By Deborah Borfitz
August 18, 2008 | Clinical investigators like the idea of collecting patient data once and using it for both medical care and research purposes. In academic settings, many faculty members adequately handle the task with some sort of homegrown electronic data collection (EDC) system, points out Stephen Woody, associate chief information officer for clinical and translational research for Duke Medicine.
Although the wholesale integration of EDC with electronic medical records (EMRs) is many years off, says Woody, it’s only sensible to improve the workflow of data collection in the interim. The business of study monitoring and site management could also use an efficiency boost. A study conducted by Eric Eisenstein, DBA, assistant professor of medicine at the Duke Clinical Research Institute (DCRI), found that 60 percent of sponsor dollars were consumed by such “coordinating” activities. The more that figure can be slashed, Woody adds, the more money will be available to do additional research and better compensate high-enrolling sites.
Information collection is financially stymied by concerns that data be pristinely clean, or at least be perceived as such by the FDA. “Early on, the DCRI didn’t want to touch clinical data,” says Woody. “It was seen as too dirty and not adequately defined.” As it turns out, “clean data means [surprisingly] little to an adequately powered study.”
 |
| Stephen Woody |
Meanwhile, progress is being made by DCRI and government initiatives such as the Cancer Biomedical Informatics Grid (caBIG) to develop a common language to more easily share information across disparate sites and information systems. The Clinical Data Interchange Standards Consortium (CDISC) also has two initiatives to enhance the healthcare-research interface.
EDC vendors have a vested interest in keeping the “dirty data” myth alive, says Woody, because the systems they sell generally have “query rules” for picking up data inconsistencies that so trouble FDA-wary sponsors. Clinical research organizations, DCRI among them, also “seem to feed off the inefficiencies of how we do clinical research. Technology needs to take over what humans do now to free them up for higher value work.”
A statistical method also needs to be developed for study monitoring in lieu of “putting people on planes all over the world,” continues Woody. “For the time spent, not much work gets done.”
Much work remains to be done before clinical and research data can have a common collection point, says Woody. The first problem is that information collected in the clinical setting is not “scoped appropriately” to meet the demands of research protocols. “Angina is a general term that could be used by doctors in different ways,” he offers as an example. “But for a protocol, it may be defined in excruciating detail.”
Whatever is ultimately developed will have to factor in the current work habits of whoever is inputting data, as DCRI learned while doing its Starbright demonstration project a few years ago, says Woody. “We watched what study coordinators did and, to our surprise, they collected CRF [case report form] data first and then populated the EMR.”
That realization triggered CDISC’s current efforts to “combine workflow” rather than attempt to collect data simultaneously for clinical and research purposes, says Woody. Study coordinators may one day be able to access a CRF while working in an EMR environment with some data items, such as patient demographics, automatically pre-filled.
One complicating factor is that “EDC vendors want to continue to own the space and provide the interface to get the data,” says Woody. The futuristic data-capture model envisioned by CDISC and partner IHE (Integrating the Healthcare Enterprise) “will turn EDC vendors into data stores.”
________________________________
This story first appeared in eCliniqua,one of Bio-IT World’s free e-newsletters. Subscribe here.