Industry First: Pfizer Submission Using CDISC

Late in 2004, with an unnamed drug, Pfizer quietly became the first company to electronically submit clinical data to the FDA in a new way. The data were delivered in adherence with specifications established by the Clinical Data Interchange Standards Consortium (CDISC). "It was real data. It was a real submission," says Pfizer's William Qubeck, associate director, electronic submissions support services.

Qubeck laughs when asked if CDISC will only benefit large conglomerates. At conferences, he's learned to anticipate questions from those who presume he commands a vast staff. Actually, the project took four people. "All of this was done in less than six months, from conception to delivery," says Qubeck. "It wasn't as if we had infinite resources or funding."

Qubeck began helping CDISC in the customary way - in his spare time, knowing CDISC might not be a part of his official job description, but that it was destined to become central to electronic submissions at Pfizer and to the industry.

Now Pfizer has mapped proprietary applications and databases into CDISC schemas. The company has a proprietary application for clinical data reporting and analysis. Thanks to Qubeck and his four lieutenants, a sub-module of that system now generates CDISC-compliant datasets; the company continues to use SAS to map the data.

Qubeck and his team developed the mini-application, and have since experimented with both one-off and more generalized solutions to the thorny problem of getting all Pfizer data into a file readable by CDISC-ready applications like those from Lincoln Technologies.

For now, Qubeck reports, ordinary clinical or data management colleagues at Pfizer cannot export SAS data into CDISC. Programmers are required. But he's working toward a day when mere mortals anywhere in the company can pull down a menu and, with a click or two, create a CDISC file.

Readers may wonder, "How does Bill know his Pfizer-formatted data and his CDISC data are identical?" An automatic validation and verification tool, Qubeck says, turned out to have a surprisingly elegant solution. Once the metadata are mapped - once the computer knew the categories of data permitted by Pfizer and CDISC - it became easy to toggle between a Pfizer-formatted file and a CDISC-friendly dataset. There happens to be a new XML format for electronic common technical documents (eCTD); Qubeck's new sub-application also supports it.

What are the benefits of CDISC? Logically, Qubeck says, if CDISC standards are only used for submitting data to the FDA, CDISC tools will chiefly benefit that agency. (Still, there may be trickle-down effects for a sponsor that expedites the FDA's handling of data.) If CDISC data models are applied earlier in the process, however, more tangible savings may appear. "Your ROI [return on investment] is highly dependent on where you implement the standard," Qubeck says. That's because sharing CDISC data between sponsor and contract research organization, or between development partners, can be radically simplified.

Qubeck says the industry wastes time talking about the structure of clinical data, and not about the facts themselves. He explains: "Are we using your standards? Are we using our standards? These conversations totally disappear. Instead, you can focus on the right thing - the data, the content. What does the data mean? What is the quality of the data? Not the labels. The structure of the data should be irrelevant."

The structure of the data, of course, is not yet irrelevant to the vendor community. Just check out the CDISC website or registered CDISC solutions providers. There are some well-known names there (Quintiles, DataTrak, IntraSphere), but mostly consultants. Major electronic data capture (EDC) companies are hard to spot.

Qubeck agrees there is a lack of tools from the major clinical technology vendors. But he says he's optimistic that the payoff from CDISC is increasingly apparent to both sponsors and vendors. "I've seen a lot more interest from both vendors and in the industry in terms of, 'What does this mean,' 'What tools can we put in place?' " says Qubeck. "People are trying to find solutions. This is an exciting time."

He says Pfizer is in the process of abandoning its own standard for laboratory data in favor of the CDISC approach. Says Qubeck: "We no longer use our own proprietary format or have those discussions."

All the same, he muses that it would be nice if major clinical data repositories and clinical data management systems had pre-configured schemas that were CDISC-compliant out of the box.

For now, Qubeck laments, he must configure Oracle Clinical databases manually, and manually load a lot of data: "Those tools and schemas are not yet readily available. That costs a lot of money. That affects adoption. Companies are looking at this as an additive cost. You need to offset that with benefits. At this point, the benefits are a more efficient FDA review. It's going to take a while to see the full benefits."

: Featured Events :

MIT Professional Institute – Summer Session 2005 TRAX Pharmaceutical Supply Chain Summit 2005 - June 6 - 8, Philadelphia, PA Personalized Medicine - June 27 - 28, Philadelphia, PA Electronic Patient Reported Outcomes (ePRO) - July 18 - 19, Philadelphia, PA Beyond Genome 2005: The Future or Medicine - June 13-16, Fairmont Hotel, San Francisco, CA   Would you like to see your conference or event listed here? Contact Shari Redan at shari_redan@bio-itworld.com for more information.

