Standardization Of Clinical Trial Data From EDC System Based On CDISC Standards

Objective It's inevitable to use electronic data capture(EDC)system and standards of Clinical Data Interchange Standards Consortium(CDISC)in clinical trial data management with the increasing requirements on the data quality from regulatories.CDISC develops a series of standards for trial design,data capture,regulatory submission,data interchange and analysis which cover whole lifecycle of clinical data management.The implementation of CDISC standards can help standardize clinical data and ensure the data quality.Protocol Representation Model(PRM)not only helps to provide standardized reference and guidance for the writing of research scheme,but also evaluate the degree of the standardization of completed protocols,which helps to control and improve the quality of clinical trails.When the external laboratory data is involved in the clinical trial process,it is necessary to carry out a unified transmission format according to the laboratory data model Laboratory Data Model(LAB)to avoid wasteful duplication of effort in programing.In China,the development of electronization and standardization in clinical trial data are still in their infancy.Despite the growing use of EDC systems,there are still many studies using paper-base data collection methods,for example Epidata.These data is quite different from CDISC standards in variable names and data structures.SAS macro programs can be used to modify these unstandard raw data into Clinical Data Acquisition Standards Harmonization(CDASH)standard for further data transformation.The linear method and hybrid method for data extraction between CDISC core standards have been studied up to now,while in some circumstances,there is no need to generate submission data,including data from clinical studies not for registeration.New data mapping path should be added from CDASH to Analysis Data Model(ADa M)to simplify the extraction.Methods A redeveloped EDC system based on the open source EDC software Openclinica is used in data capture of an example study.The protocol information and the real data of this study is what this research is based on.The core information of trial protocol can be extracted according to the study outline concept provided by PRM,and the degree of standardization can be evaluated according to the proportion of matching information.This part of the information can be stored as a separate file for other works.The standardization of laboratory data can be achieved by applying LAB.According to the data fields provided by LAB,the laboratory datasets which external laboratory data merged to,can be generated into bar delimited SAS flat files or hierarchical Extensible Markup Language(XML)files using SAS programs for transmission.The unstandard raw data captured from EDC or other data collection methods can be renamed and reorganized according to the requirements of CDASH domains by using SAS programs and manually filling parameter tables.On the basis of the linear method of our research group,a few of new SAS tool programs are added,and embedded into the original main program to maintain the consistency of CDISC data automation system.Results According to the PRM outline concepts,the relevant information in a rabies vaccine clinical trial can be found and extracted,and be stored as a separate data file for reusing and repurposing across multiple documents,databases and systems form study start-up through reporting and regulatory submissions.The SAS statement can be used to rename the variables according to the LAB field,and directly generate a flat data file separated by delimiters.By analyzing the definition of Schema elements corresponding to the XML document in the LAB model,a template of data fields from each level in the XML file can be written by SAS macro program and thus an effective XML file can be generated with all the corresponding data in the laboratory dataset.The variable names and labels can be modified and the data structure can be reorganized into standard ones by filling parameter tables in SAS macros.The ADa M datasets can be automatically generated from CDASH datasets by using the SAS automation program system with new tool macros embedded in.Discussion With the help of the outline concepts in PRM,it is convenient to evaluate the standardization of protocols and reuse the core protocol information,which improves the efficiency of information reuse and accelerates the data mapping by avoiding looking for it many times from the whole protocol.The flat data file generated according to the LAB model can be read directly by SAS statements,and because of the use of unified data fields,it is clear and easy to understand the variable information.Meanwhile,the XML format file is more suitable to be parsed in other data management systems.According to the provided Schema file,a fixed functional interface can be developed to realize automatic data docking.The CDASH and ADa M datasets can be generated quickly by using SAS macro programs and filling in the Excel parameter tables.The programs used for data standardization can be used flexibly according to the actual needs.The generality of the automatic program used to generate ADa M datasets is improved after adding tool macros.