I like to say that the intersection of health care and information technology (IT) is a great place to wake up every morning. Our environment has always been influenced by the trends, changes and pressures of both environments. Considered separately, they are two of the largest, fastest moving, most complex currents in the economy. The interplay between them creates unique challenges and opportunities. 2016 saw health care trying to assert itself—temporarily at least—as the stronger stream.
2016 was a mixed year for Cerner, with positives and negatives coexisting in a transitional health care environment. We competed extremely well in our core and emerging markets, but we also faced an uncertain health care policy and regulatory environment that decreased our clients’ urgency to buy technology within a specific timeframe. This contributed to health IT stocks in aggregate declining 32% for the year. While our overall pipeline remained strong, and we saw bookings growth in the majority of our businesses, we had disappointing performance against our own Cerner ITWorks and technology resale projections that impacted our ability to deliver all aspects of our financial plan and contributed to us missing guidance for some metrics. As a result of these industry-wide and company-specific factors, 2016 was a disappointing year for Cerner shareholders, with Cerner stock price decreasing 21% over the year. As you can see from the table on the preceding page, however, the value of Cerner stock has had an impressive run over the past 10 years and over our 30 years as a publicly traded company. It was tough to see that momentum broken during 2016, but I remain optimistic about our growth prospects. We have always focused on investing in long-term growth, doing what is right for clients, and managing the company, not the stock price, and this approach has resulted in very good returns for shareholders over time. If there is one thing I know after four decades at “the intersection,” it is that these currents move. I will use the later parts of this letter to communicate what I think is coming, and the role I believe Cerner will play.
First, here are some facts and highlights from the year. Other than the core financials, what you get out of this might depend on how well you know Cerner’s lines of business and how long you have followed our progress:
None of this changes the fact that it was a somewhat challenging year for health IT, which was already coming down off the compressed-buying-activity high of the mandated Meaningful Use era when it ran smack into late-year concerns about the potential health care impacts of an unexpected presidential election result. While we believe Cerner weathered the year better than most in our peer group, we are very critical of our own shortcomings relative to our financial plan. We can and must do better in 2017.
Before I go further, I’ll say a few things about other significant happenings that colored our year. They are a little too complex to make the highlight reel in the requisite sentence or two:
First, I want to provide an update on the Health Services acquisition. In October 2016 we completed the integration of the Health Services business with Cerner, including data centers, software development operations, field consulting, and client-facing organizations. We like our new Cerner Health Services associates a lot, and several of the associates have been integrated into the highest levels of Cerner’s leadership structures. Early on, we combined our revenue cycle teams, and a number of the more advanced Soarian® solution capabilities are now being leveraged by Cerner Millennium. We like our new clients, too. In 2016, more than 75% of clients migrating from one of the legacy Health Services platforms chose Cerner Millennium as their future platform. Those remaining on one of the legacy platforms have benefited from on-time software releases, including Meaningful Use 3 requirements for Soarian® clinical systems. Cerner Health Services hosting uptimes have hit new high water marks in line with industry-best CernerWorks hosting, and we have successfully launched new service offerings based on existing Cerner Millennium service models. There is plenty of work yet to do to responsibly serve and migrate these clients, but I consider the acquisition and integration a success.
Second, I want to comment on our project work for the U.S. Department of Defense (DoD). In 2015, together with Leidos and our other partners, we were awarded the contract to modernize the DoD’s health IT capabilities. Throughout 2016, our team worked with our partners to design, configure and test the initial phases of the MHS Genesis system, the unified electronic health record built on the Cerner Millennium platform. The system had its successful first go-live at Fairchild Air Force Base in February 2017. Additional fixed-facility deployments will occur in 2017, and work is underway on complex environments like theater hospitals, forward resuscitative sites, naval surface ships and submarines. There is a real feeling of pride in being able to bring modern health IT to our nation’s military personnel and their families. It is a huge responsibility and an honor.
And third, many of you are aware that I was diagnosed with cancer at the beginning of 2016. I gave an update on it in this letter last year. Although treatable, the treatments were physically difficult. Fortunately, I have had a career-long practice of surrounding myself with a great team that is not dependent upon me to move Cerner forward. Together with Cerner’s Board of Directors and co-founder Cliff Illig, they kept me up to date on the essentials and let me focus on treatment and healing as needed. I made a surprise return to the public eye in front of 14,000 people at our annual Cerner Health Conference in October, where I spoke about my experience.
