Utilization of Machine Learning Approaches to Predict Mortality in Pediatric Warzone Casualties.

BACKGROUND: Identification of pediatric trauma patients at the highest risk for death may promote optimization of care. This becomes increasingly important in austere settings with constrained medical capabilities. This study aimed to develop and validate predictive models using supervised machine learning (ML) techniques to identify pediatric warzone trauma patients at the highest risk for mortality. METHODS: Supervised learning approaches using logistic regression (LR), support vector machine (SVM), neural network (NN), and random forest (RF) models were generated from the Department of Defense Trauma Registry, 2008-2016. Models were tested and compared to determine the optimal algorithm for mortality. RESULTS: A total of 2,007 patients (79% male, median age range 7-12 years old, 62.5% sustaining penetrating injury) met the inclusion criteria. Severe injury (Injury Severity Score > 15) was noted in 32.4% of patients, while overall mortality was 7.13%. The RF and SVM models displayed recall values of .9507 and .9150, while LR and NN displayed values of .8912 and .8895, respectively. Random forest (RF) outperformed LR, SVM, and NN on receiver operating curve (ROC) analysis demonstrating an area under the ROC of .9752 versus .9252, .9383, and .8748, respectively. CONCLUSION: Machine learning (ML) techniques may prove useful in identifying those at the highest risk for mortality within pediatric trauma patients from combat zones. Incorporation of advanced computational algorithms should be further explored to optimize and supplement the diagnostic and therapeutic decision-making process.

Gender trends of military general surgery residency applicants.

BACKGROUND: Despite the fact that women comprise half of medical school graduates, women remain under-represented in General Surgery. We aimed to identify the trend in military general surgery applicants based on gender. METHODS: A retrospective review of the Medical Operational Data System (MODS) applicants from 2014 to 2019 was performed. Demographic data included age, gender, average board scores, applied and matched specialty. RESULTS: 204 students applied and 167 applicants matched into Army General Surgery programs from 2014 to 2019. The percentage of all students applying to General Surgery was statistically higher in females (13.4% vs 9.0%, p = 0.04), with females matching at a statistically higher rate (77% vs 57%, p = 0.02). Over the last 6 years, 55% of residents in training are men and 45% are women. CONCLUSION: Army General Surgery programs have increased the amount of women in training over the last six years. We believe that this will ultimately translate to more female surgeons contributing to military medicine.

Machine Learning for Military Trauma: Novel Massive Transfusion Predictive Models in Combat Zones.

BACKGROUND: Damage control resuscitation has become the standard of care in military and civilian trauma. Early identification of blood product requirements may aid in optimizing the clinical decision-making process while improving trauma related outcomes. This study aimed to assess and compare multiple machine learning models for predicting patients at highest risk for massive transfusion on the battlefield. METHODS: Supervised machine learning approaches using logistic regression, support vector machine, neural network, and random forest techniques were used to create predictive models for massive transfusion using standard prehospital and arrival data points from the Department of Defense Trauma Registry, 2008-2016. Seventy percent of the population was used for model development and performance was validated using the remaining 30%. Models were tested for accuracy and compared by standard performance statistics. RESULTS: A total of 22,158 patients (97% male, 58% penetrating injury, median age 25-29 y/o, average Injury Severity Score 9, with an overall mortality of 3%) were included in the analysis. Massive transfusion was required by 7.4% of patients. Overall accuracy was found to be above 90% in all models tested. Following cross validation and model training, the random forest model outperformed the alternatively tested models for precision, recall, and area under the curve. CONCLUSION: Machine learning techniques may allow for more optimal and rapid identification of combat trauma patients at highest risk for massive transfusion. These powerful approaches may uncover novel correlations and help improve triage, activation of massive transfusion resources, and trauma-related outcomes. Further research seeking to optimize and apply these algorithms to trauma-centered research should be pursued.

Gender Differences in Medical Specialty Decision Making: The Importance of Mentorship.

