WellPoint Uses Watson to Help Respond to Health Care's Challenges

With its natural-language processing skills, Watson is able to make sense of medical data in ways that traditional computing isn't capable of understanding.
Lori Beer, EVP, Specialty Businesses and Information Technology, WellPoint
IBM_watson4_600px.jpg

One of the biggest challenges -- and advantages -- of the U.S healthcare industry is data.

Just consider two astounding facts. The amount of new medical research doubles every five years. Yet, evidence-based medicine is used less than 55 percent of the time when it comes to treating patients.

There's a lot of information out there, but medical providers do not have enough time to sort through it all and apply it to patient care.

WellPoint has close ties to physician networks around the country and a broad reach: we serve 36 million Americans, or one in nine of the people in the U.S., and have relationships with more than 600,000 providers.

To tackle the deluge of data, we made an important new hire: IBM's Watson, the system that used breakthroughs in natural language computing to trump the competition -- namely two reigning human champions on the game show Jeopardy!

I was in the audience watching the Jeopardy! shows. Sitting there, I realized we could apply this new technology to solve some of health care's most pressing challenges.

Once we made the decision to use the Watson technology, our nurses spent more than 15,000 hours training Watson. The result? We developed the first commercial products based on the Watson technology. One set of products, for example, helps health plans make decisions about treatment requests more quickly and efficiently. Another, developed with Memorial Sloan-Kettering, helps physicians assess and support treatment plans for cancer patients.

IBM_injector_wellpoint.jpg
This is just a first step. By harnessing Watson and the era of cognitive computing, we can re-invent how our industry approaches care. We can provide relevant information that physicians can use, ultimately ensuring that millions of Americans have access to higher quality, more affordable care.

Everyone is trying to deliver the best possible health care. The question is how do you come up with game-changing ways to understand the data underpinning our treatment decisions, considering that 80 percent of data today is unstructured?

Watson can help answer these questions. Its natural-language processing skills can make sense of data in ways that traditional computing isn't capable of and that physicians just don't have time to sort through.

Watson does this by sifting through vast amounts of data, whether patient records or the latest medical research, parsing it to identify key pieces of information across many different cases, and then suggesting the most probable answers to a problem.

We decided to apply it first to an important area that impacts health benefits providers, patients and physicians.

When doctors ask us for authorization for a procedure or treatment, we have to assess a broad array of clinical and patient data to make a decision. We needed a way to speed up the processing of these requests and make the process more efficient, so we could save patients' time while still continuing to base our decisions on medical evidence and clinical practice guidelines that leads to the most effective treatment.

Years ago, this process relied on written authorization and manual review by a health plan's clinical staff, which could take up to a few days. We've made a lot of improvements to speed up this process, but further refinement to traditional decision support tools means that logic has to be hard-coded for each new procedure code. With Watson's cognitive computing, we saw the chance to craft a new approach.

Our nurses trained Watson using more than 25,000 cases, teaching it to interpret the meaning of queries and analyze them in the context of medical data and human and natural language, including doctor's notes, patient records and clinical feedback. Watson analyzes the data, using hypothesis generation and evidence-based learning to generate recommendations that help nurses make decisions about treatment requests.

We have to turn the corner on affordable and effective care. During the past eight years, the average median income didn't budge while the cost of health care for an average family doubled to $18,000. At the same time, we've seen a lot of new emerging technology but we haven't seen that reflected in higher quality of care.

The problem isn't that we don't know enough, it's that we don't know how to make the most of what we know. Our work with Watson is just one way WellPoint is helping transform the healthcare industry.

About the author of this Post

Lori Beer, EVP, Specialty Businesses and Information Technology, WellPoint
Lori Beer is executive vice president of Specialty Businesses and Information Technology, and a member of the executive leadership team for Indianapolis-based WellPoint, Inc. She is responsible for WellPoint’s Specialty Products, including dental, vision, life, disability and workers’ compensation, and its consumer-centric business 1-800 CONTACTS; Federal Government Solutions, which includes National Government Services (NGS) and the Federal Employee Plan (FEP); Information Technology; Information Management, and technology-driven innovation.
Sponsor Content

Join the Discussion

The Atlantic does not moderate these reader comments, except to the same extent comments are moderated pursuant to the Terms & Conditions generally applicable to all content on The Atlantic's sites. blog comments powered by Disqus

Learn More About Watson