基于3种数据挖掘模型分析16618例含附子中成药的处方用药规律

Analysis of 16618 Prescriptions of Chinese Patent Medicine Containing Aconiti Lateralis Radix Praeparata Based on Three Data Mining Models

目的:通过数据挖掘分析某院含附子的中成药处方,从中医辨证论治的角度分析该类中成药的临床用药特征规律,探究数据挖掘在中医药处方分析中的应用前景。方法:抽取2017年11月至2018年10月某院门诊含附子成分的中成药处方,从患者基本情况、中医疾病、中医证候等方面回顾分析该类中成药的临床使用情况,采用Microsoft Office Excel 2016进行基本统计,并结合Microsoft SQL Server Analysis Services 2012的3种数据挖掘算法解析处方用药特征。结果:某院使用含附子中成药共8种,14岁以下患者用该类药最多,其中小儿肺咳颗粒用量最大,中医疾病多为咳嗽、感冒、咳喘,中医证型多为寒热错杂证、风痰证、外感风邪。温补肾阳的中成药处方包含金匱肾气丸、龙鹿胶囊、尪痹胶囊、桂附地黄丸,中医疾病多为痹症、眩晕、虚劳等,证型主要为肝肾不足、肾虚血瘀、脾肾不足。决策树分析显示使用含附子的中成药处方具有年龄分布特点,其中中医疾病及证型为主要决策点;聚类分析根据该类处方患者年龄、性别、中西医诊断、中医证型将中成药处方分为10类;关联性分析显示小儿肺咳颗粒联用情况最多,但关联度不高,金匱肾气丸、芪苈强心胶囊、龙鹿胶囊和尪痹胶囊有各自的强关联用药。结论:数据挖掘算法纳入多因素分析临床处方的应用更接近中医临床辩证论治思维,结合基础统计可知某院含附子的处方整体用药符合中医辨证论治的特点。

Objective:Through data mining analysis of Chinese patent medicine prescription containing Aconiti Lateralis Radix Praeparata(CPMCA)in a hospital,the characteristic rules of clinical medication of this kind of Chinese patent medicine were analyzed from the perspective of traditional Chinese medicine(TCM)syndrome differentiation and treatment,and the application prospect of data mining in TCM prescription analysis was explored.Method:CPMCA in the hospital outpatient clinics from November 2017 to October 2018 was conducted.The clinical application of this kind of Chinese patent medicine was retrospectively analyzed from the aspects of basic situation of patients,TCM diseases,TCM syndromes and so on.The clinical application of CPMCA was reviewed and analyzed.Three kinds of data mining algorithms in Microsoft SQL Server Analysis Services 2012 were used to analyze the characteristics of prescription medication.Result:A total of 8 kinds of Chinese patent medicines containing Aconiti Lateralis Radix Praeparata were used in a hospital,and most of them were used by patients under 14 years old,among them,the dosage of Xiaoer Feike granules was the largest.The TCM diseases were mostly cough,cold and asthma,while the TCM syndromes were mostly cold-heat complicated syndrome,wind-phlegm syndrome and external contraction of wind evil.The CPMCA for nourishing and warming kidney-Yang included Jingui Shenqiwan,Longlu capsules,Wangbi capsules and Guifu Dihuang pills.The TCM diseases were mostly arthralgia syndrome,vertigo,and consumptive disease,meanwhile,the syndromes were mainly deficiency of liver and kidney,kidney deficiency and blood stasis,and deficiency of spleen and kidney.Decision tree analysis showed that the CPMCA had the characteristics of age distribution,among which TCM diseases and syndromes were the main decision points.According to the age,sex,diagnosis of Chinese and western medicine,and TCM syndromes of this type of prescription,the CPMCA were divided into 10 categories by cluster analysis.Correlation analysis showed that the combination of Xiaoer Feike granules was the most common,but the correlation was not high.And Jingui Shenqiwan,Qili Qiangxin capsules,Longlu capsules and Wangbi capsules had their own strong associated medication.Conclusion:The application of data mining algorithm in multi-factor analysis of clinical prescription is close to the clinical dialectical treatment thinking of TCM.Combined with the basic statistics,it can be seen that the CPMCA in a hospital is consistent with the characteristics of TCM syndrome differentiation and treatment.