Originally Published September 2000
In the years since National Instruments (Austin, TX) released its first version of LabVIEW software, the field of virtual instrumentation has slowly begun to attract proponents in the world of medical devices. Engineers no longer view the instrument panel solely as an array of fixed hardware components. Jon Olansen, PhD, software engineer at United Space Alliance (Houston), expects great strides to be made in developing highly useful medical product applications of virtual instrumentation technology. He concedes, however, that traditional hardware engineering still dominates. "Virtual instrumentation is making inroads in medical device testing, but to bring that to clinical instruments, we need to build a history of product development." Ultimately, Olansen and others who are exploring the possibilities of virtual instrumentation hope to cut costs in medical device design and increase the usefulness of data-gathering systems for healthcare professionals.
"The basic idea is to mimic hardware with software," says Olansen, "which allows engineers to design something that performs the function of many instruments by using a computer. Today, we can add to the hardware functions to provide much more in the way of analysis using information from the same signals providing input to fixed instruments." One goal of this work is to incorporate the most important components of different sensors into one unit. Virtual instrumentation can take products to the next level of flexibility; it extends the capability of existing instruments without an intensive research and development effort.
Instead of displaying a single data set in one format on a fixed display, virtual instrumentation allows users to view data from various sources displayed in a number of configurable formats. Olansen suggests that, building upon this concept, engineers can develop new applications by rethinking the traditional ways in which they design instrumentation.
Jeffrey Travis, director of engineering for Compuware Corp. (Austin, TX), is another developer looking seriously at virtual instrumentation issues. He notes that one of the major benefits of this avenue of development is in reducing the costs of new technology, which will drive the market. He also foresees the efficiency of software development eventually taking over many of the domains of traditional hardware development. In designing medical instruments, he explains, many components could, in theory, be made interchangeable between different devices. "We have three layers: one is data acquisition; the second is signal processing, which is where the software components largely begin; and the third is the output of information," he says. That final step, information output, has traditionally been the realm of hardware, though tools now exist to allow engineers to model the interface in software. Travis contends that the nature of software development has inherent advantages over hardware engineering, and that cost benefits will be realized by both developer and user. "If we can emulate the functions of our instruments, or many different instruments, in software, we reduce the cost of ownership drastically."
Virtual instrumentation is intended to mimic hardware functions. Photo courtesy of North American Dråger.
An individual patient's treatment course may involve many electronic medical instruments, from which specially designed devices collect data. If the configuration of electrode sets and signal-conditioning parameters can be modeled in a software environment, the software can take over many of the tasks now handled by specially designed hardware. The cost of ownership decreases, Travis explains, when information can be represented on generic information devices rather than on single-purpose, set-configuration displays.
For the instrument user, however, the benefits of the technology go beyond reduced cost. Olansen states the end-user can gain increased access to information in ways that enhance data interpretation. "We could tie the data we collect into models of what's going on in the patient. By using mathematical modeling with the signals we gather, we can characterize disease states."
DEVELOPING APPLICATIONS FOR THE REAL WORLD
Commercial efforts are already being focused on developing practical applications for virtual instrumentation. Firms that develop biomedical instrumentation have also found that the cost of designing new instrumentation can be decreased significantly if the displays are modeled in software rather than hardware components.
North American Dräger (Telford, PA), which has developed specialized display units for its biomedical instrumentation, has found that costs were indeed reduced for developing new applications. Michael Argentieri, Dräger vice president of marketing, says the technology benefits go far beyond simple cost reduction. "When you have a customer demand for special display requirements, we can get to the designing process quickly," he says. "But we also find that we can make use of the new features; the customers can get the tool they want, and we can give it to them in an enhanced way."
Argentieri recalls one specific example of software technology making it possible to customize an instrument. One client wanted an extra-large gauge on an instrument, he explains, which would have been prohibitively expensive with hardware components. "They wanted to see it from across the room. And it's so much easier to give them that if you can go to the computer screen and set up the display for variable size. This is also a good example for cost because we can't afford to make a gauge six inches across in hardware."