INTRODUCTION: While there have been many studies reviewing factors considered by medical students in deciding on a specialty, there is a relative paucity of data in how these factors differ between males and females. The aim of this study was to determine if there are differences in how male and female medical students' value various factors in choosing a career path in medicine. METHODS: A survey of basic demographic information and a Likert-based questionnaire addressing various factors in specialty choice was sent to all military medical students in the class of 2019. RESULTS: The survey response rate was 30%. Amongst the students that completed the survey, 72% were male and 28% were female. The majority of students were less than 27 years old (52%) and Caucasian (76%). Female medical students placed more importance in exposure to third year clerkships (P = 0.004) and gender diversity within the specialty (P = 0.03) than their male counterparts. There was no statistical difference in opinions of significant others or family members, desire to have a family, or having a well-balanced life between female, and male medical students. CONCLUSION: Female medical students valued exposure to third year clerkships and gender diversity more than their male colleagues when choosing a future specialty. These findings suggest that a strategic focus should be placed on mentoring female medical students in order to promote a diverse medical workforce.

Addition of neurological status to pediatric adjusted shock index to predict early mortality in trauma: A pediatric Trauma Quality Improvement Program analysis.

INTRODUCTION: Pediatric adjusted shock index (SIPA) has demonstrated the ability to prospectively identify children at the highest risk for early mortality. The addition of neurological status to shock index has shown promise as a reliable triage tool in adult trauma populations. This study sought to assess the utility of combining SIPA with Glasgow Coma Scale (GCS) for predicting early trauma-related outcomes. METHODS: Retrospective review of the 2017 Trauma Quality Improvement Program Database was performed for all severely injured patients younger than 18 years old. Pediatric adjusted shock index and reverse SIPA × GCS (rSIG) were calculated. Age-specific cutoff values were derived for reverse shock index multiplied by GCS (rSIG) and compared with their SIPA counterparts for early mortality assessment using area under the receiver operating characteristic curve analyses. RESULTS: A total of 10,389 pediatric patients with an average age of 11.4 years, 67% male, average Injury Severity Score of 24.1, and 4% sustaining a major penetrating injury were included in the analysis. The overall mortality was 9.3%. Furthermore, 32.1% of patients displayed an elevated SIPA score, while only 27.5% displayed a positive rSIG. On area under the receiver operating characteristic curve analysis, rSIG was found to be superior to SIPA as a predictor for in hospital mortality with values of 0.854 versus 0.628, respectively. CONCLUSION: Reverse shock index multiplied by GCS more readily predicted in hospital mortality for pediatric trauma patients when compared with SIPA. These findings suggest that neurological status should be an important factor during initial patient assessment. Further study to assess the applicability of rSIG for expanded trauma-related outcomes in pediatric trauma is necessary. LEVEL OF EVIDENCE: Prognostic study, level IV.

Safety code blue! Assessing the use of blue surgical sterilization wrap for homemade respirator masks during the COVID-19 crisis.

The use of surgical sterilization wrap for respirator masks during the COVID-19 crisis has become a popularized personal protective equipment alternative option due to claims supporting its ability to meet N95 standards. This study sought to assess these claims using standardized filter testing. The tested material failed to meet N95 standards and suggests its use may place medical personnel at increased risk of harm when managing COVID-19 patients.

Utilization of mobile application improves perioperative education and patient satisfaction in general surgery patients.

INTRODUCTION: Clear and accurate communication is paramount in delivering high quality surgical care. Through the development of a mobile application, we provided patients with a source of education and instruction throughout the peri-operative period. METHODS: Patients >18 years old with a smart-phone undergoing elective general surgery procedures were eligible. Patients received perioperative educational materials and text message reminders of time-sensitive events via the application. A System Usability Scale and survey was administered. RESULTS: 100 patients were enrolled; 51% completed the survey. The average SUS score was 86, correlating with >90th percentile usability. 86% of patients felt that the application improved their surgical experience, 96% said the application provided essential reminders, and 90% felt that application clarified information. 84% of patients did not identify any inconsistency between the application and surgeon. CONCLUSION: Utilizing patient's smart phones to aid in perioperative education is feasible and improves patient satisfaction. This application has a high usability score, indicating ease of use.

Dynamic trend or static variable: Shock Index Pediatric-Adjusted (SIPA) in warzone trauma.

