A54 Window based data management system for clinical trials

69s

Abstracts

and adding an extra element of randomization by using a biased coin. NERl’s system accommodates double blinded studies, multiple treatments, and weighted treatment allocations. The system can adjust for randomized patients who later are found to be ineligible and for patients who need to be randomized by mail. This interactive voice response system is ideal for enrolling patients with acute events occurring at any time and for studies with sites in different time zones. As3 USE OF ORACLE WITH AN INTERACTIVE VOICE RESPONSE SYSTEM (IVR) TO RANDOMIZE PATIENTS IN AN ACUTE CONTROLLED CLINICAL TRIAL Joseph A. McPherson and Todd R. Knuth The Cleveland Clbaic Foundation Cleveland, Ohio The randomization of patients to specific drug and/or device treatments in acute randomized controlled clinical trials has been an expensive component of such clinical research. Such randomization proceduresgenerally involve the staffing of a randomization center on a 7 day a week 24 hour a day basis. These individuals screen patients from clinical centers for enrollment into a particular study and issue the drug or device assignment for a particular patient to the clinical center. The screening process typically identifies the calling clinical center, validates the center’s ability to enroll patients in the study, asks a standard set of questions to assess the patient’s eligibility for the study, and finally, makes the actual randomization assignment. We have developed a system that has integrated an interactive voice response system (IVR) with our Oracle data base server to automate this process in less than 5 minutes in an unattended environment. This allows randomization to occur via touch tone telephone, 7 days a week 24 hours a day, in an unattended mode at significant savings. All integrity rules for validating a clinical center’s ability to participate in a study and to validate a patient’s eligibility for enrollment are maintained in the Oracle data base. Actual randomization assignments are maintained in an Oracle table and are updated with information regarding the clinical center, the patient, and the date and time the randomization assignment occurred.

WINDOW

As4 BASED DATA MANAGEMENT

Chris Boardman,

SYSTEM FOR CLINICAL

TRIALS

Ralph Harrell, Avital Cnaan, Chris Coburn and Roberta Ballard The Children’s Hospital of Philadelphia Philadelphia, Pennsylvania

In an on going multi-center clinical trial, a window based data management system was developed using INGRES and Windows 4GL. The study involves collecting information about women in premature labor, their offspring (many twins) and long-term outcomes on the babies. Information on each newborn link to their mother by the relational database. Hundreds of variables are collected on 15 forms at several time points. Due to the complexity of the study and the volume of information gathered, the database is complex with sophisticated windowing system and data management capabilities. The system incorporates several different technologies such as allowing remote access by using X-emulation software and either an Internet connection or a modem, pull-down menu driven data entry and complex relational structures. A demonstration of the advantages of this type of window based data management system will provide useful information about design, windowing and accessibility issues to the other clinical trial data management centers.