Another company with extensive virtual instrumentation projects, Premise Development Corp. (Hartford, CT), has developed a range of practical tools for the healthcare environment. Using both computer technologies and database management tools, the firm's software engineers hybridize functions of information management systems and patient-care instrumentation. Premise's CEO Joseph Adams envisions the future of medical technology as one in which information technology will play an increasingly central role.
On the clinical research front, Premise developed such tools as IntelliVent, which operates with Puritan-Bennett 7200 ventilators. The software-based system gathers data for use by researchers in determining conditions under which patients can be weaned from ventilator assistance. The software acquires data from both the patient and the ventilator to present a comprehensive profile of the patient's physical condition. With traditional display settings, the physician makes judgments based on conditions measured at the moment of viewing; however, the microprocessor-controlled ventilator actually generates a continuous line of data from the patient. The system is designed to analyze the effects of respiratory rates, carbon dioxide production, oxygen use, and respiratory system performance to maximize the amount of information available for monitoring the patient's complete ventilatory synchrony.
Premise's BioBench is designed to integrate personal computers with existing transducers and amplifiers to monitor signals ranging from electroencephalograms and electrocardiograms to pressure, volume, flow, temperature, and force measurements on arteries and lungs. The software was developed by the company for National Instruments.
Practical applications of theoretical concepts in the virtual instrumentation field have been largely restricted thus far to extensions of conventional medical device technology. These products represent a step toward more flexible generic virtual instrument systems.
ADDRESSING SAFETY CONCERNS
Having flexible technology that provides information about patients might free up clinical researchers to think about what information they want to use to improve the data flow available to them. In the present state of medical practice, which relies on hardware, physicians are often locked into thinking in terms of data that comes from the equipment rather than thinking first about exactly what they want to see and then asking the equipment to be configured to display the information desired. Some researchers who study human interaction in medical environments, however, warn that instrument displays must also help manage the access to information without inadvertently confusing the user.
Virtual instrument interfaces could potentially help practitioners in diverse fields of clinical medicine. Richard Cook, MD, professor of anesthesia at the University of Chicago Medical Center, notes that one fundamental problem that physicians encounter in the operating room involves managing the safety protocols of increasingly complex practice environments.
The technology used in medicine has gotten so much more complex than what most physicians were trained to use that practitioners need to look out for ways in which the machines introduce new safety problems. One situation that quickly becomes a safety concern, Cook notes, is when physicians begin to rely on information from an increasing array of devices to monitor the state of the patient. He suggests that sometimes situations occur when one person in the chain of professionals involved in a medical procedure doesn't have the same information as someone else in the link. The worst situations can involve potentially catastrophic failures of medical judgment.
Anesthesiologists also are constantly working under conditions that require them to monitor multiple parameters involving different systems. The working environment in the operating room has forced anesthesiologists to vary the way they interact with the display panels to avoid getting confused, Cook says. Although highly trained human operators of complex machines can learn to use them well, the more complex the environment becomes, the greater the chance for small errors to develop. In complicated procedures, small errors that would not by themselves have any significant effect on patient outcomes can nevertheless have a cumulative effect that can evolve into a critical situation.
Engineers exploring ways to build greater flexibility into display panels must also focus on alleviating the pressure on practitioners to alter their work practices. David Woods, MD, PhD, of the Ohio State University Institute for Ergonomics, says studies of human errors in complex environments can guide product development engineers in designing systems to access information. He explains that a virtual instrumentation interface that reduces the number of keystrokes needed to perform multiple tasks in combination can create complex and arbitrary control sequences that overwhelm the user. "Concatenating multiple virtual devices on a single platform forces practitioners concerned with only a single niche to deal with the complexity of all the other niches as well," says Woods. "This is in contradiction to what people are observed to do to cope with complexity."
One area of special concern is the ease with which an interface designer can proliferate modes of operation, to hide displays behind narrow viewports, and assign multiple functions to controls, Woods says. Years of study in ergonomics and human behavior have identified many of the potential pitfalls related to design, he adds, but software engineers generally do not consult such sources.