BACKGROUND: Civilian studies suggest that trending Shock-Index Pediatric Adjusted(SIPA) values can prove useful in the prediction of trauma outcomes. The purpose of this study was to evaluate the relationship between trends in SIPA and outcomes in pediatric warzone trauma. METHODS: Retrospective review of the Department of Defense Trauma Registry from 2008 to 2015, including all patients age ≤17years. SIPA was calculated both pre-hospital and upon arrival, then classified as "normal" or "abnormal" based upon previously validated thresholds. Patients were stratified into groups based on the trend of their SIPA (1-normal to normal, 2-normal to abnormal, 3-abnormal to normal, 4-abnormal to abnormal). Key outcomes including ICU admission, severe injury, mechanical ventilation, and mortality were then compared between groups. RESULTS: 669 patients were included, mean ISS 12 ± 10. The most common mechanism of injury was blast (46.5%). Overall, 43% were stratified into Group 1, 13.9% into Group 2, 14.8% into Group 3, and 28.0% into Group 4. Those patients with a persistently abnormal SIPA (Group 4) had significantly increased incidence of severe injury, ICU admission, need for mechanical ventilation, and mortality. CONCLUSION: Trends in SIPA may be used to predict trauma outcomes for children injured in warzones, with persistently abnormal values associated with worse outcomes overall.

Shock Index as a Predictor of Massive Transfusion and Emergency Surgery on the Modern Battlefield.

BACKGROUND: Shock Index (SI) has been used to predict the need for massive transfusion (MT) and emergency surgical procedures (ESP) in civilian trauma. We hypothesize that SI can reliably identify combat trauma patients that will require MT and ESP when applied to the resource-constrained, combat environment. METHODS: A retrospective review was performed within the Department of Defense Trauma Registry (2008-2016). SI was calculated using heart rate and systolic blood pressure on arrival to the initial facility with surgical capabilities. A threshold value of 0.8 was used to stratify patients into two groups (Group I, SI < 0.8; and Group II, SI ≥ 0.8). The need for MT, ESP, and mortality was compared. Regression analyses were conducted to determine the independent association of SI with MT and ESP. RESULTS: A total of 4008 patients were included. The mean age of the patients was 25.5 y, and the majority were predominately male (98%). Mechanisms of injury were blunt and blast injury (62%), penetrating injury (36.7%), and burn injury (0.5%). Overall, 77% of patients (n = 3070) were stratified to Group I, and 23% of patients (n = 938) were stratified to Group II, by SI. Group II patients had a significantly greater need for MT (8.4% versus 0.4%) and ESP (30.7% versus 6.5%), both P < 0.001. Regression analysis controlling for age, gender, Injury Severity Score, and Glasgow Coma Score confirmed that SI ≥ 0.8 was an independent risk factor for both MT and need for ESPs (P < 0.001). CONCLUSIONS: SI is a significant predictor of the need for MT and ESPs in the military trauma population, representing a simple and potentially potent tool for triage and prediction of resource consumption in the resource-limited, austere setting.

Optimal Prospective Predictors of Mortality in Austere Environments.

BACKGROUND: Prospective predictors of trauma-related outcomes have been validated to guide management in low-resource settings. The primary objective of this study was to determine the optimal prospective prediction method for mortality within combat and humanitarian trauma. MATERIALS AND METHODS: Retrospective review of the Department of Defense Trauma Registry from 2008 to 2016 was performed for adult patients. Areas under receiver operating characteristic curves (AUROCs) were calculated to assess the predictability of shock index (SI), reverse SI × Glasgow Coma Scale (rSIG), SI × Glasgow Coma Scale (SIG), Revised Trauma Score, and Trauma and Injury Severity Score (TRISS) on mortality at point of injury, arrival in emergency department (ED), and the difference in vital signs between those time points. RESULTS: A total of 22,218 patients were included. Overall, 97.1% were male, median age range 25-29 y, Injury Severity Score 9.4 ± 0.07, with predominantly penetrating injuries (58.1%), and mortality of 3.4%. ED vitals yielded higher predictability of mortality for all tests based on higher AUROCs. TRISS and rSIG demonstrated the highest AUROCs (0.955 and 0.923, respectively). The optimal cutoff value for rSIG was 14.1 (sensitivity 89% and specificity 87%). rSIG values <14.1 were significantly associated with mortality (P < 0.01; odds ratio = 5.901). CONCLUSIONS: Initial ED vital signs represented a better predictor of early mortality compared with point of injury vital signs for all predictive tools assessed. TRISS and rSIG proved to be most predictive of mortality. However, of the prospective tools assessed, rSIG may be optimal scoring tool because of its ease of calculation and its increased ability to predict mortality.

