade isolated liver injuries, 44.5% were admitted to Level II/III TC compared 55.5% to transferred to Level I TC. For low grade isolated kidney injuries, 41.2% were admitted to Low Level TC compared to 59.8% transferred to Level-1 TC. The overall rate of intervention at Lower level TC was 4.6% versus 4.8% for transferred patients (p=0.11) (Table 1.). Mortality rate was similar between the two groups (1.41% vs 1.38%) (p=0.08).

Conclusion:Few adult patients with low grade solid organ injury require intervention, even when transferred to a Level I Trauma Center. Patient with low grade solid organ injury can be safely managed at lower level trauma centers.

40.09 Geographic Variation in Under Triage of Severely Injured Older Adults: A National Analysis

Posted on January 31, 2019

M. P. Jarman1, T. Uribe-Leitz1, D. Sturgeon1, C. D. Newgard4, E. Goralnick1,3, Z. Cooper1,2, A. Salim1,2, A. Haider1,2  1Brigham And Women’s Hospital,Center For Surgery And Public Health,Boston, MA, USA 2Brigham And Women’s Hospital,Division Of Trauma, Burns, And Critical Care Surgery,Boston, MA, USA 3Brigham And Women’s Hospital,Department Of Emergency Medicine,Boston, MA, USA 4Oregon Health And Science University,Department Of Emergency Medicine,Portland, OR, USA

Introduction:  Despite CDC guidelines recommending triage of significantly injured older adults (age ≥ 65) to trauma center (TC) care, less than 50% of injured older adults in the US are treated at designated Level I or II trauma centers. We sought to determine potential variation in under triage patterns across the US and understand contributory factors.

Methods: Using 2012-2014 Medicare claims data, we identified all patients age ≥ 65 with inpatient or emergency department encounters for traumatic injury and an injury severity score ≥ 16, and classified patients as under triaged if they were treated at a non-TC without subsequent transfer to a TC. Incidence of traumatic injury and under triage were aggregated to the hospital referral region (HRR) level and linked with their Dartmouth Atlas characteristics and the American Hospital Association annual survey. We then used hierarchical logistic regression to estimate odds of under triage for each region and identify associated HRR-level factors (e.g. distribution of income, race, age).

Results: Odds of under triage increased by 6% for every 100,000-person increase in population per TC at the HRR level (OR: 1.06, 95% CI: 1.02, 1.09), and by 8% for every 1-percentage point increase in the proportion of HRR residents with income at/below the federal poverty line (OR: 1.08, 95% CI: 1.03, 1.13). HRR-level factors accounted for 71% of variation in odds of under triage. Population per TC alone accounted for 41% of variation in odds of under triage. No other HRR-level factors examined were associated with odds of under triage after adjusting for the population per TC. Figure 1 illustrates the geographic distribution of under triage by HRR.

Conclusion: Under triage for injured older adults varies substantially across the US. System-level factors account for most of this variation, particularly population size relative to TC capacity, and characteristics of the population each TC serves (i.e. income distribution). Efforts to reduce under triage of older adults should consider both the capacity of trauma systems, and population characteristics.
 

40.08 Vulnerability to Financial Hardship After Severe Traumatic Injury: The Impact of Socioeconomic Status

Posted on January 31, 2019

K. M. O’Neill3, R. A. Jean3, C. P. Gross2, R. D. Becher3, R. Khera4, J. V. Elizondo5, K. Nasir2  2Yale University School Of Medicine,Internal Medicine,New Haven, CT, USA 3Yale University School Of Medicine,General Surgery,New Haven, CT, USA 4University Of Texas Southwestern Medical Center,Cardiology,Dallas, TX, USA 5Yale University School Of Medicine,New Haven, CT, USA

Introduction: Trauma-related disorders rank among the top-five most costly medical conditions to the healthcare system. However, the impact of healthcare expenses on the families of patients suffering traumatic injury is not well studied. To address this gap in knowledge, we used nationally representative data to investigate the burden of financial hardship from out-of-pocket (OOP) health expenses for families with one or more members suffering from a traumatic injury, specifically evaluating the role of socioeconomic status and injury severity on financial burden.

