IBM Tackles Health Care
These days, however, IBM researchers
spend a good deal of time working on issues related to health care.
With this year's passage of major health reform from the federal government,
Chalapathy Neti, associate director and global leader of Healthcare
Transformation at IBM Research, has been
looking at ways to determine, among other things, how health
care organizations can meet new requirements for EMRs (electronic medical
records) and rules related to "meaningful use." This is important
because, in order for health care providers to tap into federal money, they
must demonstrate meaningful use of EMRs under federal guidelines.
"Adopting these electric systems changes your care and especially the clinical outcomes," said Neti. "What analytics tries to do, once an IT system is implemented, is determine how you measure the metrics for -meaningful use.'"
More specifically, Neti and his IBM
researchers are working on ways to enable access to data that is stored in
different proprietary systems, which has made it impossible for doctors and
health care providers to pull together all the information they need to make
EMRs work. In February, IBM
bought Initiate to bring even more analytics know-how to bear on the EMR
problem.
"This allows us to begin aggregating data from multiple systems with
different types of proprietary formats and have them all come together into a
single data model with a common format, which then lets us to analyze that
data," Neti said.
Eventually, these technologies will turn into a new discipline called
"unstructured data analytics," which will allow for the collection of more than
just patient records. It will include information from clinical notes and
pathology reports to better track diseases and recognize trends.
Health care analytics is also finding a home in the hospitals, themselves. IBM
is working with Dr. Carolyn McGregor and the
University of Ontario Institute of Technology to develop new hardware and
software that will enable the use of real-time data within neonatal care units.
Specifically, McGregor-who has doctorate in computer science and a
background in business intelligence and data warehousing-wants to take the
large amounts of data that comes in from medical devices throughout the
hospital and analyze the information in real-time instead of storing the
records in a database and examining it later.
By looking at data in real-time, physicians will be able to better monitor
the health of these babies which, quite possibly, could defend against serious,
even fatal, infections.
"In the traditional approach, you extract the data and dump it
into a data warehouse every 24 hours and then you run algorithms overnight. . .
. By the time you look at it the next day, the baby could already be showing
the clinical signs," McGregor old eWEEK
in a telephone interview.
Initially, McGregor and some other researchers had built their own system
using an event-processing paradigm to analyze data in real-time. That team
learned a lot about receiving the data from medical devices, but calculating
all that information proved very complex.
"My next hurdle was to figure out how to build an engine that
does this processing, and build one in a way that we could create a number of
different analytic-watching mechanisms to keep an eye out for different
conditions," McGregor said.
Serendipitously, at a 2007 conference in Toronto, McGregor met IBM
researchers who were already working with on a new initiative called
InfoStream, which uses IBM's DB2 database
software and third-party technology to apply algorithms to data coming in from
medical devices.
"They approached me at the conference and we talked about what my
clinical problem was. We talked about how we could take the principals of the
functionality that I needed and represent that within the Streams
context," said McGregor.
Currently, the research conducted by both McGregor and IBM
is being tested in a pilot program at the Hospital for Sick Children, in
Toronto. If all goes according to plan, a clinical trial is slated to begin in
2011.
"Adopting these electric systems changes your care and especially the clinical outcomes," said Neti. "What analytics tries to do, once an IT system is implemented, is determine how you measure the metrics for -meaningful use.'"








