Using State Data to Predict a Single Institution Mortality for Patients That Fall.

BACKGROUND: Falls are the most common cause of injury-related death for patients older than 45.  We hypothesized that a machine learning algorithm developed from state-level registry data could make accurate outcome predictions at a level 1 trauma hospital. METHODS: Data for all patients admitted for fall injury during 2009 - 2019 in the state of Pennsylvania were derived from the state trauma registry.  Thirteen variables that were immediately available upon patient arrival were used for prediction modeling.  Data for the test institution were withheld from model creation.  Algorithms assessed included logistic regression (LR), random forest (RF), and extreme gradient boost (XGB).  Model discrimination for mortality was assessed with area under the curve (AUC) for each algorithm at our level 1 trauma center. RESULTS: 180,284 patients met inclusion criteria.  The mean age was 69 years ± 18.5 years with a mortality rate of 4.0%.  The AUC for predicting mortality in patients that fall for LR, RF, and XGB were 0.797, 0.876, and 0.880, respectively.  The variables which contributed to the prediction in descending order of importance for XGB were respiratory rate, pulse, systolic blood pressure, ethnicity, weight, sex, age, temperature, Glasgow Coma Scale (GCS) eye, race, GCS voice, GCS motor, and blood alcohol level. CONCLUSIONS: An extreme gradient boost model developed using state-wide trauma data can accurately predict mortality after fall at a single center within the state.  This machine learning model can be implemented by local trauma systems within the state of Pennsylvania to identify patients injured by fall that require greater attention, transfer to a higher level of care, and higher resource allocation.

Contemporary management of traumatic cervical and thoracic esophageal perforation: The results of an Eastern Association for the Surgery of Trauma multi-institutional study.

BACKGROUND: Traumatic esophageal perforation is rare and associated with significant morbidity and mortality. There is substantial variability in diagnosis and treatment. Esophageal stents have been increasingly used for nontraumatic perforation; however, stenting for traumatic perforation is not yet standard of care. The purpose of this study was to evaluate current management of traumatic esophageal perforation to assess the frequency of and complications associated with esophageal stenting. METHODS: This was an Eastern Association for the Surgery of Trauma multi-institutional retrospective study from 2011 to 2016 of patients with traumatic cervical or thoracic esophageal injury admitted to one of 11 participating trauma centers. Data were collected and sent to a single institution where it was analyzed. Patient demographics, injury characteristics, initial management, complications, and patient mortality were collected. Primary outcome was mortality; secondary outcomes were initial treatment, esophageal leak, and associated complications. RESULTS: Fifty-one patients were analyzed. Esophageal injuries were cervical in 69% and thoracic in 31%. Most patients were initially managed with operative primary repair (61%), followed by no intervention (19%), esophageal stenting (10%), and wide local drainage (10%). Compared with patients who underwent operative primary repair, patients managed with esophageal stenting had an increased rate of esophageal leak (22.6% vs. 80.0%, p = 0.02). Complication rates were higher in blunt compared with penetrating mechanisms (100% vs. 31.8%, p = 0.03) despite similar Injury Severity Score and neck/chest/abdomen Abbreviated Injury Scale. Overall mortality was 9.8% and did not vary based on location of injury, mechanism of injury, or initial management. CONCLUSION: Most patients with traumatic esophageal injuries still undergo operative primary repair; this is associated with lower rates of postoperative leaks as compared with esophageal stenting. Patients who have traumatic esophageal injury may be best managed by direct repair and not esophageal stenting, although further study is needed. LEVEL OF EVIDENCE: Therapeutic, level IV.

One elevated bladder pressure measurement may not be enough to diagnose abdominal compartment syndrome.

Bladder pressure measurements (BPMs) are considered a key component in the diagnosis of abdominal compartment syndrome (ACS). The purpose of this observational review was to determine risk factors of ACS and associated mortality with particular focus on the role of BPM. A retrospective trauma registry and chart review was performed on trauma patients from January 2003 through December 2010. Comparisons were made between patients with and without ACS. There were 3172 patients included in the study of whom 46 had ACS. Patients with ACS were younger, more severely injured, with longer lengths of stay. Logistic regression determined Injury Severity Score (ISS) and urinary catheter days as independent predictors of ACS, whereas independent predictors of mortality included age, ISS, and ACS. Subset analysis demonstrated no association between BPM 20 mmHg or greater and diagnosis of ACS versus no ACS. Logistic regression indicated independent predictors of mortality were number of BPM 20 mmHg or greater and age. Patients with ACS are more severely injured with worse outcomes. An isolated BPM 20 mmHg or greater was not associated with ACS and may be inadequate to independently diagnose ACS. These findings suggest the need for repeat measurements with early intervention if they remain elevated in an effort to decrease mortality associated with ACS.

Repeat computed tomography for trauma patients undergoing transfer to a Level I trauma center.

Our goal was to determine the characteristics of trauma transfer patients with repeat imaging. A retrospective trauma registry review was performed to evaluate trauma patients who were transferred from referring institutions between January 2005 and December 2009. Patients were divided into those who had a duplicate computed tomography (CT) scan versus those who did not. There were 2678 patients included of whom 559 (21%) had at least one repeat CT scan, whereas 2119 (79%) did not have any repeat CT scans. Those with repeat CT scans were older (42.3 ± 27.3 years vs 37.3 ± 25.6 years), had a higher Injury Severity Score (ISS) (13.7 ± 8.7 vs 11.9 ± 8.8), and more likely to have blunt trauma (odds ratio, 4.7; confidence interval, 2.3 to 9.6) (P for all < 0.0007). Those with CT scans done only at the referring facility were younger, had a lower ISS, and shorter lengths of stay (P for all < 0.0003). ISS and age were independent predictors for repeat CT scans. Transfer patients had imaging repeated one-fifth of the time. The younger, less injured patient went without repeat imaging suggesting that they may have been adequately cared for at the outside institution.