Travis at Compuware poses the intriguing question, "How will the information technologies already developed for the Internet affect the virtual reality and instrumentation world?" He observes that technology transfer is one avenue for approaching new applications. Rapid development of Internet technologies has provided software engineers with new tools that can be applied in medical environments, he says.
"Some of these technologies are mainly applied to entertainment now, and the advances in network-capable database technology and streaming video have also been used in the financial world. We're trying these technologies on telemetry, not only on handling patient data." Travis notes that information systems for hospitals have focused on database applications for such straightforward tasks as managing patient data, recordkeeping, and documenting care, while more innovative approaches have largely been ignored.
"I see a great value in distracting the cpu in the final interface between machine and human operator," Travis contends. He explains that, if the information is made available by a computer screen—something set in a browser-like window that can be altered to suit different applications as needed—the utility of the technology to medical practice is increased. A benefit to the practitioner is improving the quality of care by raising the standards to match the needs of the medical condition. "We can mix and match the signals we want for the situations that present themselves. If you wanted to get information from ECG, EEG, and oximetry, with virtual instrumentation you can configure a software module to do exactly that and to present the precise information you need."
Eventually, virtual interfaces could provide signals to physicians and ancillary healthcare professionals through extended communications networks. "We can present the data from a patient through a Web-based system that extends the accessibility of the patient data. If the information can be distributed to any physical location reached by the server, the usefulness of the data increases." The information can be made available across an entire campus, Travis notes, wherever caregivers have the ability to log on to the information system. "It doesn't have to be restricted to terminals hooked up to the hospital information system; we can design systems to transmit signals to a pager, a PDA [personal digital assistant], or similar device." To speed response time in certain critical monitoring situations, data can be transmitted to multiple locations to alert the appropriate caregivers of changes in a patient's condition.
The application of virtual instrumentation could potentially reach beyond the boundaries of a single hospital campus. Travis speculates, for example, that an administrator of an extensive network of healthcare facilities might use the information network to help with the task of tracking quality-of-care data. If patient information is presented on virtual instruments, access to the data need not be restricted to the point of care. Both the caregivers working with individual patients and administrators looking for general trends can acquire and use the data they need from the same basic source.
Another benefit to end-users is that software modules can be downloaded when updated technology is needed, rather than through submission of requisition orders for more units of particular instruments. In addition, the equipment maintenance contract, in effect, becomes an upgrade agreement. Says Travis, "We can have these machines be smart, so they check for updates or automatically receive recall and safety alerts."
VIRTUAL REALITY TECHNOLOGY AND VIRTUAL INSTRUMENTATION
As advances in technology provide more tools for manipulating information, the virtual environment will most likely pervade the medical practice of the future. Although virtual reality systems being developed for surgical procedures do not necessarily involve the use of virtual instrumentation, they nonetheless drive the development of new conditions under which physicians will need access to data in radically different forms than those of established convention.
Jim Chen, PhD, director of the George Mason University Computer Graphics Laboratory (Fairfax, VA), believes that the efforts in developing 3-D imaging technologies for modeling the anatomy of surgical patients have driven the state of medicine to focus more intently on telecommunications. Though Chen's research centers on creating virtual objects for manipulation by surgeons, he says the direction taken by medicine in this field is pushing toward greater use of telecommunications technology to transmit medical data. In the case of treating a patient in a remote location, the medical expert needs to assess the patient's condition via information gathered by the trained technician or medic on-site.
A significant amount of development effort is still needed to develop the technology that will enable physicians to work entirely within a virtual reality environment. Says Chen, "In the current state of technology development, we have a lot of hardware dedicated to specific applications. This is a general trend. Whenever we have new applications, we have companies developing dedicated computing systems for that application." Telecommunications, however, can now reliably transmit the information needed to recreate the instrument panels of portable medical equipment virtually.
Computer graphics have already extended the functionality of conventional medical diagnostic imaging, Chen observes, by adding the visual tool of color to highlight important features. "Radiographs are black-and-white images and you need lots of training to read them, but, with color, we are able to highlight problems clearly." It may be a stretch to call this capability a virtual instrument application, but, conceptually, software technology can be used to enhance diagnostic hardware and the resulting tool has much greater functionality for the physician.
