By Catherine Varmazis
May 19, 2008 | Two speakers at the Clinical and Medical Informatics track at this year’s Bio-IT World Conference & Expo in Boston focused on the increased adoption of the CDISC (Clinical Data Interchange Standards Consortium) standards and electronic data capture (EDC) in clinical trials.
Sally Cassells, VP clinical systems at Lincoln Technologies and member of the CDISC ODM (Operational Data Model) team since 1999, said she has witnessed a dramatic uptick in adoption of CDISC standards in the past few years, citing a 2007 Tufts Center for the Study of Drug Development study in which 70 percent of respondents said they were using CDISC standards in their clinical trials.
Standards for clinical research data are a "huge component" of the FDA's plans to meet the new requirements for increased efficiency and safety monitoring that are demanded by the latest version of the Prescription Drug User Fee Act (PDUFA), passed by Congress last fall, said Cassells.
And in fact, more users are adopting the CDISC ODM in their eClinical infrastructure, she said. ODM is an XML (Extensible Markup Language) model for the interchange of archival clinical research data. It allows all the components that must be included in data interchange to be in full compliance with regulatory requirements.
In addition, ODM metadata provide a hierarchy to describe almost any data requirement, and the model incorporates a full audit trail and signature capability, thus conforming to the regulatory requirements for 21 CFR Part 11. ODM is also being used to automate the setup of case report forms.
Joking that it would take all morning to cover the entire "alphabet soup" of CDISC standards, Cassells said the important point is that all of them are in their second production release at least, and in some cases in their third or fourth, which means they include feedback from early adopters. "In most cases we have models that are fully mature and ready for production use," she said.
Some of the newer CDISC initiatives, like CDASH, support increasing levels of clinical systems interoperability. "Controlled terminology is the third leg of stool," said Cassells, "because once we have standards for content and format, then being able to control what is in that content becomes critically important."
CDISC is also part of a group called BRIDG (Biomedical Research Integrated Domain Group), which is a model that spans CDISC and HL7 standards. The BRIDG group is working to support the eventual integration of clinical research and healthcare data by building a foundation for standards interoperability and consistency using CDISC standards.
"At its core, BRIDG is a UML [Unified Modeling Language] model and a representation of shared semantics for clinical research.... It serves as a communication bridge between domain experts and users of the data, and the technical experts who have to implement tools that work with the data," said Cassells. The model "tries to define the what and how of clinical research in amazing detail, so when people go to implement tools to support those processes, they have all the details they need to ... allow the tools to interoperate."
Cassells conceded that CDISC standards are not yet perfect, but predicted that as they are implemented and user feedback is incorporated into new versions, "minor imperfections will be shaved off and [users] will begin to capitalize on their investments."
Despite the good news on the standards side, Ronald Waife, of Waife & Associates Inc., said many implementers of EDC are not using it to best business advantage. Although EDC is now used in more than 50 percent of clinical trials, Waife said most users are not really exploiting the technology to make fundamental improvements in clinical development. Instead, EDC is often used to create an “Internet veneer” over linear, paper-based workflows, he argued. This creates a mismatch with the electronic clinical trials process, which takes in data from a variety of electronic sources simultaneously.
Clinicians and data managers should be re-imagining and re-designing the entire clinical development process, Waife said. Part of the problem is a lack of understanding between the clinicians and the data managers. He suggested they educate each other by having the “difficult conversations” about what they each need, and “walk a mile in each other’s shoes.”
“How many data managers have ever visited a site or met with a patient with the disease being studied?” he asked. And clinical research associates should be asked to work out the data validation plan for an electronic case report form module so they each understand what issues the other is struggling with.
Re-imagining the eClinical workflow process is hard, but essential, Waife said. Failure to do so costs the industry millions.
This story first appeared in eCliniqua,one of Bio-IT World’s free e-newsletters. Subscribe here.