13.04 The Morbidity of Survivorship in Congenital Diaphragmatic Hernia

Posted on January 31, 2019

C. M. Miller1, K. P. Lally1, M. T. Harting1  1McGovern Medical School at UTHealth and Children’s Memorial Hermann Hospital,Pediatric Surgery,Houston, TX, USA

Introduction:  As advances in care re-define survivorship in congenital diaphragmatic hernia (CDH), particularly among infants with severe CDH, the onus is shifting to long-term management. CDH survivors are frequently challenged by pulmonary (PULM), gastrointestinal (GI), neurologic (NEURO), and orthopedic (ORTHO) morbidities. Our objective was to characterize long-term morbidity in the modern era of care and the association with neonatal risk factors as defined by CDH study group (SG) Stage (A-D).

Methods:  A single center, retrospective cohort study of survivors born 2011-2017 was performed. Patients with CDH clinic visits between 1-2 and 4-5 years of age were included. Patient demographics, prenatal, and neonatal characteristics were reviewed. The primary outcomes were morbidities at two and five year (±12 months) follow-up. Morbidities included PULM, GI, NEURO, and ORTHO.

Results: A total of 37 patients were included in the study cohort. There were 27 patients at 2 years and 10 patients at 5 years of age. Overall morbidity was 88.9% and 90% at 2 and 5 years, respectively. At two years of age, 11(41%) had PULM, 13(48%) GI, 9(33%) NEURO, and/or 9(33%) ORTHO morbidities. Of the 11 patients with PULM morbidity at two years, the majority were taking a pulmonary medication (n=8, 72.7%). GI morbidity included gastroesophageal reflux (n=4, 30.8%), supplemental feeds (n=3, 23.7), constipation (n=3, 23.1%), and GI medication (n=5, 38.5%). NEURO morbidity was dominated by neurodevelopmental delay (n=8, 88.9%). ORTHO morbidity included pectus excavatum (n=4, 44.4%), rib abnormalities (n=4, 44.4%), and spine abnormalities (n=1, 11.1%). At five years of age, 8(80%) had PULM, 2(20%) NEURO, 3(30%) GI, and/or 4(40%) ORTHO morbidities. A large portion of pulmonary morbidity at 5 years included taking a pulmonary disease related medication (n=6, 75%), asthma (n=3, 37.5%), and sleep disordered breathing (n=3, 37.5%). Only GI medication occurred more than once at 5 years (n=2, 66.6%). NEURO morbidity was confined to developmental delay (n=2, 100%). ORTHO morbidity included pectus excavatum (n=2, 50%), rib abnormalities (n=1, 25%), and spine abnormalities (n=2, 25%). Morbidity for both 2 and 5 years relative to CDHSG Stage is shown (Table). Analysis by χ2 indicated no difference in observed incidence of morbidity by CDHSG Stage.

Conclusion: Survivors with CDH continue to face significant morbidity at two and five years. At two years the spectrum of morbidity resembles that of discharge, with a predilection toward gastrointestinal morbidity. Alternatively, at five years, the primary source of morbidity is pulmonary. These preliminary data will inform multi-center collaboration in long-term CDH data collection.
 

13.03 The Role of Surgical Shunts in the Treatment of Pediatric Portal Hypertension

Posted on January 31, 2019

A. R. Cortez1,2, A. Kassam1,2, C. J. Nathan2, T. M. Jenkins2, J. D. Nathan1,2, M. H. Alonso1,2, G. M. Tiao1,2, A. J. Bondoc1,2  1University Of Cincinnati,Department Of Surgery,Cincinnati, OH, USA 2Cincinnati Children’s Hospital Medical Center,Division Of Pediatric General And Thoracic Surgery,Cincinnati, OH, USA

Introduction: Creation of a surgical shunt to divert portal blood flow plays a major role in the treatment of medically refractory portal hypertension (pHTN). Similar to the adult population with end-stage liver disease, pediatric patients suffer from intra-hepatic pHTN, but more commonly they experience pre-hepatic pHTN due to extrahepatic portal vein thrombosis (EPVT). Consequently, both portosystemic shunts (PSS) and the meso-Rex bypass (MRB) are potential interventions for these patients. Herein, we set out to evaluate our center’s experience with surgical shunts for the treatment of pediatric pHTN.

Methods: All patients who underwent PSS or MRB at a single institution from 2008-2017 were reviewed. Patient demographics, operative details, and outcomes were collected. Shunt types were compared using Wilcoxon rank-sum test for continuous variables and Fisher's exact test for categorical variables. Intervention-free shunt patency was calculated using Kaplan-Meier survival analysis with differences between groups estimated by a Log-rank test. A p-value <0.05 was considered statistically significant.

Results: 34 patients underwent surgical shunt creation during the 10-year study period. The median age was 7.7 years (IQR 4.3-12.0). 29 patients (85.3%) had pre-hepatic pHTN and 5 patients (14.7%) had intra-hepatic pHTN. The primary manifestations of pHTN among these patients were esophageal varices (97.1%) and gastrointestinal bleeding (76.5%). 18 patients (52.9%) had an MRB, 10 patients (29.4%) had a splenorenal shunt, and 6 patients (17.7%) had a mesocaval shunt. These patients experienced good short- and long-term outcomes including minimal wound complications (3.8%), re-bleeding events (11.8%), and mortality (2.9%). Moreover, there were no differences in postoperative outcomes among the groups (p>0.05 for each). Among all patients, 10 (29.4%) experienced shunt occlusion, 4 of which occurred in the immediate postoperative period and required urgent intervention. Overall, the 1- and 5-year intervention-free shunt patency rates were 75% and 63%, respectively. Although there was no statistically significant difference in shunt patency among the groups (p=0.266), mesocaval shunts appeared to have poorer rates of primary patency (Figure).

Conclusions: Pediatric patients suffer significant morbidity from the sequalae of pHTN. Our experience reinforces the feasibility of surgical shunts as an effective means of treating this disease process with low rates of morbidity and mortality. Moreover, both the PSS and MRB effectively improve patient symptoms with acceptable long-term, intervention-free shunt patency rates.

13.02 Magnet Ingestion in Children: Can We Minimize Procedures?

Posted on January 31, 2019

B. J. Slater1, T. Tran2, M. Slidell1  1University Of Chicago,Pediatric Surgery,Chicago, IL, USA, 2University of Illinois at Chicago Metropolitan Group Hospitals,Surgery,Chicago, IL, USA

Introduction: Magnet ingestion and its complications in children have been well described. When multiple magnets or a magnet and a second magnetic object are swallowed, complications can arise because they can attract each other between two loops of intestine leading to intestinal necrosis, fistulation, obstruction, or perforation. Although some algorithms have been proposed, there are no definitive guidelines for the evaluation and management of these patients and there is significant variation in practices. Multiple radiographs, CT scans, and other imaging modalities are often used to assess for position, progression, and signs of complications. In addition, some advocate for early surgical intervention to remove the magnets. The purpose of this study was to review the outcomes of patients with magnet ingestion at several institutions and determine the feasibility of minimizing the number of radiographs and invasive procedures for selected patients with magnet ingestion.

