Parallel Universes?Technology in Health Care and Clinical Development

Although the worlds of EMR and clinical trial technologies are similar, they have quite a few differences as well.
Apr 01, 2004
Although the worlds of EMR and clinical trial technologies are similar, they have quite a few differences as well.
The mind-bending premise of The Matrix is that there is a parallel world, coexisting with ours, but not apparent to any of us. This concept has long been part of science fiction, great literature (the short story “The Garden of Forking Paths,” by Jorge Luis Borges), and has now taken a place in modern physics with the gradual acceptance of string theory.

The idea of a parallel universe is a very nice metaphor for the two worlds where technology touches clinical data—our world of clinical trial technology and the parallel universe of electronic medical records (EMR). These two worlds are so close to one another that the general public assumes that they are, in fact, one and the same. The primary end user is the same (the physician and nurse), the data collected is very similar, and the requirements for privacy, security, and data quality are more or less equivalent. Once we recognize that these two worlds exist, we have to wonder—how are they the same? How are they different? And, where is the most logical point of connection between them?

Adoption technology in the two worlds has run in parallel as well. Until a few years ago these technologies were used for less than 5% of data collection, and there were tens to hundreds of vendors offering technologies that were difficult to distinguish from one another. However, with the emergence of the Internet, enterprise computing, and the general societal adoption of electronic technologies, the adoption of these technologies is rapidly rising and the spectrum of vendors is narrowing to a smaller number of larger vendors.

Despite the obvious similarities, the EMR and clinical trial technology worlds are very distinct from one another. The most obvious reason is that the end “consumer” of the data is very different in each case. In the world of the EMR the consumer is either another physician or third-party payors. Therefore, these systems are bought by hospitals, clinical practices, and insurance companies.

Clinical trial technology, in contrast, is typically purchased by pharmaceutical, biotechnology, and medical device companies who are conducting clinical trials. The types of data collected are very similar, but quite differently structured; while clinical trial data is highly structured by visit schedule and expected content, EMR systems must be able to capture a much larger collection of data types with far less standardization. The result of the differences in buyers and purposes is that EMR and clinical trial technologies have developed in parallel (driven by the availability of technology and process needs) but almost completely apart from one another.

My feet are firmly planted in the clinical trials technology universe where I typically spend my waking hours. If you don’t live and breathe clinical trials technology, you can always get a good sense of it by walking the exhibit hall floors at the Drug Information Association Annual Meeting in June. The floors are filled with technology companies from large corporations to small entrepreneurs. All companies on the floor work hard to showcase the latest technology that they have developed. If you spend some time looking at the booths, you will understand the “state of the art” of technology in clinical trials for the coming year.

The parallel of the DIA annual meeting for the EMR is the HIMSS meeting, held annually in February. I attended HIMSS this year, and had the opportunity to walk the halls and experience this parallel universe. The HIMSS meeting is truly impressive in scope and scale—in headcount alone it is over three times the size of the DIA, approaching 20,000 attendees. My expectations of the parallel but nonintersecting nature of this universe was confirmed—there were very few people from the pharmaceutical industry that I saw, and almost no presence or acknowledgement of the industry on the trade show floor. The exhibit floor itself was teeming with technology vendors displaying applications that were very similar to those shown at the DIA. Yet, the companies were almost entirely different entities than those at the DIA.

The general feel of the meeting was that of an industry in transition from individual “point solutions” to an integrated, enterprise suite of applications addressing the entire patient life cycle from registration, scheduling, patient input, physician data, lab data, record archiving, billing, referrals, third party payments, and notifications. Each of these has existed as standalone solutions, but companies like GE Medical Systems, Eclipsys, IDX, and Siemens are weaving these together into integrated solutions for practices and hospitals. The true value of these products is becoming apparent to institutions more as a suite than as individual solutions. This trend also parallels the adoption of technologies in many industries in the past—including financial services, customer relationship, supply chain management, and of course typical desktop applications.

Another notable trend was the very prominent presence of major companies in the information technology industry—companies like Microsoft, IBM, and Hewlett Packard. The news coming from the upcoming DIA show is similar. Clearly, health care IT and pharmaceutical IT have reached the level of attention of these giants. With that comes much innovation from smaller, more fleet-footed, entrepreneurial companies. This kind of attention brings long-term benefits to the buyers of this technology.

Amidst the sea of exhibits, I was able to discern a few very revealing trends in the health care industry that clearly have parallels in the pharmaceutical clinical trial market.

Enterprise portals
As more and more processes in health care become fully automated, decision makers are beginning to demand near real-time information from most of these. In the past, this meant opening each application—some legacy and some new—and using the native reporting tools available to view and explore information. This process is frustrating and painful for the average business user, and most industries are moving towards an integrated view of the various systems and data available through an enterprise portal. You can think of an enterprise portal much like a “My Yahoo,” whereby a Web page is broken down into 6–12 component sections. A user might think of this as a single Web page, but each of these sections represents a view of a particular data source, which is probably functionally distinct from the other sources, and may even be geographically distant.

An enterprise portal can tie together information from dozens of different systems and display this in a uniform environment. One section of the portal may represent a document room, with key forms and information available for viewing, printing, or editing as necessary. This could serve to provide consistent, centralized information or provide a central file storage for collaborative projects. Another section might be a window into secure email—where encrypted, secure, HIPAA-compliant email could be viewed, sent, and archived. A third section might have news and conference information, streamed from other Web sites through RSS Web services. Finally, one or more sections may be dedicated to reports from enterprise applications.

