Related Experiment Video
Updated: Jul 1, 2026
Establishing a Competing Risk Regression Nomogram Model for Survival Data
Published on: October 23, 2020
Development and validation of a nomogram model for mortality risk in burn patients: Triglyceride-glucose index as an
Anna Liu1, Yimin Zhang1, Jie You2
1Department of Burn & Wound Repair Nursing, Fujian Medical University Union Hospital, Fuzhou Fujian Province 350001, PR China; Fujian Burn Institute, Fujian Medical University Union Hospital, Fuzhou Fujian Province 350001, PR China; Fujian Burn Medical Center, Fujian Medical University Union Hospital, Fuzhou Fujian Province 350001, PR China; Fujian Provincial Key Laboratory of Burn and Trauma, Fujian Medical University Union Hospital, Fuzhou Fujian Province 350001, PR China.
Background:
To explore risk factors associated with mortality in burn patients, construct a mortality prediction model for burn patients, and expect this model to quantitatively assess the mortality risk of newly admitted burn patients, providing objective decision support for clinicians to help formulate individualized treatment strategies.
Methods:
The clinical data of 400 burn patients admitted to Fujian Burn Medical Center from January 2020 to December 2022 were retrospectively analysed and divided into the survival group (n = 383) and the death group (n = 17). Univariate and multivariate logistic regression analyses were used to investigate the risk factors associated with poor prognosis, and constructing the nomogram models. In addition, model performance was evaluated using calibration curves, receiver operating characteristic (ROC) curves, and decision curve analysis (DCA).
Results:
The results of the stepwise binary logistic regression analysis for constructing a predictive model showed that when the model included triglyceride-glucose index (TyG), absolute lymphocyte count, hemoglobin, calcium, and platelet-lymphocyte ratio (PLR), the model had the smallest Akaike Information Criterion (AIC) of 97.757. The model's C-index was 0.856. The ROC curve analysis of the model on the training set data yielded an AUC of 0.856 (95% confidence interval: 0.739-0.973), indicating high discriminative ability. The Hosmer-Lemeshow goodness-of-fit test results showed a good fit (χ²=10.905, P = 0.207).
Conclusion:
Nomograms can predict the risk of death in burn patients, providing personalized clinical decision-making for future clinical practice.