Methods: This is a retrospective study evaluating 20 patients with magnet ingestion at 4 institutions from 6/2012 until 12/2017. Demographic, diagnostic, and therapeutic procedures were recorded. Magnet ingestion features such as location, symptoms, and complications were included.

Results: The median age was 8 years (2.7-13) and weight was 27.3kg (range 17 to 48). 70% of the patients were male. 75% of patients ingested multiple magnets or a magnet and another metallic object, and the average number of magnets ingested was 4.2 (range 1 to 30). 20% of the patients were symptomatic at admission. All of the patients had plain radiographs (average number 4.3, range 1 to 10), 4 had CT scans, and 1 had an UGI study. 40% of the patients (8) had a bowel regimen in the hospital, 35% (7) underwent endoscopy, and 35% (6) underwent surgical procedures. 45% of the patients had no intervention. Of patients who underwent surgery, half were performed laparoscopically and half were begun laparoscopically with exteriorization of bowel through the umbilicus. Having symptoms, especially pain, was correlated with having an intervention (correlation factor .6)

Conclusion: Magnet ingestion in children is associated with complications. There are a subset of asymptomatic patients in whom surveillance with decreased number of radiographs and possible outpatient observation may be feasible.  

 

13.01 Early Surgery Indicated for Recurrent Spontaneous Pneumothorax in Children and Adolescents

Posted on January 31, 2019

C. Tragesser1, B. Gray2, M. Landman2  1Indiana University School Of Medicine,Indianapolis, IN, USA 2Indiana University School Of Medicine,Department Of Surgery, Division Of Pediatric Surgery,Indianapolis, IN, USA

Introduction:  Primary spontaneous pneumothorax (PSP) occurs most often in adolescent patients. There is consensus that surgical intervention plays an important role in preventing recurrence. However, the optimum timing of surgery is debated. We hypothesize that clinical and radiographic factors are associated with eventual need for surgery.

Methods:  We searched the medical record for PSP patients between ages 9 and 21 treated from 1/1/08 to 12/31/17 and collected data on chest tube management, radiographic measurements, operative management, and recurrence. We performed univariate analysis on relationships between admission events and eventual surgery or other management strategies.

Results: We identified 68 PSP admissions from 31 patients. Considering only first-time admissions, there was no association between eventual surgery and clinical factors and radiographic findings.  The single factor associated with eventual surgery was history of pneumothorax in any lung (p=0.015). For patients with prior pneumothorax who underwent surgery, operation the day after admission would have reduced hospital stay by an average of 1.5 days (min=0, max=9) and an average of 2.2 days (min=0, max=10) if performed on the day of admission, with a mean 1.85 fewer chest x-rays (min=0, max=7). Considering only first admissions, ipsilateral recurrence rate was 16.7% after surgery, 46.7% after chest tube alone, and 100% after observation alone.

Conclusion: This analysis suggests that though eventual surgery is difficult to predict, ipsilateral recurrence rate is reduced following surgery. Furthermore, earlier operation in recurrent patients could reduce resource utilization. Thus, expedited surgical treatment may merit consideration in patients with a history of pneumothorax.

 

12.20 Massive Transfusion Protocol in Geriatric Trauma Patients: Is There a Need for Stricter Criteria?

Posted on January 31, 2019

M. Hamidi1, M. Zeeshan1, T. O’Keeffe1, A. Tang1, E. Zakaria1, A. Northcutt1, N. Kulvatunyou1, L. Gries1, B. Joseph1  1University Of Arizona,Trauma And Acute Care Surgery,Tucson, AZ, USA

Introduction:
Massive Transfusion (MT: >10pRBCs/24 hours) has revolutionized resuscitation during the last decade. However, its impact on geriatric patients remains unclear. The aim of our study was to assess the outcomes in geriatric trauma patients who received MT.

Methods:
We performed a 2-year (2013-2014) analysis of the TQIP and included all adult trauma patients who received MT. We stratified the patients into two groups; geriatrics (age≥65y), and non-geriatrics (age<65y). Outcome measures were blood products transfused within 24h and in-hospital mortality. Kaplan Maier and regression analysis were performed to control for demographics, injury and vital parameters.

Results:
We analyzed 416,104 trauma patients. 4236 patients received MT, of which 12% were geriatric patients. Median ISS was 22[10-34]. There was no difference in the ISS (p=0.57), blood products received (p=0.34) or type of hemorrhage control between the two groups (p=0.77). Geriatric patients who received MT had a high mortality (63% vs. 43%, p<0.001), were more likely to have a blunt injury (93% vs. 63%, p<0.001), and less likely to receive operative intervention (61% vs. 75%, p<0.001) compared to their younger counterpart. Fig1. In the geriatric group, the aOR was highest for patients aged≥80 (OR: 4.8 [2.9-7.8]) followed by age70-79 (OR: 2.4 [1.6-3.2]) when compared to age65-69y.

Conclusion:
Two thirds of the geriatric trauma patients who received massive transfusion died. Geriatric patients requiring MT had higher rate of blunt injuries and were less likely to undergo operative intervention compared to their younger counterpart. Revision of the MT criteria in geriatric patients may improve outcomes.
 

12.19 Readmissions after Violence-related Trauma: what is the true recidivism?

Posted on January 31, 2019

M. Castillo-Angeles1, T. Uribe-Leitz1, S. Nitzschke1, D. Nehra1, A. H. Haider1, A. Salim1, R. Askari1  1Brigham And Women’s Hospital,Boston, MA, USA

Introduction:  Violence continues to be a significant public health burden. However, little is known about the causes and associated social and patient factors for unplanned readmissions among these trauma patients. Our goal was to determine the incidence and burden of violence-related trauma and to identify predictors of readmission.

Methods:  This is a retrospective study using the Florida State Inpatient Database 2010-2014. The inclusion criteria were all patients admitted for violence-related traumatic injuries. These were classified as self-inflicted, assault (violence conducted by an unknown perpetrator) or maltreatment (acts of abuse and neglect conducted by a caregiver or intimate partner). Demographics and clinical characteristics were collected. Outcomes were 30-day readmission and in-hospital mortality. Multivariable regression models were used to identify factors associated with 30-day readmission and mortality.

