While it supports a cottage industry of professionals, manually pushing piles of clinical trial data along from one system to the next isn't necessarily the most efficient way of doing things. The same company may map data for the same drug in different ways in different trials—or miss errors in one trial's data set that were fixed in another. That lack of consistency isn't necessarily noncompliant, or evidence of bad intent, but it does look sloppy to outsiders and regulators.
In a skeptical era, SAS and Medidata Solutions are trying to automate the movement of data from the systems that collect it to those that manage analysis and submission. The two companies have connected their respective products for electronic data capture and statistical analysis. The linkage is a somewhat obscure offering from the analytics firm called SAS Clinical Data Integration.
The new linkage validates the electronic pedigree of clinical data along its journey, maintaining a record of who had access to the files and what happened to them. "The information that is so important for clinical research never slips out of the protected envelope from start to finish," says David Handelsman, advanced analytics strategist at SAS.
"We've heard from a number of our customers that what they really want is a comprehensive, end-to-end approach for moving data from capture to analysis," he says "This is a way to get it done."
The SAS Clinical Data Integration product was quietly launched late in 2009. In theory, the software could be used to connect other electronic data capture (EDC) systems and clinical data repositories offered by other suppliers. But for now, out of the box, it's working only with Medidata's EDC product and depositing the EDC files into SAS Drug Development, a secured environment for collaboration and data analysis.
Handelsman notes that at some companies, clinical trial data are still routinely transferred manually, or by a physical CD. That gap between two regulated systems, Handelsman says, is "the weakest link in the chain. If information gets out of their control, and something unexpected happens, it's very difficult to document what happened or how it happened."
Medidata and SAS announced their relationship yesterday, at the 2010 Drug Information Association meeting. At the same event, Indian contract research organization QED Pharmaceutical Services, announced that it is using the SAS Clinical Data Integration software. Here's a release on that news.
The deal is hardly the first data standards initiative at either firm. But the pact is interesting for the more wholistic view of the clinical data life cycle than is typically found throughout the industry, where entire companies (or buildings filled with people at large sponsors) are devoted to a single stage of collecting or processing trial data.
SAS and Medidata have both built their franchises to support data standards. While no sensible person could stand against motherhood, kittens or clinical data standards, it's also clear that the industry's idiosyncratic, elastic definition of the word "standard," often rejiggered for each trial, have rendered the term all but meaningless in its usual English sense. The term "standard" within some large life science companies is not strictly defined in the manner that allows a USB stick or an electrical plug to easily inter-operate with devices from other companies.
But Handelsman is quick to point out that interest in clinical data standards is anything but moribund. Indeed, he says a recent FDA-Drug Information Association meeting (referenced in this story) allowed the agency to revitalize its campaign to slowly chide the industry into adopting standards. Handelsman says that while data standards are not being mandated, the agency is getting attention with its vow to give the most rapid reviews of submissions to companies that adhere to them.
"The interest in Clinical Data Interchange Standards Consortium (CDISC) has accelerated in the past few months," says Handelsman. He says the FDA's informal, nonbinding endorsement of CDISC standards probably affords the agency more latitude and flexibility than a formal regulation might. "They're the government. They are not going to say it as explicitly as we would like. If you provide data in another way, it seems like it would take longer to perform that review."
In the contract research organization (CRO) industry, Handelsman says, some firms are waking up to the fact that converting or working with CDISC data sets can be a business unto itself. "This type of solution helps them do that," he says, citing the fact that bonafide SAS programmers will not be required to use the integration software. Instead, well-trained data managers and other non-PhD, non programmer types will be able to handle the software and pass the data from the world of EDC to the statisticians. That will reduce the demands on already sought-after SAS programmers.
The Analytics Age
The amount of such projects to convert or move older data could be sizable, Handelsman says. "There is a lot of elephant to eat in terms of data standardization. We're going to see that pick up steam throughout the year."
Handelsman is taking on new responsibilities at SAS in the analytics arena. The goal is to extend the company's analytics tools for all industries into the life science research sector.
As such he is helping SAS fend off mammoth general technology firms (like IBM, HP and Oracle) that are eyeing the juicy life science sector. But SAS is not ceding any territory on that front. "This is what we do. People can keep coming at us with whatever they want. Our capabilities are real, and unmatched. A lot of people can throw around the term analytics. We actually do this. It's our business."d9A2t49mkex