Thursday, November 29, 2012

IBMs Watson Computer to take on Cancer


Tech
|
2/01/2012 @ 2:19PM |6,888 views

IBM's Watson, Cedars-Sinai and WellPoint Take On Cancer


149


20


0


WellPoint and IBM are working with oncology experts at the Cedars Sinai Cancer Institute in Los Angeles to educate Watson as a physician’s assistant. Watson, an IBM computing system proved on Jeopardy! that a computer can collect and understand vast amounts of information and provide answers.
That was fine for a game show — now Watson is being put to work digesting millions of pages of research, incorporating the best clinical practices and monitoring the outcomes to assist physicians in treating cancer patients.
“This is an opportunity to use supercomputing powers to bring vast amounts of data to the practicing physicians while making it understandable,” said Dr. Harlan Levine, executive vice president for comprehensive health solutions at WellPoint, the health benefits company. Watson can assimilate two million pages of information every three seconds.
WellPoint and IBM agreed to start with oncology because it is so complex and has so many opportunities to use supercomputing to improve the quality and consistency of health care.
“Oncology is a very complex field. The amount of information in medicine in general doubles every five years. In oncology the amount new information emerging is particularly prolific. It’s a hard area in which to keep up to date.”
Manoj Saxena, general manager for IBM Watson Solutions, explained the scale of the challenge physicians face.
“We know of 12,000 diseases, information in medicine is doubling every five years, and 81 percent of the physicians surveyed by IBM said they spend less than five hours a month reading medical journals. It is humanly impossible to get all the research, impossible to take all that information and apply it at the time of diagnosis. We have heard numbers that approximately one in five diagnoses in the U.S. is inaccurate or incomplete.”
A lot of clinical practice is rooted in a physician’s original training, which may have been a decade or two ago, and differs from one region or one institution to another. The most up-to-date are probably the physicians working in an academic medical center.
Saxena offered an example of how Watson might pull together an odd collection of information for a physician:
A patient In Texas calls his clinic to say he has the flu. Watson looks up his record and says he does this every year at this time and it turns out to be ragweed allergy. And newspapers in his part of Texas have reported a ragweed outbreak. The physician can decide to prescribe some allergy medicine.
In medical payments, Watson can look at the diagnosis, medical tests, treatments and billing and flag suspicious payment patterns.
“Eighteen percent of GDP in the U.S. is health care, and by some estimates up to $600 billion is wasted, added Saxena. “As we dig into cancer we find tens of billions that can be saved just by streamlining care with more accurate and more repeatable procedures just based on things we already know. In time, I think we will look back at medicine and draw a line at Jeopardy!
“The opportunity to use supercomputing powers to bring in vast amounts of data and make it understandable to the practicing physicians and do it consistently, accurately and quickly offers a tremendous opportunity to improve the status quo,” explained Levine. “It’s like having 1,000 consulting physicians backing up the clinician, whether the clinician is located in an academic center in a regional medical center or a community hospital, he added.
Because Watson is capable of learning, it will incorporate new data and the evolving thinking of clinicians. As cases progress, Watson will upload the patient data along with new research.
The computer will be able to draw together all of a patient’s electronic health records and sift through them for pertinent information. WellPoint is also building a longitudinal record based on data that sits in different physicians’ offices and combining the individual records with medical literature and guidelines and claims in the company’s database to develop recommendations for the clinician.
“With real life experience observing how cancer progresses over time, it can see not just what the literature recommends but the progression of the conditions so it can offer the most support possible to the clinician,” he added. “We are in a partnership with clinical experts to help us teach Watson not only content but how to think like a clinician, how to assimilate data the way a clinician should in practice.”

No comments: