Adam Gartenberg's Blog

Business Analytics and Optimization, IBM and Social Marketing

New Intelligence in Action: Predicting changes in the tiniest patients

My favorite thing about working on the Smarter Planet initiative is that I get to come across some really cool stories about how new technology can be applied to change things for the better - far beyond the measures I usually come across of "increase market share by 10%" or "improve productivity by 20%."

A perfect example of this is work being undertaken to monitor critically ill newborn "preemies" by IBM Research in collaboration with the Ontario Institute of Technology and several Canadian hospitals.  Traditionally, doctors and nurses staffing the Neonatal Intensive Care Unit (NICU) are only able to respond to critical situations after the monitors alert them of a problem; in an ideal world, of course, the staff would be able to predict when problems will occur and intervene before the situation reaches the critical stage.

Enter a research group, headed up by Dr. Carolyn McGregor, a UOIT associate professor and Canada Research Chair in Health Informatics, who have been looking at just how to do that.  They are applying advanced stream computing software developed by IBM Research (and expected to be available later this year in the form of IBM InfoSphere Streams) to collect up to 512 readings per second from monitoring devices, pulling in a constant stream of biomedical data, such as heart rate and respiration, as well as environmental data gathered from advanced sensors and more traditional monitoring equipment on and around the babies.

The researchers will then use the software to apply findings from Dr. McGregor's research to help make sense of the data and, in near-real-time, feed back the resulting analysis to hospital staff so they can predict potential changes in an infant's condition with greater accuracy and intervene more quickly.  While certain life-threatening conditions, such as infection, can be detected up to 24 hours in advance, today's manual processes are often insufficient to detect such patterns as they form.

The type of information that will come out of the research project is not available today. Currently, physicians monitoring preemies rely on a paper-based process that involves manually looking at the readings from various monitors and getting feedback from the nurses providing care.

"This research has the potential to reatly impact neonatal care through reduced mortality and morbidity rates and overall health-care costs," said Dr. McGregor. "By merging our research and technology, we are able to collect more detailed patient data in a systematic manner, do online health analysis and decision support, and get advanced early warning of emerging patterns that could predict a medical event."

Like I said, pretty cool, huh?

More: First-of-a-Kind Technology to Help Doctors Care for Premature Babies