Big Data, Personalized Medicine to Trend in Health Care in 2012

By Brian T. Horowitz  |  Posted 2011-12-28 Print this article Print

title=Harnessing Cloud Computing to Ask Complex Questions} 

Doctors want to use the large amounts of data to improve health outcomes and reduce hospital readmission rates, said Zirl. With a large enough population of data, doctors can connect the dots to make better treatment decisions.

In 2012, Nuance Communications will continue to develop its clinical language understanding (CLU) technology while integrated with IBM Watson's probabilistic data analytics to help doctors with decision support, Lake noted.

Having access to a rich set of data will allow doctors to ask questions and come up with treatment options based on the information big data provides, Shah explained.

Cycle Computing offers utility supercomputers that by storing core clusters of data in the cloud allow doctors to ask the right questions about conditions.

"Taking advantage of cloud computing capabilities, you can ask questions you wouldn't think about because you don't have 30,000 cores around in an internal environment," said Cycle's Stowe.

Shah calls the use of big data in medicine "digitizing biology." Even something as routine as a hip replacement could be personalized from the knowledge found in big data tools, Shah added.

"If you can digitize biology, then personalized medicine is possible," he said.

Data analytics technology such as Watson can help doctors make decisions on how to treat a patient similar to how makes predictions on what products customers would like to buy, said Shah.

"These guys treat every byte of data as jewels," he said. "We do not have that in health care."

Big data technology such as IBM Watson could handle the amount of data needed for personalized medicine, Shah noted.

"If a Watson can already observe tons and tons of data-and it sits there day after day, it doesn't take breaks, it doesn't take lunches-these computers when given enough data can actually find patterns that might to lead to real cures or better treatments," said Shah.

With the government pushing for meaningful use of electronic health record (EHR) applications, current EHR platforms can't handle the amount of data from all of these unstructured data sources, according to Shah. They can handle only the patient histories and basic recordkeeping.

"Future EHRs will need to be predictive and be able to say I have enough data about you to tell you when you might get diabetes," Shah noted. "Current systems are not accustomed to digitizing biology, only digitizing paper."

Making sense of big data from both medical literature and EHRs will be a challenge, Informatica's Cramer suggested.

"It's not just natural language processing and Watson and being able to mine the huge volume of medical literature and things that are out there, but it's also being able to pair that with the increasingly large volumes of discrete data that's going to be captured in electronic health records," said Cramer.

"Health care also has lots and lots of textural data that is in the clinician visit notes, and those things are going to be very hard to move to the discrete data that goes in the database, and so being able to pair that with natural language processing capabilities and bridge across those two things to us really is the landscape of big data in health care," he said.

Brian T. Horowitz is a freelance technology and health writer as well as a copy editor. Brian has worked on the tech beat since 1996 and covered health care IT and rugged mobile computing for eWEEK since 2010. He has contributed to more than 20 publications, including Computer Shopper, Fast Company,, More, NYSE Magazine, Parents,, USA Weekend and, as well as other consumer and trade publications. Brian holds a B.A. from Hofstra University in New York.

Follow him on Twitter: @bthorowitz


Submit a Comment

Loading Comments...
Manage your Newsletters: Login   Register My Newsletters

Rocket Fuel