医用耗材使用量的卡尔曼滤波预测模型构建与应用研究

Construction and application analysis of Kalman filter prediction model for medical consumables

目的:构建医用耗材使用量的卡尔曼滤波预测模型,探讨其在医用耗材管理中的应用价值。方法:选取医院临床在用的医用耗材品目,根据医用耗材使用量预测模型不同分为对照组和观察组,对照组采用单一指数平滑法预测模型进行预测分析,观察组采用卡尔曼滤波预测模型在采集医用耗材全生命周期数据基础上,搭建智能化分析平台开展医用耗材申购指导和监控管理。自制满意度调研问卷,对医院生殖医学中心、妇科和男科医护人员进行医用耗材管理临床服务质量调研,对比两组医用耗材预测效度、管理质量和临床服务水平的差异。结果:观察组医用耗材临床实际使用数量与预测采购数量呈高度相关性(r=0.842,P<0.05),预测效度平均值均好于对照组,差异有统计学意义(t=4.340,P<0.05);观察组医用耗材库存剩余率、应急采购率和质量失效率低于对照组,功能匹配度高于对照组,差异有统计学意义(t=3.130,t=2.888,t=2.622,t=2.260;P<0.05);医院生殖医学中心、妇科和男科医护人员对观察组医用耗材服务水平满意度和总认可度平均值均高于对照组,差异有统计学意义(t=2.292,t=3.303,t=2.446,t=4.951;P<0.05)。结论:卡尔曼滤波预测模型能够提高医用耗材预测准确性,降低库房管理压力和应急采购频次,提升医用耗材临床管理质量。

Objective: To construct a Kalman filter prediction model and discuss its application value in management of medical consumables. Methods: Medical consumables in clinical use in the hospital were selected divided into control group and observation group according to different predictive models. The control group adopted a single exponential smoothing prediction model for predictive analysis. The observation group adopted the Kalman filter model to build an intelligent analysis platform based on the collection of whole life cycle data of medical consumables to conduct purchase guidance and monitoring management of medical consumables. The self-made satisfaction questionnaire was used to conduct survey on clinical service quality of medical consumables management among medical staff of reproductive medicine center, gynecology department and andrology department of the hospital. The differences in the predictive validity, management quality and clinical service level of medical consumables between the two groups were compared.Results: The actual clinical use quantity of medical consumables in the observation group was highly correlated with the predicted purchase quantity(r=0.842, P<0.05), and the predictive validity was better than that in the control group,the difference was statistically significant(t=4.340, P<0.05). The inventory remaining rate, emergency procurement rate, quality failure rate and function matching degree of medical consumables in the observation group were better than those in the control group, the difference was statistically significant(t=3.130, t=2.888, t=2.622, t=2.260, P<0.05).The satisfaction and general recognition of reproductive medicine center, gynecology and andrology medical staff with medical consumables of the observation group were higher than that of the control group, the difference was statistically significant(t=2.292, t=3.303, t=2.446, t=4.951, P<0.05). Conclusion: The Kalman filter model can improve the accuracy of prediction of medical consumables, reduce the pressure of warehouse management and the frequency of emergency procurement and enhance the clinical use quality of medical consumables.