The Biostatistics and Data Management Core (BDMC) provides a central resource to the ADCS consortium for biostatistical expertise in the design, management and analysis of Alzheimer's disease clinical trials and state-of-the-art data management of these trials. The BDMC further oversees a comprehensive quality assurance program covering the spectrum of ADCS activities. The biostatistical responsibilities of the BDMC include: providing input into the experimental design of each protocol, with emphasis on the power analysis and the assessment of reliability and validity of measurement instruments; developing and generating the randomization sequences for each trial; performing preliminary and final statistical analyses on trial results; assisting in the preparation of reports and publications; developing new statistical methodology, as needed, and serving as a general biostatistical resource for investigators in the consortium. All ADCS clinical trial data is managed through the BDMC. Every aspect of the data collection process, including forms design and printing, maintaining centralized patient files, and generate special purpose and routine statistical reports on accumulating data, is coordinated through the data management operation of the BDMC. The BDMC data environment is sophisticated. Approximately 15 PC workstations are connected in a client-server architecture, using a Novell Netware local area network (LAN). Data management is conducted through the ORACLE relational database management system. The Quality Assurance program of the BDMC is extensive. At its center is the clinical and computerized monitoring of the data, but it extends to cover training, document (training manuals, manuals of operation, newsletters, etc.) and instrument development, and archiving of regulatory documents. The BDMC coordinates a clinical monitoring program through which all ADCS data forms are reviewed on-site by highly trained clinical monitors prior to inclusion in the centralized database. Computerized edit checking is done both at the time of entry, and longitudinally. A systematic audit of the database (proposed) will compare physical case report forms with their electronic images.
Showing the most recent 10 out of 294 publications
