Preliminary Study On Visualization Of Risk Management In Clinical Trials

Objective: The monitoring plays an important role in clinical trials.However,the traditional on-site monitoring is often accompanied by a large amount of human and material consumption,resulting in great pressure on clinical trials.With the development of computer softwares and network technologies,the Risk Based Monitoring(RBM)comes into being.So far,the Food and Drug Administration(FDA)in the United States and the European Medicines Agency(EMA)have both issued guidelines that allowed the risk monitoring to be used in clinical trials.Trans Celerate,a non-profit organisation founded by more than a dozen of the world's leading drug companies,is now actively exploring new models for the clinical trial monitoring.Different from the traditional monitoring method,the risk-based monitoring is to identify the risks that affect the quality of clinical trials and patients' rights,and to carry out risk assessment on the most important links most likely to go wrong,so as to conduct effective monitoring and avoid waste of resources.The Key Risk Indicators(KRIs)Monitoring,Central Statistical Monitoring(CSM)and data fraud monitoring are key technologies to implement RBM,and they have been successfully applied in foreign clinical trials' monitoring.However,domestic literatures about systematically introduce the KRIs,CSM and data fraud monitoring are relatively lack.Under the background we use SAS,R and Excel these three familiar and conveniently acquired softwares to realize visualization of the KRIs,CSM and data fraud monitoring,so as to provide methodological support for application of new concepts and new technologies in our country such as the risk monitoring and centralized statistical monitoring.This research is of practical significance to integrate with the foreign clinical trial monitoring technologies and improve the level of the clinical trial monitoring in China.Methods: For the requirements of the RBM,the key risk indicators are selected to monitor the key processes and key data in clinical trials.The centralized statistical monitoring method is adopted to investigate the quality of clinical trial data,detect abnormal data and detect data fraud.In order to satisfy the application of RBM visualization in different clinical trial personnels and to consider the accessibility of visualization softwares in practical work,the visualization of RBM is realized by writing SAS macro programs,R macro programs or using Excel graphics function directly.SAS and R graphical codes are packaged into macro programs for practical use.Results: Taking a certain vaccine phase III trial and a medical instruments' non-inferiority clinical trial as examples,using SAS,R and Excel as mapping softwares,we complete the key risk factors,CSM,data fraud these three types of graphs,write the corresponding macro programs,and preliminarily realize the visualization of RBM according to the principle of the RBM.KRIs graphs include: multicenter overall risk indicator distribution,overall risk indicator decision,multicenter missing informed consent distribution,missing informed consent decision,multicenter average deviations per randomized subject distribution,average deviations per randomized subject decision,multicenter percent deaths of randomized subject distribution,percent deaths of randomized subject decision.CSM graphs include:subject profile,polyline diagram of temperature comparison of each center,polyline diagram of temperature follow-up for each subject.The data fraud graphs include: mosaic of multicenter enrollment time distribution,bar chart of the enrollment time for subjects,scatter plot of antibody titer follow-up for each subject,profile of the comparison of subjects,first digit desicion,and clustering diagram of subjects.For each type of graphs,most of them are given the SAS macro programs,R macro programs and Excel graphical function at the same time.Conclusion: By combining RBM with visualization technologies and using multiple graphs,we can conduct comprehensive verification for data from different angles,focus verification of suspicious data and guide on-site monitoring activities,thus improving the quality and efficiency of the clinical monitoring.