At Cerner, we teach that health care ultimately becomes personal for each of us, fueling our work. We encourage each other to bring our stories to work and make them part of our mission. My sister-in-law Linda’s tragic death from sepsis in 2006 became the catalyst for a life-saving cloud-based surveillance solution we developed in 2010. A peer-reviewed study just published in the American Journal of Medical Quality in February found that the St. John Sepsis Surveillance Agent performs better than the best national protocol for sepsis detection. It identifies roughly three times the number of individuals who are on a path for sepsis to occur, and identifies them quicker, sometimes prior to the infection taking hold, and two to three hours before the national protocol algorithm triggers, on average. And because it has been deployed in the cloud to over 550 hospitals, 750,000 to 1 million patients had their care altered last year because of the agent. Each client is different, but a published report from one 284-bed hospital that implemented the agent indicated they had a 30% reduction in the risk of adverse outcomes after one year. Our clients tell us the agent helps save a lot of lives. This is incredibly meaningful to me on a personal level, because my sister-in-law’s death from sepsis is what prompted our work in this area. We can’t go back in time and save Linda, but my family is powerfully moved knowing that “her” agent has helped doctors and nurses prevent the same devastation from happening to other families.
In another personal example, my wife Jeanne’s 10-year journey with metastatic breast cancer and her experience carrying bags of her own health records around helped propel our intense focus on interoperability. This led to Cerner’s co-founding of CommonWell Health Alliance, a not-for-profit trade association of companies working together to provide patient-centered interoperability of health records, regardless of location, provider or software. Four years after launching, more than 20 care settings now are represented by its 60-plus members, which include the majority of acute care EHR companies. Up to this year, enrollment and access to patients’ health records via CommonWell has been done through the 5,000-plus doctors’ offices and other health care sites live nationwide on CommonWell services. In August, CommonWell announced that people would now be able to self-enroll in CommonWell, link to their health records in the places they receive care, and view their records on the network. We are making progress.
And now, I am using my own encounters as a cancer patient inside the gears of the health system to bring new focus to the role of the person in care delivery. The provision of health care is incredibly complex, and conditions like cancer require large teams of caregivers to work in a coordinated fashion. Sometimes the coordination breaks down, and the patient suffers the effects and is left patching up the pieces. These breakdowns can have real impacts on outcomes. It is time for the patient to be part of the team, and for their experience to be a more seamless one. All of our clients are focused on improving the patient experience and patient engagement. I believe there is much Cerner can do to improve people’s experiences with health care, as well as their outcomes.
Throughout Cerner’s history, we have grown by adding new rings around what we already do. We started in the laboratory—the nexus of a single complex, interconnected health care enterprise—and expanded. Today we still do lab, but we also do integrated delivery networks, national systems of health care, and employer- and consumer-focused work. Our pattern is to grow outward—deeper into the complexity of the multi-trillion dollar health care industry—without ever leaving our clients and their missions behind.
In the rest of this letter, I want to explore where else this journey might lead us. But to understand the direction, it’s important to discuss some realities in our health care environment. Let’s take a look.
Here is a little disclosure that will mean more to U.S. basketball fans: I make an annual practice of working on this letter during March Madness.
This particular March seems unusually mad, and it has nothing to do with basketball. That’s because here in the U.S, we are 60-something days into the first 100 days of the Trump presidency, and the political rhetoric on both sides around the potential repeal, replace or reform of the Patient Protection and Affordable Care Act (widely known as “Obamacare”) has reached a fevered pitch. As I write this, the Republican-sponsored American Health Care Act of 2017 vote has recently been pulled for lack of support, and the nation seems to be in a suspended state of uncertainty about the future legislative direction of health care. The simplified political rhetoric sometimes obfuscates the real forces that drive health care spending in the U.S., which also are the essential drivers for health IT.
