ORIGINAL RESEARCH article
Front. Med.
Sec. Intensive Care Medicine and Anesthesiology
Predicting Short-Term Outcomes of Comprehensive Nursing Care in Limb Trauma Fracture Patients Using Machine Learning Models
Minque Liao
Lan Luo
Central Hospital of Enshi Tujia and Miao Autonomous Prefecture, Wuhan University, Enshi, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Abstract
Abstract Background This study analyzes the impact of multiple factors on the short-term effects of Comprehensive Nursing Care by constructing machine learning and neural network models, providing data support for clinical decision-making. Methods We first analyzed the differences in baseline characteristics and surgery-related information between the training set and the external validation set. Then, five machine learning and neural network models - multiple linear regression (MLR), random forest (RF), support vector machine (SVM), extreme gradient boosting (XGBoost), and multilayer perceptron (MLP) - were constructed using the training set, with the dependent variables being the patient-reported outcomes measurement information System (PROMIS) Physical Function (PF) score and pain level visual analog scale (VAS) score, respectively. Ten-fold cross-validation was used for parameter tuning, and the average performance of the models was evaluated in the validation set. Finally, the best-performing model was selected, and the independent variables were ranked according to their contribution. Results The results showed that the RF model achieved the highest explanatory power, with a coefficient of determination (R²) of 0.468, and the lowest prediction error, with a root mean square error (RMSE) of 5.220, in predicting PF scores, and also performed best in predicting pain VAS scores (R² = 0.426, RMSE = 0.912), outperforming MLR, XGBoost, SVM, and MLP models. Permutation variable importance analysis (Increase in Mean Squared Error, IncMSE) showed that for PF scores, diabetes, alkaline phosphatase (ALP) level, and hypertension were the top three factors, while for pain scores, lower limb fractures, erythrocyte sedimentation rate (ESR) level, and ALP level were the top three factors. Conclusion Compared with other models, the RF model appears to be more suitable for predicting comprehensive nursing care outcomes in patients with limb trauma fractures. Notably, ALP level emerged as one of the top three predictors in both outcome prediction models.
Summary
Keywords
Fractures, Bone, machine learning (ML), Neural Networks (Computer), Nursing Care, Wounds and Injuries
Received
27 June 2025
Accepted
16 February 2026
Copyright
© 2026 Liao and Luo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Lan Luo
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.