Wayne R. Kubick

The show featured over 800 mostly exuberant exhibitors, brandishing products ranging from intelligent hospital beds, mobile computing communications and storage devices, and computers to laboratory, radiology, and surgical devices.

The wide range of software products being offered included, of course, Electronic Medical Record (EMR) and Electronic Health Record (EHR) systems, along with practice management, nursing, retail and scheduling systems, data information exchanges, telemedicine, financial accounting, human resources, and systems for cardiology, dietary management, emergency medical rooms, labs, and pharmacies.

Of course, this is not your father's DIA—it's another world on an entirely different scale. Beyond the endless aisles of exhibits generating widespread awe at such exorbitant booth giveaways as Vespas and SeaDoo jet skis, what impressed me most was how small and insignificant our world of clinical research seems by comparison.

The companies in this aggressively competitive world of health care have many things on their enterprising minds, but the quotidian problems of your average study coordinator, CRA, and data manager don't appear to loom very high on their lists.

In fact, a HIMSS position paper¹ attempting to define "meaningful use" (i.e., what the stimulus health IT money is intended to provide) has recommended a phased, steadily increasing rate of adoption based on such measures as order entry by physicians, clinical decision support, and ePrescribing—but not a word on clinical research. Though some of these—the ability to exchange summarized patient data—could factor into clinical research integration scenarios.

There were exceptions

The latest iteration of the Integrating the Healthcare Enterprise (IHE) profile (cosponsored by CDISC and various sponsor organizations) called Retrieve Form for Data-capture (RFD) shows how an EHR system can transfer existing and incremental clinical data for research purposes, such as exchanging it directly with an EDC system.

One particularly arresting early implementation of RFD is the ASTER project, sponsored by Pfizer and others, recently lauded in a Wall Street Journal article,² and currently undergoing pilot use at Brigham and Women's Hospital in Boston. ASTER has shown how an EHR system can detect a potential drug-related adverse event from a patient record and offer the physician the opportunity to automatically send an adverse event report to the FDA without the burden of re-entering all of the necessary information by hand.

Early indications are that solutions like ASTER, if widely implemented, could dramatically increase the likelihood of reporting adverse events, and allow safety reviewers and regulators a more complete picture of what's happening in the real world.

Low hanging fruit

Another example shown at HIMSS is a current live RFD study involving Outcome Sciences and Greenway Medical, indicating a significant leap ahead from proof-of-concept demos in prior years. Yet it still seems that true integration between so many different types of EHR and EDC systems with the correct mapping of so many clinical data fields would require significant effort from technical staff on both sides, and the scalability to a significant proportion of complex studies in a validated, compliant manner (including the many unanswered questions regarding the regulatory acceptability of such eSource) seems dubious.

And the current demos do not appear to be built primarily on HL7 Version 3 messaging standards, which are generally being depicted as the backbone for future health information exchanges of all kinds—including that with clinical research. Yet, it's certainly interesting to see these initial forays with boots on the ground.

Beyond this, there wasn't a lot to be seen about clinical trials on the radar screens of the major players—though a few companies such as Cerner and Allscripts did allude to some degree of clinical trial support in their exhibit booths. Meanwhile, there are other alternatives worthy of consideration that did not make obvious appearances but might merit attention as lower hanging fruit.

For example, how can we better leverage existing EHR databases to identify candidate subjects who might be interested in participating in clinical trials? How can we query and mine electronic patient data to better understand medical product safety and comparative effectiveness (another priority of the stimulus package). Or, simplest of all, what else can we do to make the conduct of clinical research easier on the sites—without resorting to major system development or integration projects?

So the question is, how do such promising demonstrations become production systems in research environments? Support for the types of features necessary for clinical trials are not necessarily high on the list of most health care providers looking to first automate their basic processes such as orders and billing or integration with existing clinical systems (often requiring use of older HL7 V2 standards), or EMR vendors vying for market share in a very competitive and fragmented marketplace.

And unless these features are installed in the first place it may be a while before they ever show up due to the extreme complexity, cost, and risks to upgrade such systems.

Meanwhile, the EHR industry is facing a grim reality simultaneous with this moment of opportunity. An article in the New England Journal of Medicine³ has estimated that only 9% of U.S. hospitals have a functioning EHR system. And with the kick of the stimulus program comes risk, as industry pundits are predicting widespread consolidation in the industry, which many of the smaller vendors won't survive.⁴