Bio-IT World Expo: Excerpts from a Reporter's NotebookRather than cobble together eCliniqua's usual weak metaphors, or fuss over precious transitions linking one cliche to another, this is merely a random list of some quotations from the May 2005 Bio-IT World Conference + Expo.
For those who missed the conference, e-clinical track organizers Becky Kush and Ken Getz (more fully identified below) once again did a masterful job of providing timely information and forward-looking insight into a wide spectrum of technologies. The sessions would not have been possible without their sage guidance and considerable labor. The other key participant, one not appropriately thanked at the conference itself, was the Expo's incomparable Kate Greene, who kept 10 billion details straight for months.
So, without further ado, here are a few sound bites from the meeting:
"At some point, we should be doing e-source." - Becky Kush, president, Clinical Data Interchange Standards Consortium (CDISC)
"The industry is faced with disconnected systems, disconnected information. There is no real connection between all of those applications in one coherent fashion. The message about new technology is not about the shiny new box. It's about building an infrastructure over which we can move forward in the coming years." -David Iberson-Hurst, Assero Ltd.
continues below
Percentage of sponsor-funded clinical trials using electronic data capture, interactive voice response, and handheld diaries in 2003: 27
Percentage of sponsor-funded clinical trials using electronic data capture, interactive voice response, and handheld diaries in 2004: 30
"In 2004, the vast majority of sites are reporting they are using a cable modem or T-1 or T-3 lines. A very small percentage, 8 percent, are using a dialup modem." -Ken Getz, Tufts Center for the Study of Drug Development and the Center For Information & Study On Clinical Research Participation
"What is unique is the amount and the depth and the breadth of the data that is part of the collaboration. This is an enormous data integration challenge." - Mike Svinte, IBM, on a collaboration with a Utah nonprofit, LineaGen, to mine 7 million patient records
"We lock databases in a median time of four days, and in one day 20 percent of the time." -Sylva Collins, Novartis
"It's a flawed solution. It will preclude the use of a lot of vendor offerings and new technologies. Using standard, open architectures is a better way of using new technologies." -Gary Lubin, Merck Capital Ventures, on the Novartis approach to EDC
"Don't obsess about the business model. At the end of the day, it's not the business model that matters. It's satisfying the customers." -Michael Martorelli, Fairmount Partners, on which EDC vendors will survive
"We can't act on paper-based data. It has to be electronic." -Jonathan Morris, President and CEO ProSanos, on how best to analyze safety data in clinical trials
"We are going to replace large, confirmatory trials with small, adaptive trials that take much less time. We'll get to market early and accept uncertainty on the safety issue." -Uwe P. Trinks, chief information officer, Sentrx
"I'm frustrated that everybody talks about EDC. EDC is not to me what we're looking for in pharma. As a customer, what I'm interested in is being able to make a decision. Data management is something we do. But it's an operational aspect and it gets way too much attention." -Jerry Schindler, Wyeth
"The market is beginning to distinguish vendors with a website versus vendors with a viable technology." -Paul Bleicher, Phase Forward
"We are building a transactional model that stores data under the HL-7 standard. We will be able to use that data, mine that data, and have it available via our clinical trial management systems." -Andy Alasso, Oracle Corp.
"The FDA wants e-source. CDISC wants e-source. The vendors don't want it. Their systems are no better than they were in 1990. That's why it costs more." -Ron Marks, Clinipace
"The sense of urgency to use EDC is clearly spiking. That's the biggest thing we're noticing. But clients are 'over' EDC in a sense. It sounds funny - adoption is so low. But [they want to know], 'What else can you do, now that you have all my data?' It's not hard to imagine that there is some sort of nexus between electronic health records [EHR], CDISC, and EDC vendors being able to talk directly to an EHR system." -Anthony Costello, Nextrials
"We're not going to have a central patient repository. There will not be a patient identification number. There will not be a central backbone to connect any of this information." -Charles Jaffe, SAIC
Debut of Post-Approval Summit at HarvardDon't call it "Phase IV." Or "post-marketing." The new lingo is "post-approval," thank you. It may be the most lucrative and competitive niche in the clinical arena. Even companies that have only dabbled in post-approval studies are professing to be specialists.
But post-approval work is also a fertile area for scientific reasons, as this reporter learned last month at Harvard's first Post-Approval Summit. Sponsored by Outcome Sciences, the meeting was just two days long. But it was an intellectually nourishing affair that described the adroit integration of pharmaceutical science and technology in a manner that was seamless and medically useful.
A speaker from Pfizer, chief medical officer Joseph Feczko, said his employer has undertaken a concerted effort to design post-approval studies to overcome the limits of spontaneously reported adverse events once drugs are on the market.
