Learn about diversity at ASEE
ASEE would like to acknowledge the generous support of our premier corporate partners.




Engineers offer ways to get American healthcare off the ‘critical’ list.

As a nurse, I’m a menace. One of my patients was a seemingly vibrant 28-year-old male whom I diagnosed as having a mild case of the flu. Clearly, it was more serious, because he’s dead now. In fact, of all the medical assessments I’ve made, only two have been right. So it’s fortunate for everyone that I’m not really a health worker; I’m just playing one in the online game HealthAdvisor. I review the symptoms, medical history, and lifestyles of men and women of various ages and states of health. I assess their individual ailments and — in theory — determine the appropriate physician and treatment. The object is to keep everyone alive without blowing my budget. But with me in control, HealthAdvisor is a more lethal gaming experience than Grand Theft Auto IV.

If HealthAdvisor sounds fun and interesting, it is. Hard, too. But the point of the game is to educate, not entertain. It’s the brainchild of engineers at the Georgia Institute of Technology’s Tennenbaum Institute, a multidisciplinary research center focused on healthcare delivery and global manufacturing. HealthAdvisor’s creators call their creation a “virtual healthcare system where new products, concepts and policies can be tried out,” intending it as an open-source portal for people who aren’t necessarily healthcare professionals. “We’re trying to harness the power of the masses, to see if we can come up with some good ideas,” says William Rouse, Tennenbaum’s executive director.

The complex issues of U.S. healthcare have long concerned medical professionals and policymakers. But given the overwhelming problem of spiraling costs and tens of millions of Americans lacking basic health insurance, experts from other fields are now being consulted, including engineers. With their problem-solving techniques and technologies — including mathematical modeling, simulations and data-mining — engineers are well-positioned to help render healthcare more efficient, effective and affordable. And as new and different models are conceived, engineers can help design and support them, analyzing if and how well they are working. “There is a great need for engineering solutions,” says Kenneth J. Musselman, strategic collaboration director at the Regenstrief Center for Healthcare Engineering at Purdue University.

That’s an understatement. In 2006, the U.S. spent $2.2 trillion on healthcare, or 16 percent of gross domestic product, according to the Kaiser Family Foundation. By 2015, that figure will likely hit $4 trillion, closing in on 20 percent of GDP. In fact, the Congressional Budget Office (CBO) estimates that by 2087, America could be spending half its GDP on healthcare. U.S. expenditure for healthcare far exceeds that of any other industrial nation: In 2005, American costs were $6,401 on a per capita basis, while France, by contrast, spent $3,374 and Britain $2,724. Nor has all this spending produced results. The 2005 National Academy of Engineering (NAE) report Building a Better Delivery System: A New Engineering/Healthcare Partnership concluded that the system is so inefficient it results in 100,000 preventable deaths and the unnecessary loss of $500 billion each year. Meanwhile, 43 million Americans have no health insurance, and a growing number of those who are insured say their coverage fails to meet rising costs. “Everyone agrees that the system is broken,” Rouse says, “and that we cannot continue the way we are.”

That was also the bottom line of the NAE report, which became an important catalyst for bringing together engineers and healthcare providers. The report chastised the healthcare industry for being too slow to apply engineering tools that have bolstered quality, productivity and performance in a number of other sectors. “The healthcare industry has virtually ignored a broad spectrum of other technologies that would radically improve the safety and efficiency of healthcare,” wrote coauthor, W. Dale Compton, professor emeritus of industrial engineering at Purdue.

No One In Charge

Obviously, the problem of inefficient healthcare delivery is nothing new, so fixing it won’t be easy. Healthcare is a fragmented system that’s grown in an ad hoc fashion, mainly run by clinicians who are focused on patients, not on the system as a whole. In addition, healthcare providers “have not, historically, done any R&D. They’ve not actively worked to improve the service,” says Ravi Nemana, executive director for services, science, management and engineering at the Center for Information Technology Research in the Interest of Society (CITRIS), a four-campus initiative within the University of California system. And before the 2005 report, clinicians were often suspicious of engineering solutions. “Engineers were not gaining a lot of traction in the healthcare industry,” says Musselman of the Regenstrief Center.

