DFUG 2003

At DFUG 2001 and DFUG 2002 Frontier Science presented a method of performing electronic data collection using DataFax. This method, now called DataFax-EDC, has matured into a product consisting of the original server component and a corresponding "electronic subject binder" (ESB) for managing electronic case report forms (eCRF's) at the clinical centers. For DFUG 2003, we will give a full demonstration of DataFax-EDC. This demonstration will cover:

1. Creating and using ESB's, including DataFax specific features, security, and hardware and software requirements.
2. Receiving and validating eCRF's from ESB's, including specific details about QC notes and reporting, and a synposis of the data transfer format.

DataFax-EDC technical documentation will also be available as a separate handout.

Efficient conduct of a trial is a key issue in any study. Therefore, the systems and processes employed to capture, clean and report clinical data become crucial success factors for any data management organization. Currently, there are several major clinical data management systems being used within the pharmaceutical industry, DataFax being one of them. Oracle Clinical, however, is one of the most widely used systems for conducting clinical trials.

This presentation will provide a comparison of Oracle Clinical and Clinical DataFax, examining the aspects of each system in regard to data management processes and resource utilization for paper-based studies. Specifically, this presentation will address the time and resources utilized under each system for the following tasks in paper based studies: CRF design and printing, query generation, resolution and tracking, overall CRF work flow, data entry processing, and reporting. This presentation is drawn from the author's experience utilizing both Clinical Datafax and Oracle Clinical over a six-year span.

On January 13-14, 2003, a group of five auditors representing four different datafax customers met at CDSI's offices in Hamilton, Ontario to review SOPs and related materials. On the basis of additional documentation provided, personal interviews and a facility tour, this group was able to complete an audit of Clinical DataFax Systems. A well-documented quality system was found. The observations and suggestions made represent issues which can be easily corrected and pose no unmanageable risk.

This report to the user's group will include a description of materials reviewed, observations made and suggestions for future group audits. A printed report will also be available.

Data management groups often must confront the predicament of modifying Case Report Forms (CRFs). This is not a problem in the initial stages of study setup, but for studies with data accrued, difficulties can arise. For example, if it were necessary for a CRF's variables to be rearranged, a discrepancy would occur between the ordering of the existing data versus future data. A solution to this would be to reorganize the variables on the DataFax plate to fit the new CRF schema, thus eliminating the imminent conflict within the data.

In response to this problem, a PERL utility has been created that facilitates the adding, deleting, and/or swapping of fields on DataFax plates. For that matter, such a tool can aid in handling general DataFax database structural changes. The current version consists of three PERL programs, add_fields.pl, remove_fields.pl, and swap_fields.pl, that can be run from the command line and have a text based interface with the user. A future release (hopefully by the meeting in Florida) will be one program with a graphical (tcl/tk) interface.

The Statistical Center for HIV/AIDS Research and Prevention (SCHARP) is a non-profit organization that conducts the clinical and statistical coordination for clinical trials with the HIV Prevention Trials Network and the HIV Vaccine Trials Network. With dozens of concurrent studies in as many countries around the world, SCHARP has an average incoming volume of approximately 70,000 pages/month.

Because of this high volume, SCHARP has decided to develop a CRF tracking system to provide information to sites when there are concerns that CRFs were not received by us. The tracking system consists of the following:

1. A cover sheet consisting of a simplified form for a seperate DataFax study. This would be used as the first page on each bundle or packet of pages being sent.

   The study, plate and visit barcode on this form provides a unique identifier and corresponding tracking number. Every cover-sheet study, plate and visit barcode will be unique, and cover sheets are assigned to sites in a manner that is systematic and documented.

2. When the forms are received by DataFax, the sender is identified through the unique combination of study, plate and visit found on the barcoded cover sheet. A human-readable form of this barcode will also be printed on the cover sheets, providing the sites with a unique tracking number for that fax packet.

   A DataFax pre-process plate-level trigger will result in the site being immediately notified by email with descriptive details of the fax bundle received (e.g. number of pages). The email response will contain the unique tracking number, allowing sites to accurately determine which fax packet the confirmation refers to. This is very important when sites may be sending dozens or even hundreds of 20-30 page fax packets throughout the day.

3. The email confirmation will contain a URL to a SCHARP web page providing details about the content of the fax packet. Initially, this will be limited to the number of pages received. But as each page is entered at SCHARP, the web site will be updated with the content details of the fax transmission.

   Many levels of detail are possible, from simply a list of study, plate and visit combinations of CRFs contained in the transmission up to including links to the images themselves, the current QC status of the record, a link to QC reports that the CRF may be mentioned in, and record status and validation level information.

   Much of the information on the web site may be "SCHARP-internal" only. Having this information centralized will be valuable when responding to the sites' inevitable "I thought I faxed that form".

Currently, this tracking system is under development and exists as a prototype. When fully implemented, it will provide valuable information to the staff at remote sites and increase their sense of participation. The details of this prototype tracking system will be discussed in the presentation.

Clinical trials often require an adjudication process to provide a standardized system for validating the potential study endpoints. In this process, case report forms (CRFs) and supporting documentation for potential endpoints are collected from the sites and sent to qualified specialists who are blinded to the study medication assignment. These specialists then confirm or refute the endpoints according to approved endpoint definitions. Accurate management is required to track each supporting document, monitor the status of each adjudication case, identify bottlenecks in the process and produce summary tables of the results in a timely fashion. We have used DataFax and a Lotus Approach database application to help facilitate this process.