Early outcomes following trauma related to sex: A matched analysis of military service members in the department of defense trauma registry.

BACKGROUND: Several studies have identified sex as a factor influencing early outcomes after trauma. With the increased representation of women in combat roles, there is a need for improved understanding of the pathophysiology of traumatic injury in women. The purpose of this study was to define sex-based differences in early combat trauma outcomes amongst military service members. METHODS: A retrospective review of the Department of Defense Trauma Registry between 2008 and 2016 was performed. A 2:1 case control match was performed to match for Injury Severity Score, mechanism of injury, and age. The primary outcome of the study was mortality. RESULTS: A total of 4,625 patients were included in the study, 2.2% of whom were women. Women were less significantly injured than men (Injury Severity Score, 7.7 vs. 11, p = 0.003) and more likely to sustain blunt trauma (81% vs. 62.5%, p = 0.01). After case-control matching, 202 men and 101 women were evaluated. There was no statistical difference in the primary outcome of mortality. There was no statistical difference in Glasgow Coma Scale score, crystalloid or colloid administration, Packed Red Blood Cells (PRBC), platelet, cryoprecipitate, or plasma usage between men and women. CONCLUSION: Contrary to the civilian trauma literature, our study demonstrated no significant difference in early mortality between male and female combat casualties in a matched cohort. This finding may represent a difference in injury patterns, resuscitation practices, or lifesaving interventions in a deployed setting as compared with civilian setting. As the proportion of women involved in combat operations continues to increase, prospective studies should be performed to better define injury patterns, as well as early and late outcomes related to military trauma in the female population. LEVEL OF EVIDENCE: Retrospective, Level IV.

Postoperative outcomes of Bascom cleft lift for pilonidal disease: A single-center experience.

INTRODUCTION: Limited data exists regarding outcomes for the Bascom cleft lift procedure for pilonidal disease. METHODS: Single-center retrospective review of patients who underwent a Bascom cleft lift from 2013 to 2018. Univariate analysis was performed to determine associations between patient-specific characteristics and post-operative complications. Postoperative complications were categorized as major or minor. Multivariate analysis was performed to identify predictors of postoperative complications. RESULTS: 235 patients were included. Forty-five percent were obese and 24% were active smokers. Minor complications occurred in 34.5% (81); major complications occurred in 19.1% (45). The recurrence rate was 4.7% (11). Smoking was not associated with postoperative complications. Obesity was independently associated with higher rates of both minor (OR 2.6, p = 0.001) and major (OR 2.3, p = 0.001) complications. DISCUSSION: Wound complications are common after Bascom cleft lift. Obesity is an independent predictor of postoperative complications. Obese patients should be appropriately counseled regarding their increased risk prior to surgery.

All trauma is not created equal: Redefining severe trauma for combat injuries.

Injury severity scores (ISS) and shock index (SI) are popular trauma scoring systems. We assessed ISS and SI in combat trauma to determine the optimal cut-off values for mortality and trauma outcomes. Retrospective analysis of the Department of Defense Trauma Registry, 2008-2016, was performed. Areas under receiver operating characteristic curves (AUROCs) were calculated for ISS and SI on mortality, massive volume transfusion (MVT), and emergent surgical procedure (ESP). Optimal cut-off values were defined using the Youden index (YI). 22,218 patients (97.1% male), median ages 25-29 years, ISS 9.4 ± 0.07, with 58.1% penetrating injury were studied. Overall mortality was 3.4%. AUROCs for ISS on mortality, MVT, and ESP were 0.882, 0.898, and 0.846, while AUROCs for SI were 0.727, 0.864, and 0.711 respectively. The optimal cut-off values for ISS on mortality, MVT, and ESP were 12.5 (YI = 0.634), 12.5 (YI = 0.666), and 12.5 (YI = 0.819), with optimal values for SI being 0.94 (YI = 0.402), 0.88 (YI = 0.608), and 0.81 (YI = 0.345) respectively. Classic values for severe ISS underrepresent combat injury while the SI values defined in this study are consistent with civilian data.

