Other Specialized Research Resources

Data Coordination Unit (DCU) (rev date 09/08)
http://dcu.musc.edu
Director: Yuko Palesch, PhD, Associate Director: Valerie Durkalski, PhD

The Data Coordination Unit (DCU) serves as the statistical and data management center for a variety of multicenter center clinical trials and clinical research studies, primarily, although not exclusively, funded by the NIH. It is a unit in the Department of Biostatistics, Bioinformatics and Epidemiology (DB2E). The current director is Dr. Yuko Palesch, Professor of Biostatistics in DB2E, and the Associate Director is Dr. Valerie Durkalski, Associate Professor of Biostatistics in DB2E.
Since its inception in May 2004, the DCU has managed many clinical studies, most involving multiple centers. Data and project management services are conducted using the DCU's user-friendly web-based clinical trials management system, WebDCUä. This system provides all the required tools for site coordination and data management in one efficient and easy to use system. Services include study design/protocol development, central registration and randomization, data management, project management (e.g., tracking patient screening logs, subject study progress/calendar, automated MedWatch forms, on-line training/certification, regulatory document collection and monitoring, study drug kit inventory and shipment tracking), biostatistical support, DSMB interface, report generation, and publications. The WebDCUä system facilitates research by maximizing the study group's productivity and efficiency, including intervention implementation across various types of centers (rural/urban private practices, community health centers and academic centers), telemedicine, collaboration with the participating centers on an electronic data capture and management process, and secure and real-time internet access to collected data. The system is efficient and reliable, and has proven to reduce the administrative responsibility at the participating centers so that more time can be spent on important scientific aspects of the studies.
The DCU maintains a full set of Standard Operating Procedures (SOPs) and staff training is updated to ensure all SOPs are followed in practice. Following the FDA guidance and DCU SOPs, the WebDCUä system is designed with the following common features as needed for this study:

• Real-time access to database information on the Internet;
• Data validation for all data entry and editing activities;
• Full data auditing, included dated tracking of audits, for all data modification and deletion;
• Subject visit/assessment schedule calendars;
• Version control – allows online access to latest version of the protocol, CRFs and other project related documents;
• Specimen and laboratory tracking – point of collection to shipping and processing;
• Registration and randomization (including drug shipments to central pharmacy);
• Fully configurable data retrieval and printer friendly reports;
• Secure Sockets Layer (SSL) encryption;
• Individual usernames and password required for secure website access;
• User group based privilege management;
• Online training on how to use the system;
• Online help message;
• Internet Explorer and Netscape web browser compatibility.

In addition to several DB2E faculty members, the DCU currently employs approximately 12 full-time staff (Database Managers, Information Technology Support, Database Programmers Project Managers, and Administrative Support). All staff activities are conducted in accordance with the DCU SOPs and Process Guidelines which are maintained and reviewed on a periodic basis.  Good Clinical Practice guidelines are followed and all staff members are certified through the Miami CITI Web-Based Training for the Ethical Conduct of Human Subjects in Research.

Computer Hardware/Software
The DB2E has a Local Area Network (LAN) with a dedicated manager, encompassing personal computers, file servers, printer servers, and other shared peripherals. Individual computers are managed centrally for software updates and network security. In addition, DB2E has in-house resources for high performance computing, including 15 dual processor servers with 32bit 3.1GHz processors, each with 4 Gb RAM processors. In addition, an 8-blade server is in the early stages of assembly that will increase the number of high performance units to 48. Finally, a grid computing initiative has been started in order to make effective use of this computing infrastructure available to non-mathematicians. The scientific software licensed to the department includes MATLAB, Splus, MathCAD, SAS, Statistica, NQuery, and STATA. The department has full-time staff responsible for the procurement, installation, operation, and basic maintenance of hardware and software.
The DCU has maintained fully networked and dedicated computing resources to provide a high performance and mission-critical system. The servers at DCU are configured to have RAID 1/5 hard drive arrays, redundant components, alternative power sources and multiple backup systems under Windows 2000/2003. This ensures the clinical trial data management system is running at all times. The database and web system architecture is scalable and system resources are measured against usage to ensure top performance.
The DCU also employs several layers of hardware firewalls in cooperation with MUSC to ensure system security. MUSC maintains firewall protection between the university and outside computer systems. The DCU maintains an additional firewall on web servers because access is not limited to within MUSC. In addition to the firewall protection, two levels of software security are in place at DCU. The first is antiviral protection: McAfee VirusScan Enterprise Edition is being used to protect all servers and workstations from infection.  Virus definitions are updated on a daily basis. The second component is password protection. MS SQL, Oracle, ClinTrial, Access, and Windows all include password protection features to prevent unauthorized access. These are all activated and kept fully functional by the DCU.
All DCU web-based clinical trial management systems adopt Secure Sockets Layer (SSL) protocol to enable encrypted, authenticated communications across the Internet and require the user to log-in at the beginning and log-out at the completion of the system session. The user authentication process is based on the combined identification of a user ID and password. The user is required to change their password periodically or their account will expire. When an internet user tries to log-in to the WebDCU™ system, the user’s browser type, IP address, log date and time will be collected for security reasons. If any suspicious log-in attempt is detected, the system begins to trace the possibly malicious attack, immediately refuses any more system access, and notifies the incident to an IS Manager or designee.