Methods: This retrospective cross-sectional study used nationally representative Medical Expenditure Panel Survey data from 2010 through 2015. We identified families in which at least one member had a traumatic injury and classified this injury based on severity; socioeconomic status was based on family income. Percent of annual family income used for OOP health expenses was used to assess for the risk of excess financial burden (OOP>20% of annual income) and catastrophic medical expenses (OOP>40% of annual income) adjusting for demographic, socioeconomic, and healthcare utilization factors in a multivariable logistic regression model.

Results: We identified 7,538 individuals with injuries, represented in 7,102 families of the total 90,964 families in the cohort. Of these, 668 families were classified as severe. Families with a severe traumatic injury experienced increased OOP medical expenses, spending on average $2,784 (95% CI: $2,106-$2833). Increased severity of injury was also associated with increased proportion of excess final burden. Overall 4.4% in the uninjured cohort experienced excess financial burden, compared to 5.6% of families with a minor injury and 12.6% of families with severe injury (see Figure). In a risk-adjusted logistic regression model, families with a severe injury were significantly more likely to experience excess financial burden (OR: 2.04, 95% CI: 1.13-3.64) and catastrophic medical expenses (OR: 3.08, 95% CI: 1.37-6.9). Families below the federal poverty line had increased odds of excess financial burden (OR: 18.1, 95% CI: 15.3-21.4) and catastrophic medical expenses (OR: 47.8, 95%CI: 35.5-64.7). 

Conclusions: Approximately 1 in 8 households with a severely injured family member experience financial hardship. These families are significantly more vulnerable to incurring catastrophic OOP health expenses than the non-injured population. This effect was most pronounced for families living below the poverty line. These results highlight the vulnerability of poor families to financial hardship from OOP expenses and reinforce the importance of injury prevention strategies.

40.07 Does CT Scanning at a Referring Hospital Increase Mortality After Reaching a Level One Trauma Center?

Posted on January 31, 2019

K. C. Gallagher1, A. J. Medvecz1, B. T. Craig1, O. C. Guillamondegui1, B. M. Dennis1  1Vanderbilt University Medical Center,Division Of Trauma & Critical Care,Nashville, TN, USA

Introduction:
With advancing technology and rapidly increasing availability, computed tomography (CT) has become a standard radiologic adjunct in the evaluation of a trauma patient. However, the appropriate timing of initial cross-sectional imaging has not yet been established and the utility of imaging at the referring hospital (RH) remains controversial. The American College of Surgeons (ACS) Advanced Trauma Life Support (ATLS) course emphasizes that imaging should never interrupt or delay resuscitation of the injured patient. Despite these guidelines, several studies have identified delays of 60-90 minutes when patients are scanned at referring hospitals prior to arriving at the treating facility. We study the effect of CT scans at referring hospitals on in-hospital mortality at a receiving trauma center.

Methods:
A retrospective cohort study was performed with adult patients transferred to a level one trauma center from non-trauma center regional hospitals between 2012 and 2017. Inpatient-to-inpatient transfers, transfers from other state-designated level one or level two trauma centers, and patients with transfer times over 10 hours were excluded from analysis. Demographics and baseline characteristics were compared with Student’s t-test and Pearson’s Chi-squared testing. The primary endpoint was in-hospital mortality. Cox regression, controlling for transfer time, was used to evaluate the effect of CT scanning on mortality.

Results:

3415 adult trauma patients were included in the analysis: 1135 (33.2%) received a CT scan at the RH prior to transfer while 2280 (66.8%) did not. Patients who received a CT scan at the RH were more likely to be older, female, white, with a higher Charlson comorbidity index, less severely injured (lower ISS and intubation rate; higher systolic blood pressure and GCS), have a blunt mechanism, and be transferred by ground. There was no difference in “as the crow flies” distance (58.3 miles versus 57.0 miles, P = 0.34), but there were significantly longer times at RH and transport times for those that did receive a CT scan (288 minutes versus 213 minutes, P < 0.005).  The unadjusted hazard ratio (HR) for mortality after receiving a CT scan at the RH is 0.32 (95% CI 0.26-0.39, P < 0.005). The adjusted model controlling for age, sex, race, mechanism of injury, transport method, ISS, heart rate, intubation status, comorbidity index, GCS, and transport time, has a hazard ratio of 0.533 (95% CI 0.42-0.68, P < 0.005).