Results: A total of 306,835 violence-related admissions were identified; 115,325 (37.59%) were readmitted within 30 days. The overall in-hospital mortality rate was 0.67%. Mean age was 42 years (SD 16.2), 135,610 (44.2%) were female and 67% were white. Self-inflicted violence was the most common type of admission (88.9%). After adjusted analysis, predictors of readmissions were age (adjusted odds ratio [aOR] 1.01, 95% confidence interval [CI] 1.01 – 1.02), female gender (OR 0.64, 95% CI 0.63 – 0.65), having public insurance (vs. private) (OR 2.45, 95% CI 2.40 – 2.51), being a racial minority (OR 0.85, 95% CI 0.83 – 0.86) and Charlson score (OR 1.26, 95% CI 1.25 – 1.28). Age, gender, being uninsured, black race and Charlson score were associated with increased in-hospital mortality.

Conclusion: Violence-related trauma have a high readmission rate and consequently high resource utilization. Multiple factors associated with disparities influence readmission among these violence-related trauma patients. Further work is needed to identify intervention strategies that may be beneficial to reduce readmissions after violence-related trauma.

 

12.18 Implications of Expanding Trauma Services Within a Health System

Posted on January 31, 2019

E. E. Freeh1, G. M. Niziolek1, R. M. Boudreau1, J. Baker1, D. A. Millar1, A. T. Makley1, T. A. Pritts1, M. D. Goodman1  1University Of Cincinnati,Department Of Surgery, College Of Medicine,Cincinnati, OH, USA

Introduction:  With the number of traumatic injuries continuing to increase in the United States, new trauma centers are opening across the country. In July 2013, a level III trauma center affiliated with the only level I trauma center in a metropolitan region was opened approximately 25 miles away. The goal of this study was to evaluate the impact and outcomes of opening a level III trauma center within a health care system while also assessing the effect this opening had on the region’s level I trauma center volume. 

Methods:  A retrospective review of the trauma registries at the level I and the level III centers was undertaken from July 2013 to December 2017. Patients were grouped by their zip code of residence and the number of patients from each zip code that presented to both trauma centers over time was examined. The median length of stay and injury severity score of patients that presented to the level III was also tracked over time.

Results: A total of 3,172 patients presented for a trauma surgery evaluation at the level III trauma center during the study period. The number of patients that presented monthly gradually rose from 36 patients in July 2013 to 80 patients in June 2017. Half of the patients (49.5%) evaluated at the level III center were residents from one of three zip codes immediately adjacent to the level III trauma center.  Admissions from those zip codes gradually increased over the study period. Another quarter (27.8%) of patients that presented to the level III center were residents of one of 8 zip codes located within a 20 mile radius of the trauma center and admissions from this demographic remained stable over the study period. Interestingly, the number of patients that presented to the level I from those 11 zip codes remained stable over the study period. A gradual increase over time in the percent of trauma service admissions at the Level III trauma center was mirrored by a decrease over time in injured patients who were discharged from the emergency department. Over the course of the study period, the median ISS at the Level III trauma center increased from 5 to 8.5 (p<0.05) between the first 6 month period and that last 6 month period. 

Conclusion: Opening a level III trauma center has allowed augmentation of the trauma services provided within a health care system. While volume related metrics may serve as an initial indicator of regional trauma volume, ensuring appropriate triage, transfer, and quality of care remain integral to the integrity of a regional trauma system as patients with increasing injury severity are evaluated at the level III trauma center.

 

12.17 Outcomes of Traumatic Brain Injury in Patients on Pre-Injury Antiplatelet Therapy

Posted on January 31, 2019

A. A. Fokin1, J. Wycech1,2, S. Mansour1,3, A. Tymchak1,3, A. Zuviv1,3, M. Crawford1, I. Puente1,2,3,4  1Delray Medical Center,Trauma Services,Delray Beach, FL, USA 2Broward Health Medical Center,Trauma Services,Fort Lauderdale, FL, USA 3Florida Atlantic University,College Of Medicine,Boca Raton, FL, USA 4Florida International University,College Of Medicine,Miami, FL, USA

Introduction: Use of antiplatelet therapy (APT) has become widespread due to an aging population and increased incidence of cardiovascular disease. While effective at mitigating the risks of cardiovascular disease, APT may increase the risk of hemorrhage in traumatic brain injury (TBI). Our study seeks to determine outcomes in TBI patients on different pre-injury APT: aspirin (ASA), Plavix and dual APT of ASA and Plavix.

Methods: This IRB approved retrospective cohort study included 346 patients with TBI between the ages of 17 and 101, who were delivered to a level 1 trauma center between 1/1/2015 and 3/30/2018. Patients were divided into 3 groups, by type of pre-injury APT received: Group 1 patients had ASA only (n=203), Group 2 Plavix only (n=56), and Group 3 had dual APT (n=87). Patients were excluded if they were also taking anti-coagulants. Age, Injury Severity Score (ISS), Glasgow Coma Score (GCS), Rotterdam computed tomography (CT) score, Marshall CT score, incidence of intracranial hemorrhage (ICH), venous thromboembolism (VTE) prophylaxis, need for neurosurgical intervention, duration of mechanical ventilation (DMV), Intensive Care unit length of stay (ICULOS), hospital LOS (HLOS), incidence of re-admission and mortality were compared.

Results: Between the groups, mean age (81.5 vs 82.2 vs 80.3 years; p=0.3), GCS (14.3 vs 14.6 vs 13.8; p=0.1), Rotterdam score (2.6 vs 2.6 vs 2.6; p=1.0), Marshall score (1.1 vs 1.0 vs 1.2; p=0.1), incidence of ICH (83.3% vs 87.5% vs 93.1%; p=0.1), DMV (4.9 vs 8.0 vs 6.14 days; p=0.7), ICULOS (3.0 vs 2.5 vs 3.9 days; p=0.2), HLOS (3.6 vs 4.0 vs 4.2 days; p=0.4), readmission rate (6.0% vs 3.6% vs 4.6%; p=0.7) and mortality (9.9% vs 8.9% vs 12.6%; p=0.7) were comparable. Mean ISS was different between groups (11.9 vs 12.5 vs 14.1; p=0.04), with a significantly higher ISS in Group 3 than Group 1 (p=0.04). The percent of patients receiving VTE prophylaxis was also significantly different between groups (8.9% vs 12.5% vs 6.9%; p<0.001), showing a significance of p<0.001 between Groups 1 and 3, and p=0.004 between Groups 2 and 3. Incidence of neurosurgical intervention was also different between groups (3.0% vs 1.8% vs 11.5%; p=0.004), showing significantly higher incidence in Group 3 than Group 1 (p=0.01) and Group 2 (p=0.03) (Fig. 1).

Conclusion: Patients on dual APT had an increase in neurosurgical interventions compared to patients taking either antiplatelet agent alone. While there was no difference in mortality among the three groups, there was a tendency toward increased incidence of ICH in patients on dual APT when compared to either group alone. VTE prophylaxis was administered less often to patients who were taking Plavix or dual APT when compared to patients taking ASA alone.

