Demystifying the association of center-level operative trauma volume and outcomes of emergency general surgery.

BACKGROUND: Recent studies have demonstrated a positive volume-outcome relationship in emergency general surgery. Some have advocated for the sub-specialization of emergency general surgery independent from trauma. We hypothesized inferior clinical outcomes of emergency general surgery with increasing center-level operative trauma volume, potentially attributable to overall hospital quality. METHODS: Adults (≥18 years) undergoing complex emergency general surgery operations (large and small bowel resection, repair of perforated peptic ulcer, lysis of adhesions, laparotomy) were identified in the 2016 to 2020 Nationwide Readmissions Database. Multivariable risk-adjusted models were developed to evaluate the association of treatment at a high-volume trauma center (reference: low-volume trauma center) with clinical and financial outcomes after emergency general surgery. To evaluate hospital quality, mortality among adult hospitalizations for acute myocardial infarction was assessed by hospital trauma volume. RESULTS: Of an estimated 785,793 patients undergoing a complex emergency general surgery operation, 223,116 (28.4%) were treated at a high-volume trauma center. Treatment at a high-volume trauma center was linked to 1.19 odds of in-hospital mortality (95% confidence interval 1.12-1.27). Although emergency general surgery volume was associated with decreasing predicted risk of mortality, increasing trauma volume was linked to an incremental rise in the odds of mortality after emergency general surgery. Secondary analysis revealed increased mortality for admissions for acute myocardial infarction with greater trauma volume. CONCLUSION: We note increased mortality for emergency general surgery and acute myocardial infarction in patients receiving treatment at high-volume trauma centers, signifying underlying structural factors to broadly affect quality. Thus, decoupling trauma and emergency general surgery services may not meaningfully improve outcomes for emergency general surgery patients. Our findings have implications for the evolving specialty of emergency general surgery, especially for the safety and continued growth of the acute care surgery model.

A National Analysis of Alcohol Withdrawal Syndrome in Patients with Operative Trauma.

Background: Alcohol withdrawal syndrome (AWS) presents with a complex spectrum of clinical manifestations that complicate postoperative management. In trauma setting, subjective screening for AWS remains challenging due to the criticality of injury in these patients. We thus identified several patient characteristics and perioperative outcomes associated AWS development. Methods: The 2016-2020 National Inpatient Sample was queried to identify all non-elective adult (≥18 years) hospitalizations for blunt or penetrating trauma undergoing operative management with a diagnosis of AWS. Patients with traumatic brain injury or with a hospital duration of stay <2 days were excluded. Outcomes of interest included in-hospital mortality, perioperative complications, hospitalization costs, length of stay (LOS) and non-home discharge. Results: Of an estimated 2,965,079 operative trauma hospitalizations included for analysis, 36,415 (1.23 %) developed AWS following admission. The AWS cohort demonstrated increased odds of mortality (Adjusted Odds Ratio [AOR] 1.46, 95 % Confidence Interval [95 % CI] 1.23-1.73), along with infectious (AOR 1.73, 95 % CI 1.58-1.88), cardiac (AOR 1.24, 95 % CI 1.06-1.46), and respiratory (AOR 1.96, 95 % CI 1.81-2.11) complications. AWS was associated with prolonged LOS, (ß: 3.3 days, 95 % CI: 3.0 to 3.5), greater cost (ß: +$8900, 95 % CI $7900-9800) and incremental odds of nonhome discharge (AOR 1.43, 95 % CI 1.34-1.53). Furthermore, male sex, Medicaid insurance status, head injury and thoracic operation were linked with greater odds of development of AWS. Conclusion: In the present study, AWS development was associated with increased odds of in-hospital mortality, perioperative complications, and resource burden. The identification of patient and operative characteristics linked with AWS may improve screening protocols in trauma care.

Interhospital Transfer for Emergency General Surgery: A Contemporary National Analysis.

BACKGROUND: Patients undergoing emergency general surgery (EGS) often require complex management and transfer to higher acuity facilities, especially given increasing national efforts aimed at centralizing care. We sought to characterize factors and evaluate outcomes associated with interhospital transfer using a contemporary national cohort. METHODS: All adult hospitalizations for EGS (appendectomy, cholecystectomy, laparotomy, lysis of adhesions, small/large bowel resection, and perforated ulcer repair) ≤2 days of admission were identified in the 2016-2020 National Inpatient Sample. Patients initially admitted to a different institution and transferred to the operating hospital comprised the Transfer cohort (others: Non-Transfer). Multivariable models were developed to consider the association of Transfer with outcomes of interest. RESULTS: Of ∼1 653 169 patients, 107 945 (6.5%) were considered the Transfer cohort. The proportion of patients experiencing interhospital transfer increased from 5.2% to 7.7% (2016-2020, P < .001). On average, Transfer was older, more commonly of White race, and of a higher Elixhauser comorbidity index. After adjustment, increasing age, living in a rural area, receiving care in the Midwest, and decreasing income quartile were associated with greater odds of interhospital transfer. Following risk adjustment, Transfer remained linked with increased odds of in-hospital mortality (AOR 1.64, CI 1.49-1.80), as well as any perioperative complication (AOR 1.33, CI 1.27-1.38; Reference: Non-Transfer). Additionally, Transfer was associated with significantly longer duration of hospitalization (ß + 1.04 days, CI + .91-1.17) and greater costs (ß+$3,490, CI + 2840-4140). DISCUSSION: While incidence of interhospital transfer for EGS is increasing, transfer patients face greater morbidity and resource utilization. Novel interventions are needed to optimize patient selection and improve post-transfer outcomes.

