HIMSS17
Orlando, Florida
February 20, 2017


Prepared opening remarks:

It is my great pleasure to be with you this morning.

This topic is very important to IBM and very personal to me. You saw a little bit about my background and some of the boards I am on with this introduction. By the time I wrap up, I hope you’ll understand how special this topic is to me.

When I became CEO five years ago, it was a really great moment. It was when we introduced Watson, our cognitive learning system, to the world in a game show called Jeopardy!. At that time, I declared to IBMers that our next “moonshot” would be in the area of healthcare. After Watson won on Jeopardy!, we decided that its first real–world grand challenge would be the hardest one we could imagine: cancer. I said I would never be arrogant enough to believe that we could change healthcare by ourselves, but we could change a small piece of it and be a part of that solution. And with a lot of friends and partners, and all of you here, I felt this was the moment.

Then, two years ago here at this HIMSS conference, our moonshot took another big step as we announced the launch of Watson Health with partners including Medtronic, Apple, and Johnson & Johnson. It’s now more than 7,000 people strong.

Now, we stand at another special moment when we as an industry need to come together to forge a path forward. Historic social and political shifts are underway and they are likely to affect healthcare. And, the role of government and regulators — in allowing mergers or possibly changing the Affordable Care Act — is on the minds of many.

Those are serious issues, but they are not the most important things happening in healthcare today. This is a profoundly hopeful moment for all of us. It’s a time when we can actually transform many parts of healthcare. We can reinvent how experts conduct research, render diagnosis, provide treatment and monitor patients. And, we can fundamentally change our understanding of personalization, privacy, epidemiology — and of medicine and health itself.

The best news of all: I think it’s within our power, and as a team, that we can shape our industry and our world for the better — as providers, doctors, researchers, patients, and citizens.

I am talking about what we at IBM call the Cognitive Era — an era that will play out in front of us. I want to explain our perspective on what this era means and what must be done to ensure that this Era emerges in the right way for healthcare.

Today, I hope to persuade you of three things:

 1. Cognitive healthcare (the idea of systems that learn) is real and it is here. It has entered mainstream life and care, it is going to live on the cloud, and it is already changing everything about the way we approach health.

 2. To scale this, two to three key architecture decisions will be made by IT professionals in the next 18 months that will affect every industry in the world — healthcare perhaps more than any other.

 3. Finally, the Cognitive Era could usher in a “golden age” — in business and society at large — if we shape it wisely. I also believe healthcare could be the leader for the world in showing what it means for an industry to shape an era wisely.

I. Cognitive healthcare (artificial intelligence) is mainstream and real. It has entered mainstream life and care; it is going to live on the cloud; and it is already changing everything about the way we approach health.

Let me start with a question for all of you. How many of you are trying to digitize something? [SHOW OF HANDS]

Okay, almost everyone. So, no matter the industry, this supports the idea that digital is the foundation for everything. But, competitive advantage is going to come from being cognitive — systems that learn massive amounts of data.

We are seeing a land rush into artificial intelligence — including in healthcare. We know AI is growing rapidly, but not all AI is the same. This is not about consumer applications like devices that answer questions in your home; this is not speech to text on the front of a search engine. It’s about augmenting intelligence. It’s about serious applications in financial services, retail, energy, transportation, education — and importantly, in healthcare.

Having worked for several years with thousands of clients and partners to apply Watson to healthcare and business, I’d like to start with five lessons that we have learned about what cognitive for enterprises and societal systems requires. It is not enough to think of artificial intelligence or natural language; you need vision, you need deep learning, and you need many more algorithms. It must:

1. Provide a range of powerful cognitive services — from machine vision to natural language processing and deep learning.

2. Provide transparency — so doctors, patients, researchers and other decision makers can see the data sources and training methods and thus have confidence in the recommendations and insights produced by your cognitive systems.

3. Be designed for and by experts in your industry — it must be domain specific. Healthcare is not like retail or banking. It has unique dynamics, lexicons, information, expertise and requirements … from regulation to competition

4. And, it must live on a cloud platform that is ubiquitous and can handle cognitive workloads that are enterprise–strength in reliability and security. It must be a cloud that is built for big data and for security, and with hybrid architecture in order to connect across the wide healthcare spectrum. For this wide ecosystem, you must be able to connect people — from an academic center to a community hospital, as an example.