 

12.16 Repeat CT in Patients with Head Trauma on Pre-Injury Antiplatelet Therapy with Negative Initial CT

Posted on January 31, 2019

S. Mansour1,2, A. Tymchak1,2,3, A. Zuviv2, M. Crawford2, J. Wycech2,3, A. A. Fokin2, I. Puente1,2,3,4  1Florida Atlantic University,College Of Medicine,Boca Raton, FL, USA 2Delray Medical Center,Trauma Services,Delray Beach, FL, USA 3Broward Health Medical Center,Trauma Services,Fort Lauderdale, FL, USA 4Florida International University,College Of Medicine,Miami, FL, USA

Introduction:
There is a debate on the efficacy of repeat head computed tomography (RHCT) in traumatic brain injury (TBI). There is also a trend to increase precautions taken with elderly TBI patients, especially those on antiplatelet therapy (APT). Our study seeks to assess the need for RHCT in TBI patients on pre-injury APT who had negative findings on the initial head CT (IHCT).

Methods:
This IRB approved retrospective cohort study included 58 TBI patients on pre-injury Aspirin, Clopidogrel or both between the ages of 17 and 101, who were delivered to a level 1 trauma center between 1/1/2015 and 3/30/2018 and had a negative IHCT. Patients were excluded if they were also taking anti-coagulants. Patients were divided into 2 groups: Group 1 with those who did not receive RHCT (n=31) and Group 2 with those who received RHCT (n=27). Age, Injury Severity Score (ISS), Revised Trauma Score (RTS), Glasgow Coma Score (GCS), Rotterdam CT score, Marshall CT score, incidence of intracranial hemorrhage (ICH), platelet count on admission, Prothrombin Time (PT), Partial Thromboplastin Time (PTT), platelet transfusion, venous thromboembolism (VTE) prophylaxis, need for neurosurgical intervention, Intensive Care Unit length of stay (ICULOS), hospital LOS (HLOS), incidence of re-admission and mortality were compared.

Results:

Group 2 had statistically higher mean Rotterdam score (2.0 vs 2.4; p<0.001), incidence of ICH (0.0% vs 48.1%; p<0.001; Fig. 1) and need for platelet transfusions (37.0% vs 9.7%; p=0.01) than patients in Group 1. Furthermore, Group 2 also had a higher percentage of patients who were taking Clopidogrel than Group 1 (25.3% vs 59.3%; p=0.01).

Between Groups 1 and 2, age (80.7 vs 79.6; p=0.4), ISS (8.1 vs 9.0; p=0.6), RTS (7.8 vs 7.8; p=1.0), GCS (14.6 vs 14.5; p=0.9), Marshall score (1.0 vs 1.0; p=1.0), platelet count (246.0 vs 191.0; p=0.2), PT (11.1 vs 10.9 seconds; p=0.1), PTT (26.2 vs 26.6 seconds; p=0.7), VTE prophylaxis (3.1% vs 9.4%; p=0.3), rate of neurosurgical interventions (0.0% vs 6.3%; p=0.2), ICULOS (4.2 vs 2.6 days; p=0.4), HLOS (4.5 vs 2.6 days; p=0.2) and readmission rate (3.2% vs 3.7%; p=0.9) were similar. No patients died in either group.

Conclusion:
In patients with head trauma on APT who had negative IHCT, repeat CT showed ICH in 48.1% of them, however there was no difference in outcomes between patients who did and did not receive RHCT. This can be explained by a relatively small volume of hemorrhage in these patients as demonstrated by mean Marshal Score of 1.0 in these patients. Physicians were more likely to repeat CT scans in patients on Clopidogrel.

12.15 Do Advanced Directives Alter the Outcome of Care in the Trauma Population?

Posted on January 31, 2019

C. J. Rust2, S. Agarwal1, K. Haines1  1Duke University Medical Center,Critical Care Surgery,Durham, NC, USA 2University Of Wisconsin,Madison, WI, USA

Introduction: Advanced directives (AD) give patients more control over their medical management when they cannot voice their own decisions. However, end-of-life planning is not equal across the population. Serious injuries happen suddenly and unpredictably, and advance directives provide a framework from which families and caregivers may better understand patient’s wishes. This analysis aimed to determine what populations have advanced directives prior to admission in trauma and how those directives affected outcomes.

Methods:  Adult patients admitted to the hospital and recorded using the National Trauma Data Bank were reviewed. The primary outcome was presence of AD. Secondary outcomes included mortality, disposition and length of stay (LOS) among AD patients. A multivariate logistic regression model was developed for DNR status controlling for age, sex, race and ethnicity, insurance status and BMI classification. Multiple multivariate linear and logistic regressions for secondary research outcomes controlled for the same set of covariates and AD.

Results: From 2013-2015, 2.1 million inpatients were identified and 33,768 had AD present on admission. Average age of patients with AD was 78.2 ± 11.2 years and 60.6% were female. Most were Caucasian (91.5%) followed by African American (2.6%). Medicare was the most common insurance (76.1%) followed by private (12.5%). Asian and African American patients were less likely to have AD compared to their Caucasian counterparts (0.526, 0.404; p<0.001). Patients with Medicare were more likely while uninsured patients were 0.6 time less likely (1.249, p <0.001) to have AD. Patients with AD had increased likelihood of mortality (4.069; p<0.001) being discharged to SNF and hospice care (1.214, 4.825; p<0.001). Routine discharge, homecare, and leaving against medical advice were all negatively associated with AD (0.676, 0.421, 0.213; p<0.001). AD was positively associated with use of mechanical ventilation (1.581; p<0.001) however patients spent less time intubated (-1.146; p <0.001). Patients that died or were discharged to hospice and had an AD had shorter LOS (-1.160, -1.740; p<0.001). Of patients with AD, uninsured patients had higher mortality rates (2.775; p<0.001) were less likely to be sent to SNF and receive home care (0.482, 0.319; p<0.001). African American patients (1.248; p<0.001) had increased LOS, as did Medicaid patients (0.757; p<0.02). Medicare and uninsured patients spent fewer days on mechanical ventilation than privately insured patients (-1.519, -1.652; p<0.05).

Conclusion: Patients with AD were predominantly elderly white females.  Mortality rates for those with AD were greater than the general population, however, AD did not limit initial care as this population was more likely to be on a ventilator. Those with AD had higher rates of mortality, discharge to hospice and shorter hospital stays, possibly indicating palliative measures were taken. 

12.14 Octogenarians with Blunt Splenic Injury: Not All Geriatrics are the Same

Posted on January 31, 2019

A. Grigorian1, M. Lekawa1, V. Joe1, S. D. Schubl1, T. Chin1, A. Kong1, J. Nahmias1  1University of California, Irvine,Surgery,Orange, CA, USA

Introduction: Geriatric trauma patients (GTP) (age≥65 years) with blunt splenic injury (BSI) have up to a 6% failure rate of non-operative management (NOM). GTPs failing NOM have a similar mortality rate compared to GTPs managed successfully with NOM. It is unclear if there is a linear relationship with mortality in patients ≥  65-years old presenting with blunt splenic injury and undergoing NOM. We hypothesized the failure rate for NOM in octogenarians (aged 80-89 years) would be similar to patients aged 65-79 years, and risk of mortality in octogenarians who fail NOM to be higher than octogenarians managed successfully with NOM.

