Role

The primary role of the Statistical Center in the PCPT is to ensure the successful implementation of the trial design through effective data management. This includes the following task.

Evaluate statistical aspects of the study design and develop appropriate statistical methodology.
Modify the study design as required, in conjunction with investigators and the Operations Office.
Consolidate and analyze data collected by the study sites.

Develop and maintain the PCPT data base.
Develop and maintain the Study Manual, data collection forms and procedures.
Develop, monitor and implement the enrollment and randomization system.
Review and enter data submitted and report discrepancies to the study site.
Act as study site liaison for procedural and data management issues.
Implement and monitor quality control procedures.
Train study site personnel in policies and procedures required to implement PCPT.

Monitor activities of central laboratory facility for technical error and secular trends.
Receive, interpret and distribute results of serum analyses from central laboratory facility.

The data management plan for this study features the use of two technologies designed to minimize support staff costs and improve data timeliness and quality: (1) data transmission using facsimile (fax) transmission into the Statistical Center computer; the digitized fax images are subsequently manipulated using a computer; and (2) a computerized participant randomization system using touch tone telephone input and pre-recorded sound bites.

All PCPT data are sent from the study sites via fax transmission and are received at the Statistical Center via the DataFax, an integrated image management and data management system. Dedicated toll-free 800 lines are provided for faxing data. Using optical character recognition (OCR), the DataFax reads data items submitted on specially designed forms into its data base.

The PCPT runs DataFax on a network of 22 Sun workstations. The DataFax also requires the use of HylaFax (as its fax modem software) and Framemaker for forms design.

By using an image management system, the Statistical Center will be "paperless" with respect to data management for this trial. It is estimated that this trial will generate approximately 5.5 million data form images and require approximately 125 gigabytes of computer storage space. Current storage needs are met by a Sun SPARC Storage Array, which stores user directories, software and ancillary studies, and a DEC Storage Works 410 RAID array which stores all data form images.

Each eligible and willing participant was randomized following a three month enrollment period. Randomizations were completed in May 1997. The Statistical Center used technology which enabled the randomization to take place during a telephone call to a Statistical Center computer. This computer was configured to process touch-tone input and to output pre-recorded instructions.

The hardware used was 2 Gradient DeskLab boxes. Programmed subroutines were used to access required tables in the data base and perform the dynamic treatment assignment (as is done with Southwest Oncology Group therapeutic trials). After the participant's eligibility was determined, an appropriate study drug bottle was identified and its identification number transmitted to the caller.

The Study Manual describes all study-specific procedures for randomization and follow-up including the study calendar, defines terms, and provides instructions for completing and submitting the data forms. Site-specific master forms sets are distributed to each site.

Faxed data forms are electronically identified and sorted into discrete files (by form, visit and participant identification number) for subsequent processing. The image and the stored data are compared and validated by data coordinators and data control technicians. Problems are identified and annotated for inclusion in quality control reports prepared for each study site.

The focus of the quality control (QC) review is to ensure that the participant is eligible for both enrollment and randomization, and that participant follow-up is completed per requirements specified in the protocol and the Study Manual.

Each data coordinator initiates QC review with a check of the contents of the participant's file to ensure all required records were received. Each form image is then reviewed for accuracy, consistency, and compliance with the requirements of the protocol. Problems identified during the initial validation process are addressed and resolved if possible.

All quality-related problems identified by the data coordinators are also annotated on the DataFax data record created from the form's image. A quality control report, detailing each of these annotations, is created and faxed to the study site for resolution of the problems. This quality control report also lists those required forms not yet received at the Statistical Center. In response to the report, the study site faxes an amended version of the original data form. The data coordinator will access these records, resolve the identified problems, and validate the data record appropriately to indicate the quality of the data it contains.

PCPT staff training is conducted in conjunction with each Southwest Oncology Group Meeting. Each training workshop is developed to instruct new staff in the implementation of the PCPT and to address current issues impacting study and data management. A portion of each workshop focuses on the development and implementation of participant adherence programs at each study site. Training materials are developed to support the agenda of each workshop and are distributed to all attendees.

The Study Manual serves as the primary training tool. All updates to procedures and forms are distributed to all PCPT staff as study bulletins, memos, and Study Manual revisions. These are mailed to each Study Site. Any form revisions or procedures changes take a primary focus during each training workshop to ensure that these changes are communicated clearly.