5. Finally, it must be an open platform because this is an industry all about innovation — innovation that doesn’t come from one person but from an ecosystem of people.

You would not be surprised if it told you that this is the basis of the Watson and Cloud platform, which has emerged as the AI platform for business … and for healthcare. Watson solutions are being built, used and deployed today in more than 45 countries and 20 industries. Watson is being trained in nine languages. And, Watson is on a path to touch a billion consumers by the end of this year — many of those through many of you.

I’ve said that this is mainstream. While there are others with solutions, let me use us as an illustration to think about what a world in healthcare looks like with Watson — with cognitive. Let’s talk about a few dimensions of healthcare that I think are ripe to change.

1. Discover new medicines and solutions:

With Watson’s help, the Barrow Neurological Institute in Phoenix, Arizona recently sped a breakthrough in ALS research, discovering five genes associated with ALS that had never before been linked to this disease. Watson ranked from 1 to 1,500 the genes that could affect ALS and provided evidence–based predictions regarding which genes might be associated with it. The Barrow team examined Watson’s predictions — and found eight of the top 10 genes affect the disease. Importantly, these evidenced based predictions were all illuminated as to why — that’s what matters.

2. Personalized care:

Personalization of care is a challenge and there is a worldwide shortage of physicians — notably in cancer care.

Let’s consider oncology. In 2012, there were 14 million new cancer cases introduced. Estimates are that will increase to 22 million within the next two decades. Seventy percent of these new cases are in developing countries, and countries like India and China and those in Southeast Asia simply do not have enough oncologists. Think about 1 doctor for 1,600 patients in India, for example. Now, hospitals in Thailand, India, Finland, China and the US are using Watson for Oncology to personalize care for patients.

For example, India’s Manipal Hospitals — a large private hospital chain — deployed Watson for Oncology in late 2015 to help oncologists identify better treatment options for the more than 200,000 patients who receive care there every year. Validation studies compared the amount of time taken to capture and analyze data to generate recommendations with and without Watson. Studies showed that it took an average of 12 to 20 minutes when done manually and just 20 seconds when done with Watson.

Starting next month, doctors at Jupiter Medical Center here in Florida will use Watson’s cancer–focused insights to help match patients with the most effective personalized treatments in a very busy community oncology setting.

3. Precision medicine:

Precision medicine is another innovative approach to cancer diagnosis, treatment and management that takes into account a patient’s genes, family history, environment and other factors. When I say precision medicine, I think genomics, our environment and how we live.

At the University of North Carolina’s Lineberger Comprehensive Cancer Center, Dr. Ned Sharpless used Watson for Genomics to analyze more than 1,000 recent cases. To me, this is the great promise of the era we’re walking in. In many instances, there was “no standard of care” for these hardest cases, according to Dr. Sharpless. Watson identified the same potential therapies as their tumor board 100 percent of the time. More importantly, in nearly 300 patients, Watson found clinically actionable insights that had not been identified by the humans. So, one–third of the patients potentially had more options available to them. To me, that’s what this is all about.

We’re scaling our work in genomics now and bringing it into local communities. Building on years of work with academic centers, in 2016 we launched IBM Watson Genomics from Quest Diagnostics — precision cancer medicine powered by Watson — to make cognitive genomic analysis available to any patient in the U.S. through their doctor. In addition, we have partnered with Illumina to integrate Watson for Genomics into their tumor sequencing process. The aim? To expand access to genome data interpretation and help standardize and simplify genomic data interpretation.

And in a move to bring Watson to the frontlines of patient–centered care, today we are announcing that IBM Watson Health has entered into a partnership with Massachusetts–based Atrius Health, an organization renowned for its leadership in value–based care. Together, we will develop a cloud–based offering that provides physicians cognitive insights about a patient’s health status and care plans right within their electronic health record system — to help make each health decision personalized to that individual.

Precision medicine is coming alive.