Methods: The Trauma Quality Improvement Program (2010-2016) was queried for patients with BSI. Those undergoing splenectomy within 6-hours were excluded to select for patients undergoing NOM. Two groups were compared: age 65-79 vs. 80-89. A multivariable logistic regression model was used to determine risk for failed NOM and mortality.

Results: From 43,041 NOM BSI patients, 3,600 (8.5%) were aged 65-79 and 1,236 (2.9%) were 80-89. Both groups had a similar median injury severity score (ISS) (p=0.10) and failure rate of NOM (6.6% age 65-79 vs. 6.8% age 80-89 p=0.82). From those failing NOM, octogenarians had similar units of blood products transfused (p>0.05), however, a higher mortality rate (40.5% vs. 18.2%, p<0.001), compared to patients aged 65-79. Independent risk factors for NOM failure in octogenarians included ≥ 1 unit of packed red blood cells (PRBC) (p=0.039) within 24-hours of admission. Octogenarians that failed NOM had a higher mortality rate compared to octogenarians managed successfully with NOM (40.5% vs 23.6% p=0.001), which remained after adjusting for severe ISS (OR 2.25, CI 1.37-3.70, p<0.001). Late failure of NOM ≥  24 hours (vs. early failure) was not associated with increased risk of mortality (p=0.88) but ≥ 1 unit of PRBC transfused had higher risk (OR 1.88, CI 1.20-2.95, p=0.006).

Conclusion: Compared to patients aged 65-79 with BSI, octogenarians have a similar rate of failed NOM. Octogenarians with BSI who fail NOM have over a two-fold higher risk of mortality compared to those managed successfully with NOM. One or more units of PRBC transfused increase risk for mortality. Late failure of NOM did not increase risk of mortality in octogenarians with BSI.

 

12.13 Trends in Firearm Injuries among Children and Teenagers in the United States

Posted on January 31, 2019

O. A. Olufajo1, A. Zeineddin1, E. De La Cruz1, M. Williams1, E. Cornwell1  1Howard University College Of Medicine,Department Of Surgery,Washington, DC, USA

Introduction:  Gunshot violence in the United States occurs at a magnitude that has drawn increasing national attention. Children and teenagers form a vulnerable group who are often victims of firearm injury. Although there are targeted efforts to reduce firearm injuries in children and teenagers, the national trends of injury in this age-group relative to the adult population is not well studied. This study seeks to measure trends in firearm injuries evaluated in US hospitals comparing children and teenagers to adult victims.

Methods:  Data from the National Trauma Data Bank (2010-2015) was used in selecting patients evaluated for firearm injury. Based on their age at presentation, patients were classified as children and teenagers (< 20 years old) or adults (≥ 20 years old). Using International Classification of Disease, 9th Edition, Clinical Modification, external cause of injury codes, firearm injuries were categorized as assault, self-inflicted, unintentional or undetermined. Changes in the proportion of firearm injuries among children and adolescents relative to the overall population was determined using trend analyses. Patients were sub-stratified by race/ethnicity and by insurance type to evaluate differences in trends by socioeconomic factors.

Results: There were 37709 children and teenagers (18.9%) and 162282 adults (81.1%) in the study cohort. Compared to adults who had firearm injuries, children and teenagers were more likely to be Black (61.1% vs 52.8%) and more likely to be insured (79.6% vs 61.3%) (both P <0.001). Between 2010 and 2015, the number and proportion of children and teenagers who had firearm injuries decreased significantly (6769, 21.7% vs. 6118, 17.3%; P for trend <0.001) (Table). Sub-stratification by race/ethnicity and insurance type showed a similar trend across all groups, more marked among Hispanics (1304, 26.4% vs. 848, 19.8%; P for trend <0.001) and those on Medicaid (2581, 39.7% vs. 2940, 28.1%; P for trend <0.001). Analyses of injury categories among patients with firearm injuries showed a decrease in the number and proportion of children and teenagers with assault between 2010 and 2015 (5411, 22.7% vs. 4558, 17.3%; P for trend <0.001) but a similar trend was not seen in self-inflicted injuries (252, 8.7% vs. 311, 9.5%; P for trend= 0.266).

Conclusion: Despite reductions in the numbers and proportions of firearm injuries among children and teenagers, there remains a significant burden of injury among this age group, particularly among Blacks. Variations in socioeconomic factors are associated with trends in firearm injuries. Continued efforts are necessary to ensure safety and reduce firearm injuries among children and teenagers in the United States.

 

12.12 Multidisciplinary Family Meetings Facilitate the Use of Comfort Measures in Dying Trauma Patients

Posted on January 31, 2019

J. K. Bhangu1,2, B. T. Young1,2, S. E. Posillico1,2, H. A. Ladhani1,2, S. J. Zolin1,2, C. W. Towe2,3, J. A. Claridge1,2, V. P. Ho1,2  1MetroHealth Medical Center,Division Of Trauma, Critical Care, Burns And Emergency General Surgery,Cleveland, OH, USA 2Case Western Reserve University School Of Medicine,Cleveland, OH, USA 3University Hospitals Cleveland Medical Center,Thoracic & Esophageal Surgery,Cleveland, OH, USA

Introduction:
American College of Surgeons guidelines suggest that a structured family meeting in trauma patients with high risk of mortality or permanent disability should be performed to align care with patient goals and avoid life-sustaining care inconsistent with patient values. Multidisciplinary family meetings (MDFM), rather than meetings with a single team, may facilitate decision making by allowing multiple specialists to provide concordant perspectives on prognosis. We hypothesized that use of MDFM would be associated with higher utilization of comfort measures for dying trauma patients.

Methods:
All trauma patients who died at an academic adult level I trauma center (December 2014 to December 2017) were reviewed. Patients who died within 24 hours of arrival or who were transferred to non-trauma services were excluded. Age, injury mechanism, length of stay (LOS), and use of tracheostomy or gastrostomy tube were collected. Code status was categorized as Full Code, Do Not Resuscitate (DNR) or Comfort Care (CC). DNR allowed escalation of care and aggressive measures until cardiac arrest. For CC, only interventions which would maximize the patient’s comfort were instituted.  A family meeting was defined as any documented discussion with family addressing prognostication and/or goals of care. A MDFM required the presence of least 2 disciplines, including caregivers from different specialties (typically trauma, neurosurgery, or palliative care), social workers, or chaplains. Comparisons were made between patients with and without MDFM, via Wilcoxon rank sum or Fisher’s exact test. 