HealthAdvisor, an online game developed by researchers at Georgia Tech, allows players to assume the role of medical professionals and 'treat' patients. The game's virtual healthcare system is intended to try out new products, ideas and policies.

HealthAdvisor, an online game developed by researchers at Georgia Tech, allows players to assume the role of medical professionals and 'treat' patients like the man complaining of a fever in the bottom left frame. The game's virtual healthcare system is intended to try out new products, ideas and policies.

The result, says Tennenbaum’s Rouse, is a system so convoluted that many experts concluded it wasn’t really a system at all. But, he adds, they were wrong. It is a system -- complex and adaptive, and one composed of myriad interdependent agents that must coordinate with one another, even though each has its own agenda. Such massive adaptive systems are not easily changed, as they can’t be ordered to act in concert. “They’re like a huge federation of entrepreneurs with no one in charge,” Rouse comments. Yet it may be possible to effect change by poring through the data to identify what has worked, then applying incentives and prohibitions to produce better results.

Fortunately, healthcare is an industry that generates tremendous amounts of data — it simply hasn’t been fully exploited. As James M. Tien, engineering dean at the University of Miami, told a spring, 2008 conference on healthcare and engineering in Washington, healthcare exists in a “DRIP” environment: data-rich, information-poor. But if information derives from processing data, the knowledge gained from that information can — and should — be put to use. For example, many urban clinics serving low-income populations are hard-hit by patients who fail to show up for appointments, sometimes averaging around 30 percent. That means that doctors are left idle by no-shows, even as other patients calling for appointments are told they have to wait days or weeks for one. “Obviously,” Musselman says, “there’s a disconnect here.” Yet clinics are reluctant to overbook in case their assumptions about no-shows are wrong.

Now, researchers at the Regenstrief Center are using data-mining technologies to study patient records. They’re hoping that a clinic’s daily no-show rate might be predicted by mathematically scouring reams of data that provides such information as patients’ age, distance from the clinic, medical history, and timing of their previous appointments. If the prediction software proves accurate, clinics might be able to do more impromptu rescheduling every day, more patients will get in to see a doctor, and doctors will have less wasted time on their hands.

While re-configuring the healthcare delivery system is a job tailor-made for industrial and systems engineers, the skills of mechanical, electrical and many other types of engineers are needed. Civil engineers are essential for building new hospitals. Bioengineers can contribute with new devices or therapies. One such CITRIS-funded bioengineering project is a small Magnetic Resonance Imaging (MRI) scanner priced at only $50,000, instead of the usual $1.5 million. Not only could the machine be used on specific parts of the body, its small size and low cost should make MRI scans more available in remote areas. Telemedicine — the delivery of medical data or services via phone lines, the Internet or wireless connections — is another growing area of interest for healthcare delivery engineers, one that often revolves around the creation of new devices. Yet engineers involved in healthcare delivery are often more interested in processes than in new equipment. “No matter how innovative something is, if you don’t get the overall system right, you can’t afford it,” Rouse says.

In fact, advanced technologies may be fueling the higher healthcare costs. According to the CBO, the major reason costs have tripled since the 1960s is the explosion of therapies for ailments that were once impervious to treatment. Some of these breakthrough devices, therapies and techniques are very costly to use. And though others aren’t as expensive, their popularity is also contributing to escalating costs. Bone marrow transplants or stem cell therapy are now used to treat a range of cancers, while life-saving coronary procedures have soared over the last 30 years. As more people live longer and survive once-fatal diseases, the system is becoming increasingly overburdened. Nonetheless, the added clinical benefits of new technologies are rarely inspected for cost-effectiveness. The CBO study points out that sometimes, “older, cheaper alternatives . . . offer comparable outcomes for patients.”

