Data MATRIX on the SDRG data analysis guide for the FDA

Data MATRIX on the SDRG data analysis guide for the FDA

The symposium gathered experts who discussed various subjects on data management and biostatistics, pharmacometrics, particularities of adaptive design and data collection. The presenters also covered the main differences between the frequentists’ and biologists’ viewpoints, but eventually agreed on the fact, that the bias based approach is more fundamental for clinical trials allowing not only to cut study timelines and costs but also use all the data obtained throughout the study more effectively.

Andrey Myslyvets, Data MATRIX biostatistician, raised the topic of the need for standardization of basic concepts and bringing the glossary to a standardized version. This is the foundation by which the Russian professionals will be able to communicate with the world leading experts about complex statistical methods of data processing using the same language. The SDRG (Study Data Reviewer's Guide) in the Study Data Tabulation Model was taken as a specific example - a necessary document for submission to the FDA, which facilitates the work of the reviewer from the standpoint of regulatory authorities, and therefore speeds up and simplifies the overall procedure.

Andrey spoke about the steps to be taken to compile a complete FDA package of documents and demonstrated the most efficient algorithm for collecting, storing and preparing of documentation. Data collection in SDTM format is the first important step for the clinical trials report, including not only datasets but also related documentation (aCRF, SDTM spec, SDRG and define.xml). Part of his speech, Andrei dedicated to the structure of the SDRG based on real cases encountered in Data MATRIX.

“Unification and standardization of fundamental concepts is important in order to keep up with the current trends and communicate with the world industry leaders in the same language. In addition, it will reduce the time for developers and CROs to correct errors, since most of them can be avoided at the of the protocol design, CRFs and the Data of Management Plan stages,” commented Andrey Myslivets.