Development and validation of nomogram for prognosis of adult sepsis patients: a retrospective observational cohort study

Abstract Objective To develop a new nomogram for the bedside evaluation of patients with sepsis. Method This was a retrospective study. The clinical information of patients with sepsis from 2020 to 2022 was collected at the First Hospital of Jilin University. Associated risk factors of patients with sepsis were identified by univariate and multivariate logistic regression and used to develop a nomogram. The nomogram validated the area under the receiver operating characteristic curve (AUC), the Hosmer-Lemeshow test, and the calibration plot. Furthermore, decision curve analysis (DCA) was used to evaluate the effect of the prediction model in clinical application. Result 168 patients were included and divided into a training cohort (n = 123) and a verification cohort (n = 45) by the year. Multivariate logistic regression revealed that age, surgery, solid tumor, lactic acid, and coagulation index (CI) were associated risk factors for in-hospital mortality.The AUC was 0.83 (95%CI 0.76–0.91) and 0.81 (95%CI 0.691–0.937) in the training and validation cohorts, respectively. The calibration plot and Hosmer-Lemeshow test prove that the nomogram has good accuracy.DCA shows that the nomogram of the predictive model has good clinical value. Conclusion Age, surgery, solid tumor, lactic acid, and CI are associated risk factors for sepsis. Physicians could use the new nomogram in making clinical decisions and predicting patients’ prognosis.