Chen Li-Ping, Cai Yong-Min, Li Jian-Sheng
Zhang Zhongjing College of Chinese Medicine, Nanyang Institute of Technology, Nanyang 473004, China.
Institute of Literature in Chinese Medicine, Nanjing University of Chinese Medicine, Nanjing 210046, China.
Zhongguo Zhong Yao Za Zhi. 2017 Apr;42(8):1609-1616. doi: 10.19540/j.cnki.cjcmm.2017.0052.
To explore the medication rules of famous veteran traditional Chinese medicine (TCM) doctors in treatment of chronic bronchitis, a structured medical record database for famous veteran TCM doctors in modern clinical books was established. First, Lantern 3.1.2(Kongmin light) implicit structure analysis software was used to build an implicit structure model and make an implicit interpretation. Then, SAS 9.1 was adopted to mine herb-herb, herb-symptom and herb-syndrome association rules. Through the mining, 1 274 commonly used herbs for chronic bronchitis were found, including liquorice, bitter almond, pinellia, dried tangerine or orange peel, poria cocos. The medicine types included antiasthmatic medicine, qi-tonifying medicine, and heat-phlegm removing medicine. The medicine tastes included sweet, pungent and bitter. The meridian distributions included lung, spleen and stomach channels. The famous commonly used prescriptions included Xiaoqinglong decoction, Maxing Shigan decoction and Erchen decoction. The 147-herb implicit structure model for the first diagnosis was built to get 44 hidden variables, 88 hidden classes, 7 comprehensive clustering models, 9 dual herb associations, 50 triple herb associations and 89 quadruple associations. Totally 297 medical records for the second diagnosis were compared to obtain 24 herb-symptom associations, which reduced ephedra, bitter almond, pinellia and added poria cocos, atractylodes, dangshen, 20 dual herb associations and 8 triple herb associations. A model for the top 83 symptoms and top 96 herbs in the first diagnosis was built to get 50 hidden variables and 101 hidden classes. The commonly used herbs of famous veteran TCM doctors included bitter almond, pinellia, dried tangerine or orange peel, poria cocos, which feature mild property and sweet favor and enter lung, spleen, kidney meridians; the commonly herbal pairs included atractylodes and poria cocos, asarum, fructus schisandrae and pinellia, as well as poria cocos, dangshen, pinellia and dried tangerine or orange peel. The herb-syndrome associations involved ephedra, bitter almond, cassia twig and scanty sputum, pinellia and external cold and internal fluid syndrome, turbid phlegm obstructing lung syndrome and spleen deficiency syndrome, bitter almond and turbid phlegm obstructing lung syndrome, phlegm-heat accumulated in the lung and wind-cold invading the lung. The implicit structure model can be used to quantify tacit knowledge of TCM, extend to the herb and syndrome level for the first time, and quantify the major, minor and compatible statues of different variables in the same data. This study concluded that the medication rules of famous veteran TCM doctors in treatment of chronic bronchitis have certain reference value for the clinical diagnosis and treatment. The data sources, analysis methods and mining results were relatively reliable and objective, and can provide a reference for rules of other disease treatment based on syndrome differentiation.
为探讨名老中医治疗慢性支气管炎的用药规律,建立了现代临床医案中名老中医结构化医案数据库。首先,运用兰灯3.1.2(孔明灯)隐性结构分析软件构建隐性结构模型并进行隐性诠释。然后,采用SAS 9.1挖掘药-药、药-症、药-证关联规则。通过挖掘,发现治疗慢性支气管炎常用中药1274味,包括甘草、苦杏仁、半夏、陈皮、茯苓等。药物类别有平喘药、补气药、清热化痰药等。药味包括甘、辛、苦。经络分布包括肺经、脾经和胃经。名老中医常用方剂有小青龙汤、麻杏石甘汤、二陈汤等。构建初诊147味中药隐性结构模型,得到44个隐藏变量、88个隐藏类别、7个综合聚类模型、9个药对关联、50个三味药关联和89个四味药关联。对297例复诊医案进行比较,得到24个药-症关联,减少了麻黄、苦杏仁、半夏,增加了茯苓、白术、党参,20个药对关联和8个三味药关联。构建初诊前83个症状和前96味中药模型,得到50个隐藏变量和101个隐藏类别。名老中医常用中药有苦杏仁、半夏、陈皮、茯苓等,其药性平和、味甘,归肺、脾、肾经;常用药对有白术与茯苓、细辛与五味子与半夏、茯苓与党参与半夏与陈皮等。药-证关联涉及麻黄、苦杏仁、桂枝与痰少,半夏与外寒内饮证、痰浊阻肺证、脾虚证,苦杏仁与痰浊阻肺证、痰热壅肺与风寒袭肺等。隐性结构模型可量化中医隐性知识,首次拓展到中药与证候层面,量化同一数据中不同变量的主次及配伍关系。本研究结论为名老中医治疗慢性支气管炎的用药规律对临床诊疗有一定参考价值。数据来源、分析方法及挖掘结果相对可靠、客观,可为其他辨证论治疾病的用药规律提供参考。