Helping newborns, infants fight their fight

The NICU could be one of big data's most meaningful proving grounds

Anyone looking for concrete examples illustrating the benefits of big data--and the difference between big data and other forms of analysis and diagnosis--should consider the examples of personal heartbreak and professional determination of Carolyn McGregor, PhD, professor at the University of Ontario Institute of Technology, and Annamarie Saarinen, principal at Ainsley Shea and founder of the Newborn Foundation, as well as the IBM (NYSE: IBM) technologists they worked with to begin improving the analysis and best practices in natal intensive care units across the country.

In a webinar presented by InformationWeek yesterday, these two woman, both touched by tragedy and heartbreak in the loss or suffering of their newborns, shared their experiences and explained how they were inspired to seek change in the way technology, particularly the data that high-tech medical equipment produces, is used to diagnose and treat infant and adult patients. They were joined in their efforts by Charlie Schick, PhD., director of big data solutions, healthcare & life sciences at IBM and Tim Dinger, PhD., of IBM's T.J. Watson Research Center.

McGregor began her career in computer science steeped in business intelligence and was no stranger to analytics technology when she found herself in the hospital delivering a premature daughter, with a rare chromosome abnormality, who soon passed away. Though no technology could have saved her daughter, McGregor said, she did notice that a lot of data, and ultimately information critical to the care of both her and her child, were not being fully utilized.

Saarinen's daughter survived, but not before a lot of tense moments, risky transports and serious surgeries. Saarinen said technology did help save her daughter, but that technology still could have been used much more effectively.  As a self-described health IT nerd with a background in  public policy and health economics, and mother of three, she said newborns have been part of her "ecosystem." So when things went well after her daughter was born and she was about to head home with her, she wasn't expecting that a final manual check by the pediatrician (listening to her daughter's heart) would detect a murmur. And she also wasn't thinking about how a better use of all the high-tech medical equipment may have detected it earlier and saved a lot of anxiety and pain. But she thought about it quick enough after the diagnosis.

Her daughter had congenital heart disease and required a transplant. It was successful and when things settled down, Saarinen formed her nonprofit organization to promote screening and better diagnostic tools for this disease.

It took almost two years, but her organization finally got a federal advisory committee to recommend every newborn get screened for this condition. She didn't stop there. During the ordeal, she noticed a lack of communication between the hospitals they were transported to and from, and added that to the list of things to fix. "The doctors at hospital three asked me what beta blocker she was on at hospital two. I had just given birth and my daughter was fighting for her life. That they would even take my word for it, even if I thought I knew, was scary," Saarinen said.

Both women were amazed at the amount of data being generated by all this high-tech equipment and the lack of good information that was being used. "Babies are telling a story but we don't have the tools to watch them closely enough," McGregor said.

There was equipment monitoring the breathing, heart rate, movement of oxygen, nutrition and other critical functions, but there was no way of correlating that data to identify patterns. McGregor said there are thousands of readings that can be taken in that time that may help detect a situation before it becomes critical. She, Saarinen and IBM's Dinger and Schick, worked together to develop and deploy those tools. In less than two years, they launched the Newborn coalition and started the remote echo project for early diagnosis.

Now 4.2 million babies born in the U.S. will be screened for the early diagnosis of heart defects, which could increase the detection rate ten-fold. 20 to 35 percent of infant deaths from CCHD are attributed to delayed or missed diagnosis, according to Saarinen.

Last year, New Jersey Governor Chris Christie signed Newborn Pulse Oximetry Screening into law and Indiana and Maryland took steps to include screening in their legislation. Today, nine states have passed legislation for the screening, six have implemented it and 13 states have legislation or statutory language introduced or pending.

Their efforts also reduced the typical time for determining the onset of infection from 24 hours to a matter of minutes by analyzing the heart rate in sub-second intervals. Dinger said the NICUs were suffering from catastrophic data loss, which may be acceptable in some industries, but should not be in this case. IBM stepped in with its high-performance computing platform called InfoSphere Streams, which allows clinicians to leverage its ability to ingest, analyze and correlate information from thousands of real-time sources in many data types such as text, images, audio and sensors.

Dinger said one thing institutions need to do is understand what kind of data they are already recording, but may not be analyzing. It will give them a tremendous opportunity to think about applying real-time analytics. "I recommend they do a survey and look at data architectures to see what they have and don't have," before moving forward, he said.

McGregor said they are just scratching the surface to show how early intervention can happen in a timely and efficient way.

For more:
- see InformationWeek webinar

Related Articles:
Spotlight: Cancer center wins leadership award in big data category
Challenges for big data in health care
Blue Cross Blue Shield of Tennessee scales its analytics