Risk prediction model and validation of respiratory failure in patients with sepsis
10.3969/j.issn.1006-2483.2026.04.019
- VernacularTitle:脓毒症患者呼吸衰竭风险预测模型构建及验证
- Author:
Jianwei ZHAO
1
;
Qi WANG
1
;
Chengkang LU
1
;
Xia LI
1
Author Information
1. Ganzi Tibetan Autonomous Prefecture People's Hospital, Department of critical care medicine; Kangding, Sichuan, 626000, China
- Publication Type:Journal Article
- Keywords:
Sepsis;
Respiratory failure;
Risk prediction model
- From:
Journal of Public Health and Preventive Medicine
2026;37(4):91-95
- CountryChina
- Language:Chinese
-
Abstract:
Objective To analyze the influencing factors of respiratory failure in patients with sepsis, and to construct and validate a risk prediction model. Methods A total of 308 patients with sepsis in Ganzi Tibetan Autonomous Prefecture People's Hospital from August 2021 to September 2024 were enrolled and divided into a modeling group and a validation group. According to whether respiratory failure occurred during hospitalization, the patients were divided into a respiratory failure group and a non-respiratory failure group. Single-factor and binary logistic regression analysis were used to construct a risk prediction model, and the ROC, calibration, and DCA curves were used to assess the model's predictive performance. Results The incidence of respiratory failure was 20.83% (45/216) in the modeling group and 20.65% (19/92) in the validation group. Albumin, D-dimer, serum calcium, procalcitonin at admission, and sequential organ failure assessment were independent influencing factors of respiratory failure in patients with sepsis. The ROC curve analysis nomogram model predicted that the AUC of respiratory failure in the modeling group was 0.994 (95%CI: 0.985-1.000), and the AUC of the validation group was 0.897 (95%CI: 0.826-0.968). The calibration curve analysis showed that the calibration curve fit well with the ideal curve. DCA curve analysis showed that the net benefit was greater than 0 across all threshold probability ranges. Conclusion The risk prediction model constructed in this study is highly effective and can provide reference for medical staff to evaluate respiratory failure in patients with sepsis.
