Tip:
Highlight text to annotate it
X
LANINGHAM: Back at Information on Demand 2011.
I'm Scott Laningham with Todd Watson.
We're joined right now by Craig Rhinehart, who is director, ECM Strategy
and Market Development, IBM Business Analytics.
Welcome, Craig.
RHINEHART: Well, thank you.
LANINGHAM: Thanks for taking a moment.
RHINEHART: Thanks for having me here.
LANINGHAM: You probably sprint in every direction, right?
RHINEHART: It's a little busy here.
There's a lot of...well, there's 11,000 of our closest friends here.
LANINGHAM: And you've got your leather couch here
for a good 10 minutes or whatever you want to spend.
So just relax, and good to have you.
I wanted to ask you, clearly, the big thing today is the announcement,
IBM Content and Predictive Analytics.
Right? RHINEHART: Yes.
LANINGHAM: Can you tell people again...we talked about it a little bit, but reiterate what it is.
RHINEHART: Well, IBM Content and Predictive Analytics for Healthcare is a new solution
that we announced here at the event.
And what it basically is, it's the first solution
that IBM has brought forward that's ready for Watson.
It complements and leverages the same Watson technology that we've announced
that we're commercializing for healthcare.
LANINGHAM: Coming from that team, I mean, with that technology and working together...
RHINEHART: Well, we share technology with Watson.
If you think about it, the real value to these kinds of technologies is that most
of the world's information is unstructured, doesn't live in the database.
LANINGHAM: Right.
RHINEHART: Text, images, you know, you name it, video, audio, it's all unstructured information.
With over 80 percent of that out there, it seems kind of foolish if we only do analytics
or data management on, you know, one-fifth...
LANINGHAM: Right.
RHINEHART: ...of the available information.
So the goal of this new solution that we announced today as well as Watson is
to help our customers get more value out of that other 80 percent.
And you need technologies like this natural language processing,
which is basically what understands -- the form that we as humans communicate,
it understands that -- and that's, you know...we're moving down that path,
and today's a big announcement for us as a result.
RHINEHART: Right.
Todd? WATSON: So Seton Healthcare was featured as the customer...
RHINEHART: Yes.
WATSON: ...at least today.
I'm curious like what kind of scenarios or business problems they're going to be able
to leverage this tool to try and solve using the predictive analytics.
RHINEHART: Well, that's a great question.
And natural language processing itself, and while it's sort
of the engine that's driving this, that's not new.
We've been doing that for some time.
What's new is...in this solution, is that we're combining that core natural language processing
and content analytics with predictive analytics.
And what Seton found valuable is they're in the care delivery business.
And in their case, the healthcare industry, as probably everyone knows,
is undergoing a transformation from a volume-based business to a value-based business.
There are incentives; there are also penalties.
Now, Medicare, next year, is going
to start penalizing providers with high readmission ratings.
You discharge a patient, they come back within 30 days,
it could have been prevented in a lot of the cases.
If...for those providers who have high readmission rates,
Medicare is going to start penalizing them on the form of reimbursement penalties.
New England Journal of Medicine says that one in five of these readmissions are preventable.
So what Seton is trying to do is they want to use content analysis and predictive analysis
as a way to identify across an entire population of patients what are the factors
that are driving readmission, and then within our patient population,
who are the highest risk...based on these, you know, factors,
who are the high risks of being readmitted.
And if we can use this solution to do that, then that then puts them in the situation
where they can intervene through care coordination and take the necessary steps
to reduce those readmissions because it's costly for them, it's not optimal care for the patient.
LANINGHAM: For the patient, right.
RHINEHART: And so this is key.
If you can...it's the holy grail.
If you can really improve patient care, top objective,
at the same time reduce unnecessary costs
and further avoid those penalties, then that's a home run.
LANINGHAM: And work within a system that's increasingly complex
and there's a lot more integration.
So you can kind of get it coming and going, solve both problems.
RHINEHART: Absolutely.
You know, this whole industry, healthcare, is going through a transformation.
And you know, many organizations, at least the smart ones like Seton,
want to put their information or their insights to work for them, really make them actionable.
And in their case, right, they can now deliver better care to a population that is,
you know, chronic in terms of disease.
LANINGHAM: Is this technology designed really for an enterprise level,
I mean, large healthcare organizations?
Or are there some mid-market applications for...
RHINEHART: Oh, no, there are absolutely mid-market applications.
It spans a number of different use cases, both....
And the real interesting thing here is it leverages clinical information alongside
of operational information.
And by combining the two is where you can unlock those insights that allow you
to deliver better care and at the same time have better operational outcomes.
LANINGHAM: And adaptable across the different aspects
of the healthcare industry, I'm assuming.
No? RHINEHART: Yes.
Payers, providers, research, drug manufacturers, medical device manufacturers.
There are use cases across a broad spectrum.
WATSON: So what about expansion into other industries?
Where do we expect to see this go next?
RHINEHART: Well, you know, we announced for healthcare today and around use cases
like what Seton's working, but this is the sort
of thing that's easily appliable to other industries.
Every industry has the same opportunity to leverage
that 80-plus percent that's unstructured.
The question is, are they doing it effectively?
And most cases, the answer is not.
Now, there are some industries that are, in government
or public sector, particularly in law enforcement.
Lots of unstructured informations locked away in investigative reports that, you know,
we unlock that unstructured, correlate it across an entire district,
trends and patterns start to emerge.
One customer...one policing organization has used this effectively
in the past to solve crimes faster.
They actually solved two cold *** cases in the first week of deploying a system.
WATSON: Hmm.
LANINGHAM: Wow.
RHINEHART: And they unlocked it...it sounds kind of funny.
It's going to sound like an episode of SVU.
But what unlocked it was they found a pattern in this information of, within a write-up --
so this is text -- you know, of a tattoo that had been described on...from a certain gang,
you know, on the neck and whatever the name was.
But they found that pattern within the context of other events related to these open cases.
They were quickly able to see a connection...
LANINGHAM: Wow.
RHINEHART: ...that they had never been able to see before,
and arrests and convictions followed.
LANINGHAM: Do you think Todd and I could get access to this to ensure that we don't get
to the airport any earlier than we have to on our way home?
[ LAUGHTER ]
We can hang here longer; spend less time there?
RHINEHART: Well, I guess that depends on how well this interview goes.
[ LAUGHTER ]
LANINGHAM: Craig, thank you so much for your time, really.
WATSON: Yes.
LANINGHAM: Appreciate it very much.
WATSON: Great stuff.
LANINGHAM: Craig Rhinehart, again, is director ECM Strategy
and Market Development, IBM Business Analytics.
I'm Scott Laningham with Todd Watson.
We'll be back at Information on Demand 2011.