Conclusion:
There is a survival advantage for patients that receive CT scan prior to transfer to level one trauma center despite having significantly longer transport times. This finding contradicts both current ATLS recommendations and previous data, necessitating further investigation and discussion.
 

40.06 Rates and Correlates of Risky Firearm Behaviors among Youth Treated in an Urban Emergency Department

Posted on January 31, 2019

C. A. Mouch1, J. Goldstick3, M. A. Walton3,4, M. A. Zimmerman5, R. M. Cunningham2,3, P. M. Carter2,3  1Michigan Medicine,Department Of Surgery,Ann Arbor, MI, USA 2Michigan Medicine,Department Of Emergency Medicine,Ann Arbor, MI, USA 3University of Michigan,Injury Prevention Center,Ann Arbor, MI, USA 4Michigan Medicine,Department Of Psychiatry,Ann Arbor, MI, USA 5University of Michigan,School Of Public Health,Ann Arbor, MI, USA

Introduction: Firearms are the leading cause of trauma-related mortality among urban youth and emerging adults (EAs). However, little is known about patterns of risky firearm behaviors (i.e., firearm carriage in high-risk situations; firearm discharge in high-risk situations; firearm threats or use against another person) that may increase the likelihood of subsequent injury or death. This analysis examines rates and correlates of risky firearm behaviors (RFBs) to inform future prevention efforts, including current hospital-based violence prevention programs.

Methods: Youth/EAs (age 16-29) seeking ED treatment in an urban Level-1 trauma center completed a computerized survey. Validated measures assessed past 3-month risky firearm behaviors (RFBs), including firearm carriage in high-risk situations (i.e. carriage under the influence of alcohol/drugs, during a drug deal, while commiting a crime, or while hanging with friends), firearm discharge in high-risk situations (i.e. firearm discharge under the influence of alcohol/drugs, during a drug deal, while committing a crime, during a fight, while fleeing police, to scare someone, or while hanging with friends), and/or partner/non-partner firearm aggression (i.e., threats or use of firearm against another person). Sociodemographics, substance use, mental health, prior violence involvement, peer delinquency, and community violence exposure were also assessed. Descriptive statistics, bivariate analyses, and multivariate logistic regression identified rates and correlates of RFBs.

Results: 1,312 youth/EAs completed the survey (mean age 22; 30% Male; 51% Black; 56% receiving public assistance), with 136 (10%) youth/EAs reporting RFBs during the past 3-months. Of those reporting RFBs, 51% endorsed firearm ownership, 74% reported firearm carriage in high-risk situations, 40% endorsed firearm discharge in high-risk situations, and 31% reported partner/non-partner firearm aggression. Of note, while only 9% of youth/EAs reported either carriage or discharge for self-defense in the prior 3-months, 79% also engaged in RFBs. Multivariate logistic regression identified significant correlates of RFBs, including older age (AOR=1.08), male sex (AOR=1.84), Black race/ethnicity (AOR=1.62), full-time employment (AOR=1.88), substance misuse (AOR=2.34), ED visit for violent injury (AOR=2.29), attitudes favoring retaliation with a firearm (AOR=1.46), friends who own or carry firearms (AOR=4.73), higher levels of neighborhood collective efficacy (AOR=1.05), and higher levels of perceived community violence exposure (AOR=1.04).

Conclusion: Youth/EAs seeking treatment within an urban trauma center report elevated rates of RFBs that potentially increase their risk for subsequent injury and/or death. Prevention efforts should focus on enhancing elements of hospital and community violence prevention programs that address risky firearm behaviors among high-risk populations.

Financial Support: NIH/NIDA K23DA039341; NIH/NCATS UL1TR000433.

40.05 Racial and Socioeconomic Disparities In Prehospital Pediatric Firearm Injury Mortality

Posted on January 31, 2019

M. Hoof1, J. Friedman1, A. A. Smith1, K. Ibraheem1, D. Tatum2, J. Duchesne1, R. Schroll1, C. Guidry1, P. McGrew1  1Tulane University School Of Medicine,Surgery,New Orleans, LA, USA 2Our Lady of The Lake,Surgery,Baton Rouge, LA, USA

Introduction: Prior studies have shown that African American (AA) and uninsured patients have worse outcomes among pediatric victims of firearm injury. However, whether racial and socioeconomic disparities exist in prehospital mortality in pediatric firearm injury (PFI) is unknown. The objective of this study was to determine if the disparities that exist in overall mortality are already present on-scene.

Methods: The National Emergency Medical Services Information Systems (NEMSIS) database was queried for all pediatric (age 0-18) firearm incidents from 2010-2015. A linear regression was carried out using variables associated with mortality on univariate analysis to identify factors associated with prehospital mortality.

Results: 16,808 firearm injuries met study criterion, though information on race and insurance status was only available for 12,268 and 3,080 events, respectively. Victims were more likely to be AA and uninsured. White (CA) children made up a higher proportion of suicide attempts and accidental injuries, whereas most assault victims were AA. CA children had a higher mortality on both univariate and multivariate analysis (MVA). Uninsured children had higher mortality on univariate analysis. On MVA, the odds ratio (OR) of death for uninsured children was 2.169, though this was not significant (p = 0.051). Other factors associated with mortality included anatomic location of injury and intention of injury, with suicide attempts having the highest mortality.

Conclusion: To our knowledge this is the largest study of prehospital mortality in pediatric firearm injury. Uninsured children had higher on-scene mortality on univariate analysis. On MVA, uninsured victims were twice as likely to die as their insured counterparts, though this only approached significance (p = 0.051), perhaps due to insufficient sample size. This trend supports the idea that poor preinjury health status may affect outcome following trauma. However, we did not see higher prehospital mortality in AA children, which has been observed in studies on overall PFI mortality. Therefore, it is possible that the outcomes disparities noted in those studies are related to in-hospital factors. Further studies are needed to examine this topic.

40.04 Medicaid Expansion Has No Effect on Mortality but May Improve In-hospital Outcomes

Posted on January 31, 2019

D. Tatum1, S. Taghavi1,2, A. Smith2, S. Baker2, R. Schroll2, C. Guidry2, P. McGrew2, C. McGinness2, T. Jacome1, J. Duchesne2  1Our Lady of the Lake Regional Medical Center,Trauma Specialist Program,Baton Rouge, LA, USA 2Tulane University School Of Medicine,Department Of Surgery,New Orleans, LA, USA

Introduction:
As part of the Affordable Care Act (ACA), Medicaid in Louisiana was expanded to include non-disabled, low income, non-elderly adults, providing a unique opportunity to examine the effect of insurance status on trauma patients in a historically low socioeconomic status (SES) region. How Medicaid expansion has influenced outcomes in an urban population with a high volume of uninsured patients has yet to be examined on the local level in Louisiana. We hypothesized that there would be a significant decrease in uninsured patients without a concomitant decrease in mortality.

Methods:
All patients aged 18 – 64 were identified in the trauma registries at two major regional trauma centers. Medicaid expansion in Louisiana began July 1, 2016. As such, the pre-expansion study period was July 1, 2015 – June 30, 2016. The post-expansion period was July 1, 2016 – June 30, 2017. Chi Square and Mann-Whitney U test were used for statistical analysis. Significance was set at P < 0.05.

Results:
A total of 8,088 patients met study criteria. Of these, 3,607 (45%) were pre-expansion, 4,481 (55%) were post-expansion. After expansion, Medicaid rates doubled from 22% to 44% (P< 0.001), and the uninsured rate decreased from 40% to 21% (P<0.001). Though number of patients increased post-expansion, median (IQR) ISS decreased significantly (see table). There was no difference in race (P = 0.782), injury type (66% blunt pre-expansion vs 66% blunt post expansion; P =0.757), or overall mortality (4.7% pre vs 3.7% post; P = 0.194); however, patients in the post-expansion cohort had shorter hospital and ICU length of stay.

Conclusion:
While Medicaid expansion in Louisiana has not reduced mortality in nonelderly trauma patients, it may have resulted in better in-hospital outcomes. In addition, expansion may explain increased patient volume and decreased ISS, as formerly uninsured patients with mild injuries may have refused care in the past. Future studies are needed to further delineate the impact of Medicaid expansion in the trauma population.

40.03 Post-Discharge Outcomes in Penetrating Pediatric Trauma at an Urban Level-one Trauma Center

Posted on January 31, 2019

G. M. Siegel3, T. Lee3, C. Wakefield3, A. Katrikh3, D. Webster3, J. Poirier1, J. Mis2, A. Shah1, M. Kaminsky2  1Rush University Medical Center,Pediatric Surgery,Chicago, IL, USA 2John H. Stroger, Jr. Hospital of Cook County,Trauma,Chicago, IL, USA 3Rush Medical College,Chicago, IL, USA

Introduction: In Chicago, gun violence significantly impacts individuals younger than eighteen. In 2017, 16% of shootings involved individuals younger than eighteen years of age with 115 resulting in a fatality. It is well-established that pediatric patients are at greater risk for failure to follow-up (FTF) after a penetrating trauma. Existing literature on trauma follow-up does not explicitly examine FTF among pediatric trauma patients at a safety net, level one trauma center that serves an area with high incidence of gun violence. A better understanding of how to serve this complex population may aid in reducing recidivism, adverse outcomes post-discharge, and the perpetuation of violent crime. We sought to assess patterns and rates of FTF across specialties, emergency department (ED) utilization, and re-hospitalization for surgical pediatric patients post-penetrating trauma at a publicly funded level-one trauma center in Chicago.

Methods: A retrospective cohort study was conducted on surgical patients (ages 0-18) with penetrating trauma admitted from 2008-2016. To examine potential predictors of FTF, nested linear models were created and compared from EMR data and U.S. Census Bureau data. 

Results: We reviewed the records of 216 patients (average age of 15.8 ± 3.6) that suffered a penetrating traumatic injury (73.1% assault with a firearm) with an injury severity score (ISS) of 11.6 ± 13.9. These individuals reside in ZIP Codes with a median household income of $37,372, which was significantly below the city of Chicago median of $66,020 in 2016. Overall follow-up compliance rate was 65.9% with a 16% post-discharge ED visit and 9.3% re-hospitalization rate within 12 months. Ethnicity (p=0.01) and ethnicity, sex, and race (p=0.04) in combination were found to be independent predictors of follow-up compliance. ICU length of stay, hospital length of stay, and ISS did not predict follow-up compliance (p-values 0.09 to 0.78).

Conclusion: We found marked differences in overall patterns and rates among the pediatric trauma population in terms of ED recidivism and FTF than cited in previous literature. Further study is necessary to expand our sample population and determine characteristics driving follow-up, ED utilization, and re-hospitalization among pediatric trauma patients in low-socioeconomic communities with a high burden of gun violence. Additionally, decreasing FTF and improving the quality of post-discharge care is a potential mechanism for cost-containment and reducing long-term recidivism. Establishing benchmarks for pediatric trauma follow-up and assessing post-discharge outcomes for pediatric patients should be a priority for all trauma centers.

 

40.02 Point of Care Ultrasound by EMS in Rural Settings can Identify Injuries and Change Triage

Posted on January 31, 2019

M. Helm2, E. Villegas2, P. Mammen2, N. Ellis2, S. Joseph1  1Valley Health System,Las Vegas, NV, USA 2Texas Tech University Health Sciences,Department Of Surgery,Odessa, TX, USA

Introduction:

In rural settings, timely access to medical care is the main contributor to complications. In West Texas, the average time from injury to EMS arrival is approximately 80 min and arrival to the definitive site of care is over 167 min. 

