Causes Of Adverse Reactions Based On Gray Theory And The Associated Rules Of Traditional Chinese Medicine Injection Study

Object:This dissertation brings in Gray System Theory and Data Mining, study the application mode of the Gray Correlation Analysis and Association Rules in Chinese Medicine Injection (CMI) ADR causes analysis in the clinical application process. Through case analysis, aimed to quantitatively and intuitively reflect the correlation between CMI ADR factors and ADR in clinical application process, and select the key ADR factors and key main elements, then mining association rules which can reflect CMI ADR occurrence and association among ADR factors. This dissertation aims to provide new idea to existing CMI ADR causes analysis, and provide meaningful references for objectively understanding CMI ADR and rational clinical medication.Methods:(1) Documentary Research:fully consult published CMI ADR causes literature and books in the last 10 years, so that to grasp the status of CMI ADR causes analysis, and classify them according to the methods they use, then summarize their advantages and disadvantages. (2) Gray Correlation Analysis:study the application mode of Gray Correlation Analysis in CMI ADR causes analysis in the clinical application process. Use it to quantitatively reflect the correlation between the ADR factors and ADR, according to the correlation sort, select the key factors and key main elements. (3) Data Mining:study the application mode of data mining in CMI ADR causes analysis in clinical application process. Use it to mining rules which can intuitively reflect CMI ADR occurrence and the association among the ADR factors. (4) Case study:fully consult the published CMI ADR case reports, select a representative CMI as the case, of which ADR case reports is more than the majority. Use the application mode of Gray Correlation Analysis and Association Rules Mining in CMI ADR causes analysis in clinical application process, to analyze the data of the right CMI ADR case reports.Results:(1) Documentary Research:divide the existing CMI ADR causes analysis into three categories according to the methods they use, which are experimental research, review research and epidemiological research. Because of the objective condition restriction or methodology restriction, there are still some disadvantages in existing CMI ADR causes analysis, just as not systematic, long-term, lack of large database support, without taking correlation among ADR factors into account, and so on, it is which make the existing CMI ADR causes analysis hard to achieve a major breakthrough. (2) Application mode study of gray correlation analysis and association rules mining in CMI ADR causes analysis in clinical application process:briefly analyze principle of Gray Correlation Analysis and Association Rules Mining, determine the required information from CMI ADR case reports, and briefly describe the problem to be analyzed, besides these, advance Gray Correlation Analysis steps of CMI ADR causes and the key points in Association Rules Mining. (3) Case study:collect a total of 520 published Shuang Huanglian injection ADR case reports, after data cleaning, left 517 effective samples. The results of Gray Correlation Analysis: in the clinical application process of Shuang Huanglian injection, the ADR factors correlation sort is " whether the initial medication, with or without allergic history, liquid type, patient sex, liquid volume, age, liquid injection dose, solid injection dose, original illness, drug combination". The main elements sort in ADR factors is "not initial medication> initial medication, with allergic history> without allergic history, GS> NS> GNS, male> female, liquid injection> solid injection, original respiratory illness> others, combination with antibiotics> combination with non-antibiotics, and in factor "age" the correlation of adolescent is maximum, in factor "liquid volume" the less the correlation larger, in factor "dose" the more the correlation larger. Results of the Association Rules Mining:when the min_sup and min_conf is 10% and 50%, obtain 41 rules which well reflect the Shuang Huanglian ADR occurrence. Then evaluate interestingness of the rules and select them according to the Gray Correlation Analysis and Documentary Research results, with 18 interesting rules left. These rules did clearly and intuitively reflect the association among ADR factors, possible ADR risk combination and ADR risk probability.Conclusion:(1) The application of Gray Correlation Analysis and Association Rules Mining in CMI ADR causes research in clinical application process is significantly effective. On the one hand, effective and interesting results were obtained in case study. On the other hand, the results of Gray Correlation Analysis and Association Rules Mining did well confirm each other, in the association rules the most frequent factors are which with large correlation in Gray Correlation Analysis. And the result of Gray Correlation Analysis was also well used in interestingness evaluation of association rules. Both of the methods have got a good combination. (2) Based on the results of case study, conclusions were obtained after discussion, which are:patient factor determine the initial risk difference and the possible ADR type, medication factor can largely avoid or promote the ADR risk, and most importantly, the difference and instability of the CMI quality are the most fundamental factor which leads to the CMI ADR.