4. Manage chronic disease:

Last September, Medtronic showed a first version of the Medtronic Sugar.IQ with Watson app — a first–of–its–kind cognitive app powered by Watson and designed to help make daily diabetes management easier and more effective. The app uncovers important patterns and trends to help show how behavior affects glucose level in real time. And, early research shows the technology can predict hypoglycemia 2 to 3 hours in advance with an accuracy of 85 to 89% — never done before.

5. Address high–needs individuals and vulnerable populations:

This is an area where much help is needed. IBM is working with 50 health and human services agencies in 18 countries — in the US, in 38 states and with every federal agency. We’re analyzing data for 44 percent of all Medicaid beneficiaries today.

This idea to help across all of these dimensions to me is a different world. It’s already starting to happen — it’s mainstream and it’s here.

II. Key decisions will be made in the next 18 months that will affect every industry — perhaps healthcare more than any other.

All of us as professionals are at the intersection of health and IT in some way. You will have a great influence on three important decisions — these critical decisions will be made to ensure scaling, security, reliability, and privacy. They’ll address three key platforms: cloud, data and AI. Let’s address what you’d look for in each one of them. Hopefully in the Q&A we can talk about the downside of not implementing this approach.

CLOUD

The first decision to make is about the cloud. You’re going to need a cloud that’s optimized for all this cognitive data — one that’s hybrid and secure. Businesses and institutions need hybrid cloud architectures — it’s a reality. Those clouds must be optimized for data and AI. And, they must be built for advanced security — because it must be.

That is why we have built the IBM Cloud that we have today. All of the examples I shared are possible because the solutions are built on the IBM Cloud — with proven operational performance, storage, and access. It is the industry’s only GxP, HIPAA–enabled cloud built within a Quality Management System. It is healthcare–ready.

Blockchain will do for trusted transactions what the Internet did for information. It creates a shared ledger for the secure transfer of any asset. We are building a complete blockchain platform. In fact, we recently announced that the FDA is working with IBM Watson Health to explore blockchain technology in oncology — integrating data from multiple sources — EHR to PGHD — to provide a complete view of patients. I hope in the Q&A we talk more about what blockchain can layer on and do.

You will make a decision on a cloud platform with these attributes.

DATA

Next, you will have to make big decisions about your data. It’s been a long time since we talked robustly about changing a data architecture. But, it matters now because we have to have to pull together so much different data, a full spectrum of it — including structured and unstructured data — to create the world I just described.

Consider this unstructured kind of data. Images are estimated to make up as much as 90 percent of all medical data today. It has been difficult for physicians to glean important information from images. But, it can be done.

You must maintain control of your data and your insights — these are your most differentiating assets. 80 percent of the data in the world is not searchable on the Web. Think about that. That is where the greatest insights are — and they’re yours. The idea of having a data platform that allows you to combine all of this data and the insights are yours is going to be an important question to ask.

Again, these requirements are why we have created the Watson Health platform the way we have. We already have a vast trove of data: 300 million medical records, 30 billion images, and 40 million PubMed and other research abstracts, and every patent.

All with secondary use rights for you to combine data … and the insights are yours.

You will make a decision on a data architecture platform.

AI

Finally, you’ll have to make a decision about AI in a cognitive platform. To capture the deep value in your data, you will need to look for a platform that provides a range of services, incorporating technologies like deep learning, machine learning, neuro–linguistic processing, patient similarity and other cognitive technologies specifically geared to healthcare.

You’ll have to provide transparency so your users and decision makers can see the data sources and training methods and thus have confidence in the recommendations and insights produced by your cognitive systems. You and your doctors will want to know how these systems were trained and how answers were derived.

Watson has been trained by the world’s greatest researchers, physicians, educators and policy leaders — from Memorial Sloan Kettering, to Cleveland Clinic, to Mayo Clinic and more. It has ingested the world of medical literature. It reads and learns and has the experts correct it. It is trained and tested, repeatedly. Watson is not a “black box” that simply spits out an answer. It shows how and why an answer was given. Watson shows the doctor the basis for its recommendations and its level of confidence in them.

These are all the things I’d look for in a cognitive platform. Think about a cognitive platform and what you can do with it.