Technology and Culture: One Assessment of Pharma

For those inside pharma, shortcomings in the clinical trial vendor community are traditionally wrapped up into a bundle about the size of a soccer ball. This bundle is dragged out and waved around on a stick whenever it is necessary to justify why certain databases or applications must be developed internally, in-house, by companies in which the core competency is manufacturing and marketing billions of small chemical tablets.

The functionality and usefulness of pharma's internal software is unknown. That's frustrating. One hears the clinical trial vendor offerings maligned from time to time, yet has no balancing perspective on the homegrown alternatives inside pharma.

A recent exchange with Scot Silverstein, a physician and assistant professor of healthcare informatics and IT at Drexel University, was helpful in this regard. Silverstein spent time at one of the biggest companies in the industry. It's clear he is pained by criticism of pharma. He wrote to chastise eCliniqua, as most readers do.

In a new drug safety database, he pointed out, the quality of its spontaneously reported data on adverse events will be severely scrutinized by any sophisticated customers. He's right. We began chatting about other topics - not specific applications, or specific companies, but the psychological environment in which technology is selected and operated.

Why does it take so long to make technology decisions? Silverstein replied: "It takes bold and ideally cross-disciplinary people to make bold and wise investment decisions. What I've seen in pharma (especially after past searches and my most recent year-and-a-half of seeking new opportunities) is that the HR process has so ultra-granularly defined the skill sets of positions that there is great inflexibility in hiring those who have transferable skills. My experience outside pharma in development of datasets, case report forms and analytics, and of having developed and implemented electronic medical record systems, seemed to be viewed as of negligible value towards working in clinical data management. The concept of transferable skills has been lost in the industry. This creates inbreeding of thought, and reduces the creative thinking required for innovative boldness (as opposed to ill-informed recklessness)."

Why does it take the industry so long to process data? "My only explanation is one of comfort," he says. "Ten, twenty, thirty years ago, it took years to do the slicing and dicing. IT has advanced much more rapidly than intensely bureaucratic data management process engineering." During one job interview, Silverstein reports, he learned adverse event data at the company was tightly controlled by the IT department; safety and regulatory people had to beg for it and frequently could only receive extracts of the data on paper printouts.

Why does the industry get bogged down? "I spent significant time answering requests for McNamara-like metrics in areas where such metrics were, I felt, ill-conceived and not appropriate," Silverstein says. "Developing appropriate metrics on enhancement of scientific creativity and long-term ROI [return on investment] of informatics assets is difficult. Doing so calls for flexible, actuarial/probabilistic thinking and consideration of the costs of lost opportunity, not the rigidity of assessments based on the equivalent of 'body counts' and simplistic costs-per-user-of-information that are favored by bean counters."

Why is pharma not doing what insurers and tobacco companies and other reputationally challenged industries do? In brief, engaging in sincere dialogue with patients (who feel overcharged) and physicians (who feel misled by industry's selective publication of trial data). Silverstein's answer: "This behavior is caused by personal characteristics reinforced by 'corporate culture,' (which can be defined as the pathologies of the few inflicted upon the many). Intense arrogance, greed, reactionary thinking, and stupidity are among the factors I've observed. In any industry, ultimately many decisions about strategy come from the highest management levels. It's a situation where the pathologies of the few can contribute to the behaviors of the many."

McClellan on Drug Safety, Pharma, IT at 2005 Post-Approval Summit

Two IT domains - clinical trials and hospitals - are colliding. Companies like eResearch Technology have been studying this and working out the details. So has Michael Barrett, a consultant.

Now there is a high-level federal push to tie the world of ordinary patients to that of clinical trials. The speaker who triggered this realization was Mark McClellan, former FDA head and current administrator of Medicare and Medicaid.

McClellan was speaking at the Post-Approval Summit, a new conference dedicated to what used to be termed Phase IV or post-marketing trials. The meeting was organized by a vendor with ties to Harvard University, Outcome Sciences, which assembled a singularly distinguished roster of speakers from industry, academia, and government.

If Dr. McClellan accomplishes half of what he has set out to do, he will deserve to be elected president of the United States. Otherwise, the Guinness Book of World Records should salute him for having spoken the longest, most complex sentence in the history of the English language. That sentence began on the Longwood campus of the Harvard Medical School at 3:10 p.m. on May 9, 2005. It ended roughly an hour later.

For those who have not heard McClellan's 400-word-per-minute vision of quality and technology in healthcare, it is beguiling and persuasive. Seamlessly connected with networks and software, physicians and patients, hospitals and insurers, pharmacies and drug companies toil to shrink costs and raise the quality of medical care.

"Your conference is an extremely timely one," McClellan told the crowd. But he quickly challenged his audience with the hype of the genomic age. "You can count on one hand the drug products where a patient's genetic profile really matters." Noting that private and public sources spend $100 billion annually to research new drugs and devices, he said: "The cost, the time, the effort that is going into these sciences has never been greater. We don't have a lot to show for it."

