Friday, October 05, 2012

Transforming Health Care through Data

How many of you, when you think of Watson and the epic Jeopardy win, associate it with solving some of the most complex problems of our time? I for one, didn't. I had no idea that Watson is currently being used by IBM Research together with organizations that have complex problems to solve to find innovative solutions, but it makes a lot of sense.

Lori Beer, Executive Vice President, Enterprise Business Services, WellPoint, Inc., serving 34 millions of people is one of the poster child for such an application of Watson.

WellPoint, which mission is "to improve the lives of the people we serve and the health of communities", is currently in a partnership with IBM making history by applying Watson's capacities in an initial pilot for utilization management: treatment recommendation approval.



Health Care industry biggest challenges at this time are the issues of affordability, accessibility and consumer experience. Right now, the customer satisfaction for Health Care is below Airlines, which says a lot of the state of things.

As information doubles every 5 yrs., there are 21M articles in PubMed/Medicine and 1M are added every year. Lori raises the question of how to keep up? what tools could help? And this is how Watson, who reads millions of pages in a few seconds, can have an important role. The reality today is that 81% of physicians report spending only 5 hours or less per month reading medical journals.
The end consumers also have a poor engagement with their own health and the costs involved.
When you have huge amounts of un-structured data, the problem becomes how to know what's relevant. We need ontologies to organize the information and understanding what's more relevant at the point it needs to be used.


So, they set up a small pilot trying to tackle the issue of approving treatment requests from physicians.


Watson lends its capabilities of probabilistic understanding of dense amounts of data, natural language processing and the ability to learn interactively are being used towards the resolution of this
learning to work, in order to digest a vast amount of information and spit out a probabilistic number of accuracy. If the number is below a certain threshold, the decision needs to be reviewed by a physician. The physician looks at the research that lead Watson into computing its output and can discard and re-iterate as many times as needed.

Some of the requirements they had to get out of the way in order to get Watson to work for them, are the issue of single master profile per consumer (all the versions of me who are indeed me) and the issue of getting all the data for a patient organized. which is the holly grail of Health 2.0. Even when nowadays some of your doctors enter your medical history in a computer, it doesn't mean that that computer can talk to other computers even across the same organization. The solution to this problem is the longitudinal patient record and this was the second step in order to be able to feed Watson appropriately.

Watson goes through the same cycle as physicians do (just in a slightly different way :).
First, it goes to medical school, it has to ingest and digest medical facts and policies.

Now, Watson goes on what would be the equivalent of Residency, it's in training. Doctors are looking very careful to the responses that Watson is giving,

Then, as any student, Watson needs to be scored according to the accuracy of its results.

Watson accuracy has major improvement as it learns from its mistakes and re-iterates.
When the score is not so good, the questions are: Does Watson need to ingest more data? 

The teach cycle is crucial, until it reaches the accuracy level required.

As part of the process of testing Watson, a Dr gave his hardest case ever, one which took him years to get to the right diagnosis, and Watson identified the condition amont its top 3 recommendations.

Since the use of Watson, the approval process went from 48 hours to a few minutes. Imagine the value of that as a provider.

The use case #2 for Watson is in Oncology, first int the area of breast and then lung cancer.

The future of Watson in Health Care would be to go into diagnosis, tele diagnostics, genomics analysis, among many other areas. In this industry, all conversations need to consider privacy, security and liability as they reside at the core of Health Care.


Someone in the audience brought up the issue of how people could be fearful of a machine biasing doctors towards certain answers, but when we compare the traditional physicians methods that are estimated to be fact based only in 45% of the cases, it just makes sense to adopt tools like Watson.


Other questions on Lori's and WellPoint's mind are in the area of Consumer Experience where there is so much room for improvement:
How to engage the consumer?
How to have patients think of health as their asset?
How to have apps that bring the connection and deliver information in the point of care as well as follow the customer home?
How to create the next generation customer experience? including the possibility of telemedicine.
The future of Medicine, might be to bring back the doctor to where the patient is at, like it was in the past. avoiding ER costly visits. Solutions such as remote weight measurement alert clinicians of potential issues ahead of time.

Other solutions mentioned were part of the Mobile Health Wallet, that includes electronic id card, QR code, customer service # connects to phone call, voice video chat with point of care, mobile ways to locate nearest in network provider.

All this solutions need to create a space for technologists and clinicians to dialog and come together to provide solutions and Lori has a relevant role in creating this space.

Now, we can think and dream of other areas and ways in which technologies like Watson can help and build a future for Health Care that is affordable, accessible and a great customer experience.


7 comments:

Michael Holmes said...

Great post. Thanks. This is Michael Holmes, Program Director in IBM's Watson Solutions group. It's a real honor to work with innovative organizations like WellPoint and it's exciting to think about the positive impact we're working toward in people's lives. You might be interested in this short update on work we're doing with WellPoint http://www.youtube.com/watch?v=7d6c-5kJdGk&feature=plcp as well as this short vid on our work with Memorial Sloan Kettering training Watson to fight cancer http://www.youtube.com/watch?v=8DBqLTdPolI

Cecilia Abadie said...

Hi Michael!!
Thank you very much for the videos.
It's amazing to be a witness to these times in technology and Watson makes our world even more exciting by helping in practical ways, but also in answering the ancient question: can machines think?
I can't wait to see what the combination Watson with mobile, sensors and context can bring to the world.
Thanks again for your comment!

Paige Newton said...

Above all things, health care around the world, for the sick and elderly should be widely discussed. The live in care in east sussex should not be a problem.

Kimberly Mullin said...

Staffs of urgent care in phoenix az - are pleased to hear this news about Watson and IBM's willingness to face innovations and changes just to address to people's need in terms of medical attention.

Jordan Marcus said...

Now there's more to look forward in Watson, especially the Mobile Health Wallet that can assist more patients as soon as possible. I can see the Consumer Experience as a survey from patients to reach patients' concerns faster.

Louisse Campbell said...

The new Mobile Health Wallet caught my attention, with its benefits, a lot of people would subscribe to it. It will make their lives easier; I just hope the ny rehab center would also provide a solution like this one.

Home Care Atlanta GA said...

Very interesting info!Perfect just what I was looking for!