Diversity, Volume of Clinical Data Will Make Standards Matter, Says Burke of SAS

Editor's note: We recently mused about data warehouses and pleaded for help from readers. We're still seeking feedback on that point. But the following short, sweet, lucid email came from SAS's Jason Burke, director of life sciences strategy and solutions.

The industry should not just worry about choosing between a centralized or federated design, Burke says. Rather, either type of system should easily handle growing volumes and types of data. Let's not forget, he says, that data analysis-not data storage-is the real goal. To facilitate such analysis in the future, he argues, the work of the Clinical Data Interchange Standards Consortium will only become more important.

By Jason Burke

Any organization adopting a centralized data storage model will have to deal with two realities: the exponential growth in data volume and complexity over the next decade, and the diversity of data sources that will drive such growth.

Federated information repositories are already a reality. Any organization adopting a centralized repository strategy will still need to work in the federated world. So the question is not whether to use centralized data storage or federated data storage. The question is how to make intelligent use of information, given that:

1) the data you want to exploit may not be data that your organization collected (i.e., third-party research repositories),

2) the volume of data related to your research is growing such that it may not be easy to identify what data you should care about,

3) the business needs for making better decisions earlier require that you move closer to real-time exchange of research information, and

4) any organization's ability to standardize both information and business processes will be challenged by the growing heterogeneity of this information landscape.

These issues cannot be addressed by debating data storage models. New standards are needed to help provide the "glue" that will allow organizations to let go of the idea of data location, and to focus on data analysis. As our industry continues to invest in CDISC-related standards activities, the role of metadata-driven IT architecture standards, not database standards, will be the new imperative for organizations of all sizes.

Please Sign Here: New Digital Tools from ARX to Combat the $21 Signature

In an age when large transactions occur online, and total strangers do business virtually, it's odd how important a paper signature remains. ARX Inc. is trying to change that.

Based in Israel, with offices in the U.S. and Europe, ARX has been around for almost 20 years, trying to help the world manage the snowdrifts of paper that entomb most offices and corporations. They're not removing paper, as such, but one of the reasons people like paper: You can sign it, spill coffee on it, prove that it passed through your hands. Most of us don't feel the same way about files on our computers.

Electronic signatures, in many implementations, have been proprietary in nature. A company buys a system. It works. But just try to take a document from that system to another organization.

Then there is the core ARX technology, CoSign, which uses hardware and software to leverage the portable, cross-platform Adobe Acrobat format. Small life science companies, ARX has found, recoil at the price of Adobe Acrobat Professional, which can securely sign documents electronically.

But the Adobe Acrobat Professional package can cost several hundred thousand dollars for an enterprise license. In some cases, smaller pharmas simply license a few key employees. That saves money. But it also creates contorted and inefficient workflows if everything in the company has to go through Dianna over in Regulatory.

ARX, on the other hand, allows every employee to sign documents digitally. The company has a new (free) application, OmniSign, which works with any standard program that accesses a printer.

OmniSign is even versatile enough to sign readouts from laboratory instruments.

In addition, ARX has special plug-ins for widely used applications like Microsoft Word and Excel. Best of all, the ARX platform works on old computer hardware-mainframes in the basement of a hospital or academic medical center, perhaps.

ARX installs a box that attaches to your network, and software that runs the box. For those who care about the security terminology, ARX's CoSign uses public-key infrastructure (PKI) to encrypt documents and signatures. For the rest of you, PKI is an open, highly robust and nonproprietary approach.

Why use anything? Well, ARX cites third-party data that suggest the actual cost of (as the company puts it) "wet signing" your John Hancock with a 29-cent ballpoint pen comes to $18 to $21 per document. Ouch. When one ponders the thousands or hundreds of thousands of pages signed for a single trial, paper signatures may be one of the more glaring extravagances in industry-sponsored research.

The ARX system, by contrast, can pay for itself in a few months. DataLabs and ClinPhone are both using ARX technology to expedite the movement of clinical trial paperwork. "The signature is embedded inside the PDF document and travels with the document to the receiving end," says Gadi Aharoni, president and CEO of ARX. "The receiving end doesn't need to install any software to read the signature. You will be able to see the graphical signature, the image of my signature."

Aharoni adds that European privacy directives, far more strict than anything in the U.S., are becoming the world standard. Global companies, after all, find it easier to learn the toughest rules and adopt them as company policies. "Europe is about half a step ahead of the U.S. in terms of legislation and regulation," he says. "The number of regulations that specifically require digital signatures is higher than in the U.S." With governmental and corporate privacy lapses reported in newspapers on a regular basis, he would not be surprised by tighter regulation in North America in the years ahead.

Within the specialized confines of the pharmaceutical industry, ARX says, there is widely enabled functionality that enables entire systems to authenticate users. But it's another matter to prove who signed a particular case report form (CRF) or a specific 1572 form. "You can't send your entire database or electronic data capture (EDC) system to a business partner," says Rodd W. Schlerf, ARX's life sciences practice manager. CoSign's ability to allow users to sign partial sections of documents is increasingly attractive, he says, as is letting multiple individuals sign the same document electronically.

