A new lasso-logistic regression model was developed and validated, identifying nine independent risk factors for postpancreatectomy hemorrhage among 9,631 patients across two cohorts. The model demonstrated strong predictive accuracy, achieving an area under the receiver operating characteristic curve of 0.87 during external validation, significantly outperforming previous prediction models (p<0.001). With 9.0% and 8.5% of patients experiencing hemorrhage in the training and validation cohorts, respectively, this model shows promise in enhancing surgical safety and patient outcomes.
• Why it matters: Enhancing predictive accuracy addresses risk of postpancreatectomy complications.
Journal Article by Duan Y, Du Y (…) Wang C et 5 al. in Int J Surg
Copyright © 2024 The Author(s). Published by Wolters Kluwer Health, Inc.