Research On Clinical Data Mining Of Insomnia Medication Based On Complex Network And Association Rules

Objective:This study aims to investigate the commonly used drug pairs by Professior GuoRongjuan from DongFang Hospital and explore new prescriptions to treat insomnia through analyzing some clinical data concerning to rules of prescription, drug properties, drug pairs, etc.Methods:The basic data about the patients treated by Profrssor Guo was obtained statistically from the HIS database of DongFang Hospital. The characteristics such as population distribution, drug preference and so on were summarized. Some information such as drug frequency, four properties, five tastes, channel tropism was analyzed by Traditional Chinese Medicine Inheritance support system. The correlation, combination and administration of drugs were studied with entropy clustering algorithm based on the analysis of commonly used drug pairs and their association.Results:Among 3084 patients with insomnia treated by Professor Guo, we found that the distribution of suffering was normal.83.4% chose TCM therapy.68.16% was female. The ages of victims in both male and female peaked during 50-59. By analyzing their initial prescriptions, we found that,59 herbs, prescribed more than 100 times among 207 herbs used by Professor Guo to deal with insomnia, were dominantly warm, cool and neutral on the one hand. And majority of them were sweet, acrid and bitter on the other hand. The top ten applied in the prescriptions were Semen Ziziphi Spinosae(1089), Radix Rhemannia Praeparata (1049), Radix et Rhizoma seu Caulis Acanthopanacis senticosi (1034), Spica Prunellae (929), Cornu Saigae Tataricae powder (908), Margarita powder (845), Pericarpium Citri Reticulatae (754), Herba Epimedii (732), Radix Curcumae (718), Radix Astragali (644) respectively. The dosage of the first three herbs was as follows:120g (Semen Ziziphi Spinosae),15g (Radix Rhemannia Praeparata),30g (Radix et Rhizoma seu Caulis Acanthopanacis senticosi). The highest-ranked channels affected by the prescriptions were liver, kidney, spleen, heart, lung respectively. According to the association rules,21 drug pairs involving 12 different herbs were commonly used by Professor Guo. By means of entropy clustering algorithm,41 drug pairs involving three or four different herbs were found with high frequency of application and thus we explored 7 new prescriptions.Conclusion:This paper investigated and analyzed the disciplines between signs and therapies performed by Professor Guo to treat insomnia through one dimensional frequency diagram and two dimensional network diagram, making the data about prescriptions more objective and clarified. The drug pairs commonly used by Professor Guo and the new prescriptions had the effect on reinforcing the kidney, spleen as well. The findings about the new prescriptions and drug pairs may provide the reference for fundamental research and clinical practice. However, the results still need to be evaluated by the clinical practitioners and further experimental research. And also the data mining employed in this study could be considered by clinical data mining of other diseases and data technicians engaging in the clinical research.