Vetting New Technologies

That’s one reason why engineers involved in healthcare delivery want to determine the overall feasibility of new products before putting them to work. Modeling and simulations are good ways to vet new technologies, testing when and how they can best be used. One key advantage of simulations is that all sorts of options can be explored without endangering anyone. “In healthcare, you can’t experiment with real systems. It’s too risky,” says Michael Kuhl, an industrial and systems engineer at the Rochester Institute of Technology and an expert in modeling and simulations. When seeking to improve operating or emergency room procedures, “you want to make sure the procedure you implement works before you do it.”

Kuhl got involved in healthcare issues during his dissertation research, when the techniques he developed to model arrival processes were used in a groundbreaking study to match organ donors and recipients. More recently, he has worked with a graduate student using algorithms to determine which hospitals might benefit from switching from film to a digital diagnostic imaging machine. Their research demonstrated that as patient volumes get larger, digital machines are the better option — in part because storage of extensive film records can become burdensome.

Regenstrief researchers similarly have employed simulations to implement the kinds of logistics used in warehousing to help a hospital consolidate and better staff its operations. This approach ensures that each surgical suite is properly supplied with instruments, which after every use have to be cleaned, inspected, sterilized and repackaged. Using data collected from another hospital, Kuhl has produced simulations to organize operating room schedules, so that patients are ready for surgery as soon as a room becomes available.

Other efforts to improve hospital productivity go beyond the medical. One grad student study Kuhl supervised examined a hospital’s food service to find ways to pare costs. It determined that a room-service method, letting patients pick the timing of their meals, made for a more efficient kitchen and staffing operation than serving everyone at set times, and also saved money.

Of the $600 billion spent on lab tests each year in the U.S., 70 percent of that money pays for paperwork, says Shankar S. Sastry, engineering dean at the University of California, Berkeley, and director emeritus of CITRIS. And paperwork is prone to costly errors. Sastry argues that huge savings can be had by more and better use of electronic recordkeeping, employing software that can detect mistakes and issue prompts. “This is, pretty simply, low-hanging fruit,” he observes.

Building design is another area in which data-mining can be beneficial. Many American hospitals were built shortly after World War II, and their archaic designs can greatly reduce productivity. So, as older structures are replaced, engineers are investigating design options. What is the optimal layout for a nurse’s station — circular, oval, star? To collect data, Regenstrief researchers placed GPS monitors on nurses at an Indiana hospital for a two-week period. “The question was, how much of their time was spent at patients’ bedsides,” says Musselman. The answer, very little — only 15 to 35 percent. Instead, nurses spent much of their days foraging for equipment, some clocking as many as nine miles per shift. Clearly, more effective design could dramatically reduce wasted time.

Engineers are also hopeful that telemedicine can make healthcare delivery not only more effective and less costly but available to remote areas. Monitoring devices used by patients could help clinicians determine the best times to see them, instead of relying on set schedules. That could allow for earlier interventions with greater success and less costly treatments. For example, patients who undergo chemotherapy need to take two sets of expensive drugs to boost their red and white blood cell counts, which are depleted by the chemo drugs. Patients take these drugs on a regular schedule, no matter what their cell counts are. But if cell counts could be monitored daily, they could take the drugs only when necessary. “In theory, you not only lower costs, but get better efficiency,” Nemana of CITRIS says. “It’s just like just-in-time inventories.”

Another CITRIS telemedicine project is the Cellscope, a device that turns an ordinary cellphone into a handheld, high-resolution microscope. Originally conceived by a student team in a campus-wide contest at the University of California, Berkeley, the Cellscope has many applications, ranging from home care to primary care in far-flung areas. It’s capable of imaging blood cells, snapping a picture of the image, then sending the data to a hospital or lab for diagnosis. The prototype of the snap-on device costs $75, but researchers estimate that the cost could fall to $10 to $20.

