Abstract: Nosocomial infection is an important index of medical service quality. The existing literatures mainly study infection transmission based on the compartmental model, but fails to consider the status of discharged patients, which often overestimated the risk of systemic infection. To solve this problem, this paper proposes a research on the propagation and control of nosocomial infection based on multidimensional continuous-time Markov chain (MCTMC), so as to accurately depict the infection status and evolution process of patients and healthcare personnel. According to the characteristics of nosocomial infection transmission, infection vectors (healthcare personnel) were integrated into the mechanism of SIS (Susceptible-Infected-Susceptible ) to build a MCTMC-SIS model. By analyzing and comparing the state evolution rates of these two models, it is proved that the MCTMC-SIS model is advanced. Finally, the effectiveness of the control strategy is verified by numerical simulation, and the importance of parameters is evaluated. The results show that the rate of pathogenic bacteria on admission has a higher influence on the transmission of infection. When the rate is high (≥ 0.05), it is ineffective to reduce the risk of infection by improving the discharge rate of patients and the frequency of hand-washing. In actual infection control, infected patients need to be monitored in advance to reduce the risk of infection.

Key words: nosocomial infection, infection vectors, multi-dimensional continuous-time Markov chain (MCTMC), numerical simulation

摘要： 医院感染（院感）是医疗服务质量的重要指标。现有文献主要应用仓室模型研究院感传播，未考虑出院病人的感染状态，往往高估系统感染风险。针对这一问题，提出多维连续时间马尔科夫链的院感传播及管控策略研究，精准刻画病人及医护人员的感染状态及其演变过程。根据院感传播特点，将感染媒介（医护人员）集成到SIS传播机制，构建MCTMC-SIS模型。通过分析对比两类模型的状态演变率，证明MCTMC-SIS模型先进性。最后，通过数值仿真验证管控策略的有效性，并评估传播参数的重要性。结果显示，病人入院携带病原菌率对院感传播影响较高。当入院病人携带病原菌概率较高（≥0.05）时，提高病人出院率及医护人员洗手频率，无法有效降低病人感染风险。在实际院感管控中，需提前监测感染患者，降低病人感染风险。

关键词: 医院感染, 感染媒介, 多维连续时间马尔科夫链, 数值仿真