Blog Post

After years of hype, IBM wants to sell Watson Health. What does that mean for health AI?

By Andrew Rebhan

February 23, 2021

    If you ask most people what they think of when they hear "IBM Watson," the answer is probably tied to the game show Jeopardy! IBM's breakthrough computing platform crushed its human competitors on the public stage in early 2011, creating a media sensation and reinvigorating research and development around artificial intelligence (AI). 

    Weekly line: Big Tech is driving even deeper into health care amid Covid-19

    A decade later, The Wall Street Journal has released a story indicating that IBM executives are now exploring options to sell Watson Health. What happened?

    IBM Watson launches to much fanfare, but troubles soon mount

    IBM announced Watson Health in 2015, marking the first move of Watson technology into a specific industry. Watson Health was pitched as a cloud-based service that could gather and process big data to deliver customized insights across various customer segments, including health care providers, payers, and life sciences. Watson Health aimed to leverage AI (primarily natural language processing) to process medical records, peer-reviewed journals, medical texts, drug information, and more in its attempt to revolutionize clinical decision support, most notably around cancer diagnosis and treatment.

    To support these capabilities, IBM made a series of acquisitions to build up its data repository and cognitive computing functions, spending over $4 billion to acquire Merge Healthcare, Explorys, Phytel, and Truven Health Analytics

    But it didn't take long for Watson Health to face its share of criticisms for high-profile setbacks and challenges, with reports noting that the platform often struggled with interoperability, high-quality data collection, faulty decisions, and limited training mechanisms. Its failed partnership with MD Anderson Cancer Center in 2017 brought a fresh wave of criticism to the business, and in April 2019, IBM announced it was winding down Watson's work on AI-enabled drug discovery due to poor financial returns. The company also hasn't shown significant payback from its various business acquisitions.

    IBM seeks to pivot its legacy business for future growth

    But Watson Health isn't an anomaly in IBM's overall portfolio. Over the past decade, IBM has seen consistent drops in revenue while accumulating substantial debt, frustrating its shareholders who have been waiting for the company to turn things around financially.

    IBM now has a new CEO in Arvind Krishna, who has made it clear that AI and hybrid cloud computing will be growth segments for the business moving forward. IBM acquired the open-source software company Red Hat for $33 billion in 2019 to try and compete in the cloud computing market, a space where its competitive profile has dramatically shrunk compared to Microsoft, Amazon, and Google. IBM is also spinning off its managed IT services division, which accounts for roughly a quarter of its total sales.

    Given the change in leadership and the need to split off legacy business lines to support higher-value opportunities, it makes sense that Watson Health is facing renewed scrutiny. While health care has been seen as a major market for growth by Big Tech firms, profitability has been an elusive goal.

    What Watson Health's demise could mean for health care's AI adoption

    But would a Watson Health collapse hurt health care's AI adoption overall? Hardly. AI has made tremendous progress in the past few years. Many leading health systems are now realizing quantifiable front-line improvements to the quality and efficiency of care delivery. However, these improvements come through the judicious application of these technologies. We are still a long way away from having a general or "strong" AI that can be applied to handle broad, complicated health challenges (much like how Watson was marketed as being a solution for just about anything).

    Successful deployment of these powerful technologies demands maturity from both the vendors involved and the health system's governance processes. No decision support or prediction system is perfect—the key to responsible use is a solid understanding of the capabilities, weaknesses, and appropriate use of the AI models. This is why many health care stakeholders have initially begun their AI journey with applications in lower-risk administrative or operational areas to build experience and process maturity before tackling higher-risk clinical processes. IBM came swinging out of the gate, focusing on the complexities of oncology and genomics—two pillars of precision medicine that are still largely aspirational for most health care organizations.

    In the scramble of AI vendors entering the health care market, concerns about the quality of data, limitations of models, or expectations for ongoing management and supervision of these tools are often overlooked. Health care organizations adopting AI should ask several questions in the early stages of their strategic planning:

    • What is the specific outcome we are trying to improve? Do we measure and closely monitor that outcome today?
    • Does the vendor solution have a track record with a similar population we serve or in a similar environment that we operate in?
    • Can a vendor demonstrate its efficacy using our actual data under realistic conditions before we make a full commitment?
    • Who will manage the training, re-tuning, and ongoing monitoring of the models in production?
    • How will AI-provided insights be incorporated into existing workflows and applications? How do workflows need to change?
    • What are the risks and fallback processes if the technology produces misleading or incorrect answers?

    Inevitably, the adoption of AI in health care will result in some failures, but the benefits will be enormous as we learn to take advantage of its novel, powerful capabilities.

    The double-edged sword of being in the spotlight

    This story touches on a challenge that many health care vendors have faced for decades (especially with AI): the over-marketing and premature deployment of solutions before they're ready for mainstream adoption.

    As we've seen with other companies (looking at you Haven and Theranos), sometimes overhyping a business early on can drive up early investment but become detrimental in the long run. Rather than operate in stealth mode to set up appropriate data governance, test results, and publish studies in partnership with health care stakeholders, Watson Health sought to make a splash through a buzzy marketing campaign, showcasing its AI vision before it was ever a reality.

    It's that type of promotion that can certainly gain the attention of the general public, but can simultaneously irk researchers and clinicians who want to make sure a technology supports evidence-based medicine before deploying it outside of controlled settings—especially in a sensitive service line like oncology where people's lives are on the line.

    Give IBM credit, though. It made AI sexy, and it had a part in kickstarting the era of big data computing, owning the public mindshare of what AI could become for years. And trying to use AI to eradicate cancer is a moonshot effort that everyone can get behind.

    It's also worth stating that while Watson Health may be struggling in its broader ambitions, it has consistently managed to establish partnerships with many health care organizations across the globe, even announcing a new partnership with Humana just last week. While many industry skeptics have focused on Watson Health's struggles, it's still a billion-dollar business, and if acquired by the right company, it could be a valuable asset in building an AI portfolio.

    Follow the path to Artificial Intelligence at your organization

    AI

    A new wave of AI-powered capabilities is likely to improve—and even—transform health care operations, but success with these new technologies requires a strong foundational analytics program.

    Health care organizations must ensure they have the necessary data sources, architecture, governance, talent, leadership, and data-driven culture for their programs to deliver consistent value. With all of the necessary assets in place, health care organizations will then be ready to put analytics and AI into practice.

    Download Now

    Topics

    X
    Cookies help us improve your website experience. By using our website, you agree to our use of cookies.