Development and validation of a nomogram for predicting PICC catheter-related bloodstream infection among patients with hematologic malignancies.
To identify risk factors for PICC-related bloodstream infections (PICC-CRBSI) in patients with hematologic malignancies and to develop and validate a predictive nomogram for clinical risk assessment.
This retrospective cohort study included 764 patients with hematologic malignancy who had PICC from a tertiary center (2021-2024). LASSO regression identified key predictors after addressing class imbalance via SMOTE-NC in the training set (n = 534). A multivariate logistic regression was used to construct a predictive nomogram, subsequently validated in a validation set (n = 230). Model performance was assessed by area under the receiver operating characteristic curve (AUC), calibration curves with bootstrap resampling, and decision curve analysis (DCA).
In our cohort, the occurrence of PICC-CRBSI infections was 6.02% (46/764) and Gram-negative bacteria was the major causative pathogen. Multivariate regression analysis showed that History of diabetes, Age, Time of PICC placement, PICC insertion attempts, Catheterized Diameter, Catheterized vein, ANC, ALC and D-dimer were independent risk factors associated with PICC-CRBSI in hematologic malignancy patients. Our nomogram model demonstrated a good calibration and discrimination in both training and validation sets, with AUC values of 0.883 and 0.822, The DCA suggested potential clinical utility.
This validated nomogram integrates patient, catheter, and biomarker-specific factors to individualize PICC-CRBSI risk stratification in hematologic malignancies, potentially guiding targeted prevention strategies.
This retrospective cohort study included 764 patients with hematologic malignancy who had PICC from a tertiary center (2021-2024). LASSO regression identified key predictors after addressing class imbalance via SMOTE-NC in the training set (n = 534). A multivariate logistic regression was used to construct a predictive nomogram, subsequently validated in a validation set (n = 230). Model performance was assessed by area under the receiver operating characteristic curve (AUC), calibration curves with bootstrap resampling, and decision curve analysis (DCA).
In our cohort, the occurrence of PICC-CRBSI infections was 6.02% (46/764) and Gram-negative bacteria was the major causative pathogen. Multivariate regression analysis showed that History of diabetes, Age, Time of PICC placement, PICC insertion attempts, Catheterized Diameter, Catheterized vein, ANC, ALC and D-dimer were independent risk factors associated with PICC-CRBSI in hematologic malignancy patients. Our nomogram model demonstrated a good calibration and discrimination in both training and validation sets, with AUC values of 0.883 and 0.822, The DCA suggested potential clinical utility.
This validated nomogram integrates patient, catheter, and biomarker-specific factors to individualize PICC-CRBSI risk stratification in hematologic malignancies, potentially guiding targeted prevention strategies.