Related Experiment Video
Updated: May 23, 2026
Establishing a Competing Risk Regression Nomogram Model for Survival Data
Published on: October 23, 2020
Development and internal validation of a nomogram for predicting 30-day mortality in older patients with
Jun Zhou1,2, Jingping Liu1,2, Hua Yin3
1Department of Laboratory Medicine, the First Affiliated Hospital with, Nanjing Medical University, Nanjing, Jiangsu, 210029, China.
Objective:
To develop and validate a prognostic model for predicting 30-day mortality in older patients with hemophagocytic lymphohistiocytosis (HLH).
Methods:
This retrospective cohort study enrolled 204 HLH patients aged ≥ 65 years from January 2015 to November 2023. We divided the cohort into development and validation cohorts in a 7:3 ratio. Then we used logistic regression analysis and the least absolute shrinkage and selection operator regression (LASSO) to develop a prognostic model. Performance was assessed using the area under the receiver operating characteristic curve (AUC) and decision curve analysis (DCA).
Results:
The 30-day mortality rate was 40.7%. Multivariate analysis identified five risk factors independently associated with 30-day mortality of older patients with HLH: Age, platelet (PLT), alanine aminotransferase (ALT), UREA, and ferritin. The model has good discrimination and calibration ability (AUC: 0.828 (0.755-0.886) for the development cohort and 0.773 (0.654-0.891) for the validation cohort). The model showed excellent calibration and clinical utility. Kaplan-Meier survival curve analysis showed that patients with the nomogram value > 0.3851 were positively correlated with higher 30-day mortality (P < 0.001).
Conclusion:
The model incorporating age and four routine clinical parameters accurately stratifies 30-day mortality risk in older HLH patients demonstrating strong discriminative ability and clinical applicability, thereby providing a basis for clinical decision-making.