Both simulations and data-mining may help streamline the development and delivery of medication. Trials are a major reason why it takes so long and costs so much to develop drugs, and considerable time and money are lost on trials that end badly. Through simulations, engineers may be able to better predict which trials will fail so that they can be brought to an end much earlier. Data-mining can be used to personalize drug treatment, and the technology could also be applied to the so-called lab-on-a-chip nanotechnology that identifies specific proteins that indicate disease. “This promises to personalize medicine,” says Sastry.

Fee for Service or Success?

Data Mining and Simulations make clinics and operating room schedules more efficient.Rouse, of Georgia Tech’s Tennenbaum Institute, advocates changing how providers are reimbursed. Currently, they’re paid for their activities and time. What might work better is paying for successful treatments. Fees could be set on a risk-weighted basis, so that doctors in areas with large geriatric populations and low cure rates wouldn’t be competing against those in areas with large youthful populations, like college towns. Physicians tend to balk at such suggestions, saying that outcomes are partly up to patients, and you can’t force patients to take medicines or follow advice. “But that’s true of a lot of professions,” Rouse argues. A variation of the outcomes model would place greater responsibility on both doctors and patients. Last year a Milken Institute study reported that treatment of the most chronic ailments cost the country $277 billion in 2003, but the lost productivity of those patients cost it $1 trillion — a 10 percent hit to the nation’s GDP. Both Rouse and Sastry argue that a better system would make greater use of incentives to promote wellness and lifestyle changes — healthier diets, more exercise, no smoking — so that fewer people get sick.

Are enough dollars available to properly fund engineering research into healthcare delivery? So far, the funding has a clean bill of health, according to Musselman, with money coming in from the usual sources: federal and state governments, industry and private foundations. “It is an area that a great number of foundations are interested in,” he says. CITRIS, which has a remit that extends beyond healthcare, also has many industrial backers, including Hewlett-Packard, Intel and Microsoft. At the federal level, however, funding for healthcare delivery may fall through the cracks — The National Science Foundation may see it as within the purview of the National Institutes of Health, and the NIH may see it as engineering. But the boom in biomedical engineering has helped break down many barriers, and Musselman notes that the 2005 NAE report “was a tremendous boost.”

Opinion is split on whether healthcare delivery engineering will emerge as a separate discipline, or if that is even necessary. Musselman would like to see it happen, and Georgia Tech’s College of Engineering already offers a master’s in health systems. But Nemana thinks it will remain part of a larger emerging discipline, service engineering. That, he says, would enable graduates to apply their skills or look for jobs in any number of service sectors, from healthcare to retail. One thing everyone agrees upon is that there is plenty of need for engineers in healthcare delivery.

Meanwhile, despite the high mortality rate of my HealthAdvisor virtual game clinic and my ineptitude at diagnosing patients, the nurse whose part I play is doing okay financially. I’ve accumulated revenues of $80,000, and spent just $1,859. Luckily, I’m not paid for results. Inadvertently, I’ve nearly followed the game plan set by Rouse’s son, Will. When Rouse first created the game, he asked Will, then 11, what he thought of it. Will said he wouldn’t treat any of the patients. Why? Because some would live for awhile anyway, Will explained, and that way, he wouldn’t have to pay out for expensive therapies. Smart kid. Of course, doing nothing is not a viable option in the real world. But as Rouse’s HealthAdvisor game demonstrates, there are no easy solutions for addressing America’s chaotic healthcare system.

Still, the engineers are optimistic. Kuhl believes that, working together with healthcare providers, “we can accomplish a lot. Progress will be made through collaborations, so I’m very hopeful.” He adds wryly, “If I wasn’t, I wouldn’t be doing it.”


Thomas K. Grose is a freelance writer based in the United Kingdom.




© Copyright 2008
American Society for Engineering Education
1818 N Street, N.W., Suite 600
Washington, DC 20036-2479
Telephone: (202) 331-3500