The transfer of clinical trial data from a study database (MS SQL or Oracle) begins with downloading the data into common portable data formats. These common format data files serve as the source data for all reporting, analyses, merges and transformations. The eXtensible Markup Language (XML) dataset format has been adopted as a default electronic data transfer format for WebDCUTM among common portable formats based on XML’s globally supported, platform-independent and machine readable data characteristics, which enable data exchanges between different computerized systems in the medical research area. The permanent SAS dataset format can be used, when the study sponsor or the data requestor specifies the SAS data format.

Data Security
Rapidly changing computer and communications technology requires that security issues be assessed on a regular basis and modified as indicated. It is important that all personnel involved fully understand the importance of handling data in a proper manner. Some specific measures to be implemented to assure data integrity include the following:

• All changes to data files made will be recorded utilizing an audit trail.
• Microsoft SQL Server provides multi-level password and security systems. A limited auditing trail of certain events relating to security, authentication procedure for log-in password, and access confirmation to the database are included in the system. As another precaution, passwords are immediately changed and access restricted for any employee who is terminated.
• During the course of database’s existence, files with study outcomes will be restricted to core staff. The back-up copies of the database will also be maintained in locked files. There will be password protection for the network, further assuring that confidentiality will be maintained.
• Although every reasonable precaution will be taken, the safeguards cannot guarantee a virus free environment. However, the extensive long-term back up system should allow recovery of files in the unlikely event that files are destroyed by a virus, if files were created before the infection.
• In addition to the use of passwords and other security measures, all documents containing identifying information on individuals or physicians are considered confidential materials and will be safeguarded to the greatest possible extent. No information which identifies a specific person, hospital, or physician will be released to or discussed with anyone other than study staff members. All employees will be informed of this policy before being hired and required to sign a pledge to maintain the confidentiality of all information. In addition, a set of guidelines will be developed for the staff dealing with the specific office and data processing procedures to assure the confidentiality and security of all data.
• The database will be stored on the file server of DCU using a separate file system.

Back-up Schedule and Archiving

The data backup schedule consists of full-verified backups performed Friday’s and the last Saturday of every month; differential verified backups are performed Monday - Thursday.  “Verify” is an option selected in the tape backup setup menu, which ensures that data has been copied accurately onto the tape. The full backup consists of recording a new copy of all operating system (OS), application software, and data contained on the server onto a digital tape cassette. The differential backups consist of recording all data (including OS and application files) that has changed since the last full backup onto a digital tape cassette.  As an example, on Monday all data (OS and application inclusive) that has changed that day or over the weekend is backed up.  On Tuesday, all data that changed that day or Monday or over the weekend is backed up.  This pattern continues until Friday, when a full backup is performed, resulting in a complete new copy of all OS, application, and data files.  Under this schema, a system failure on Thursday which requires a complete rebuild of the hard drive would be remedied by an initial reinstallation of Windows Server 2003, followed by a restore of the full backup from the previous Friday’s tape, and completed with a sequential restore of the differential data from Monday -Wednesday’s tapes. The DCU SOPs address prevention plans to be followed in case of weather emergencies.

<back to outline>