• Chongwei Li, Linghui He, Junwei Xu, Lili Wang, Xiaoli Cao, Hui Zhang, Pingping Ma, and Yongmei Yuan.
• School of Public Health, Nantong University, Nantong, Jiangsu, China.
• Brit J Hosp Med. 2024 Oct 30; 85 (10): 1111-11.

AbstractAims/Background Patients receiving treatment in specialized cancer hospitals are particularly susceptible to multidrug-resistant organisms (MDRO) infections due to factors such as weakened immune systems caused by intensive treatments and prolonged hospital stays. This study aims to investigate the risk factors for MDRO infections in the cancer specialty hospital setting and to develop a corresponding risk prediction model. Methods Patients diagnosed with MDRO infections were selected for the MDRO infection group (n = 238), and those without for the non-MDRO infection group (n = 238). Non-parametric tests, chi-square tests, and multivariate logistic regression analysis were used to identify the primary risk factors for MDRO infections. With the aid of analysis utilizing R software 4.4.1 (R Foundation for Statistical Computing, Vienna, Austria), we developed a nomogram prediction model, which was evaluated using the receiver operating characteristics (ROC) curve, calibration curve, and decision curve analysis (DCA). Results Age, antibiotic application time, and central venous catheterization were independent risk factors for MDRO infection (p < 0.05). The constructed nomogram prediction model for patients with MDRO infection has a C-index of 0.8640. The ROC curve results showed that the prediction model has a specificity of 0.7700, a sensitivity of 0.8800, and an area under the curve (AUC) of 0.8800. Conclusion This study identifies significant risk factors for MDRO infections in a cancer specialty hospital setting and offers a clinically useful prediction model, which may aid in targeted preventive measures and optimization of antibiotics usage, thereby potentially reducing the incidence and impact of these infections.

29 keywords...

hide…