CULLEN T. VOGELSON
Moving into the 21st century
Digital source documents and casebooks lead the way.
Clinical research in the United States has been conducted more or less the same way for decades. Regardless of the type and phase of the study, the therapeutic area, and the specifics of a trial, the nuts and bolts of human research are well established and rigidly consistent. A clinical research study is conducted according to the following general outline:
An investigative site is chosen to recruit patients into an approved trial.
After patients give their consent, the investigative centers (sites) enroll them into trials.
The sites conduct assessments, evaluations, and laboratory analyses according to the study protocol.
Data are recorded in source documents and transcribed later by a coordinator into the study-specific case report form (CRF).
A monitor compares each data point in the CRF with the source record.
The monitor then ships a paper copy of the CRF to a central data management center (per FDA regulations, the site maintains the original source documents and a copy of the CRF).
For accuracy, data-entry technicians double-enter each data point into a clinical database.
A computerized review of the database then generates queries about data inconsistencies, missing information, and incomplete entries. The monitor resolves the queries with the site.
Data-entry technicians double-enter the resolutions.
Another computerized review occurs, and the reviewâ?”queryâ?”resolution loop continues until the data are â?ocleanâ?ť.
Finally, the database is transferred to the clinical statistician for analysis.
The clinical research process just outlined is slow, tedious, and expensive. As a result, pharmaceutical companies seek ways to reduce the time and expense mandated by research practice. One such attempt involves the use of new, computerized recording and transmitting devices. These technologies are increasingly sophisticated and offer the satisfying possibility of cost and time savings. They also introduce new concerns for research monitors, auditors, and record keepers.
Source documents can be entered electronically as a result of the advent of inexpensive portable handheld devices. These devices contain simple software that can capture the notes a researcher would normally scribble on paper during a patient examination. Many of them also contain study-specific software to prompt the investigative personnel with fill-in-the-blank style questionnaires.
The advantages of maintaining source documents electronically are obvious: Researchers can immediately review, search, and organize the entered data. Virtual documents also accurately store information about when the data were collected and by whom. (This latter point means that an audit trail is ensured.) Finally, the data can be printed or stored with relative ease.
Unfortunately, the use of portable electronic devices raises two issues of concern about data integrity. The first issue is the possibility of data errors. For example, a researcher is more likely to transpose or mis-enter the numbers in a blood pressure reading when keying into an electronic notebook than when writing onto a paper one. The second integrity issue relates to storage. The FDA requires that each site maintain the original source documents for a study, and so keeping data on a battery-powered device may be problematic.
The solutions to these two problems require a certain level of diligence on the part of the investigative site. Data errors can be caught if the data are reviewed immediately, but patient appointment schedules and other time constraints may make this difficult. Some software programs, however, check for odd entries and query them in real time (e.g., entering a patientâ?Ts date of birth as May 32 would be disallowed).
In terms of record maintenance, the FDA considers a computer hard drive sufficient for â?opermanentâ?ť storage, so researchers must routinely download the patient data from the handheld devices.
Many research trials include a patient diary, diary card, or other form of participant-recorded evaluation. Here again, portable handheld devices are a possibilityâ?”although their use is still minimal because of associated costs.
The advantages and disadvantages of giving patients electronic diaries are identical to those related to computerized source documents. Some additional advantages exist, however. Patients are less likely to mis-enter the kind of personal data commonly collected in diaries (e.g., patients may be asked to rateâ?”on a graphical scaleâ?”how they feel each morning immediately before dosing); and errors that result from a coordinator misunderstanding, misinterpreting, or incorrectly transcribing patient responses into the CRF are eliminated.
The digital casebook
The most significant advance in clinical research practice is the use of computerized CRFs. These fall into two categories: stand-alone and Web-based. The former involves entering data into a computer-based casebook that is subsequently collected on disk or transmitted electronically to the data management center for review. The second category involves a virtual CRF, maintained on a secure Web site in which investigative centers enter and submit their data directly. In both cases, copies of the completed CRFs can be maintained as hard copies or electronic files.
