故障模式影响及危害性分析与机器学习在生命支持设备风险管理中的研究OA

Objective:To establish a risk-stratified management system that was driven by failure modes effects and criticality analysis(FMECA)and machine learning,and to explore the application effect of that in the operation and management for hemodialysis equipment of hospital.Methods:The FMECA method was adopted to conduct a risk assessment that combined qualitative and quantitative assessment for equipment,which introduced a machine learning model to predict and classify the risk levels of the equipment,thereby realize risk-stratified management for equipment.A total of 46 hemodialysis machines in clinical use of Shanghai Seventh People's Hospital of Shanghai University of Traditional Chinese Medicine from 2023 to 2024 were selected as the study subjects.The conventional management method was adopted to manage equipment during January 2023 and December 2024,and the strategy with risk-stratified management system(risk-stratified management method),which was driven by FMECA combined with machine learning,was adopted to manage equipment during January and December 2024.The differences of the management effect in clinical use,the management effect of technical support for equipment,and safety indicators of operation and maintenance for equipment between the two kinds of management methods were compared.Results:The average operation rate of equipment that adopted risk-stratified management method was(94.63±4.95)%,which was significantly higher than(91.79±5.24)%of conventional management method,and the difference was significant(t=2.672,P<0.05).However,the averagely relative operation rate and the turnover rate of equipment of risk-stratified management method were respectively(8.47±2.39)%and(9.58±2.67)%,which were lower than those of conventional management method,and the differences were significant(t=11.739,11.313,P<0.05).The average failure rate of equipment that received the risk-stratified management method was significantly lower than that of the conventional management method,and the difference was significant(t=12.098,P<0.05).The averagely self-repair rate and compliance rate of scrapping of the risk-stratified management method was higher than those of the conventional management method,and the differences were significant(t=3.752,5.361,P<0.05).The incidence of averagely pressure failure,incidence of electrical failure,incidence of failure in circulation system,and averagely defective rate of cleaning and disinfection of equipment that adopted risk-stratified management method were significantly lower than those that adopted conventional management method,and the differences were statistically significant(t=6.004,10.261,10.407,14.324,P<0.05).Conclusion:The risk-stratified management system that is driven by FMECA and machine learning can significantly improve the management effect,the effect of technical support for life support equipment in clinical use,which can reduce failure risk of equipment.