Blog Posts for Robert M. Califf, MD

Perverse Incentives, Congenital Heart Disease, and American Medicine


July 9, 2019

Robert M. Califf, MD
Director, Duke Forge

As a parent of a child with congenital heart disease (CHD), I felt a mixture of grief and horror upon reading a recent New York Times story that laid bare serious problems with a CHD surgery program at a nearby health system – and in doing so, also captured the agonized discussions that took place among medical staff as they tried to figure out how to respond to a worsening crisis.

But I’m not just any parent of a child with CHD. My child is now 40 years old, having been lucky enough to have surgery as an infant for complex CHD in 1978 at a high-volume institution with an acclaimed “go to” surgeon and surgical team. I’m also a cardiologist who has spent a career assessing the quality of cardiovascular care and a former Commissioner of the Food and Drug Administration. I now work at a university and its health system and at Verily Life Sciences, an Alphabet company whose mission is to make health information useful. You might say that this story hits me deeply on both a professional and personal level.

The Times report focuses on problems at a single surgical program and how the health system responded to those problems. But the larger story here is about more than just the shortcomings of a single person or even an institution. It’s about an approach to health and healthcare in this country that has gone terribly wrong at a time when much better solutions are in plain sight....Read more


Please Do Not Call Us Providers

January 3, 2019

Robert M. Califf, MD
Director, Duke Forge


Mary Norine Walsh, MD
Medical Director, Heart Failure and Cardiac Transplantation
St. Vincent Heart Center, Indianapolis, IN

As the healthcare professions are increasingly absorbed into complex business models, new tensions arise between the older ideas of healthcare as science, art, or profession on the one hand, and the modern concept of healthcare as big business on the other. While there are beneficial and detrimental aspects of this tension, where we land on that balance can have far-reaching consequences—consequences that are both shaped and reflected by the names and terms we choose to apply.

Clinical disciplines at their best involve the humane application of knowledge in the service of individuals and communities. Those who care for patients (and we will discuss the term “patient” and alternatives in an upcoming blog) know that their calling entails much more than a starkly transactional provision of services. At every level, it involves a complex dialogue informed by individual goals, preferences, hopes, and fears, as well as deeply personal (and private) decisions shared among family members. All of this human complexity must then be integrated with deep professional knowledge and experience drawn from relevant fields (medicine, nursing, physical therapy) and the overarching context supplied by medical, behavioral, social, and economic issues. The skills and knowledge required to navigate these interactions are hard-won and deserve recognition. Given this, it’s perhaps understandable that we and many of our colleagues find ourselves irked at being lumped under the flattening and misleading term “provider.”...Read more

Resources Reconsidered

December 20, 2018


Robert M. Califf, MD
Director, Duke Forge

Academic medical centers and the universities that house them have become big businesses. To take two examples: Duke and Stanford University (together with their associated health systems) are relatively modest in size by today’s standards for nonprofit healthcare systems. Still, both reported annual revenues of in excess of $5 billion in 2016. And for both institutions, the largest share of expenses on the balance sheet comes from salaries and benefits for employees and faculty.

In other words, when we talk about the healthcare industry, we are in essence talking about people. 

Over the course of a career in which I’ve become increasingly involved with healthcare and research administration, I’ve noticed the ubiquitous habit of referring to people as “resources.” The Oxford English Dictionary defines “resources” in this context as: “A stock or supply of money, materials, staff, and other assets that can be drawn on by a person or organization in order to function effectively.” For the narrow purposes of financial aggregation, lumping personnel costs with inanimate equipment makes sense. But inanimate resources and people are very different. When a manager or leader reduces the number of copy machines or consolidates purchasing for more efficient use of technology, such a decision seldom triggers strong feelings, and money saved can be redirected to laudable purposes. However, when an organization “downsizes” or “right-sizes” a workforce in order to adapt to change, there are direct human consequences....Read more

The Power of Words

December 13, 2018

Robert M. Califf, MD
Director, Duke Forge

Dictionary page

One thing that I quickly learned during my time as FDA Commissioner is that words can have an enormous impact on people’s actions. In theory, it should be true that given the right perspective, “sticks and stones may break my bones, but words will never hurt me.” However, the human condition is geared toward action motivated by beliefs about how those actions will affect the future, and specific words both reflect and shape those beliefs in remarkable ways. Partly in an effort to stimulate discussion about the words we commonly use, and partly as personal therapy, I’m introducing a weekly blog post on words that from my perspective are encOKoding and amplifying biases, hindering real understanding, or even prompting the wrong kinds of action.

For example, I’ve grown wary of a number of terms currently in common use in healthcare:

  • “Provider” to describe clinicians
  • “Owned lives” to describe insurance coverage of people
  • “Two-sided risk” to describe a situation in which clinicians or health systems agree to a contract for an amount of money, keeping some if less is spent or losing some if more is spent
  • “Resources” to describe people in any organization.

I believe that all of these terms reflect and evoke mental images that are detrimental to the goal of humanistic healthcare as expressed in the Berwick’s famous “triple aim” of better health, lower cost, and better patient experience of healthcare, and in the expanded “quadruple aim” first articulated by Bodenheimer and Sinsky, which adds the well-being of clinicians as a key component...Read more

The Duke Clinical Research Institute: A Perspective on Leadership

August 21, 2018, 2018

Robert M. Califf, MD
Director, Duke Forge

The DCRI is currently experiencing its fourth transition in leadership since its founding in 1996. It seems like only yesterday that we were first imagining what could be accomplished if we organized an entity that convened talented staff and faculty leadership within an operational infrastructure—one that was purposefully designed to enable faculty in leading and conducting multisite clinical research in concert with remarkably creative staff professionals.