Early identification of injury reduces triage time, enhances appropriate utilization of resources, and improves outcomes. Point of Care Ultrasound (POCUS) is used to visualize major injuries and identify sources of shock. However, image acquisition and interpretation in the prehospital setting is difficult, requires extensive training, and may be time consuming. We set out to see if first responders could acquire adequate images with minimal training and quantify the added triage time needed for POCUS. 

Methods:

EMS from rural counties were trained using a 2-hour hands-on course in the use of POCUS. Images included neck, pulmonary, cardiac, abdomen, and pelvic windows. Surveys of first responders were used to assess skill acquisition and usefulness of training sessions.   

EMS were then given POCUS for a 3 month field test. Images were saved for evaluation. Surgeons reviewed images to assess image quality and identification of target structures.

After completion of the field test, first responders were surveyed to assess ease of use, technical complications, and total added time.  

Results:

63 first responders were trained on POCUS. All trainees felt that POCUS would be beneficial in triage. 82.5% (52/63) trainees wished to have further training. 30% (19/63) had difficulty with the pulmonary images.

57 patients were recorded by the trained EMS, with 19 patients for trauma.  

Cardiac and abdominal images were good quality and target structures were identified in 93% of all patients. 2 patients had images concerning for free intraperitoneal fluid. Tracheal imaging was done on only 6 patients.

39 EMS completed follow up surveys. All reported the POCUS was easy to use. The average time to complete POCUS was 3 minutes. Technical difficulties included the length of the probe cord, charging of the devise on an ambulance, and ambient lighting within the ambulance. Interest in expanded use included transfer of images, vascular access, imaging for other conditions, and interpretation of images while in transit.       

Conclusion:

EMS can perform POCUS after a short hands-on session. Images obtained appear to be adequate to identify major injuries. Triage decisions and field resources can be maximized using POCUS. We believe the cost of adding POCUS to rural EMS is offset by improved triage time, reduction of wasted resources, and improved survival.

We note the expanded use of POCUS by EMS for patients with other conditions may further reduce the upfront cost of this program. Finally, we noted great interest from the rural hospital emergency room to have access and training to this tool.   

We recommend Wifi capabilities to allow image interpretation while in transit and direct communication between EMS and trauma surgeons in rural areas.

40.01 National Application of the ACS NBATS Tool: Geographic Distribution of Trauma Center Need

Posted on January 31, 2019

M. P. Jarman1, T. Uribe-Leitz1, Z. G. Hashmi1, A. Salim1,2, A. H. Haider1,2  1Brigham And Women’s Hospital,Center For Surgery And Public Health,Boston, MA, USA 2Brigham And Women’s Hospital,Division Of Trauma, Burns, And Critical Care Surgery,Boston, MA, USA

Introduction: Trauma centers (TCs) in the US are geographically concentrated in urban regions and are not distributed according to population needs. This unequal distribution has created barriers in access to trauma care for large segments of the population. The objective of this study is to identify geographic regions in need of additional trauma centers in the US and demonstrate national application of the American College of surgeons (ACS) Needs Based Assessment of Trauma Systems (NBATS) areas using readily available secondary data.

Methods:  Using 20 State Inpatient Datasets from the 2014 Healthcare Cost and Utilization Project, we identified severely injured patients (Injury Severity Score ≥ 16) treated at TCs (Level I/II/III) and non-TCs. We then aggregated injury incidence to the Hospital Referral Regions (HRR) level, and linked incidence with data from the American Hospital Association Annual Survey, the Dartmouth Atlas of Health Care, and the American Trauma Association Trauma Information Exchange Program. We then used multinomial logistic regression models to estimate HRR-level injury incidence in the remaining 30 states. We applied the ACS NBATS tools to the 306 HRRs in the 48 Contiguous United States and identified the number of additional Level I/II TCs needed in each region to care for the expected volume of severely injured patients.