Machine learning based predictive modeling of readmissions following extracorporeal membrane oxygenation hospitalizations.

Background: Despite increasing utilization and survival benefit over the last decade, extracorporeal membrane oxygenation (ECMO) remains resource-intensive with significant complications and rehospitalization risk. We thus utilized machine learning (ML) to develop prediction models for 90-day nonelective readmission following ECMO. Methods: All adult patients receiving ECMO who survived index hospitalization were tabulated from the 2016-2020 Nationwide Readmissions Database. Extreme Gradient Boosting (XGBoost) models were developed to identify features associated with readmission following ECMO. Area under the receiver operating characteristic (AUROC), mean Average Precision (mAP), and the Brier score were calculated to estimate model performance relative to logistic regression (LR). Shapley Additive Explanation summary (SHAP) plots evaluated the relative impact of each factor on the model. An additional sensitivity analysis solely included patient comorbidities and indication for ECMO as potential model covariates. Results: Of ∼22,947 patients, 4495 (19.6 %) were readmitted nonelectively within 90 days. The XGBoost model exhibited superior discrimination (AUROC 0.64 vs 0.49), classification accuracy (mAP 0.30 vs 0.20) and calibration (Brier score 0.154 vs 0.165, all P < 0.001) in predicting readmission compared to LR. SHAP plots identified duration of index hospitalization, undergoing heart/lung transplantation, and Medicare insurance to be associated with increased odds of readmission. Upon sub-analysis, XGBoost demonstrated superior disclination compared to LR (AUROC 0.61 vs 0.60, P < 0.05). Chronic liver disease and frailty were linked with increased odds of nonelective readmission. Conclusions: ML outperformed LR in predicting readmission following ECMO. Future work is needed to identify other factors linked with readmission and further optimize post-ECMO care among this cohort.

Association of Interhospital Transfer With Outcomes of Extracorporeal Membrane Oxygenation: A Contemporary Analysis.

BACKGROUND: Extracorporeal membrane oxygenation (ECMO) has emerged as a life-sustaining measure for individuals with end-stage cardiopulmonary derangements. An estimated one-third of patients must be transferred to a specialized center to receive this intervention. Therefore, the present study sought to characterize the impact of interhospital transfer (IHT) status on outcomes following ECMO. METHODS: The 2016-2020 National Inpatient Sample was queried to identify all adult (≥18 years) hospitalizations for ECMO. Patients were stratified based on transfer status from another acute care hospital. Multivariable regression models were developed to assess the association between transfer status and outcomes of interest. Patient and operative factors associated with IHT were identified using regression. RESULTS: Of an estimated 61,180 hospitalizations entailing ECMO, 21,410 (35.0%) were transfers. Annual transfer volume doubled over the study period, from 2915 to 5945 (nptrend < .001). The predicted morality risk of non-transfers decreased between 2016 and 2020 but remained similar in transferred patients. Following adjustment, transfer was associated with increased odds of in-hospital mortality, complications, duration of stay, and hospitalization costs. Patients experiencing transfer were less likely to be of black race and private insurance status. CONCLUSION: Despite increasing transfer volume and utilization of ECMO, IHT was associated with significant mortality and hospital complication risks. Further work to reduce adverse outcomes, resource burden, and socioeconomic differences within IHT may improve accessibility to this life-saving modality.

Machine Learning Based Prediction of Postoperative Pancreatic Fistula Following Pancreaticoduodenectomy.

OBJECTIVE: To develop a novel machine learning (ML) model to predict clinically relevant postoperative pancreatic fistula (CR-POPF) following pancreaticoduodenectomy (PD). SUMMARY BACKGROUND DATA: Accurate prognostication of CR-POPF may allow for risk stratification and adaptive treatment strategies for potential PD candidates. However, antecedent models, such as the modified Fistula Risk Score (mFRS), are limited by poor discrimination and calibration. METHODS: All records entailing PD within the 2014-2018 ACS NSQIP were identified. Additionally, patients undergoing PD at our institution between 2013 and 2021 were queried from our local data repository. An eXtreme Gradient Boosting (XGBoost) model was developed to estimate the risk of CR-POPF using data from the ACS NSQIP and evaluated using institutional data. Model discrimination was estimated using the area under the receiver operating characteristic (AUROC) and precision recall curve (AUPRC). RESULTS: Overall, 12,281 and 445 patients undergoing PD were identified within the 2014-2018 ACS NSQIP and our institutional registry, respectively. Application of the XGBoost and mFRS scores to the internal validation dataset revealed that the former model had significantly greater AUROC (0.72 vs. 0.68, P<0.001) and AUPRC (0.22 vs. 0.18, P<0.001). Within the external validation dataset, the XGBoost model remained superior to the mFRS with an AUROC of 0.79 (95% CI 0.74-0.84) versus 0.75 (95% CI 0.70-0.80, P<0.001). In addition, AUPRC was higher for the XGBoost model, compared to the mFRS. CONCLUSIONS: Our novel ML model consistently outperformed the previously validated mFRS within internal and external validation cohorts, thereby demonstrating its generalizability and utility for enhancing prediction of CR-POPF.