We live in the age of information, where the sheer volume, velocity and variety of data being created necessitate their being processed in a new way. The data and its related technology offer more than just health insights.
Nothing is quite as vulnerable as a premature baby fighting for its existence, alone in a neonatal intensive care unit. That tiny body lies tethered to numerous monitoring devices that provide a continuous feed of vital signs – such as heart rate, breathing and blood pressure – at a rate of a thousand readings per second. The exhaustive volume of data is too much for physicians and nurses to absorb. After 24 hours, it is discarded.
Such data, however, could provide vital signals about a premature baby’s fragile condition. Computer scientist
Dr. Carolyn McGregor, whose first child died after an early birth, hopes to provide premature babies with an extra chance.
She has devised an online health analytics platform called Artemis, named after the Greek goddess of childbearing. As Canada Research Chair in health informatics at the
University of Ontario Institute of Technology (UOIT), McGregor and her team of researchers are collaborating with IBM at a number of hospitals to test software that tracks vital signals in premature babies.
They needed a computer platform to take all the different signals from lots of different babies and environments Carolyn McGregor McGregor’s work follows research showing that babies who develop infections display changes in their heart rate, or heart-rate variability, 24 hours before the infection sets in.
“People had already identified this trend but didn’t have a way to watch it in real time,” she explains. “They needed a computer platform to take all the different signals from lots of different babies and environments to continue research on infection, but also use the same platform to research many other conditions.”
The software processes the newborns’ vital signs in real time, tracking 16 different data streams – such as heart rate, breathing, blood oxygen levels and blood pressure – which together amount to 1,260 data points per second per baby. It also seeks patterns in the data, then stores the information. The hope is that Artemis will allow doctors to identify subtle changes in a baby’s condition that may signal the onset of infection or another medical condition.
Taking over from a doctor’s “intuition,” the software provides signals that the naked eye would miss, and allows clinicians to administer medical treatment before symptoms deteriorate. “We monitor premature babies’ heart rate and respiration,” McGregor continues, “and can delineate whether heart rate variability happens shortly before the onset of an infection or because the baby is being given certain drugs, which can also trigger heart-rate variability.”
Looking beyond infection
Artemis was first introduced at the Hospital for Sick Children in Toronto in 2009. In 2010, a cloud-computing version went live at the neonatal unit of the Women & Infants Hospital in Providence, Rhode Island, where readings are fed to the UOIT. The project was extended in December 2012 to China’s Children’s Hospital of Fudan University in Shanghai.
“In China, they don’t use morphine,” McGregor explains, “allowing us to carry out cross-cultural studies and see the different heart-rate variability changes without the use of morphine.” McGregor and her team have now moved beyond looking only at infection to examine various other conditions, such as retinopathy of prematurity (an eye disease that causes some premature babies to lose their sight, notably afflicting blind pop star Stevie Wonder), or premature babies who forget to breathe, as their brain stems aren’t yet fully developed.
McGregor is also trialing an algorithm that classifies different types of conditions, such as low oxygen levels or gaps in breathing. Other plans include a study on adults. She also intends to continue testing with the Apollo project, which provides home-based monitoring.
“When these premature babies go home,” she says, “the Apollo platform would alert a medical person if there is a change in the babies’ condition.” While McGregor is still in the process of publishing and confirming the findings from the project, she is hopeful that Artemis will be introduced in neonatal intensive care units (NICUs) worldwide. She is equally confident about the beneficial effects of big data in the medical arena. “Big data has the potential to be the next disruptive technology after genomics. We are at the cusp of a whole new wave in clinical research.”
There is a human factor in trying to keep an eye on 250 people. You can’t continuously monitor each one for every second Timothy Buchman Whose information is it?
With the vast amounts of personal data collected, the question of moral responsibility urgently needs to be addressed. Who has the right to collate and publish all this information about a person’s body and its functioning? How does it affect the individual’s right to privacy? McGregor concedes, “We still need a framework for this. It needs to be on the mandate for public policy.”
Opinion surveys so far are positive. While people have concerns about health insurers’ use of big data, most are in favor of its use if it can provide insights that might mean the difference between life and death for premature babies. As Dr. Timothy Buchman, professor of surgery and anesthesiology at Emory University in Atlanta, Georgia, and director of the Emory Critical Care Center, points out, the data involved in such projects are hardly sensitive or high-risk. “We’re talking straight physiology values, such as your blood pressure,” he says, “which most reasonable humans aren’t going to get sensitive about. Can you remember what your blood pressure reading was two years ago even, and are you bothered about it?”
He adds, “There is a human factor in trying to keep an eye on 250 people. You can’t continuously monitor each one every second.” Fortunately, computers lend a hand where humans could fail. Buchman and colleagues have been using software from IBM and Excel Medical Electronics since early 2013 to monitor intensive care patients using real-time streaming analytics.
The system can analyze more than 1,000 real-time data points per patient per second and identify patterns that could indicate serious complications, such as atrial fibrillation, an abnormal heart rhythm triggered by a lack of blood oxygen or drugs. Buchman is certain that the research project will be deployed quickly at the hospital. “Instead of looking at single, six-second snapshots of ICU patient data, this system lets us see new views and trends of data that are being processed in real time.”
Big data and disease detection
Dr. Craig Feied, professor of emergency medicine at Georgetown University, believes
big data will transform medicine for the better. The inventor of a system known as Azyxxi, which provides real-time access to patients’ medical history at the touch of a button, is convinced big data will help identify patterns that can help in the early detection of diseases, such as cancer.
In his opinion, big data technology and real-time analytics will ultimately transform the world of critical-care medicine. “Big data will not make a diagnosis for us but will act as an early warning for caregivers to show us who is heading in the wrong direction.” He envisages a future where the same predictive capabilities possible in weather forecasting, for example, will also be employed in medicine.
This new world of healthcare will include better preventative care. “We are going to look after patients by keeping an eye on them not just in the ICU but also in their daily lives via tracking devices,” says Buchman. “There is a lot of information we need to know. If you are old, infirm and forgetful, we need to know that you are taking your medicine. Or if you get lost frequently, we need to know if you are straying off your usual routes.” The next generation of caregivers will need a broad skill set, including being adept at technology. With ground-breaking devices to monitor, collect and assess data, the ability to manage these tools will be vital in assisting healthcare practitioners to allow all of us to live longer, high-quality lives.