Study On The Comprehensive Evaluation And Administration Optimization Strategy Of Meropenem Original And Generic Drugs

Objective: To evaluate the safety,effectiveness and economics of meropenem original and generic drugs,and to apply machine learning methods to establish efficacy prediction models and to provide technical tools for optimizing drug delivery strategies.Methods: First,Bayesian network Meta-analysis was used to compare the effectiveness and safety of Meropenem generic drug(trade name: Beineng)and generic drug(trade name: Mepem)in the treatment of infectious diseases.Second,diagnosis and treatment data of adult patients using Mepem and Beineng in a large tertiary hospital was retrospectively collected to compare the effectiveness and safety of Mepem and Beineng in a real world situation using propensity score matching.Third,based on the results of effectiveness and safety comparison in the real world study,cost-effectiveness analysis is conducted to compare the economic of Mepem and Beineng.Finally,efficacy prediction models based on multiple machine learning algorithms for Mepem and Beineng were established,and the predictive performances of the models were evaluated.Results: First,no statistically significant difference between Beineng and Mepem in terms of clinical effective rate and bacterial clearance rate in the treatment of infectious diseases was observed through network Meta-analysis.In terms of the incidence of adverse reactions,there was no significant difference between Beineng and Mepem.There was also no statistically significant difference in the incidence of abnormal digestive tract function and rash between the two drugs.Second,real-world study revealed no statistically significant difference between the Beineng group and the Mepem group in terms of clinical effective rate,degree of reduction in CRP,PCT,WBC,NEUT%,and lung imaging efficacy.The microbial clearance rate of the Mepem group was higher than that of the Beineng group,and there was a statistically significant difference between the two groups(P<0.001).Stratified analysis of the types of infection strains showed that the clearance rate of Escherichia coli in the Mepem group was remarkably higher than that in the Beineng group,with a statistically significant difference(P=0.002).There was no significant difference in the clearance rate of the other strains between the two groups(P>0.05).And no statistically significant difference was observed in the incidence of adverse reactions,abnormal incidence of Cr,ALT,AST,TBi L,Hb,PLT,eosinophil count between the two groups.In terms of economy,the cost of the Beineng group was99752.22 yuan,while the cost of the Mepem group was 122631.77 yuan.The Beineng group had the lowest cost-effectiveness ratio(1338.96),while the Mepem group had the highest cost-effectiveness ratio(1592.62).Third,the Beineng efficacy prediction model based on KNN infill method,elastic network for feature screening combined with random forest algorithm had the best performance(AUC=0.663,AUPRC=0.855);the Mepem efficacy prediction model based on random forest infill method,elastic network for feature screening combined with Bernoulli naive bayes algorithm had the best performance(AUC=0.781,AUPRC=0.914).Conclusion: Beineng is similar to Mepem in terms of safety and effectiveness,but the economic advantage of Beineng is more obvious.The clinical efficacy of both drugs can be predicted via machine learning methods,and this study can provide a reference for the development of intelligent aids for clinical drug selection.