Abstract
PURPOSE: This study aimed to develop and validate the Carbapenem-Resistant Acinetobacter baumannii Risk Scoring System (CRAB-RSS), a novel predictive model designed to assess the risk of carbapenem-resistant A. baumannii (CRAB) infection in intensive care unit (ICU) patients. METHODS: A retrospective cohort analysis was performed on 412 patients (315 with CRAB and 97 with carbapenem-susceptible A.baumannii [CSAB]) from 2020 to 2024. Three independent risk factors were identified: mechanical ventilation (adjusted odds ratio [aOR] = 3.2, 95% confidence interval [CI]: 1.8-5.6), prior carbapenem exposure (≥48 hours; aOR = 1.89, 95% CI: 1.32-2.71), and hospitalization duration exceeding 14 days (aOR = 1.67, 95% CI: 1.25-2.23). RESULTS: The model demonstrated robust discriminative ability, with an area under the receiver operating characteristic curve (AUROC) of 0.887 in the derivation cohort and 0.918 in the validation cohort, along with satisfactory calibration (Brier score: 0.094 versus 0.088). Its performance was significantly superior to that of the Sequential Organ Failure Assessment (SOFA) score (ΔAUROC = +0.21). CONCLUSIONS: CRAB-RSS is a quantitative risk stratification tool derived from a single-center cohort for CRAB infections in intensive care unit patients, demonstrating superior performance to SOFA in local validation. External multicenter validation is warranted before broad clinical implementation. Its innovative features include: (1) a fixed-weighting design (e.g., assigning a baseline score of 2 points for mechanical ventilation), and (2) reliance on only three readily obtainable clinical variables to complete the assessment. Decision curve analysis revealed that the application of CRAB-RSS could reduce unnecessary carbapenem use by 28%-42% across probability thresholds of 10%-30%, with a maximum reduction of 38% achieved at the 20% threshold.