Enhanced CBET To Optimize Sample Size And Precision Analysis

With the increasing difficulty of the development of new drugs, bioequivalence (BE) studies in the research and development area of new drug are attached more and more emphasis by pharmaceutical companies globally. This situation promotes the increasing number of BE studies and bioavailability (BA) studies to be initiated. CBET (Clinical Biostatistics Executable Tool) is a web-based statistical calculation tool in pharmaceutical company "A", which is mostly used to calculate sample size in BE studies and precision in BA studies. It plays a very important role in the designing stage of these BA and BE studies. However, even though CBET uses SAS, a famous and well recognized statistical calculation software, as its calculation tool in the background, it still has several shortcomings, such as possible mistakes can be made when related parameters were manually input; lacking of flexibility of some outputs; unable to generate statistical graphs and tables.The feasibility of using visualized interface with SAS at background to do statistical analysis and calculation has been verified. In the designing stage of clinical studies, utilizing SAS to accomplish some calculations and analysis and making use of its property of user friendliness and high accuracy, visualized interface can not only effectively avoid mistakes made in the process of manual input and misuse of different results, but also increase the proportion to use the right study design and assure the quality of the clinical study.Our research uses Microsoft Visual Studio as the interface designing environment and successfully worked out a new application --- Advanced CBET. It uses interfaces between software to transfer data with SAS. It realized the function of calculating sample size and generating graphs in BE studies and calculating precision and generating graphs and estimation tables in BA studies. For one thing, all these functions make up the shortcomings in the old CBET. For another, it added more functions that can promote calculation efficiency and accuracy in the designing stage of BA and BE studies.Our software contains total 29 interfaces, in which 9 functions involves statistical calculation, analysis, graph and table output and the rest of them are subsidiary interfaces for the software. This application was validated by 15 results from protocol of really studies and the accuracy is 100%, with no obvious calculation incorrectness. The validation process will continue because this application will be put into real work. We hope that this application can play its role in helping clinical statisticians to reduce calculation mistakes and improve work efficiency and after all, assure the quality of clinical studies.