二值响应模型与决策树在门诊失约行为研究中的应用  

Application of Binary Response Model and Decision Tree in the Study of Missed Outpatient Appointment

作　　者：吴越[1] 徐丛剑[1] 程子桐 李斌[1] 苏毓 李梦斐 WU Yue;XU Cong-jian;CHENG Zi-tong(Obsterics and Gynecology Hospital of Fudan University,Shanghai,200090,China)

机构地区：[1]复旦大学附属妇产科医院,上海200090 [2]中南财经政法大学统计与数学学院,湖北武汉430073

出　　处：《中国医院管理》2018年第10期36-38,共3页Chinese Hospital Management

基　　金：上海市医院协会医院管理研究基金(201801049)

摘　　要：目的通过将二值响应模型与决策树相结合,对门诊预约失约患者的特征进行鉴别和归纳,以期为减少患者失约行为的发生,提高门诊预约管理质量提供参考。方法将二值响应模型与决策树联合应用,鉴别出对失约行为具有显著效应的特征,并将其作为分类节点,构建决策树,最终获得失约患者的特征组合。结果在样本医院的妇科门诊中,预约提前时间≤1天的患者,预约提前时间>1天且≤2天并且预约新院区就诊的外地患者及预约提前时间>20天的外地患者更容易发生失约行为。结论可利用该模型,对预约患者的失约率进行预测,并采取针对性措施,减少失约行为的发生。Objective To summarize and identify the characteristics of outpatients who missed the appointment by combined use of Binary Response Model and Decision Tree in order to reduce the missed appointment ratio and improve the quality of outpatient appointment management.Methods Through Binary Response Model and Decision Tree,the characteristics of significant effect on the missed appointment were identified and used as the classification node of Decision Tree.Then the Decision Tree can help to obtain the characteristic combination of the outpatients who missed appointment.Results In Gynaecological Clinic of sample hospitals,outpatients with an appointment≤1 day in advance,ecdemic outpatients who make an appointment>1 day and 2 days in advance and choosing new campus,and ecdemic outpatients with an appointment>20 days in advance are more likely to miss the appointment.Conclusion The model can be used to predict the rate of missed appointment,and targeted measures should be taken to reduce the occurrence of this phenomenon.

关 键 词：二值响应模型 决策树 门诊预约 

分 类 号：R197.323.2[医药卫生—卫生事业管理]