Font Size: a A A

Real World Study Of Chinese Patent Medicine

Posted on:2023-03-30Degree:MasterType:Thesis
Country:ChinaCandidate:L X LiFull Text:PDF
GTID:2544306614998429Subject:Integrative basis
Abstract/Summary:PDF Full Text Request
1 BackgroundThe progress of big data and computer science technology has promoted the development of real-world study,in the past few years.Under the background of the combination of artificial intelligence and medical big data,machine learning has been widely used in disease diagnosis,theraphy and prediction of prognosis,which promotes the progress of medical treatment.This makes it possible for real-world study to find the applicable population,disease and usage of Chinese medicine,and even find the appropriate treatment regimens that can achieve good clinical efficacy based on the above three factors.Pneumonia is the most common cause of death among children under 5 years old over the globle.Xiyanping injection is extracted from andrographis paniculata and refined by modern technology.It has the effects on clearing away heat and toxic material,relieving cough and dysentery.A large amount of evidence-based medical evidence supports its use in treating pneumonia.Based on real-world data,this study realized the recommendation of medication regimens,by the case of Xiyanping injection in the therapy of pediatric pneumonia with different disease condition in the help of machine learning technology.2 ObjectiveWe analyzed the combined medication mode by real world data,and compile the medication scheme recommendation software to recommend the best medication scheme for patients on the premise of clarifying the efficacy of Xiyanping injection in the treatment of pneumonia.3 Methods and Results3.1 The evaluation of evidence based medicine of Xiyanping injection in the treatment of childhood pneumoniaWe comprehensively searched the published systematic evaluation and meta-analysis of Xiyanping injection in the treatment of childhood pneumonia,and a total of 12 studies were included.PRISMA 2020 and AMSTAR2 were used to screen the studies with high report quality and methodological quality,GRADE was used to evaluate the evidence quality.A total of 4 outcome indicators were evaluated of bronchopneumonia,including fever relief time and so on.For the level of evidence,we got B for 6 outcome indicators,C for 12 outcome indicators,and D for 1 outcome indicators.A total of 5 outcome indicators were evaluated of mycoplasma pneumonia.For the level of evidence,we got B for 1 outcome indicators,C for 4 outcome indicators.A total of 7 outcome indicators were evaluated of viral pneumonia.For the level of evidence,we got B for 1 outcome indicators,C for 6 outcome indicators.3.2 Research for the medication therapeutic regimens analysis of Xiyanping injection in the treatment of childhood pneumonia based on real world dataThe data was from the real world electronic health data warehouse established by institute of basic Research in Clinical Medicine,China Academy of Chinese Medical Sciences.The cases were extracted of Xiyanping injection in the treatment of childhood pneumonia,and were coded as well as standardized by international general standards.After determing the high frequency drugs needed for cluster analysis,SPSS 25 software was used for hierarchical cluster analysis.There were 8902 patients were included in the study with an average age of(5.46±2.82)years old.We found that the most common type of pneumonia was bronchopneumonia,and the most common complication was myocardial damage,The most commonly used combination drug was ambroxol.The most commonly used antibiotic was the third generation cephalosporin,followed by azithromycin.The most commonly used antiviral drug was ribavirin.We finally obtained 10 core medication regimens by hierarchical clustering,of which 5 prescriptions were typical.The medication regimens 1 were Xiyanping injection+Ambroxol+budesonide selectivity β2 adrenergic receptor agonists+montelukast.The medication regimen 2 were Xiyanping injection+vitamins+the third generation cephalosporins+glucocorticoids+compound amino acids.The medication regimen 3 were Xiyanping injection+azithromycin+Chinese medicine clearing the heat+the first generation cephalosporins+adenosine arabinoside.The medication regimen 4 were penicillins+ipratropium bromide+xanthines+paracetamol+ephedrine.The medication regimen 5 were immunostimulant+ibuprofen+furosemide+immunoglobulin+adrenaline.3.3 Research for recommendation of the medication regimens of Xiyanping injection in the therapy of childhood pneumonia realized by the technology of machine learningWe screened the characteristics affecting clinical effectiveness,and finally chose total 13 characteristic variables,included gender,age,type of pneumonia and complications(myocardial damage,upper respiratory tract infection,bronchitis,congenital heart disease,mycoplasma infection,liver damage,asthma,fever and diarrhea)as well as which core medication regimens patient taken,based on the real-world data of 8902 patients.With the outcome of whether the patient was cured as the target variable,the Naive Bayes algorithm was conducted to construct the clinical efficacy prediction model.5533 patients were included after the screening of real-world data that met the requirements of characteristic variables and target variables.We divided the data that the training set to test set was roughly equivalent to 7 to 3.The accuracy of the training set of the prediction model was 73.76%,the test set was 73.63%,and the area under the AUC curve was 0.60[95%CI(0.57,0.61)].Python programming language was used to compile the medication regimens recommendation software based on the efficacy prediction model.The user input the information of a patient’s gender,age,type of pneumonia and complications,and the software automatically matches all core medication regimens to predict the outcome.Then,the medication regimens were recommended which the predicted clinical outcome was cure.4 Conclusions(1)We evaluated the report quality,methodology quality and efficacy evidence level of systematic evaluation and meta-analysis of Xiyanping injection in the treatment of childhood pneumonia.The result showed that the clinical efficacy of Xiyanping injection was definite in treating children with pneumonia,and the efficacy varied according to different populations,disease subtypes and medication regimens.(2)Cluster analysis was carried out to find five core medication schemes on the real world data of Xiyanping injection in the treatment of children’s pneumonia.The prediction model was constructed by using naive Bayesian algorithm in machine learning,and the recommendation software was realized which could recommend the best medication regimens.Our research had good generalization and certain reference significance for deepening the rational use of Chinese patent medicine and improving clinical drug application level.
Keywords/Search Tags:Chinese patent medicine, really world study, evidence evaluation, cluster analysis, machine learning, individualized therapeutic schemes recommendation
PDF Full Text Request
Related items