For those interested in reading the tea leaves about health IT, I suggest it’s good to look past the rhetoric and “FUD” (fear, uncertainty and doubt) for a moment and lay out the base case for a growing future business:
1. We have an unsustainable cost curve. In the U.S., growth in health care spending has outpaced the overall growth in the economy for a half century. Stated differently, health care is taking a bigger piece of the American pie every year. The Centers for Medicare and Medicaid Services (CMS) indicates that almost 18% of the U.S. gross domestic product (GDP) is already spent on health care, and that percentage is only going up. As a point of reference, the $3 trillion we spend as a nation on health care far exceeds, say, the $600 billion we spend on defense. Out-of-pocket spending is high and rising. The Medicare population spends well above the average of $10,000 per person, nearly all of which is paid for by the federal government. And no economics expert alive today has been able to reverse the trend, regardless of which political party has been in power. It is not sustainable, and yet so far it has been unstoppable. It will create the demand for fundamental change. The heart of the cost issue is neither access nor insurance reform, although those get a lot of attention. The heart of the cost problem is how to actually stop spending as much money to achieve the desired results. Although the U.S. has the highest level of spending, other developed nations face a similar predicament, so it really is a global challenge.
2. The populous baby boom generation (born 1946-1964) is reaching retirement age. The first boomer turned 65 in 2011. As has been well-chronicled, the impact of Boomers on health care delivery and cost will continue to grow at the same time their contributions to Medicare (through the tax rolls) will decline. Further stressing the system, people are living longer lives than they were when the system was designed.
3. Chronic conditions—ranging from diabetes to heart disease—account for 85% of health care spending, and obesity-related chronic conditions have increased at alarming rates. The top six cancers are said to be preventable. Prediction, prevention and intervention can play a real role in reducing costs.
4. The broken health care payment system is driving costs. In the U.S., doctors still mostly get paid based on the volume of visits and procedures they generate, not based on keeping people well. As Cerner Board member Denis Cortese, M.D., has written, we have “a system inherently and unavoidably designed to drive and incentivize high cost, not great value or customer care.” Our payment system drives workflow pain points such as excessive documentation to achieve E&M codes and patient care pain points such as professionals “practicing at the top of license” and limiting any form of patient contact that can be performed by a lesser professional.
5. Any stakeholder with a brain will raise the, “What are we getting for our money?”, a/k/a “Prove the Quality”, argument. The U.S. has the most expensive health care system in the world but is ranked behind many other nations on measures of infant mortality, mortality amenable to medical care, and mortality at age 60.
6. Over the last several decades (and particularly the last seven years of Meaningful Use), health IT has been recognized by members of both political parties as a means of increasing quality while reducing cost. The core content of U.S. health care has been digitized through EHRs. While there is grueling work still to be done—particularly for interoperability, usability and patient safety—digitization can now be assumed. Throughout the rest of the world, varying states of digitization exist.
7. The second and third order impacts of achieving a fully digitized state will be profound and create far bigger opportunities than any we have seen in the current EHR era. Think of similar impacts that occurred after music, books and banking were digitized. The largest impacts were on the middlemen. In the music industry, there is more music being created and consumed today, but with almost no brick-and-mortar music stores. This is what I expect to happen in health care.
8. Even though health care processes are now automated in a basic sense, the current systems still contain many inefficiencies—such as fragmentation of care and high administrative costs—as well as opportunities to make better use of the explosion of scientific and medical knowledge.
9. While health care is ripe for major transformation, system-level innovation—changing both health and care—requires scale. Doing it around the world requires global scale.
10. Following the trend of other industries, health information systems of the future will become more natural to use, intelligently anticipating and simplifying workflows, and cognizant of all factors—clinical, social, environmental and geographic—that affect health and personal engagement in managing health.
Regardless of where you stand, consider that the dialogue around Obamacare and its Republican alternatives is actually quite alike in that its main focus seems to be access and insurance reform, not care delivery reform. The most discussed cost elements are related to the lack of a shared pool to spread the healthy in with the sick. Other than pricing on medications, there isn’t much of a discussion around cost of care delivery. My point is, whether Washington decides to replace or revise core elements of Obamacare, the cost implications of the ten forces above will continue to pile up, even accelerate. To think otherwise would be the real March Madness. Absent rationing care or patently denying care to the sick and vulnerable—two things most politicians claim to be against—the State and Federal deficits will continue to mount. The hunger for new solutions—and new enterprise entrants—will grow with it.
To an entrepreneur, the opportunity to elevate quality and decrease cost is massive. In my view, the race is on. It is a race between technology-driven transformation—and budgeting to limit care through a single-payer system. For the sake of future generations, I know which of the two alternatives I want to win. The money to transform the system is in the system itself. Information technology is the single biggest lever to drive cost down and quality up. An innovative health information technology company—with deep clinical knowledge, proven platforms, global reach, and an enormous technological moat—could make a real impact.