McClellan said that the large, hemorrhaging federal health insurance programs are starting to pay for better IT for hospitals and physicians, and to reimburse providers not simply for services, but for provable improvement in patient health. This should not be a radical idea. Yet it is. Better IT is one of the tools he's using to effect the change.

When it comes to hospital and clinical trial IT standards, McClellan offered few specifics but said, "We're supporting standards that have the potential for broad public adoption." It's clear that technology will be used aggressively to save the federal government money. "Electronic systems can do a lot to help doctors make effective decisions," he said.

McClellan called the present FDA system for monitoring drug safety inadequate. "In 2005," he said, "we should not be dependent on spontaneous adverse events reports from overworked providers. We are not learning as much as we should about the effectiveness of treatments. It should not be costly one-off studies that help us get there."

McClellan's vision is of a unified digital infosphere that includes hospitals and pharma. It also encompasses patients, physicians, pharmacies, insurers and outcomes researchers poring over which surgeries/drugs/treatments work, and which don't. Scientific and personal data are used to proactively minimize costs and improve health.

This vision is admittedly futuristic. With small percentages of doctors now using electronic prescription pads, with electronic data capture (EDC) in clinical trials still a novelty and paper-dependent, and with harmonization between hospitals and clinical trials a pipe dream, McClellan's world is not going to arrive tomorrow.

But, wow - there is a grandeur, a sweep to all this that is enthralling.

Indeed, McClellan says critical elements of what he's talking about are already happening at forward-thinking, IT-savvy places like the U.S. Veteran's Administration and the Cleveland Clinic. McClellan's vision is doable, and not as far as away as doubters and skeptics might believe. When it comes, it could have highly positive impacts for society. The impact on pharma is less clear. But it could include higher, broader utilization of any generic and patent-protected drugs that can be rigorously linked to better patient health.

: Featured Events :

Upcoming Industry Events MIT Professional Institute –Summer Session 2005 The Massachusetts Institute of Technology is currently accepting registrations for more than 40 intensive, week-long programs designed especially for professionals. Courses taught on campus by MIT faculty from June through August. Offerings include biotechnology, chemical engineering, energy, IT, mechanical engineering, MEMS, nanotechnology, physics, technology policy, system architecture, telecommunications, and other topics of vital interest to today's engineers and scientists. Please visit our website http://web.mit.edu/mitpep/pi/bioit.html TRAX Pharmaceutical Supply Chain Summit 2005 - June 6-8,  Loews Philadelphia Hotel, Philadelphia, PA With no clear pharmaceutical supply chain leader creating the path to follow, we are on the upslope of an enormous shift and are uncertain as to where to look for guidance and requirements.  TRAX 2005 aims to relieve some of this confusion by providing you with solid case study examples, both successes and failures, on attempts to increase supply chain security and efficiency, updates from the leading regulatory authorities, and the perspectives of the retail pharmacies, all providing guidance as to how to optimize and secure your manufacturing and distribution channels. Visit us at www.traxsummit.com/index.cfm/newsection=yes/brochurekeycode=XE3500BIOIT Personalized Medicine  - June 27 - 28, Philadelphia, PA   https://www.cbinet.com/show_conference.cfm?confCode=HB545  CBI's conference is designed to help companies to assess theranostic market opportunities, develop and execute a model of business development and overcome the regulatory hurdles in bringing both a diagnostic and a drug to market. Key topics include: The FDA Office of In-Vitro Diagnostics (OIVD) guidance on companion product development and a regulatory infrastructure that supports theranostics growth; An industry-wide response to biomarker validation and development ; Strategic alliances to coordinate the development of drugs with diagnostics ;Emerging applications and the clinical utility of new genotyping tests ; Public policy and legislative issues impacting theranostics ;And many more. For more information or to register, please contact the Center for Business Intelligence toll free by phone at 1-800-817-8601 or via e-mail at cbireg@cbinet.com. Electronic Patient Reported Outcomes (ePRO) - July 13-19, Philadelphia, PA https://www.cbinet.com/show_conference.cfm?confCode=HB531  This first of its kind summit provides details on cutting-edge technology, regulatory requirements, validation, security considerations and cost/benefit analysis. The event gives you the opportunity to actively participate in industry dialogue on where ePRO technology can benefit you most, as well as hear in-depth information from a number of blind industry case studies. You receive the critical information necessary to actively apply ePRO technology to your clinical trials. Beyond Genome 2005: The Future or Medicine - June 13-16, Fairmont Hotel, San Francisco, CA.  This event Features programs on Bioinformatics, RNA Interference, Proteomics, Systems Biologym, and new to 2005 Genomic Variation. We anticipate more than 1,000 scientists and over 40 exhibiting companies. visit www.beyondgenome.com for more information.