The Coalition Against Major Diseases started one and a half years ago to develop tools that could improve drug development. These tools are to be pre-competitive, i.e., open to all; hence, CAMD forms a consortia to build them. The idea is that everyone has much at stake, but no single player can build these tools or get the field at large to embrace them. Importantly, no single company can elicit nearly the same degree of regulatory engagement that a neutral coalition enjoys. So despite a certain “consortia fatigue” that can set in with large collective efforts, collaborative work is CAMD’s bread and butter. “Consortia are the model to effect outcomes in the pre-competitive area. You can’t do it with any one group,” Rick Myers of C-Path said at a CAMD conference held 30 November 2010 in Washington, DC. For its projects, CAMD has recruited representatives of seven patient groups and foundations, 12 biopharma companies, leading academic scientists, scientists at the Food and Drug Administration (FDA) and the European Medicines Association (EMA), as well as the National Institute on Aging (NIA) and the National Institute of Neurological Disorders and Stroke (NINDS). More companies will soon join, said CAMD’s executive director Marc Cantillon.
Besides herding these cats, what is CAMD doing, exactly? The coalition has five working groups: on data standards, a disease model, on biomarkers, statistics, and on government submissions. To start with number one, a clinical data standard for all AD trials is the “rock on which all this is built,” said Cantillon. [Editor’s note: Yes, dear reader, here now a few graphs on standardization. Read on, it’s not so bad!] Standardization is key to getting information to reviewers at the FDA/EMA and to ensure they digest it efficiently. Even though about half of submissions these days come to the agencies electronically, that does not mean they are standardized. Even banal things like age and sex vary from one submission to the next, slowing down the reviewers as they try to make sense of the material.
Data standards are soporific to most people, but speakers insisted they deserve an image makeover. “Standards are far and away the most important goal with respect to savings and understanding AD clinical trial data,” a Pfizer scientist said. “Data standards allow for collecting and interchanging data across programs. Our field is an intellectual power of Babel. Each investigator has a few trials and speaks from that. The variability seems huge. Everybody is coming to the table with a little piece of the truth. If we can come to the table with standard data, CAMD has been a victory already.” Marc McClellan, former head of the FDA and now at the Brookings Institution, weighed in with this: “Data standards in Alzheimer’s and Parkinson’s are really really—and I’ll add another really here—important for the efficiency of the regulatory process itself. We have to create momentum behind these standards.”
CAMD has developed a data standard for AD trials together with the Clinical Data Interchange Standards Consortium. With its more than 250 member organizations, this global nonprofit establishes freely downloadable industry standards in many areas of technology and science. Bron Kisler of CDISC showed a slide of five different global outlet plugs to symbolize the problem of disparate datasets that connect only when each is hooked up to a universal adaptor plug. CDISC and CAMD scientists have developed such a universal adaptor, i.e., an AD clinical data standard. They will publish this standard, along with a guide on how to implement it, for free download and comment on the CDISC website in January 2011. These standards use defined terminology and data structure (for more on these, read up on STMD and CDASH on CDISC’s website. The CDISC website lists a series of data standard training workshops, too.
This voluntary standard is useful in several ways, scientists said. Existing clinical trial data can be mapped to this standard and in this form flow into a pooled database (see below). Going forward, trialists can opt to use the CDISC data standard from the get-go for new trials. This is preferable, speakers agreed, because aligning existing data from old trials to a new standard is difficult. Leading the pack, the Alzheimer’s Prevention Initiative, which intends to start pre-symptomatic trials in high-risk populations as early as 2011, has indicated it will adopt the CDISC data standard, said Klaus Romero of CAMD.
Kisler emphasized that a disease-specific data standard improves collaboration in many ways. It makes it easier to collect data, to exchange data, and to aggregate it so scientists can analyze much larger datasets. “CDISC’s open-access philosophy made it our key partner on standards,” Cantillon added. “With them, we are developing downloadable Alzheimer’s and Parkinson’s case report forms in this new standard for download by fall 2011.”
As a neutral convener, CDISC has become the FDA’s partner on standards, as well, said Josh Benner of the Brookings Institution. The agency not only welcomes submissions in CDISC standards to speed internal review, it also calls for their use in one of the main FDA projects launched with stimulus funding from the Recovery Act. The Partnership in Applied Comparative Effectiveness Science (PACES) aims to facilitate meta-analysis across studies to pinpoint subgroup-level differences in effectiveness. This goal echoes in the mind of AD clinicians, who at times spotted tantalizing hints for promising subgroup effects (i.e., in the less advanced, in ApoE4-negative patients), but those hints generally lacked sufficient power to inform next steps. For PACES, the FDA needs a clinical trial repository of datasets collected using common CDISC standards. “This project is now underway,” said Benner, adding that PACES serves as an example of how CAMD is aligning its work with that of the FDA.
For its part, the NINDS has begun its own initiative to develop common clinical data elements that the institute’s grant recipients can use in their clinical research. Petra Kaufmann from NINDS noted that this effort currently covers traumatic brain injury, spinal cord injury, epilepsy, stroke, frontotemporal dementia, and other conditions. Standards for Parkinson’s are currently out for public review at the agency’s website. In discussion, Steve Broadbent of C-Path cautioned that having separate standards—a CDISC standard, a NINDS standard—might confuse investigators. Kaufmann said that NINDS will ensure that its data elements are compatible with CDISC’s standards so extensive remapping for data-sharing hopefully won’t be necessary down the road.
Scientists at the conference debated how best to get individual researchers to adopt these standards. Science is enough like art that investigators, given their druthers, tend to work like individualists, pursuing their own ideas and methods. Some attendees urged NIA and NINDS to use its power of the purse to mandate use of standards so individual investigators stop spending scarce federal dollars on developing yet more separate data collection instruments. Other preferred the gentler power of persuasion. “We want the standards to be really good so people will want to use them,” Kaufmann said. Others cautioned that experience has shown individual investigators to resist using even excellent common standards, and that years might be lost until investigators adopted them “organically.” Short of a mandate, adoption will require at least teaching effort, scientists agreed. “We all love our own way of doing things,” said Dan Van Kammen of the Cure Huntington’s Disease Initiative. “Our task is to bring people together to see that the standard is better for them.”
For its part, the CAMD-CDISC data standard is already being put to work in a pooled clinical trials database for all researchers to peruse. For the latest on that, see Part 3.—Gabrielle Strobel.
- Phoenix: Vision of Shared Prevention Trials Lures Pharma to Table
- DC: Shared Pain Is Lessened—Open-Trial Data Gain AD Model
- DC: CAMD Convenes Stakeholders to Reform Alzheimer’s Trials
- DC: Biomarkers, Parkinson’s—CAMD Needs All Hands on Deck