Even though Obamacare and its alternatives have become heavily politicized, there is one bit of bipartisan reform passed and signed in 2015 that has a chance of reshaping care delivery in a powerful way.
To explain the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA), a little background is needed. Some 55 million U.S. citizens are on Medicare, the Federal safety net insurance program for people over retirement age and younger people with disabilities. In the 1990s, after years of unsustainable growth in Medicare spending, the Balanced Budget Act of 1997 attempted to fix the problem by introducing the Medicare Sustainable Growth Rate (SGR), a formula that tied Medicare spending growth to U.S. GDP growth. Unfortunately, because health care spending routinely outpaced growth in the GDP, the SGR formula eventually resulted in unreasonable annual payment cuts for the doctors who cared for Medicare patients. Seventeen times between 2003 and 2014, Congress passed temporary “Doc Fix” legislation to prevent the cuts from being applied. In 2015, Congress passed MACRA, also known as the “Permanent Doc Fix”, with broad bipartisan support.
MACRA repealed the flawed SGR, but it also introduced significant changes in the future payment model for Medicare physicians. Instead of authorizing traditional fee-for-service payments, MACRA mandates that future payments to physicians who serve Medicare patients be based on one of two value-based payment schemes.
Inside Cerner, we believe that MACRA could be the government’s next Meaningful Use—in other words, a program that will impact buying behavior. Unlike Meaningful Use, which mandated EHR adoption without requiring proof of better outcomes, MACRA is a competitive mandate that will reward providers who achieve better outcomes and penalize providers who do not. Most providers are thin-margin organizations, and a 2% penalty on half of their revenues is a major event. As the payer for roughly 50% of care, Washington knows how to change health care using the carrot and the stick. Now that the core system is digital, they can demand the measurements to continuously move the meters toward higher quality, lower cost care.
Broadly, CMS has been using its status as the largest health care buyer in the world to move the market toward outcomes-focused, risk-based payment, with former Secretary Burwell calling for 50% of all Medicare payments to be value-based by 2018.
With the change in presidential administration, we will see a predictable review and perhaps even a little drama. However, policymakers seem confident that MACRA will continue as planned. It is difficult to predict the direction of policymakers, but based on the forces and pressures that drive health care spending ever upward, it is hard to imagine ignoring the power of shifting from fee-for-service to a value-based system inside the competitive marketplace.
There is a reproducible formula for how value and results are created at Cerner, one that starts with vision and then flows downstream to mission, strategy, structure, process, tools, effort and finally results. The “water” can be muddied at any point along the stream, but can ultimately only be as pure as what preceded it. Whenever we don’t like our results, or perhaps their trend or variance, we stop and examine every level in the formula, starting at the top.
I consider the most important executive functions to be the ability to develop vision, mission and strategy. Together, they’re the art of working on tomorrow, today. Vision is based on contemplating a desirable future state, of connecting dots to mentally solve a problem that exists in the future. Once we have clear vision of the compelling future state and have identified an actionable mission to get there, it is time to choose the strategies that will accomplish the mission. One of our dependable growth strategies is to build world-class information technology platforms that address the future need. There is art in the timing.
Last year, I spent considerable space in this letter on the evolution of population health management and why value-based-payment was necessary to drive it. Without rehashing all of that, at a very simplified level, population health management is a model of care provision where a group of providers take on financial risk and responsibility for the health of a defined population, and they actively work to improve the health of every person in that population so that costs are controlled and outcomes are improved. In our view, population health management is going to have to be the answer to the question of how to curb the drivers of health care spending. Payment reform adds a missing ingredient, the immediate “Why now?” that gives providers the incentive to transition to risk-based, population-based, outcome-based models of care.
As the market emerges for population health management systems, we like where we’re at and what we’ve got. We started with a compelling vision for population health management. We defined the mission in 2012, and we began building it. Our HealtheIntent platform is entirely cloud-based and is designed to manage the health of defined populations whose boundaries are not limited to a single health care enterprise. Its sources of data are boundless, including numerous EHR platforms, insurance claims, prescription data, consumer-contributed social data, GPS and environmental data. The data aggregation platform went live in 2013 and now has a growing family of solutions sitting on top of it. By the end of 2016, we had more than 100 clients, including some of our primary competitor’s EHR clients, and the database contained more than 6PB of aggregated data from more than 88 million persons. Another core growth strategy after building world-class information technology platforms is to add high-value services to leverage the technology and extend its benefits. The era of population health management will provide many opportunities for value-added services that make use of the power of the HealtheIntent platform. Like with Cerner Millennium, it could very much end up being about being at the right place at the right time with the right stuff.