Viscoelastic testing in combat resuscitation: Is it time for a new standard?

BACKGROUND: Traumatic hemorrhage and coagulopathy represent major sources of morbidity and mortality on the modern battlefield. Viscoelastic testing (VET) offers a potentially more personalized approach to resuscitation. We sought to evaluate outcomes of combat trauma patients who received VET-guided resuscitation compared with standard balanced blood product resuscitation. METHODS: Retrospective analysis of the Department of Defense Trauma Registry, 2008 to 2016 was performed. Multivariate logistic regression analyses of all adult patients initially presenting to NATO Role III facilities who required blood products were performed to identify factors associated with VET-guided resuscitation and mortality. A propensity score matched comparison of outcomes in patient cohorts treated at VET versus non-VET Role III facilities was performed. RESULTS: There are 3,320 patients, predominately male (98%), median age ranges from 25 years to 29 years, Injury Severity Score of 18.8, with a penetrating injury (84%) were studied. Overall mortality was 9.7%. Five hundred ninety-four patients had VET during their initial resuscitation. After adjusting for confounders, VET during initial resuscitation was independently associated with decreased mortality (odds ratio, 0.63; p = 0.04). Propensity analysis confirmed this survival advantage with a 57% reduction in overall mortality (7.3% vs. 13.1%; p = 0.001) for all patients requiring blood products. CONCLUSION: Viscoelastic testing offers the possibility of a product-specific resuscitation for critically injured patients requiring transfusion in combat settings. Routine VET may be superior to non-VET-guided resuscitation for combat trauma victims. LEVEL OF EVIDENCE: Therapeutic study, level IV.

Big problems in little patients: Nationwide blunt cerebrovascular injury outcomes in the pediatric population.

BACKGROUND: Blunt cerebrovascular injuries (BCVI) are uncommon but potentially devastating. The epidemiology, outcomes, and screening criteria are well described in adults, but data in pediatric patients are extremely limited. The purpose of this study was to characterize pediatric BCVI in a large nationwide sample. We hypothesized that outcomes of BCVI in the pediatric blunt trauma population will vary by age. METHODS: We conducted a retrospective cohort study of the Kids' Inpatient Database for pediatric BCVI from 2000 to 2012. Epidemiology, associated injuries, outcomes (including stroke and mortality), and the utility of standard screening criteria were analyzed. RESULTS: There were 1,182 cases of BCVI identified, yielding an incidence of 0.21%. Patients were predominately male (69%; mean age, 15 ± 5 years). Injuries were 59% carotid, 13% vertebral, and 28% unspecified, with 15% having bilateral or multivessel BCVI. Although younger patients (<11 years) had significantly lower ISS and decreased severe associated injuries (all p < 0.01), they had a similar mortality rate (10%) versus the older cohort. Additionally, the stroke rate was significantly higher among the younger patients versus their older peers (29% mortality for <11 years vs. 15% for ≥11 years, p < 0.01). Only four of seven commonly utilized risk factors were associated with BCVI overall, but none were significantly associated with BCVI in younger children (<11 years). CONCLUSION: This represents the first nationwide assessment of BCVI in the pediatric population. Pediatric BCVI carry considerable mortality and stroke risk. Despite being less severely injured, younger children (<11 years) had similar a mortality rate and a significantly higher stroke rate compared with older pediatric patients. Furthermore, commonly utilized adult screening criteria had limited utility in the younger cohorts. These findings suggest pediatric BCVI may require screening and treatment protocols that are significantly different than currently utilized adult-based programs. LEVEL OF EVIDENCE: Prognostic/Epidemiological Study, level III.