Results: Of the 306 HRRs examined, we identified 115 (37.6%) as needing additional TCs (Figure), including 18 (5.9%) without existing Level I/II TCs. Of the HRRs in need of additional resources, 43 (37.4%) needed one additional TC to meet projected demand for trauma care, 51 (44.3%) needed two additional TCs, 15 (13.0%) needed three additional TCs, and 6 (5.2%) needed four additional TCs. HRRs in need of additional TCs were often geographically clustered, indicating the presence of complex trauma service areas with substantial geographic barriers to TC care. HRRs in need of additional trauma resources were found in all regions of the US, and included both large population centers with multiple existing TCs and rural regions with small patient populations.

Conclusion: This is the first study to demonstrate the feasibility of using public data sources to apply the ACS NBATS tool at a national level. We identified several regions in the US where increased trauma care resources are essential to meet population-based demand for trauma care. By examining the national distribution of trauma care needs, we can identify regional patterns in the distribution of TC resources, and support trauma system organizational decisions that optimize access to care. Based on our findings, we propose the use of HRRs to standardize projections of trauma service need when using the ACS NBATS tool.

17.20 ROTEM as a Predictor of Mortality in Trauma Patients with an Injury Severity Score Greater than 15.

Posted on January 31, 2019

A. R. Smith1, S. Karim1, R. J. Reif1, W. C. Beck1, J. R. Taylor1, B. Davis1, A. Bhavaraju1, M. K. Kimbrough1, K. W. Sexton1  1University Of Arkansas for Medical Sciences,Department Of Surgery, Division Of Trauma And Acute Care,Little Rock, AR, USA

Introduction:  The Injury Severity Score (ISS) is an important tool for grading the severity of injury to trauma patients. Major trauma is commonly defined using an ISS threshold of 15 and has been shown to correlate with mortality likelihood, length of hospital stay, and the need for major surgery. Assessing hemostatic function in a timely manner is crucial for these patients in order to reduce the risk of mortality. Rotational thromboelastometry (ROTEM) is a whole blood viscoelastic hemostasis analyzer that allows for the detection of and differentiation between coagulopathies. The purpose of our study is to retrospectively evaluate trauma patients with an ISS greater than 15 who also received ROTEM to determine if ROTEM is a better predictor of mortality than conventional coagulation testing. 

Methods:  We performed a retrospective review of all trauma patients with ROTEM and an ISS greater than 15 admitted to the emergency department between November 2015 and August 2017. A total of 301 patients met the aforementioned criteria and were included in this study. Univariate and bivariate statistics were performed using JMP Pro (Cary, NC). Each patient was sorted into groups based on their coagulation phenotype (hypocoaguable, normal coagulation, hypercoaguable) for both ROTEM and conventional coagulation tests (Partial Thromboplastin Time [PTT], Prothrombin Time [PT], International Normalized Ratio [INR]) and compared the test results to their relation with mortality. Nominal logistic regression was performed.

Results: For the 301 patients included in the study, significant predictors of mortality included ROTEM APTEM Clotting Time (CT), ROTEM APTEM Alpha Angle, and ROTEM INTEM Clot Formation Time (CFT), with ROTEM APTEM CT being the most significant. On nominal logistic regression, APTEM CT (p=.007), APTEM Alpha Angle (p=.028), and INTEM CFT (p=.037) were the only significant predictors. PTT (p=.059), PT (p=.141), and INR (p=.634) were not significant predictors of mortality in this data set. 

Conclusion: ROTEM APTEM Clotting Time, which is the time from start of measurement until initiation of clotting in the presence of aprotinin, a fibrinolysis inhibitor, is a significant predictor of mortality in trauma patients with an ISS greater than 15. ROTEM APTEM Alpha Angle and ROTEM INTEM Clot Formation Time are also significant predictors of mortality, whereas conventional coagulation tests did not have a significant contribution to predicting mortality in this patient population.
 

« Previous Page
Next Page »