Health care is incredibly complex. Consider for a moment the fact that, in 1950, the total body of published medical knowledge likely progressed at a rate sufficient to double every 50 years. By 2020, however, the doubling rate for new medical knowledge is projected to be every 73 days. That’s crazy, but it might not be too much of a problem as long as all of the new knowledge gets diffused into practice quickly. Unfortunately, it doesn’t. A famous study by Balas and Boren showed that it takes 17 years for a mere 14% of new evidence published in medical journal articles to work its way into practice 50% of the time. With the help of information technology, though, these and other hard realities can become opportunities. IT is the only lever strong enough to change health care.
The British science fiction writer Arthur C. Clarke once wrote,
“The Information Age offers much to mankind, and I would like to think that we will rise to the challenges it presents. But it is vital to remember that information —in the sense of raw data—is not knowledge, that knowledge is not wisdom, and that wisdom is not foresight. But information is the first essential step to all of these.”
He’s right. What he’s describing is a sequence, a series of transformations. Solving big problems in health care requires this exact sequence, and that is what is at the heart of the most sophisticated branches of IT—artificial intelligence (AI) and cognitive computing. I want to share some examples of what we are already doing using cognitive techniques, and reflect on the tremendous promise they hold for the future.
The very name “Cerner” is related to the concepts of discernment and understanding. We created our first AI-based solution, Discern Expert®, in 1988. Discern Expert was (and remains) an event-driven, rule-based decision support software application that allows its users to define clinical and management rules that get applied to ongoing clinical event data captured in other parts of the system. Its uses are really only limited by a clinician’s imagination; for example, it can be used to send out alerts if a relevant lab result has returned a certain predetermined value so that a physician can determine whether to stop a surgery from occurring. The HTML-based Discern Advisor® solution was added in the mid-2000s to focus on medication usage criteria, automation of complex decision trees, patient scoring systems and clinical calculators in support of complex clinical decision-making.
Today we have numerous Discern®, CareDecisions, Cerner Math and HealtheIntent solutions that use a variety of machine learning techniques. Some are auto-adaptive, while others require more human curation to tune their effectiveness. Some are aimed at accomplishing difficult data mapping tasks at tremendous scale, while others are laser-sighted on assisting in the prediction and prevention of a specific cost, consequence or condition. In aggregate, they are aimed at helping to solve big problems in health care.
Prior to her sudden and tragic death from strep pneumonia sepsis, my sister-in-law Linda was a kindergarten teacher in rural Kansas. A challenging feature in her case, not unusual in rural environments, was her movement between different care venues as her condition deteriorated. It was only days later in the final location, an ICU an hour away from home, that her true condition was recognized at last and antibiotic treatment was initiated—too late, unfortunately, to save her. Each venue—whether rural primary care, rural emergency room, ambulance, larger hospital, or ICU—had generated its own data about her, but there was no common connection among the venues sufficient to create wisdom about her condition, much less foresight about what to do about it. When Cerner engineers sought to create a solution capable of helping to save the next “Linda,” they had to solve the big problem of how to collect various types of data from different venues, gather it into the cloud, and transform it into a single understandable record that could become the basis for taking further actions such as monitoring and generating alerts. They did this with help from cognitive techniques such as ontology mapping and adaptive knowledge processing. Inspired by the cross-venue complexity of Linda’s story and others like it, the standards-based longitudinal person record is now an important foundational building block within HealtheIntent, our platform that accepts data from an unlimited number of sources, including multiple venues using disparate EHRs.
As of today, we don’t make any cognitive systems meant to replace humans, but we use cognitive techniques to automate tasks that are too big or repetitive for humans to handle, and we use AI-based machine learning to provide adaptive, embedded forms of assistance we call “cognitive moments”—discrete information-based predictions and interactions that help advise and assist a physician or other clinician at the moment of decision. They are adaptive in the sense of reading contextual clues present in the myriad combinations of roles, actions, events, venues and conditions; and then consulting models to identify which patterns exist that might inform what should happen next. When appropriate, our CernerMath data scientists leverage machine learning to build predictive mathematical models for use within CareDecisions and elsewhere. To date, they have built more than 50 predictive models.