Others involved in the development of virtual reality technologies also concur that the trend is toward an increasing use of computer-based systems in medicine, which will drive a change in the way physicians interact with instrumentation used with medical equipment. Jonathan Silverstein, MD, assistant professor of surgery and codirector of the Virtual Reality in Medicine Laboratory at the University of Illinois at Chicago Health Sciences Center, notes that technology developers will need to consider how healthcare professionals use information.
Established procedures can sometimes prompt practitioners to resist embracing a new technology. Radiologists, for instance, might question the need for 3-D images when they were trained to conceptualize what is going on from 2-D images, says Silverstein. As a result, one of the challenges to developers is to generate the patient information in a form that is consistent with the way physicians are used to seeing it and using it. A critical role for Silverstein in his development projects is to guide the engineers in presenting the generated information in a useful and familiar manner. He explains that, "I can say that surgeons are familiar with it this way, so if we can get the software to do it this way, we have one less obstacle to acceptance."
Silverstein is involved in developing a system that incorporates a wearable computer display, which allows the surgeon to see the 3-D image of the patient's anatomical map without obscuring the real-world view of the operating room. The technology will allow the computer image to be superimposed on the real view of the patient in a kind of heads-up screen. "We're talking about augmented reality more than virtual reality," Silverstein says.
One system capability envisioned by Silverstein and his collaborators involves developing data subsets that provide highly detailed information about structures within the surgical field. "We are trying to connect the data obtained from imaging the patient to standard anatomical information that lets the computer know where the structures are." By using a high-resolution data set from the Visible Human Project and aligning those images with the patient's images, the group hopes to enable its system to display specified parts of the patient's image.
Rather than rely on the complete image of the liver and surrounding organs, for instance, a surgeon can program the display to highlight specific tissues or substructures. "I might want to see what the vascular tree in the liver looks like so I can avoid vessels as I enter with the laparoscopic instruments," says Silverstein. In such applications, 3-D imaging can construct views of important substructures that will be encountered during particular procedures, and displayed when they are needed.
Such advanced working environments for operating room personnel are expected to begin to reintegrate the virtual realm of images with the material world of the patient; however, the advantages will also come with an increasing transparency between information systems. With well-developed virtual instrumentation technology that can be accessed through a networked system, it will be possible to program virtual instrument panels to appear on different screens as needed.
In fact, many of the goals of virtual reality technology developers actually mirror those involved in virtual instrumentation work. "Why do all this stuff?" Silverstein asks rhetorically. "But consider that surgeons had a great way to get information before, and now that's been taken away by minimally invasive surgery." The instruments that have become accepted in modern surgery have broken up the broad-picture information that the surgeon once had. Instruments like the videoscope limit the field of view. In addition, surgeons now receive their information from many different sources located throughout the working environment. The advantage of current research efforts is that they begin to collapse the sources of critical information back to a manageable location that allows practitioners to concentrate on their immediate work.
Cost and efficiency are often the driving forces behind new technology, so it is not surprising that advocates of virtual instrumentation emphasize the potential for reduced cost of ownership and savings in the design process. As Argentieri at Dräger says, however, medical device developers also get access to more advanced engineering thinking by exploring software approaches to new product concepts. "When you go to this kind of development, you get a greater assortment of talented software engineers to choose from; the really good ones want to advance in their work to using graphical user interfaces and related types of platforms." Engineering trends are thus driving the development of the skills needed to adopt new technologies grounded in information science, he says.
The medical device developer needs to keep up with those trends to retain the interest of the best engineering talents. Compuware's Travis believes that software will never replace hardware completely, stating that "we're never going to get rid of hardware; we need hard components at some point." He agrees, however, that software technology is advancing into areas that were the exclusive realm of hardware only a decade ago. If the basic hardware around a patient is set up for generic data acquisition, software modules can be configured for the multitude of diagnostic and treatment monitoring functions that require many different machines. As such advances continue, the flexibility of software-based instrumentation will undoubtedly direct clinicians to new ways of thinking about the patient information they need to view.
William Loob is a medical writer living in Brooklyn, NY.