Development and validation of a nomogram for predicting carbapenem-resistant organism infection and prognostic risk factors analysis in elderly hospitalized patients

Abstract

Background: With the extensive use of carbapenem-resistant organism (CRO), CRO infection is constantly detected clinically which has limited available drugs. There have many mechanisms of CRO resistance and rapid horizontal transmission, while the elderly with low resistance is more likely to acquire nosocomial infection. We aimed to screen elderly patients for nosocomial CRO infection and potential risk factors for prognosis. Methods: A total of 177 patients with CRO and carbapenem-sensitive organism (CSO) infection were included in this study. A least absolute shrinkage and selection operator (LASSO) analysis was used to select variables. The nomogram was constructed with multivariable logistic regression. The performance of the model was assessed by the receiver operating characteristic(ROC) curve, calibration, decision curve , and clinical influence curve. Using X-tile to stratify the prognosis risk, Kaplan-Meier curve for risk assessment. Results: Respiratory diseases, mechanical ventilation, indwelling urinary catheters , and APACHE II over 20 were selected as predictors of the CRO infection model. The model showed good discrimination and consistency in the training set and the validation set. The area under the ROC was 0.840 (95% CI: 0.773-0.900)in the training set and 0.822 (95% CI: 0.678-0.936) in the validation set. Decision analysis and influence curve showed that the model was clinically useful. Hepatobiliary diseases, indwelling urinary catheters, and hospital stays longer than 20 days were used as prognostic predictors. After analysis, the prognostic model demonstrated good discrimination of 0.817 (95% CI: 0.729-0.893) and consistency. Risk stratification showed the high-risk group had a poorer prognosis. Conclusion: Predicting clinically relevant risk factors for CRO nosocomial infection and prognosis in elderly patients. This may help the treatment of clinical drug-resistant infections.