Keeping Ahead of the Technology Curve

Technology is surging ahead in Europe—and not for the first time, the regulatory system is having to run to keep up.

The increasing use of computers in the conduct of clinical trials now extends way beyond data management, analysis and reporting at the site of the sponsor, contract research organization, or laboratories, and storage of medical records, and is reaching far into the capture of clinical data. One of the current clinical trials areas where technology has held out the promise of a breakthrough is in electronic patient records. But turning the promise into reality is proving complex.

Peter O'Donnell

Theory vs. practice

Everyone is now familiar with the potential of electronic patient reported outcomes (ePRO) in acquiring data, where patients record observations, rating scales or investigational medicinal product use on electronic devices that are supplied to them, either for primary efficacy or for supportive data.

It can be particularly valuable in those quality-of-life areas of therapy, such as seasonal rhinitis, depression, pain management, or asthma, where the only end-point is often subjective. If the patient can record subjective responses on a regular basis so they can be transmitted directly to the investigating physician, either in real time or at the end of a period, authenticity and accuracy should—in theory—be guaranteed, with savings of time and greater convenience all around. But that is only the theory.

In practice, the wide diversity of approaches bristles with technical complexities, from the variety of software, hardware, and systems—PC, LAN, WAN, laptop, email transmission, Web-based systems, and interactive digital voice response systems. Large-scale exploitation is impeded both by this proliferation of (often still unproven) possibilities, and—just as importantly—by the uncertainties over whether and how far regulators will accept data derived through such systems. Unsurprisingly, therefore, the use of handheld devices as an electronic form of patient diaries is one of the subjects that is currently exercising drug regulators in Europe.

Some of the physical attributes of paper records require rethinking in the context of electronic media. One of the dilemmas is what to do about the concept of a source. In European orthodoxy, the basic concept of source data is that it permits not only reporting and analysis but also verification at various steps in the process for purposes of confirmation, quality control, audit or inspection. With electronic data from a patient diary, new challenges emerge.

Unlike a handwritten patient diary or the traditional case record form, the Palm Pilot (or whatever similar handset) where the patient has entered data is not in the same way a source: the individual records—be they five, 50 or 500—that it contains are uploaded to a central database, but the devices themselves are not retained. In such circumstances, the concept of "source changes" perhaps has to be redefined, to take account of the fact that the only source that is then available is what can be downloaded from the database.

Similar questions arise over authenticity. It is possible, of course, for a user identity code to be required for each original entry. But arrangements must be made for that stamp of authenticity to accompany the piece of data each time it is moved from the device to a database, and from there into any other system.

Some devices allow the entry of adverse event reports too, which again is valuable in theory, but can cause problems in practice. Without foolproof processing systems, there is a risk that such reports may get lost, delayed or even overlooked—unlike a paper report that comes across the investigator's desk. If the report concerns a significant event, there could be serious consequences. And unless the system of transmitting data from the individual patient's device to the database is perfectly designed, there is also the risk that the investigator might not see the individual reports at all.

Agency guidance

In a bid to cope with some of these questions, the European Medicines Agency is expected to come out shortly with a draft guideline on expectations for electronic source documents used in clinical trials. The aim is to offer general guidance that can then be adapted to the specific situation of a particular electronic data capture (EDC) instrument or to the context in which it is used—and it will be, EMA officials admit, not easy to provide guidance without tying everyone down to technology that could rapidly become outdated.

It will cover the use of electronic source data in clinical trials conducted in any of the 30 countries belonging to the European Economic Area (that is, the 27 EU member states, plus Norway, Iceland, and Liechtenstein) or where clinical trial reports are submitted as part of marketing authorization applications to regulatory authorities in any of these countries. This will move European thinking forward on a subject that has slipped out of the headlines for a couple of years.

It was in late 2007 that the EMA published a reflection paper on the subject, after its GCP inspectors working group had reviewed the growing need for a common approach. With the increasing use of information technology in pharmaceutical development, it recognized there was a need for clear guidance on the use of electronic source data and on the principles that should apply. This was necessary, it said, "to ensure that the processes can be used and accepted with confidence when such requirements are complied with, and that the benefits that these systems offer can be fully utilized."

Areas of focus

The comment period closed just over two years ago, and since then the group has been refining its thinking. Central to the approach is that—in line with Good Clinical Practice—source data and documents form a cornerstone of record keeping in clinical trials. The fundamental issues, such as traceability or change-control, remain common in many cases to both paper and electronic systems. But electronic systems present additional challenges in providing an adequate level of confidence in the data.

Much of the EMA focus is on considering who provides and controls the electronic tool being used. And much of it is on protecting the notion—irrespective of media employed—of source data and the records that hold those data. These include that the data and records are accurate, legible, contemporaneous, original, attributable, complete, consistent, enduring, and also available when needed.

Translating GCP principles into the domain of electronic records, the EMA is sure to require that sponsor's operating systems ensure there is adequate validation, with standard operating procedures in place, and arrangements for audit trails of data changes, to ensure that there is no deletion of entered data. The audit trail will have to be adequate to provide evidence that changes made are attributable, date and time stamped, explained, and cannot be amended or deleted. They will also have to maintain a security system to protect against unauthorized access, which will involve a list of individuals authorized to make data changes.

Data backup and safeguards for the blinding of the study will be necessary. And if data are to be transformed during processing, it must remain possible to compare the original data and observations with the processed data. An unambiguous subject identification code will be needed that allows identification of all the data reported for each subject.

It will also set out some further principles, such as that any instrument used to generate, capture, transfer, manipulate or store source data will have to be an accurate representation of the protocol, so that the data specified within the protocol is captured correctly. The validity of the data capture process will be fundamental to ensuring that high-quality data is produced as part of a clinical trial, so it will need to ensure that all data required is captured, and that the same data are captured in a consistent manner.

Study subjects capturing the data will have to be trained in the accurate use of the device, and assurance will be needed that the subject has recorded the data at the time stated.

In the end, what EMA—and what the clinical trials community—wants is a workable consensus on the scope and limitations of ePRO. For the sake of both, the risks of abuse have to be satisfactorily eliminated: credibility is an asset neither can afford to put at risk. When the draft guideline appears, the clinical trials community will have four or five months to comment. To ensure they do not miss this opportunity to patch a potentially valuable mechanism into their work, they could start reflecting right now on just what they want to see.

Peter O'Donnell is a freelance journalist who specializes in European health affairs and is based in Brussels, Belgium.