Results:
177 patients met inclusion criteria. 68% of patients were male; median age was 70 (IQR 58-83). Most patients were admitted after blunt trauma (90%). The median hospital LOS was 6 days (IQR 4-12). 49 patients (28%) had at least one MDFM, 117 (66%) of patients had meetings with individual teams only, and 11 (6%) had no documented meetings. Patients with and without MDFM had similar age and LOS. At the time of death, 73% of patients were CC, 18% were DNR, and 9% were full code.  Patients with MDFM were more likely to be CC at the time of death (88% vs 68%, p<0.05), and less likely to be DNR (8% vs 23%, p<0.05). (Table) 

Conclusion:
In our center, families commonly discussed prognosis and goals of care with single teams. MDFM were less commonly performed but were associated with a higher use of comfort measures.  We encourage the use of MDFM in the trauma setting and are adopting a protocol to identify patients appropriate for MDFM in our institution.

12.11 HIV Screening And Early Referral In The Trauma Population. The Experience Of A Large County Hospital

Posted on January 31, 2019

L. R. Taveras1, T. Turner-Wentt2, S. W. Ross1, H. B. Cunningham1, T. D. Madni1, J. B. Imran1, M. L. Pickett1, M. Zhou1, S. Park1, H. A. Phelan1, M. W. Cripps1  1University Of Texas Southwestern Medical Center,Department Of Surgery, Division Of General And Acute Care Surgery,Dallas, TX, USA 2Parkland Health and Hospital System,HIV Department,Dallas, TEXAS, USA

Introduction:
A significant proportion of HIV-infected individuals are unaware of their serologic status. The trauma population has a higher HIV prevalence. An HIV screening program was implemented in partnership with the state health department. Our objective is to evaluate the use and results of the program in our trauma population.

Methods:
Patients evaluated in the emergency services department are screened for HIV. This is a retrospective analysis of prospectively collected data from July 2015 to February 2018. Counseling is given to all positive screens and referral to specialty care is offered. Patients were divided based on screening status. Significance was set at α = 0.05 on two-tailed testing. Student's t test, and a chi-squared test were performed where appropriate.

Results:

6175 patients were evaluated in the trauma department. Of those, 449 (7.3%) patients were previously screened the prior year and were excluded. The remaining 2024 (35.3%) patients were screened and 27 (1.3%) screened positive (table 1). The majority of the patients that screened positive were male (100%), white (77.8%), not-hispanic (63.0%) and had no insurance (70%). 25 (92.6%) patients were counseled regarding their results; median days to counseling was 0 [IQR 0 – 1.5 days]. Six (22.2%) patients were lost to follow-up, the remaining 19 patients were offered referral to specialty care but 14.3% declined.  Age, gender, race, ethnicity, injury severity score, trauma activation level, and payor type were not significant predictors for a positive screen on logistic regression analysis. 

Conclusion:
Despite a significantly higher rate of HIV in the trauma population, barely a third of patients are screened. Majority of patients that screened positive were offered referral to specialty care. Such high diagnosis rate justifies the existence of this screening program; however, steps must be taken to increase screening rate.
 

12.10 Age-Related Changes in Coagulation After Trauma

Posted on January 31, 2019

C. R. Reed3, T. Fitzgerald2, H. Leraas3, R. Kamyszek5, C. Vatsaas1, C. Ray4, J. Otto2, E. Tracy2, S. K. Agarwal1  1Duke University Medical Center,Trauma And Critical Care,Durham, NC, USA 2Duke University Medical Center,Pediatric General Surgery,Durham, NC, USA 3Duke University Medical Center,General Surgery Residency,Durham, NC, USA 4Duke University Medical Center,Pediatric Critical Care,Durham, NC, USA 5Duke University Medical Center,School Of Medicine,Durham, NC, USA

Introduction

Trauma-induced coagulopathy (TIC) is associated with worse outcomes. The coagulation system normally undergoes development throughout life. TIC has been described in adults and sparsely for children, but no studies have directly compared incidence and outcomes of TIC among infants, children, adults, and older adults.

Methods:

A single-institution retrospective review of all level 1 and 2 trauma activations from 2013-2017 at a Level 1 trauma center was performed. Inclusion criteria were complete records and an injury severity score (ISS) >= 9. Data were collected and analyzed for patient age, ISS, mechanism, mortality, prothrombin time or international normalized ratio (PT or INR), partial thromboplastin time (PTT), and fibrinogen. The patient cohort was divided into age groups reflecting normal development to evaluate both patient factors and lab values in infants (=< 2 years), children (3-13 years), adults (14-64 years), and older adults (>= 65 years). Using bivariate statistics, we compared the adult control group to the other age groups. Coagulopathy was defined as abnormal PTT (>= 38s) or INR (>= 1.3). Normal fibrinogen was defined as plasma levels 150-400 mg/dL. Statistical significance was declared if p < 0.05.

Results:

During the study period, 1,983 patients met the inclusion criteria and were analyzed. Overall mortality in all groups was 12% and rates of coagulopathy by abnormal PTT and INR were 7% and 10% respectively. Adults had a slightly higher median ISS compared to the other age groups. Despite their lower ISS, infants and children were more likely to have coagulopathy on arrival using both PTT and INR definitions, and older adults were more likely to be coagulopathic via INR (p < 0.05). Nearly 20% of patients =< 2 years presented with abnormal coagulation studies and 32% of these infants were also hypofibrinogenemic on arrival compared to 9% of adults (p = 0.0019). Median fibrinogen on arrival did increase with each age group, and older adults were more likely to be hyperfibrinogenemic compared to adults (19% vs. 5%, p < 0.0001). Both coagulopathy and hypofibrinogenemia were associated with mortality in subgroup analyses of all age groups with the notable exception of older adults, where abnormal PTT and INR were not associated with mortality.

Conclusion:

The incidence of TIC differs by age group and may suggest differences in hematologic development of the coagulation system that should be considered when caring for these patients, especially when evaluating patients for transfusions and hemostatic adjuncts. Further study should focus on outcomes of resuscitation, correction of coagulopathy, thrombosis, and other outcomes at the extremes of age.

12.09 Unplanned ICU Admission Is Associated With Worse Clinical Outcomes in Geriatric Trauma Patients

Posted on January 31, 2019

H. Mulvey1, R. Haslam1, A. Laytin1, C. Sims1  1University of Pennsylvania,Trauma, Surgical Critical Care, And Emergency Surgery,Philadelphia, PA, USA

Introduction:  Unplanned ICU admissions are associated with increased morbidity and mortality. The impact of these events on geriatric trauma patients, however, has not been previously investigated. We hypothesized that unplanned ICU admission is associated with negative outcomes in geriatric trauma patients and sought to identify predictive risk factors. 

Methods:  All trauma patients over the age of 65 admitted to an urban, academic, level I trauma center from January 1, 2012 to June 31, 2018 were identified. A prospectively collected administrative database was queried for demographics, co-morbidities, injury characteristics, and vital signs on admission. Outcomes including ICU days, overall length of stay, and mortality were assessed. The timing and incidence of complications was confirmed by review of the medical record. Univariate analysis was performed using Chi-square, Mann-Whitney U, and Student t-tests, where appropriate. Variables found to be significant underwent a binary logistic regression analysis. *p<0.05 =significant. 

