A study on the recent 20 years’ herbal medication rules and syndromes distribution in TCM treatment of Diabetic Nephropathy by Data mining

Aim: To study the recent 20 years’ herbal medication rules and syndromes distribution in TCM treatment of diabetic nephropathy based on modern literature. Methods: We used SPSS18.0 and SPSS Modeler14.1 for data analysis. The descriptive analysis and cluster analysis were performed by SPSS 18.0; the association rules analysis with the Apriori algorithm was performed by SPSS Modeler14.1. Results: 274 TCM formulas were obtained after screening. The 5 most frequently used Chinese herbal medicines are milkvetch root, Indian bread, common yam rhizome, danshen root, Asiatic cornelian cherry fruit; The 5 most common syndromes of DN are deficiency Qi and Yin, obstruction of collaterals by blood stasis, deficiency of liver and kidney, damp-heat stasis obstruction, Yin asthenia generating intrinsic heat; the 5 most common pathological factors are Yin deficiency, blood stasis, Qi deficiency, endogenous heat and phlegm. In association rules analysis, we displayed association rules among herbs of confidence greater than 85% and support greater than 15%. By clustering, we obtained 12 sets of core couplet herbs. Conclusion: The syndrome differentiation of DN is not fixed but dynamic development progress. The syndromes and prescription regularities in the current 20 years are systematically summarized by data mining, which provides data support for clinical differentiation and further researches.