主 办:北 京 中 医 药 大 学
ISSN 1006-2157 CN 11-3574/R

北京中医药大学学报 ›› 2018, Vol. 41 ›› Issue (8): 681-688.doi: 10.3969/j.issn.1006-2157.2018.08.011

• 临床研究 • 上一篇    下一篇

基于隐结构模型的名老中医辨治京津冀地区慢性支气管炎用药规律研究*

陈丽平1, 李建生2#, 蔡永敏3, 卞华1   

  1. 1 南阳理工学院 河南省张仲景方药与免疫调节重点实验室 河南 473004;
    2 河南中医药大学 呼吸疾病诊疗与新药研发河南省协同创新中心;
    3 南京中医药大学基础医学院
  • 收稿日期:2018-02-03 出版日期:2018-08-30 发布日期:2018-08-30
  • 通讯作者: 李建生,男,博士,教授,博士生导师,研究方向:呼吸疾病的中医药防治,E-mail: li_js8@163.com
  • 作者简介:陈丽平,女,在读博士生
  • 基金资助:
    国家自然科学基金青年科学基金资助项目(No.81704200),河南省科技攻关项目(No.172102310422),河南省博士后科研项目(No.00104257),河南省中医药研究专项(No.2015ZY02003)

Senior prestigious TCM practitioners’ experience in the treatment of chronic bronchitis: analysis of medicinals based on latent structure*

Chen Liping1, Li Jiansheng2#,Cai Yongmin3, Bian Hua1   

  1. 1 Nanyang Institute of Technology, Henan Key Laboratory of Zhang Zhongjing Formulae and Herbs for Immuoregulation, Henan 473004, China;
    2 Respiratory disease diagnosis and treatment and new drug research and development synergy innovation center of Henan Province, Henan University of Traditional Chinese Medicine, Henan 450046, China;
    3 School of Basic Medicine, Nanjing University of Chinese Medicine, Jiangsu 210046, China
  • Received:2018-02-03 Online:2018-08-30 Published:2018-08-30
  • Supported by:
    Young Scientists Fund of National Science Foundation of China(No.81704200),Science and Technology Tackling Project of Henan Province(No.172102310422),Postdoctoral Scientific Research Project of Henan Province(No.00104257),Traditional Chinese Medicine Research Initiative of Henan Province(No.2015ZY02003)

摘要: 目的 基于隐结构模型探讨名老中医辨治京津冀地区慢性支气管炎的病因病机、证候分布及用药规律。方法 提取已构建的现代名老中医肺病数据库中京津冀地区慢性支气管炎医案346例,先用Lantern 3.1.2软件分别构建慢性支气管炎症状和药物的隐结构模型,再进行隐类诠释;最后用SAS 9.1软件对药-药、药-症不同字段进行关联规则挖掘。结果 对病案的110个显变量(症、舌、脉)建模,得到17个隐变量,35个隐类,6个综合聚类模型;对137味药物建模得到27个隐变量,55个隐类,7个综合聚类模型;对药-症进行关联规则挖掘,并设定最小的支持度5%、置信度60%,得到二联项74组,三联项273组,四联项411组。京津冀地区慢性支气管炎患者的体质特点为“上燥中湿下虚”,常见痰热蕴肺证、肺阴虚证、风寒犯肺证,核心症状为咳嗽、气喘、咯痰、痰白,核心药物为半夏、苦杏仁、麻黄,核心方剂为三子养亲汤、二陈汤和小青龙汤。结论 隐结构模型综合应用多维聚类和贝叶斯网络的模式,能解决中医辨证所面临的缺乏客观、定量的问题,且符合中医的非线性思维,可用于挖掘名老中医的学术经验。

关键词: 隐结构模型, 京津冀地区, 慢性支气管炎, 名老中医, 医案, 数据挖掘

Abstract: Objective To explore causes, mechanisms, symptoms, and commonly-used medicinals for chronic bronchitis from the viewpoint of senior prestigious traditional Chinese medicine practitioner in Beijing-Tianjin-Hebei Region. Methods 346 chronic bronchitis cases in Beijing-Tianjin-Hebei region were obtained from the database of lung disease cases treated by modern senior TCM practitioners. The Lantern software 3.1.2 was used to construct the latent model of chronic bronchitis symptoms and commonly-used medicinals. The association of medicinal-symptom, and medicinal combination was explored by using SAS software 9.1. Results 110 observable variables (pattern, tongue and pulse) in first contact of the 346 cases were collected and constructed as a latent model, and 17 latent variables, 35 hidden classes and 6 comprehensive clustering models were obtained. 27 latent variables, 55 hidden classes and 7 comprehensive cluster models were obtained from 137 medicinal models. Associations between medication and disease were identified at 5% minimum support degree and 60% confidence level, and 74 groups of 2 medicinals, 273 groups of 3 medicinals, and 411 groups of 4 medicinals were obtained. Constitutions of chronic bronchitis patients from the Beijing-Tianjin-Hebei region were featured by“upper dryness, middle dampness and lower deficiency”. Common patterns of chronic bronchitis include phlegm-heat pattern, lung yin deficiency pattern, and wind-heat invading the lung pattern; Core symptoms of chronic bronchitis involved cough, panting, phlegm, and white sputum; Core medicinals includes pinellis rhizome, bitter apricot kernel (Kuzingren, Semen Armeniacae Amarum) and ephedra (Mahuang, Herba Ephedrae); Core formula includes Sanzi Yangqi Tang (Three-Seed Filial Devotion Decoction), Er Chen Tang (Two Matured Substancess Decoction), and Xiao Qing Long Tang (Minor Green Dragon Decoction). Conclusion Implicit structure model integrated application of multidimensional clustering and Bayesi network It can solve the lack of Objective and quantitative preblems in TCM syndrome differentiation, and it conforms to the nonlinear thinking of TCM. therefore, it can be used to explore the academic experience of famous old Chinese medicine.

Key words: Latent structure model, Beijing-Tianjin-Hebei region, chronic bronchitis, senior prestigious TCM practitioner, medical records, data mining

中图分类号: 

  • R256.1