Results: Of the 2925 geriatric patients admitted during the study period, 96 (3.28%) patients experienced an unplanned admission to the ICU.  Patients with unplanned ICU admissions were older (80.2±9.4 v. 78.4±9.0, p=0.048), had a higher ISS (10 [5-17] v. 9 [4-13], p<0.001), more comorbidities (3.8±1.8 v. 3.4±2.2, p=0.033), and a higher Charlson Comorbidity Index (5 [4-6] v. 4 [3-5], p=0.024). On logistic regression, however, only ISS was predictive of unplanned ICU admission. Interestingly, 68 of the unplanned ICU admissions (70.8 %) were unplanned ICU readmissions (4.7% v. 1.9%, p <0.001; OR 2.3, p=0.001).  Patients with unplanned ICU admission experienced longer hospital stays (20.1 ± 19.5 v. 6.7 ± 7.4, p<0.001), more days in the ICU (10.1 ± 11.7 v. 2.3 ± 5.6, p<0.001), more complications (2.8 ± 2.0 v. 1.8 ± 1.4, p<0.001), and higher rates of mortality (11.5% v. 5.0%, p=0.019). 

Conclusion: Geriatric trauma patients who require an unplanned ICU admissions experience significantly higher morbidity and mortality. Although ISS strongly predicted the need for unplanned ICU admission, the median ISS for these patients was only 10 suggesting that “at risk” elders could potentially be under-triaged to the floor. Given that unplanned admissions occur more frequently in patients previously admitted to the ICU, strategies that provide an extra layer of care such as step-down units or geriatric consultation could potentially improve outcomes.

 

12.08 National Impact of Updated Field Triage Guidelines on Older Adult Injured Patients

Posted on January 31, 2019

T. Uribe-Leitz1, M. P. Jarman1, D. J. Sturgeon1, J. W. Scott4, G. Ortega1, S. R. Lipsitz1, A. Salim1,3, C. D. Newgard2, A. H. Haider1,3  1Brigham And Women’s Hospital,Center For Surgery And Public Health,Boston, MA, USA 2Oregon Health & Science University,Center For Policy And Research In Emergency Medicine,Portland, OREGON, USA 3Brigham And Women’s Hospital,Division Of Trauma, Burns, And Critical Care Surgery,Boston, MA, USA 4Harborview Medical Center, University of Washington,Department Of Surgery, Trauma & Surgical Critical Care,Seattle, WA, USA

Introduction: The CDC published new guidelines for field triage of injured patients in 2012. This update includes special considerations for older adults and specify the need for this vulnerable population to receive high level (Level I or Level II) trauma center (TC) care. The effect of this policy on older adults has not been analyzed. We sought to determine if the updated guidelines have improved field triage for severely injured older adult patients at a national level.

Methods: We queried Medicare claims data from 2008-2014, to identify older adults (age ≥ 65) diagnosed with traumatic injury. Under-triage was defined as Non-Trauma Center [(NTC) Level III-V, and non-trauma centers] care for patients with an injury severity score (ISS) ≥ 16, per the American College of Surgeons Committee on Trauma (ASC-COT) benchmark. We used a difference-in-difference (DID) study design, with patients transported by EMS as the exposure group compared to those not transported by EMS as the no-exposure group. The pre-exposure period was from 2009 to 2010, the post-exposure period from 2013 to 2014, and 2011 to 2012 as a washout period. We used multivariable logistic regression models to estimate the statistical significance of the association between the updated guidelines (policy change) and under-triage adjusting for age, sex, race, region, mechanism of injury (falls), Charlson Comorbidity Index, in-hospital death, ISS, and TC proximity. 

Results: A total of 180,436 severely injured older adult patients were included in our analyses, from which 85,030 (47.1 %) were transported by EMS and 95,406 (52.9 %) were not transported by EMS. Older adults were distributed evenly in pre-exposure and post-exposure groups. Mean age was 81.4 (SD, 7.8) and 81.7 (SD, 8.0) respectively, 54% were female and 90% white in both groups. Pre-/Post-policy rates of under-triage were 50.2% and 47.5% for the exposed and 44.5% and 43.4% for the unexposed, in unadjusted analyses. Pre-/Post-policy rates of under-triage were 32.7% and 29.7% for the exposed and 28.6% and 27.2% for the unexposed, after adjusting for covariates of interest. The DID model revealed a policy-associated change in under-triage rate of 1.6% among older adults, p= 0.0016. 
 

Conclusion: The release of the 2012 updates to the CDC guidelines was associated with a small significant reduction in under-triage rates among older adults. Additional multilevel strategies are required to improve access, and continue to reduce under-triage to truly serve this vulnerable population.
 

12.07 Determining Trajectory to Predict Injury: The Use of Abdominal X-Ray Imaging in Gunshot Wounds

Posted on January 31, 2019

A. G. Goldenberg1, J. Badach1, C. Arya1, J. San Roman1, J. Gaughan1, J. P. Hazelton1  1Cooper University Hospital,Division Of Trauma,Camden, NJ, USA

Introduction: Cavitary triage is important in treating patients with gunshot wounds of the torso, as it allows the surgeon to define the sequence of clinical management. The practice of marking wounds with radio-opaque markers and obtaining X-rays of various body regions has been done in an attempt to determine the trajectory of missiles and help identify which organs may be injured. We hypothesized that such X-rays do not alter the clinical decision of the surgeon in regards to emergent operation vs. further diagnostics, and that hemodynamic parameters would be the most crucial piece of information for the surgeon. 

Methods: We developed a 50-patient (89 injury sites) PowerPoint survey based on cases seen at our Level-1, urban trauma center from 2012 through 2014. X-ray images were de-identified and cases were selected so that none of the survey participants would have had contact with the study patients during their initial resuscitation. Images of a silhouetted BodyMan (BM) with wounds marked, X-rays of the neck and torso (XR), and vital signs (VS) were shown in series for 20 seconds each. Surgeons were asked after each image to record which organs they thought could be injured and to document their next step in management (emergent operation vs. further diagnostics).

Results: Ten surgeons with varying clinical experience completed the survey (>6y in practice, n=3; 1 to 6y in practice, n=4; in-training Fellow, n=3). Data was analyzed to determine the inter-rater reliability (agreement, ICC) for each mode of clinical information (BM, XR, VS). Predicted vs. actual injuries were compared using absolute agreement and kappa statistics. We found that no one piece of information (BM, XR, or VS) was helpful in allowing the surgeon to accurately determine the predicted vs. actual organ injuries. Overall, the most experienced surgeons (>6y in clinical practice) were better than in-training Fellows (PGY-6) in accurately predicting actual injury (93% vs 78%, p=0.021). Pulmonary injury, as evidenced by the chest X-ray, had the highest agreement amongst all potential injuries (ICC=0.727). VS had the highest ICC across all groups in determining the clinical plan for the patient (ICC=0.342), while both BM and XR had low ICCs across all groups in determining clinical plan (ICC=0.162, 0.183).