When designed well, the views can have three levels. First is an “alert” level which has little information, but can keep managers abreast of overall status of a project. Think of a company that manages a group of hospitals across the country and needs to monitor bed availability. The alert level might be a set of gauges indicating percent occupancy of individual hospitals colored red, yellow or green to indicate the presence of an issue.

Next is a “report” level with information summarized in tabular form, for example a list of hospitals and units within these hospitals with bed occupancy, and types of patients (male/female, medical/surgical, etc.). This more detailed view may be available to an upper-level administrator by clicking on an alert level indicator, or it may be the opening screen for an operational project manager. Finally, a click on an individual unit could lead to the third level—the data level. Here you would view individual patients, with ID numbers, diagnoses, etc. This information might only be available to some users, depending on their rights within the system (and identified by their login authentication username/password).

The development of enterprise portals is an intermediate step in the true integration of systems and processes. These portals don’t require systems to be fully integrated; they only need the ability to query and display information from a variety of disparate applications. Whether they are 20-year-old accounting systems or state-of-the-art EMR systems, they can appear to be part of a single integrated system through an enterprise portal.

Enterprise portals are being designed and implemented for pharmaceutical applications as well.

Integration brokers
The next trend is true system integration, an ideal that is becoming more achievable through the development of standards; in the case of health care this is being led through the HL-7 effort. Beyond the standards, though, is the development of “integration broker” technologies, often based in XML standards. In the past, connecting individual software applications required direct “plumbing” between the two applications—a manual job that required adjustment or even a complete redo every time one or the other application changed. When considering a hospital system with dozens of applications, the work of integrating these applications grows at an exponential rate. Enter the integration broker.

Integration brokers work through managing messaging through a common information bus. The broker “listens” to data coming from one application, transforms it into a data schema that is understandable by another application, and routes the newly transformed data to its target. By providing bidirectional messaging between applications, the integration broker makes it possible to weave together software solutions so that they work together. Thus, a physician order entry system can communicate with a drug safety system, a clinical pharmacy robot, and a billing system through an integration broker, ensuring that drug interactions are noted and that the proper drug is dispensed and billed to the patient.

The advantages of integration brokers are not lost on pharmaceutical clinical trial technology. More on that in a future column.

Computerized physician order entry
Most of us are focused on the collection of clinical data, and when we think about bringing technology to the overall health care arena, we are usually thinking about EMR. However, if you follow the money you will realize that the area of greatest activity and progress is in software that supports the flow of funds in health care.

Another obvious area for the application of technology is in computerized physician order entry (CPOE). The advantages of such systems are overwhelmingly obvious upon brief consideration. In the narrowest definition, such a system would correct the most common serious errors in medication—illegible handwriting, drug interactions, and inappropriate doses. These address the major issues in the widely quoted report, “To Err is Human,” from the Institute of Medicine. Despite this significant value, general estimates indicate a penetration of 10% or so in hospital settings. Most hospitals are waiting to integrate such systems into an overall lab ordering and chart reviewing system, allowing a complete feedback loop in patient management.

The growth and strength of CPOE systems on the HIMSS floor was quite apparent. One of the drivers in their adoption is the increasing availability of Tablet PCs that accept handwriting. One such fully functional Tablet PC (with a built in keyboard and wireless functionality as well; is not much larger than a Palm Pilot. At this size/feature point, the Tablet PC may finally enable fully mobile CPOE and even electronic medical records in hospitals and physicians’ offices.

Observers of the overall technology arena have noted an explosion of interest in radio frequency identification (RFID). You can think of RFID as a next-generation bar code that doesn’t require line of sight or any specific orientation. This technology involves a tiny chip with an antenna that can be attached to packages or incorporated in personal IDs. The chip can be programmed to contain detailed digital information. When the chip is “interrogated” by an RFID reader it acts as a transponder, broadcasting its information in a local area. The information is gathered by the reader and can be modified as well. Thus, a box containing packaged pharmaceuticals with an RFID tag can be shipped across the country from manufacturer to distributor and pharmacy, carrying and transmitting details of the journey.

The potential uses of this technology are only limited by our imaginations. Certainly the FDA is thinking about these issues. With the recent bar code regulations requiring bar codes on prescription drugs and blood products, and a patient-matching system in hospitals, the FDA places a lot of confidence in Auto ID technology. More recently, the FDA issued a report on counterfeit drugs, with the conclusion: “The adoption and common use of RFID as the standard track and trace technology, which is feasible in 2007, would provide better protection (against counterfeiting of drugs).”

One particularly interesting use of RFID was prominently demonstrated at HIMSS. People in one booth wore ID badges containing an RFID chip. In another booth it was possible to follow the movement of these people on a floor diagram. The value of this for following staff in a hospital, patients in a Phase I unit, and other people is worthy of consideration.

Best of both worlds
In most parallel universe stories, the two worlds often may approach and even cross one another. Our worlds of health care technology and clinical trial technology are much like this, sometimes coming very close, but always advancing in parallel. The most obvious connection between the worlds is the direct transfer of data from one to the other. This is currently a very difficult effort whenever it is attempted. It is not scalable to the extent that would be needed to cause a paradigm shift in approach to the conduct of clinical trials.

Efforts in this direction are beginning through the combining of standards efforts from CDISC and HL-7. However, much of the connectivity will likely come through the development of sophisticated integration brokers, perhaps using the CDISC/HL-7 standards as a base. Many assume that these two worlds will eventually grow together over time to form one common future. While this seems likely, I must close with the obvious question: “How long will it take?”