Schlerf explains how the ARX system works. "Instead of printing a document to the regular printer, you print it into a CoSign printer. The CoSign printer converts the document into a PDF, and then it applies the signature to the PDF. There are a lot of legacy systems and proprietary or homegrown applications that people want to be able to sign. They print out a document and put a wet signature on it."

Not using paper probably offers at least three cost savings. The first is the sheer cost of the transporting and handling the document. The second is the delay incurred as pieces (or piles) of paper get lost or forgotten in the course of a trial. The third is the cost of storing paper. In the case of ARX's e-archiving, the documents' "signer authenticity is guaranteed, and that accountability, data integrity and non-repudiation of documents and transactions is ensured," the company website notes.

How much does the ARX approach cost? The price would vary based on the needs of the client. But a one-time charge of $18,000 (plus annual maintenance) might cover a typical installation for 100 users. The most receptive customers, says Schlerf, are those with geographically distributed operations and companies with large numbers of partners and customers.

Clinical Uses (and Users) of Technology at September "Bridging" Conference

Melissa Newcomer has an interesting perspective on the development of clinical trial management systems (CTMS). "I worked at Pfizer for five years," she notes, and while there, became acquainted with that company's homegrown, internally developed CTMS. Now, as Allergan's manager of global clinical quality and training, she's using an unnamed, commercially supplied CTMS out of the box.

Both of her CTMS perspectives, as it happens, will be rolled into her talk at the Bridging Pharma and IT conference at the Sheraton in Philadelphia, Sept. 25-27. She's part of the clinical track (Track 3). But there are two additional tracks devoted to effective science/IT combinations in drug discovery and development, respectively.

The conference is a unique place to hear about the synergies between drug discovery and technology. In the case of speakers like Newcomer, the "Bridging" conference is not Introduction to Technology 101. It's an advanced-level seminar with lessons from the forefront of skillfully combining science and IT in pharmaceutical research.

Newcomer says she is looking forward to helping the Philadelphia audience understand some of the nuances of all varieties of CTMS. Consider, as an example, the complexities of serving both internal and external users. The latter may be overlooked at first. But some partners and data monitors may not be using company laptops. That, in turn, may affect what software external users can use to view and alter data inside the CTMS. The resulting issues around firewalls can be solved, but take time and thought.

Another concern, she says, is that some big-picture executives may have unwarranted expectations about magical visibility into clinical trial data. Depending on how a CTMS is built, that may be possible. Or not.

Newcomer phrases the matter diplomatically: "There is a misconception as to what upper management thinks a CTMS can do and what it actually does." In part, managerial expectations may be too high because different trials use electronic data capture (EDC) to varying degrees, or even connect to different clinical data management (CDM) systems or processes within the company.

At the risk of stating the obvious, a CTMS can only provide data that is in the system. With different inputs and subsystems, different trials necessarily will have different levels of data freshness. Newcomer gives an example. Some companies may have different flavors of EDC or CDM that allow relatively rapid access to data-or to data that has been taken off paper forms. "By the time the CRFs are pulled and put into Oracle Clinical, you may not be seeing enrollment numbers for 90-150 days," Newcomer notes.

In short, different technologies or processes used in different trials can make the 30,000-foot view of a company's portfolio of trials irregular. There may be perfect acuity into some trials but limited insight into others. There is a workaround, naturally.

Suppose that under special circumstances, a clinical project assistant manually keys in the fresher enrollment data. Problem solved? Partly. What happens a week later? By then, of course, a monitor may be tapping in the same patients using the usual process. "You've just created a duplicate," Newcomer notes. "How do you recognize that?"

Some therapeutic areas, she says, may need additional flexibility built into a CTMS. You may want to see if your CTMS provides the ability to add additional treatment cycles on an ad hoc basis. With diseases like cancer and HIV, patients may need unexpected rounds of therapy. "You have to have that flexibility," Newcomer says, adding that correctly setting up the patient visit schedules and the billing is one of the most vital steps in building or implementing any CTMS.

Newcomer's presentation, as best we can tell, will presume that many companies are threading together a variety of data streams in a CTMS. Everyone worries about managing investigator payments. A more neglected component of a CTMS, however, is software that accelerates the work of site monitors. "If you're going to do a CTMS," Newcomer advises, "do it in its entirety. You need to think about all the pieces of the puzzle."

She is particularly enthusiastic about interactive voice response (IVR) tools that directly update a CTMS with live enrollment data. But Newcomer also says she tends to assume that most sophisticated sponsors will have a growing percentage of trials that use EDC-and that EDC systems will tend to populate the CTMS at many companies. "Anything you can do to get the enrollment data in real time, or close to real time, is imperative," Newcomer says. Her presentation in Philadelphia later this month should be fascinating.

Upon reflection, it's clear EDC has become the industry's bedrock-the prosaic slab of granite upon which the rest of the house of clinical trial data analysis will be built. Some sponsors may look down their noses at one EDC solution or another. So be it. But without a well-grounded foundation, other data analysis structures will be rickety at best.

Editor's note: The Bridging Pharma and IT 

meeting is presented by Cambridge Healthtech Institute, which publishes Bio-IT World and this newsletter.

 

Contact Mark Uehling at: markuehling@earthlink.net.