"What we need to do is structure post-approval management plans to collect much more prospective data and place less emphasis on spontaneous adverse event reporting," Feczko said. "Until recently, all of us were lax about setting up programs that would look at signal detection. Once there was a signal, it was too late to put in place a reasonable program to evaluate that signal."
After unexpected safety issues with the antibiotic Trovan, he noted, Pfizer began using post-approval research more assiduously. Being proactive, he suggested, meant trying to anticipate a wide variety of questions that might arise. It cannot have been cheap: he suggested that one-fifth of Pfizer's entire R&D spend is on post-approval work, perhaps double the industry average.
The good news? When regulators have questions, usable data are already in hand. "We're making better decisions based on harder data than we would ever have been able to make from the spontaneous event reporting," Feczko said.
The FDA also sent a speaker. Anne Trontell, deputy director of the office of pharmacoepidemiology, did not try to finesse the fact that some of the government's systems to collect and analyze safety data have never been properly funded by scientifically illiterate lawyers serving life terms in the U.S. Congress. No, of course, she never said anything impolitic.
What Trontell did say was completely neutral: "The spontaneous reporting system has many limitations. But we still believe it serves as an important cornerstone. To increase the quality and volume of the reports would be desirable."
So what is the agency doing now, with limited resources? For starters, it's beginning to tap data from other federal agencies, including the U.S. Department of Veterans Affairs and the Centers for Disease Control. In the latter case, using emergency room visits, the first year of data is already being analyzed.
GlaxoSmithKline's Ronald Krall, senior VP of worldwide development, gave the attendees a peek at software from Lincoln Technologies. It displays post-approval data. Large quantities of data could automatically pinpoint even faint signals of potential interest to a drug company. Medical reviewers can even click on a particular square or area of a heat map and learn more about the case, Krall said.
"We are in a position to do much better than spontaneously reported adverse events. It won't be essential that we get physicians and healthcare providers to recognize and report an event," Krall said, explaining that his company planned to forge partnerships with hospitals to get "direct" access to data. "This is an example of how our analytic methodology and information technology allows us to do much richer analysis of these types of events."
If there was one speaker at the Post-Approval Summit who was an especially hard act to follow, it might have been Hal Barron, a cardiologist and statistician at the University of California, San Francisco. He's also working in industry as senior VP of development at Genentech.
Barron stressed that post-approval data should not be viewed as a poor stepchild to randomized clinical trial data. "Observational data is not just a type of research you can do if you can't afford a randomized clinical trial," he said. "There is a huge value that can only be obtained from observational data."
Barron sketched Genentech's National Registry for Myocardial Infarction (NMRI), which since 1990 has collected data from more than 1,600 hospitals and 2.3 million patients in five separate stages. A heart attack occurring at 3:00 a.m. might never be included in a clinical trial, he suggested, simply because of hospital staffing schedules. But a registry could include such a heart attack.
In the case of the NMRI, the database has accumulated far bigger numbers of patients than any clinical trial. The sheer size of the dataset, Barron said, has allowed scientists to understand heart attacks in women and the elderly in ways that could have taken decades or unthinkable sums of money using Phase III studies. "A randomized clinical trial," Barron said flatly, "would never tell you what is going on."
A final example of the caliber of the Post-Approval Summit was C. Geoffrey McDonough, VP of global medical programs at Genzyme. McDonough described patient registries for virtually all of the company's therapeutic areas, which afflict comparatively small numbers of patients. There are only hundreds or dozens of patients in some of Genzyme's registries.
Especially after the Genentech talk, with millions of records, the paradox of small registries was not immediately obvious. But it's obvious that registry data can be just as valuable when patient populations are tiny. Each record becomes more precious -- more scientifically interesting.
At times, the registries help Genzyme decide whether and how to proceed with a full-fledged trial. "The most important thing was to establish a database to establish a longitudinal view of how patients were responding to this therapy," said McDonough, again something that is difficult in chronic disease and in clinical trial that conclude on a fixed date.
In some cases, McDonough reported, the database itself becomes a laboratory apparatus, illuminating previously unknown connections between genotype and phenotype, between two subpopulations of patients -- or related diseases.
Not incidentally, he said, the fact that Genzyme does not completely control the data works in its favor. Outside investigators use the database to publish their own articles in the scientific literature, indirectly bolstering the company's reputation.
Surprisingly, McDonough did not dodge the most contentious issue surrounding post-approval studies. Anti-pharma activists claim that many Phase IV studies are scientifically dubious, undertaken primarily to bribe naive physician-investigators. That notion may apply to selected sponsors in rare cases, but it was gravely wounded by the end of the first day of the Harvard conference.
McDonough allowed that pressure to use registry data for marketing purposes does exist. But he said it's possible (and advisable) to resist it. "The pressure to use registries for marketing purposes is enormous," he said. "That is not what this should be for."
