医院获得性耐碳青霉烯类肺炎克雷伯菌血流感染预测模型的构建与验证OA

Objective To analyze the risk factors for hospital-acquired bloodstream infections(BSIs)caused by carbapenem-resistant Klebsiella pneumoniae(CRKP)and to develop a nomogram prediction model.Methods A retrospective study was conducted at the Affiliated Hospital of North Sichuan Medical College from January 1,2020,to December 31,2025.A total of 147 cases of hospital-acquired BSIs caused by Klebsiella pneumoniae(KP)were included.Cases were divided into CRKP(n=59)and carbapenem-sensitive KP(CSKP)(n=88)groups based on susceptibility to carbapenem antibiotics.Potential risk factors were initially screened using LASSO regres-sion,followed by multivariate logistic regression to identify independent risk factors.A nomogram prediction model was then constructed based on these factors.The model's performance was evaluated using Brier score,area under the receiver operating characteristic curve(AUC),Hosmer-Lemeshow test,calibration curve,and decision curve analysis(DCA).Bootstrap internal validation with 1 000 resamplings was performed to assess model stability.Results Multivariate logistic regression analysis identified the following as independent risk factors for hospital-acquired CRKP BSIs:use of carbapenem antibiotics,use of quinolone antibiotics,admission to ICU,and co-infection with multidrug-resistant organisms(MDROs)at other sites(all P<0.05).The nomogram model demonstrated good discrimination and calibration:Brier score was 0.100,AUC was 0.923(95%CI:0.880-0.967),the Hosmer-Lemeshow test was not statistically significant(χ2=4.552,P=0.208),the calibration curve closely approximated the ideal curve,and DCA indicated a clinical net benefit at a threshold probability of approximately 0.1-0.8.After bootstrap internal validation,the AUC was 0.913(95%CI:0.882-0.967),with an AUC decay of 1.08%,which further confirmed the model's stability.Conclusion Carbapenem use,quinolone antibiotic use,ICU admission,and concomitant multidrug-resistant organism infection at other sites are independent risk factors for hospital-acquired CRKP BSIs.The nomogram prediction model constructed based on these factors exhibits good discriminative ability and clinical utility,and may inform the early identification and intervention of patients at high risk of hospital-acquired CRKP BSIs.