基于数据挖掘的保肝药物专项点评实践

Data Mining-based and Targeted Reviews of Prescriptiosn for Hepatoprotective Drugs

目的构建数据挖掘模型对2所医院6种保肝药物处方进行专项点评,为药师快速精准开展药物临床应用专项点评提供新思路。方法创建保肝药合理用药点评标准知识库,应用SPSS Modeler 18.0软件,基于知识库构建聚类分析与关联规则分析模型,对2022年1-12月共2所医疗机构的6种保肝药物的相关诊疗数据实施专项处方点评,并通过人工复核验证其准确。结果6种保肝药的总体使用合理率不低于95%,但均存在无适应证用药情况;无高危人群用药情况;双环醇片和异甘草酸镁注射液存在用法用量不合理行为;水飞蓟宾胶囊和熊去氧胆酸胶囊在相互作用方面存在不合理用药行为。结论采用数据挖掘模型对6种保肝药进行点评,同时对不合理用药病例进行人工确证,验证了该模型的实用性和准确性,提高了药师点评处方的效率,有效节约医疗资源,为医院药学处方点评提供了一种新的思路和方法。

Objective To establish data mining models for targeted reviews of prescriptions for six liver-protecting drugs in two tertiary-level hospitals in order to help pharmacists to review clinical medications precisely.Methods A knowledge base of standards was established for review of rational medications with liver protective drugs.SPSS Modeler18.0 software was used to build models for cluster analysis and association rules before targeted reviews of prescriptions involving six liver-protecting medications were conducted using data collected in 2022 at two hospitals.and the accuracy was re-confirmed manually.Results The percentage of rational use of the six liver protection drugs was 95%or above,but each of the drugs was occasionally used for wrong indications.None of the six drugs was used by high-risk groups.Improper usage of Bicyclol Tablets and Magnesium Isoglycyrrhizinate injections was detected.Problematic medications with Silibinin capsules and Ursodeoxycholic Acid capsules due to drug interactions were also found.Conclusion This study has constructed a series of data mining models for evaluation of six liver protection drug prescriptions.The practicality and accuracy of the models have been verified through manual confirmation of cases of irrational drug use.These models can improve pharmacists'efficiency of prescription evaluation,help save medical resources,and provide a new approach to reviewing prescriptions at pharmacies.