Today, we have the pleasure of announcing new initiatives for healthcare ecosystems to unlock insights hidden in data silos and help communities at scale. The Central New York Care Collaborative will deploy cognitive to create the nation’s first population health platform that connects an entire region — 1,500+ care providers serving citizens of Central New York who are on Medicaid. It will integrate data across primary, behavioral, community and acute care settings, updated daily, to create a holistic view of each patient, enabling clinicians to provide higher quality, individualized care while reducing Medicaid costs.

So, think about the platform thought and what you can do with it.

III. Finally, the Cognitive Era can usher in a golden age — in business and society at large … if we shape it wisely. Healthcare could be the leader here.

This world that I’ve described is one that’s on the cusp of changing — one in which you’ll have tremendous influence on the decisions you will make. As I mentioned, my last point is that it could be a golden era. Every new era of technology raises amazing aspiring dreams, yet it also comes with questions.

This is one of the reasons why I returned to the World Economic Forum Annual meeting in Davos this past January; IBM hadn’t attended in decades. I returned this year because there was much discussion about AI. Is it good or bad for the world? What would be its impact, particularly on jobs? I feel very confident to answer these kinds of questions because all of us that have such an impact on this must take it seriously. And when a new era comes — which doesn’t often happen — it’s our responsibility to guide that technology into the world in an ethical and really enduring way.

As I prefer to say, IBM is 105–years young — not old. I feel we’ve been blessed to work with many of you on what have been some of the world’s toughest problems. This work has ranged from Excimer laser surgery technology to the DNA transistor, from setting up the US Social Security System to landing a man on the moon. We call it a moonshot, as we really did work on landing a man on the moon.

So, I thought we’d have a little fun with this and show you a short video clip. How many of you saw the new movie Hidden Figures? I’m not a producer nor am advocating it for personal reasons! This clip is about the dawn of NASA’s space program. [SHOW VIDEO]

It is a great movie — I am a little biased. IBM has participated in every US manned space effort in history. When we say we believe in moonshots, I really believe in it! And we know that achieving that kind of goal requires not only technology … but also values.

So now, let me return to what I said about returning to Davos. Just last month we did something we hadn’t done in over a decade. We introduced a letter to the whole IBM community called, Transparency and Trust in the Cognitive Era.

I want to end on this thought about these times being a Golden Era — if we usher it in right. This letter has a number of main tenets that apply to cognitive systems in any industry. I hope you embrace these principles. I’m going to be sure IBM does. There’s three of them, and they’re in everything we build, what we do and what we bring to the world around this technology.

1. Purpose: Make no mistake that we believe AI’s purpose is to augment human intelligence and what man does — not replace it. We are here to augment what man does. This is man and machine working together. We are building an ‘aide’ or a colleague that supports a doctor, a nurse, a radiologist, an IT professional in security, etc. This is about augmenting intelligence, not about fear. Everything we build is built that way.

2. Transparency: We will be clear with you and our clients about when and where AI is being applied, about the data and training that went into its recommendations, and the business model. Don’t believe in a world where all of the insights and intellectual property belong to one person. This is a world where we’ll bring insight and you’ll bring insight — and the training algorithms go to you. It’s a world of transparency.

3. Skills: As far as I can remember, with every new era comes a set of new jobs. Some jobs will change and some will be new. Therefore, it’s going to be incumbent upon all of us to train for the new skills required in the new era. In fact, we coined a term called “new collar.” It’s not blue collar, it’s not white collar. It’s a future job that’s going to bring data and technology to almost anything that you do. And, it starts by preparing our kids as they graduate school and retrain all around what this world does.

Those are our three tenets around purpose, transparency and skills.

Conclusion

I conclude where I began. I couldn’t be more excited to be here today. This is an historic moment that I will always remember.

This is a world of forces at work — some we can control and others outside of our control. If I could make one recommendation: don’t be tentative. This is a time to play offense by building the future on cognitive platforms. We can create a world that is healthier … more secure, less wasteful, more productive, and more personalized. In the end, it’s a world that is fairer, more diverse, and more just. I believe this is a world we all want to live in.

Thank you for letting me keynote this morning. I wish you a great conference.

[Discussion continues with Ginni Rometty and Steve Lieber, President and CEO of HIMSS.]