Premature babies require constant monitoring to survive, and it’s a challenge for doctors to analyze the massive amounts of data on their vital signs. The ability to make fast medical decisions is especially important in remote areas, where doctors must decide whether to transport a baby to a facility with a neonatal intensive care unit. That’s where IEEE Member Carolyn McGregor comes in. The associate professor of health sciences at the University of Ontario’s Institute of Technology, in Oshawa, Canada, is working with new software to improve the way doctors can analyze large amounts of information. The following article describing her efforts appeared in the June edition of IEEE Women in Engineering Magazine.
Each year, more than half a million premature babies are born in the United States, meaning they come into the world after less than 37 full weeks in the womb. They are tiny and fragile. They often have underdeveloped lungs and weak immune systems. They can have problems digesting food and gaining weight, and many develop necrosis or sepsis, a type of blood poisoning. When these babies are born in rural or remote towns, often hours away from the nearest neonatal intensive care unit, the situation is even more dire. The doctors who care for these babies are faced with extremely difficult and complex decisions about how to treat them, whether to transport them, and how best to monitor their condition and stave off infection.
The volume of biomedical data that emerges from monitoring these babies every second—information on things like heart rate, blood pressure, breathing, and oxygen levels—is enormous. To be manageable, readings from each monitor are typically summarized and printed out on a large piece of paper every 30 minutes. It’s a method that results in a huge data loss, says Carolyn McGregor, a Canada Research Chair in Health Informatics at the University of Ontario Institute of Technology (UOIT) in Oshawa. Modernizing the system—improving the way doctors use computer and engineering solutions to analyze and store a larger amount of information—could improve survival rates for premature babies, she says. In other words, it could save lives.
McGregor and her team at UOIT’s new state-of-the-art Health Informatics Laboratory are now working with prototype software developed by researchers at IBM’s Watson Research Center in New York (made available under IBM’s First of a Kind program, which is designed to link innovative technologies with real-world problems). McGregor says the computing software can help doctors and nurses make better and quicker decisions about how to treat preemies.
McGregor, who has a doctorate in computing science, arrived at UOIT in 2007 from Australia, where she established and led the health informatics research group at the University of Western Sydney. Early in her career, she worked on developing new ways for companies to use intelligent decision support systems to monitor their business performance. Later, she used her experience with intelligent systems to develop the first on-demand simulated neonatal intensive care unit, which showed how neonatologists could use computer and information technology to remotely support doctors caring for premature babies in rural areas of Australia. Like business executives, these doctors “also have to manipulate large volumes of data,” says McGregor, but “the analysis is more complex when analyzing something like an ECG wave, beat to beat. We’re looking for more complex behavior than in a stock market wave.” While presenting her research in Hong Kong, McGregor met Bernadette Schell, founding dean of the faculty of business and information technology at UOIT, who said she believed the technology could help revolutionize neonatal care in Canada’s most isolated northern hospitals.
Often, these hospitals don’t have neonatal experts on staff, and pediatricians and obstetricians have to rely on telephone conversations to get expert opinions. Next-generation teleconferencing technology, on the other hand, would allow immediate face-to-face discussions with these specialists, who can look at video streams and X-ray images and help sort through patient data. “It gives them a much better picture of the viability of survival or transport,” McGregor says. Eventually, the hope is that extraction technology will allow off-site specialists to receive information from all the machines monitoring the baby in real time so that they can virtually assess the infant’s condition.
Preemies are always connected to monitors and sensors that detect changes in a baby’s condition and environment. Data are produced constantly and rapidly. The IBM software will be able to process the 512 readings per second that some of the monitors generate. Ultimately, the goal is to integrate all the physiological readings that are taken—at birth, during transport, and in the intensive care unit—and give neonatologists the computing tools and intelligent support systems they need to analyze this data almost immediately. The data will also be stored so that doctors can look for commonalities that might enable them to make more accurate predictions about a baby’s condition; currently, data are usually discarded after 24 hours.
The ability to constantly watch and find patterns in the data can make a huge difference. Consider the case of sepsis, which 25 percent of preemies acquire. By the time sepsis shows up in a blood analysis, many of these babies are already very sick; about one-quarter of them die. But research shows that the heart rate of a baby who has sepsis will change subtly hours before other symptoms appear, and McGregor’s intelligent monitoring system and information analysis can help doctors spot this early warning, allowing them to administer antibiotics sooner and reduce any impact on the baby’s organs. “My daughter couldn’t be saved for other reasons,” says McGregor, who has two young children, “but there are lots of babies who can, by changing the way neonatologists are aided by technology.”
More accurate data collection and information sharing can also help lower health-care costs by reducing the length of time an infant has to stay in the intensive care unit. The average hospital stay for a premature baby costs about US $150 000; when a baby develops sepsis, for example, the length of stay can double, meaning the cost goes up.
Today, McGregor continues testing and research with the help of the undergraduate and graduate students she supervises at UOIT’s simulation lab, which has several of the medical devices that are used in neonatal intensive care units internationally. She’s also collaborating with neonatologists at the Hospital for Sick Children in Toronto to test the computer software and technology in a real-world environment.
“One of the misconceptions is that engineers are all in back rooms without windows and working with equipment,” she says. “As someone recently said to me, doctors may help people one on one, but engineers have the ability to help save many lives by building an infrastructure. We really have the opportunity to utilize technology to help a lot of people.”