CRFs specifically designed for capturing adjudication results and tracking the adjudication process were created and added to the existing study CRFs. The characteristics of each adjudication case (such as case number, adjudicators assigned, type of endpoint, date of endpoint, language of supporting documents, adjudicators assigned, date sent/received to and from adjudicators) were entered in the Lotus Approach application and the relevant fields were exported and entered in to DataFax. One can then track the supporting documents, the location of each adjudication case, the final adjudication decision and any other relevant study-specific variables. Queries for missing or incorrect supporting documents and data are incorporated in the Quality Control reports and faxed to sites and monitors regularly. Using a DataFax Study Specific Report, an adjudication summary sheet listing all potential endpoints with dates of onset and adjudication status can be easily prepared combining data from the study CRFs and the adjudication CRFs for each patient. Time from receipt of endpoint CRFs to adjudication decision can be tracked and problems can be easily determined (such as slow reply from sites/adjudicators or requests for additional information). Summary tables of the adjudication decisions, status of each case and number of agreements and disagreements between adjudicators can also be created and transmitted quickly to the required parties (sponsor, steering committee, DSMB). The use of an accurate and efficient system such as the integration of DataFax with a more traditional database application streamlines the labour-intensive process of adjudication.

The validation process in DataFax requires a shift in clinical work practice and is feasible only with high-speed connectivity. As a supplement to visual validation ("VV"), we have developed paper Monitoring Report Forms ("MRFs") as an alternative method for validating images beyond data entry. The MRF allows the clinical research associate (CRA) to report whether a case report form (CRF) is "clean" or "dirty" (discrepant with source).

As with CRFs, once an MRF has been completed it is faxed to DataFax. The majority of data fields on the MRF are intelligent character recognition compliant, requiring only limited manual data entry. The data on the MRF are exported to SAS for processing, using the standard DF_export.rpc procedure. In SAS, the CRFs are parsed according to various criteria and are then grouped into one or more DataFax retrieval files (drf). These files are passed back to the DataFax server where batch edit checks further process the records (e.g. generating missing page QCs; adding QCs; moving records to a higher level).

As a paper-based system, MRFs provide a familiar medium for communicating with DataFax and require only nominal training for those familiar with CRFs. This solution is accessible to CRAs that are not comfortable with electronic data review as well as to contract CRAs or CROs that are not cleared for internal access to the DataFax system. Additionally, MRFs provide a more efficient way of moving CRF images through the review process where direct access is impractical due to slow connectivity (cross-continental or international remote access).

More and more pharmaceutical companies and contract research organizations are realizing the importance of adopting common standards for data exchange, e-submission, and archiving. After searching for a common standard for a while, we have decided to use the standards from the Clinical Data Interchange Standards Consortium including operational data model, analytical data model, and submission data model.

The CDISC data models support the end-to end data flow of clinical trials, from the source(s) into an operational database, through analysis to regulatory submission. By adopting CDISC standards, we can exchange clinical data easily with our sponsors and partners. We are currently adopting the Operational Data Model and exploring other CDISC standards. The ODM refers to the standard data interchange models that are being developed to support the data acquisition, interchange and archiving of operational data. The other CDISC models include lab data model, submission data model, and analytical data model. The SDM refers to the standard metadata models being developed to support the data flow from the operational database to regulatory submission.

Based on our DataFax to Oracle ETL (extract, transform, and load) engine, we are building a data engine and web interface to wrap DataFax clinical data in the ODM. We will demonstrate some of our preliminary results and share our future plans. The main goal of this presentation is to share our understanding about CDISC standards and our approach to integrate DataFax with our integrated clinical web portal and operational trial management systems, and make our systems CDISC compliant.

During the course of a study, clinical monitoring and review verifies that source data have been accurately transcribed on to Case Report Forms (CRFs) and that CRFs have been transmitted to the database (e.g. DataFax). Within DataFax, accuracy is monitored interactively by range checks, edit checks and visual verification and by running integrity check (DF_IC) reports and batch edit checks. Post-processing and statistical analysis (e.g. in SAS) provides another layer of confidence that the database is an accurate representation of the clinical profile.

When the last patient visit has been completed and a study or site is ready for closeout, one must "ensure that the all of the expected data are present and clean". Those data that are collected on a continuous basis (e.g. Adverse Events; Concomitant Medications; Comments) on "running" or "log" type forms are a potential source of error. These data are not always discrete and so they cannot be reported in neat installments nor can log pages be monitored and signed off in the same way as visit-based CRFs. Some data require end-of-study resolution (e.g. a "stop date" or a denotation that the AE or medication is continuing past the point where data are being collected). One must also ensure that all pages in a log series have been received.

Methods for closing out a study range from reviewing one hundred percent of the data fields/ CRFs to a review of critical fields and/or a random sample to... no action at all! At PDL, we determined that visual verification in DataFax or the submission of Monitoring Report Forms provides an adequate assurance that data are present and accounted for with the exception of log pages. This presentation will review the study closeout process currently used at PDL including the use of electronic-Data-Over-Image ("eDOIs"). At site closeout visits, the CRA visually compares a composite of the data-over-latest-image for each DataFax log page against the original CRF and source document as a final accuracy check of the database.

Managing contract payments to investigators in a large long-term clinical trial with many sites can be a nightmare on the one hand, or, if administered well, can contribute greatly to the success of a study. The clinical sponsor would like to reimburse investigators promptly and fairly for their conscientious efforts, and hold back payments when investigators underperform. Investigators see an efficient payment system as an incentive to get subjects in for visits and send their CRF data back to the sponsor on time.

DataFax, because of its expeditious and automated processing of CRFs, provides the foundation for an efficient and easily managed contract payment system. For more than three years, VaxGen has relied on such a system, based on DataFax and SAS processing, to pay investigators in a large trial with 61 sites and more than 5,000 subjects. In this presentation, I will outline the basic concepts of our system, and provide detailed examples of how it works.