Virtual CRFs have significant advantages, and for simplicity and cost-effectiveness, Web-based casebooks are best. In fact, the only real disadvantages involve security and storage issues. But the reasons in favor of using Web-based CRFs are significant and include the following:
Coordinators can receive instant feedback when entries are incomplete, inconsistent with previously recorded data, or do not match the required form of the database (e.g., a year is entered as 98 when the database mandates a four-digit entry). In these cases, instant feedback reduces the number of reviewâ?”queryâ?”resolution loops and eliminates the need for data-entry technicians to double-enter all collected research data.
Computerized CRFs are easily searched, reviewed, and audited.
If electronic source documents are used, then an algorithm that inserts the data into the appropriate places in the CRF can reduce the amount of time a coordinator spends on the project and improve the accuracy of the captured information.
Complex queriesâ?”those generated by human CRF reviewers at the data management centerâ?”can be forwarded electronically to the sites, and the coordinators can resolve them directly. This process avoids the need for data-entry technicians and guarantees that the changes a coordinator makes to the dataset are consistent with the existing database (i.e., changes to the database are reviewed and then queried, if necessary, upon submission).
Virtual CRFs can capture vital accountability information, such as when and by whom an entry is made or changed. Furthermore, new regulations make it possible to submit electronic signatures that guarantee that the FDA-required investigator review and approval process occurs for each dataset.
A monitoring perspective
Monitors, who must travel to the sites to review the CRFs for accuracy and completeness, are big winners in the transition to Web-based clinical trials. In particular, sites that combine electronic devices for patient self-evaluations with computerized source documents and virtual CRFs reduce the need for extensive visits. In addition, the work that a monitor must accomplish during a site visit is dramatically altered; monitors need no longer spend time separating copies of CRFs and shipping them off for analysis. Furthermore, the elimination or reduction of the tedious and time-consuming reviewâ?”queryâ?”resolution loop is a blessing. (Note that even with data collected virtually, travel is still a part of the monitorâ?Ts job because other types of in-person review and site training are necessary.)
Many Web-based trial systems also offer important diagnostic and (rudimentary) statistical data reviews, which can help monitors identify problems before they become significant. For example, knowing the number and type of queries that each site addresses may help monitors assess an individual siteâ?Ts training needs. Furthermore, these diagnostics can help pinpoint research centers that may become targets for FDA audits (e.g., centers with high patient enrollment or significant participant dropout rates).
Most important, having a wealth of summary information available for each site can help identify those that may be engaged in negligent or fraudulent behavior. For example, a warning message may be sent automatically to the monitor of a site if the data submitted by that site are entered at unusual times, such as 2 a.m. on a Sunday morning.
The big picture
Electronic CRFs were first designed roughly ten years ago. Their implementation, however, met with numerous problems: Sites needed to be provided with computer systems that were often costly, security issues were significant, data storage was a concern, and no means of obtaining an electronic signature was legally allowed. Thus, the few companies who made the technological leap were disappointed. They found that new layers of documentation and paperwork were introduced, cost outlays undercut most potential benefits, and time savings were insignificant.
Changes in regulations, coupled with technological advances, however, have dramatically altered the playing field. Today, companies like Phase Forward, eResearch Technology, and Databean offer researchers the beneficial features of electronic information gathering as well as functions such as investigator referrals, patient schedulers, and individualized consulting services. The benefits of electronic systems thus include lower costs, enhanced data quality, and improved research speeds. As companies that specialize in gathering data electronically for clinical trials expand and make deals (e.g., in May 2001, Eli Lilly chose Phase Forward as its standard platform provider for Web-based trials), the future of virtual research is assured.
(For more information on the clinical uses of personal digital assistants (PDAs), see “Rx for PDAs”, Modern Drug Discovery, January 2002.)
Cullen T. Vogelson is a former assistant editor of Modern Drug Discovery. He writes and teaches in northern California. Send your comments or questions regarding this article to [email protected] or the Editorial Office by fax at 202-776-8166 or by post at 1155 16th Street, NW; Washington, DC 20036.