Conclusion: In this pilot study we found that marking wounds and obtaining X-rays, other than a chest X-ray, did not result in accuracy in predicting injury, nor agreement among participants. Further, these X-rays did not alter the decision making of the surgeon relative to need for operative intervention. Patient vital signs were the only piece of information found significant in determining clinical management. We conclude that marking wounds for radiographic localization is an unnecessary step during the initial resuscitation of patients with gunshot wounds to the neck and torso.

 

12.06 Cost of Care Associated with Code Status Among Advanced Cancer Patients with Acute Care Surgical Consultation

Posted on January 31, 2019

L. Marcia1, Z. W. Ashman2, E. B. Pillado1, C. Hines1, D. S. Plurad2, D. Y. Kim2  1David Geffen School Of Medicine, University Of California At Los Angeles,Los Angeles, CA, USA 2Harbor-UCLA Medical Center,Surgery,Torrance, CALIFORNIA, USA

Introduction:  Advance directive (AD) and do-not-resuscitate (DNR) orders are expected to improve end-of-life care in a cost-effective manner. Our objective was to describe the relationship between cost of care and timing of DNR orders among stage IV cancer patients with Acute Care Surgery (ACS) consultations.

Methods:  Single institution retrospective review of all ACS consultations over an 8-year period in stage IV cancer patients with ACS consultation. Cost estimates were calculated from the Healthcare Cost and Utilization Project (the Nationwide Inpatient Sample), Centers for Medicare and Medicaid Services in California and other published cost estimates. Statistical analysis included univariate analysis to characterize data, ANOVA and chi-square tests to assess correlation

Results: Two hundred three patients were identified; mean age was 55.3 ± 11.4 years old, 48.8% were male. Fifty patients (24.6%) underwent exploratory laparotomy for gastrointestinal obstruction and/or perforation and 26 (12.8%) underwent other types of surgery. Twenty-one patients (10.3%) had a DNR order on-admission, 54 (26.6%) became DNR post-admission and 128 patients (63.1%) remained full-code. DNR post-admission was associated with longer mean length of stay (LOS) (19.6 days) vs DNR on admission (7.0 days) and full code (10.5 days; p<0.01). This was similar for ICU LOS: 7.7days in the DNR post-admission, 1.7 in the full code and 0.9 days in the DNR on-admission groups (p<0.01). DNR post-admission was associated with higher total cost of hospitalization ($76,133) compared to DNR on-admission ($25,114) and patients that remained full code ($36,857; p<0.01). DNR post-admission was associated with higher cost for ICU stay compared to DNR on-admission and full-code groups ($11,747 vs $1,304 vs $2,520, p<0.01). Procedural/surgical costs were higher for DNR post-admission compared to DNR on-admission and full-code groups ($610 vs $381 vs $535; p = 0.34). Hospital mortality was higher in the post-admission DNR group (82.1%) in comparison to the DNR on-admission (10.7%) and full code groups (7.1%, p< 0.01). The mean estimated cost after inpatient death was $88,662. The full-code group had the lowest cost after inpatient death followed by DNR on-admission and DNR post-admission groups ($6,351 vs $10,491 vs $106,016, p<0.01). In the post-admission DNR group, the mean interval between admission and DNR was 19 days and from post-admission DNR to in-hospital death was 4 days.

Conclusion: DNR post-admission is associated with higher cost-of-care than DNR on-admission and full-code groups. This group experiences prolonged hospital and ICU LOS, which are the primary drivers of cost in all groups. These patients elect for aggressive (and costly) care at the end-of-life despite a known life-limiting diagnosis and poor immediate prognosis. These interventions may not extend length of life meaningfully and may worsen the quality of life before death. Improvements should be made to identify such patients, with the goal of providing high quality care sensitive to the context of a terminal diagnosis. 

 

12.05 Multistate Study of Disparities in Trauma Care: Are People of Color at Higher Risk of Under-triage?

Posted on January 31, 2019

D. A. Alber1, T. Uribe-Leitz1, G. Ortega1, A. Salim2, A. Haider1, M. P. Jarman1  1Brigham And Women’s Hospital,Center Of Surgery And Public Health,Boston, MA, USA 2Brigham And Women’s Hospital,Trauma, Burns, And Surgical Critical Care,Boston, MA, USA

Introduction: Racial disparities in trauma care are well-established. Disparities in the treatment of severely injured patients at hospitals without designated trauma services, defined as under-triage, have been observed for certain interest groups (e.g. children, and older adults). The relationship between race/ethnicity and under-triage remains unknown. We sought to examine if differences by race and ethnicity exist in trauma center care following severe traumatic injury.

Methods: We performed a retrospective cohort analysis of patients with severe traumatic injury in 21 geographically diverse states from the 2014 HCUP State Inpatient Databases, linked with hospitals’ trauma center status from the American Hospital Association Annual Survey. Under-triage was defined as treatment at a Level III/IV or non-trauma center for patients with an injury severity score > 15. Multivariable logistic regression models were used to compare the likelihood of undertriage between White, African American, and Hispanic patients, adjusting for sex, age, injury severity, injury mechanism, and patient urban/rural residence. Our primary analysis included all severely injured patients. Sub-analyses of patients with penetrating injury and rural residents were also conducted.

Results: A total of 83,817 severely injured patients were identified, 55.4% of patients were under-triaged. Unadjusted rates of under-triage were highest among White patients (58.5%), followed by Hispanic patients (56.8%), and African American patients (43.4%). Compared to White patients, adjusted odds of under-triage were higher for Hispanic patients (Odds Ratio [OR]=1.16, 95% Confidence Interval [CI]=1.11, 1.22), and lower for African American patients (OR=0.68, 95% CI=0.65, 0.72) and patients of other races (OR=0.70, 95% CI=0.67,0.72). Sub-analysis of patients with penetrating injuries showed a similar pattern for African American and Other Race. Under-triage rates did not vary by race or ethnicity in sub-analysis of rural patients.

Conclusion: Urban Hispanic patients possess significantly higher rates of under-triage compared to their White counterparts, while African American patients experience significantly lower rates of under-triage. Under-triage rates for rural populations do not vary by race or ethnicity. Increased under-triage rates for Hispanic patients may be the result of the geographic distribution of Hispanic populations relative to trauma centers, or language barriers inhibiting prehospital providers’ ability to effectively assess patient status. Future studies should attempt to include larger populations of Hispanic patients, consider additional geographic measures, as well as cultural and language barriers to care for Hispanic populations.

« Previous Page
Next Page »