Given the Arthur C. Clarke quote, it shouldn’t be a surprise that the raw data you feed your AI models matters a great deal.
Last year, Microsoft launched Tay, an AI-based Twitter chatbot that had been built using “relevant public data” that was said to be “modeled, cleaned and filtered.” That certainly sounds like a good diet. But once it was turned on and fed a supplemental diet of sarcastic and abusive tweets, it went from saying “Humans are super cool” to voicing racist and misogynistic opinions in less than 24 hours. Needless to say, it was turned off.
At Cerner, we believe we have a near-perfect diet for our clinical AI-based models. The discovery, development and validation of our CernerMath predictive models have primarily been performed using Cerner’s own Health Facts® data warehouse.
This will get a little technical, but bear with me. Health Facts is a HIPAA-compliant, de-identified, privacy-protected, EHR-derived, ontology-mapped, longitudinally Enterprise Master Patient Index (eMPI)-linked repository of the serial care-episode health records of the patients cared for at nearly 700 U.S. health institutions that have established data-rights agreements with Cerner. Begun on January 1, 2000, and expanded daily since then, Health Facts currently contains more than 150 million distinct persons’ longitudinal records, more than 400 million episodes of care, and more than 5 billion clinical events. Health Facts is not a context-poor “claims” database. Instead, it includes the majority of the content in patients’ EHR records, including medications, lab results, procedures, problem list entries, diagnoses, and claims—with each data element or transaction time-stamped with minute-level precision and each successive episode for a given person longitudinally linked via a key that is encrypted from the eMPI.
When we want to “teach” one of our machine learning based AI models, we might begin with a cohort of 20,000 or more cases and a comparable number of controls, and give it several hundred input variables. From there it gets even more technical. Compare this to a traditional clinical trial where you have a cohort of perhaps 100 or 200 patients and you hold all variables constant except one. This is a very different way of finding patterns and evidence. We are studying what is actually occurring versus a staged activity, and looking at large numbers of cases. This type of access to big data can bring to light patterns that have never previously been known, leading to further investigation and modeling. If we can use that new pattern to accurately predict the next cases, then we know we have found something of real value.
We have known for a long time that cognitive techniques hold a lot of promise for health care. They are beginning to deliver on the promise. As we look to the future, we see a number of logical next steps. Many of these are aimed at helping caregivers deal with the increasing complexity of health care and the rising expectations to be aware of “everything,” even when “everything” is beyond the ability of one person to comprehend.
A natural follow-up to the standards-based longitudinal person record will be the so-called “contextualization” of the record so that each member of the care team gets a perspective of the person’s record that meets the needs of their role and venue. We can manually curate content for single-condition needs, but the comorbid cases—those involving two or more coexisting medical conditions—require a significant knowledge discovery and curation effort. We plan to use machine learning techniques to take a more systematized approach to discover such content, yet verified by our informatics and clinical personnel. This type of data-driven approach to derive knowledge will help create more cognitive moments for providers and researchers that are both contextual and adaptive in nature. We can also use techniques similar to those used to create the longitudinal person record to help assimilate the rapidly expanding body of new medical knowledge into what we call model practice and model outcomes.
The cognitive techniques that are transforming other industries are also transforming the provision of health care. Given the relative complexity of health care, it will take time to figure out which techniques will be most impactful and build them out. It is an exciting time to be involved in health IT.
Cerner believes in innovation. A large part of our value proposition to our clients is that we establish and maintain our technology leadership, remaining contemporary so that our clients receive benefits from meaningful advances in computing. Never in the history of IT has the rate of change moved this fast nor had such a ubiquitous impact on almost everyone in our society. The old concept of rising expectations is at work, down in the complexities of health care. The cognitive era of health care will happen, and we will lead.
To our shareholders new and old, thanks for your continued confidence. I continue to be amazed by the amount of growth potential there is in health IT. The future is not guaranteed, but I like our chances of being the ones who can positively impact the delivery of care.
Neal L. Patterson
Chairman of the Board, Chief Executive Officer & Co-founder