基于Apriori算法的超标医疗费用关联规则挖掘  

Using Association Rules Data Mining to Explore over Proof Medical Costs Based on Apriori Algorithm

作　　者：陈梅[1] 刘忠[1] 陈国英[1] 周娟[1] 

机构地区：[1]成都市龙泉驿区第一人民医院,610100

出　　处：《中国卫生经济》2013年第10期73-75,共3页Chinese Health Economics

基　　金：四川省卫生厅科研课题基金资助项目(110108)

摘　　要：目的分析医院病人医疗费用构成,探讨医疗费用项目结构的合理性,挖掘超标医疗费用超标的原因,使医院管理者有针对性的控制医疗费用。方法以各疾病诊断相关组医疗费用的3/4分位数作为该组超标费用阈值,区分标内费用和超标费用;利用数据挖掘中的关联规则挖掘方法中的Apriori算法挖掘与超标费用有强关联的条件属性。结果在呼吸系统疾病中,1/4的超标医疗费用患者消耗掉了超过40%的医疗资源。关联规则模型共挖掘出4条有意义的关联规则,监测到发生费用超标的病例占超标费用人数的10.7%。4条规则中具有此条件使得超标发生的可能性依次是原来的2.47、2.36、2.30和2.29倍。结论:医院质量控制部门对医疗费用管理时可以参考关联规则挖掘出的强关联规则,对入院病人中满足规则的应给予重点关注,在满足病人医疗需求的前提下尽量避免超出标准费用的情况发生,加强医疗费用支出的事前预防。Objective： To analyze the composition of patients＇ medical costs, to explore whether the composition of medical costs is reasonable, to find the causes of over proof medical costs and to help the hospital managers to control medical costs. Methods： Use 3/4 percentile of the diagnosis related group medical costs as the threshold of over proof medical costs, distinguish the standard medical costs and over proof medical; use Apriori Algorithm in the association rules method of data mining, to find the condition attributes which have strong association with the over proof medical costs. Results： In the respiratory system disease, 1/4 of pationts with the over proof medical costs had occupy over 40% medical resources. There were 4 significant association rules in the association rule model, they could monitor 10.7% patients have over standard medical costs. The possibilities of developing over standard medical costs with the 4 rules were 2.47, 2.36, 2.30 and 2.29 times more than before, respectively. Conclusion： The quality control department in the hospital may use the strong association rules to control medical costs, to pay more attention to the patients suit with these rules. In the precondition of satisfying the medical demand of patients, avoid situation of over proof medical costs and enhance the prevention of medical costs from happening.

关 键 词：超标医疗费用 关联规则挖掘模型 APRIORI算法 

分 类 号：R197.3[医药卫生—卫生事业管理]