By Deb Borfitz

September 21, 2010 | If clinical trial data were sourced electronically rather than on paper, pharmaceutical companies could reap huge savings—most notably by eliminating site monitoring for purposes of source data verification (SDV). Yet e-sourcing remains a relatively rare practice, in part because sites find it inconvenient to enter data directly into electronic data capture (EDC) systems. Double data entry is the norm even at sites utilizing electronic health records (EHRs), since they have no way to automatically populate electronic case report forms (eCRFs) with trial-relevant data. Although integration of EDC and EHR systems has been a topic of discussion for two decades, the economic downturn has helped confine the idea to a handful of low-risk, early-phase study pilots, says Jim Rogers, CEO of integration champion Nextrials.

Jim Rogers

Although both the FDA and EMEA have voiced their support of efforts to collect source documents electronically and auto-populate the database, U.S. pharmaceutical companies have been “more conservative” than their European counterparts in interpreting regulations, says Douglas Bain, vice president of operational excellence at Medidata Solutions. A number of European-based companies support the use of EDC for e-source data capture provided the approach is backed by standard operating procedures. “The industry needs to get to the point where most if not all data is entered directly in EDC systems, either through electronic transfer [from lab or EHR systems] or keyed… directly as e-source. Sites do not like transposing data from worksheets or other source documents into EDC.”

Electronic transfer of study data from an EHR to an EDC system, though technically feasible, remains problematic, says Bain. CDISC standards are aligned with Health Level Seven (HL7) standards for information interchange among healthcare-oriented computer systems. But broad utilization of CDISC and EHR remains patchy. EHR/EDC data interchange in any case requires comprehensive mapping and ground rules between the two systems as well as a good deal of patient trust, especially as it pertains to informed consent. Trial volunteers consent to share only study-relevant data, as spelled out in the protocol, not necessarily their entire health history. 

Douglas Bain

The bigger problem is that EHR systems themselves are differentially standardized, except in Denmark where data is stored in a common format as a matter of national policy. “[Standardization] makes it easier for companies like us,” says Bain. “Systems integrators have only to create a single interface.” Ultimately, a “point and click” tool will emerge to do high-level mapping between different sources and destinations, making it applicable to increasing numbers of studies and accelerate adoption of the enabling CDISC standards.

Nextrials has used CDISC’s RFD standard to become one of only two EDC vendors that has demonstrated (at both DIA and CDISC meetings) that it is realistic to “surface” a study-specific CRF in the EHR environment, with appropriate fields auto-populated from existing EHR data. Nextrials in also utilizing another CDISC standard (RPE) allowing the two systems to incorporate protocol definitions into the configuration, addressing both regulatory and privacy concerns, says Rogers.

GE Healthcare IT is among a number of leading EHR vendors—multiples the size of the largest EDC player—that has joined the integration crusade. “When it comes to EDCs, we recognize the FDA holds source information as critical,” says Mark Dente, MD, chief medical informatics officer. Combining the technological prowess of EHRs and EDCs can also “streamline and bring cost savings to the drug development process,” allowing “certain cost-prohibitive therapies… [to] see the light of day.”

Two Ways to Reduce SDV
A key challenge to direct data entry, particularly in hospitals, is getting an EDC system in proximity to visiting study participants. Even with the availability of tablet PCs and iPads, e-sourcing doesn’t always make sense, says Bain. Certain psychiatric patients, for example, might grab an unsecured device and throw it across the room.

Typically, direct data entry is done “sporadically” by sites whenever it is expedient to open a laptop in patients’ presence as opposed to delving into their medical charts and patient histories for information, says Rogers. EDC systems currently don’t allow studies to be designed so sites can specify which data points are being e-sourced.

At some point, sponsors will probably start factoring the ability and willingness to do direct data entry into how they select sites for studies, says Rogers, because it makes trial data available “sooner and cleaner.” But if portable devices are going to be used in lieu of EDC, for both data entry and storage, the information should be encrypted on the hard drive. That’s more the exception than the rule today.

On studies not currently being e-sourced, sponsors are already thinking about limiting SDV, says Bain. The Targeted SDV module within Medidata Rave, the company’s platform for EDC and clinical data management, allows companies to establish and execute their own rules about this. “Provided appropriate control procedures are in place and justified, we believe a targeted SDV solution meets applicable regulatory requirements.”