Early Successes, Early Challenges

Our early experience was shaped by significant success with research done in acute care cardiology and undergirded by the decades of development that went into the Duke Databank for Cardiovascular Disease. Despite this promising foundation, building the DCRI nevertheless entailed substantial risk. We endured many sleepless nights as we tested the capacity of an academic medical center to work with other academic leaders and clinical research sites around the world to generate large volumes of high-quality evidence in conjunction with industry, the National Institutes of Health, and other federal agencies. However, the result of these early efforts was not only a large and productive organization, but also the development of a generation of leaders in clinical and translational research, as highlighted recently in Forbes....Read more

CTTI, Mobile Health, and a Continuing Mission for Transformation

July 15, 2018

Robert M. Califf, MD
Director, Duke Forge

A decade ago, many observers were warning that the U.S. clinical research enterprise was on an unsustainable path, as the already substantial cost, complexity, and administrative and logistical burdens of clinical research continued to mount. Although attempts to address this incipient crisis often devolved into exercises in fiddling around the margins, the Clinical Trials Transformation Initiative (CTTI) was distinguished by its scope and ambition. As its name implies, CTTI is devoted to enabling transformative changes. But while progress has been made, much work remains, and overall the clinical trials enterprise has not yet “turned the corner.”

What started as a public-private partnership between Duke and the FDA has now grown to encompass more than 80 partners representing academia, government, industry, and patient groups. Through this multi-stakeholder approach, CTTI remains focused on working to bring the kind of change to the design, conduct, and reporting of clinical trials that will ultimately make trials more informative at a lower cost. This commitment to transformation will be on full display this Monday as CTTI unveils a wide-ranging, comprehensive set of recommendations for the use of mobile technologies in clinical trials.

Mobile tech represents both promise and peril for biomedical research. On the one hand, wearable sensors and phone apps will open a floodgate of data that until recently has been difficult, impracticable, or impossible to access. Clinicians, patients, and researchers will be able to tap into a torrent of rich information that provides a complex, nuanced, and more holistic portrait of individuals and interactions among people as they exist in their daily lives – not just the tiny and unrepresentative segment that can be observed in the artificial confines of the clinic or research facility....Read more

The U.S. Opioid Crisis: Digital Tools, Data-Driven Approaches, and New Ways of Thinking About Public Health Challenges


March 1, 2018

Robert M. Califf, MD
Director, Duke Forge

Grim headlines about the ongoing U.S. opioid epidemic have become dismayingly familiar. Recently released data from the Centers for Disease Control and Prevention show deaths related to drug overdose – a category substantially driven by opioid abuse – are growing at an alarming rate. A total of 63,600 people died from drug overdose in the U.S. in 2016—a 21% increase from 2015, itself a record year for overdose deaths. It’s difficult to overstate the enormity of this crisis: for the second year in a row, U.S. life expectancy has fallen, in part due to deaths caused by the misuse of powerful, highly addictive painkillers. Beyond the epidemic’s toll on individuals and families, it is also hollowing out communities and placing unprecedented strains on social services.

As the trajectory of the crisis grows steeper, the need for effective action grows increasingly urgent. Getting a handle on the problem, however, is both complex and difficult. The impact of the opioid epidemic is not uniform across the country, but is concentrated in particular geographic regions and correlated with specific socioeconomic and demographic characteristics. For these reasons, strategies for tackling the crisis need to be fine-tuned to meet different circumstances. Caution is called for: responses to drug-abuse epidemics have often created unintended consequences as attempts to curb one problem give rise to others, and for many people, opioids represent one of the few effective options for managing severe pain....Read more


The Other 99 Percent

Geographical variation in US healthcare systems, 2016

February 19, 2018

Robert M. Califf, MD
Director, Duke Forge

Ever since the publication of the Flexner Report in 1910, U.S. medical schools have largely focused on human biology as the cornerstone of the medical profession. This massive investment in biological sciences has in many ways paid off, both in terms of expanding fundamental knowledge and in developing effective therapies. Meanwhile, schools of public health have focused on sampling methods to understand population health and develop effective policies and implementation strategies—a difference in emphasis encapsulated by the refrain: “social determinants are the primary determinants of health.” Of course, there is substantial overlap between these two disciplines, and both have contributed to dramatic improvements in survival for U.S. and global populations. Likewise, both are essential to our growing ability to precisely determine which treatments are best for a given person (“precision medicine”). 

Until recently, attempts to fully integrate the contributions of clinical medicine and population health faced a major roadblock: namely, the fact that essential data were stored in paper medical records that were largely inaccessible to researchers. However, an enormous national investment has ensured that almost every American now has a digital health record....Read more


Tapping the Brakes on Machine Learning

February 9, 2018

Robert M. Califf, MD
Director, Duke Forge

Amid a great deal of current excitement, shading into hype, surrounding biomedical applications for machine learning, Vanderbilt University professor of statistics Frank Harrell sounds a timely note of caution.  At his Statistical Thinkingblog, Dr. Harrell examines a recent study by Stanford researchers who used a deep-learning approach to develop an algorithm for predicting patient mortality. The ultimate aim of the predictive tool, which draws on data from millions of electronic health records, is to identify patients for whom palliative care might represent an appropriate option.

In the post, Dr. Harrell painstakingly dissects some of the key limitations and sources of potential bias in the Stanford researchers’ approach and contrasts them with other mortality prediction models that rely on methods that, while less eye-catching than newer machine learning approaches, yield results that may be both more reliable and more amenable to inspection and evaluation....Read more

Flipping the System

February 9, 2018

Robert M. Califf, MD
Director, Duke Forge

Years ago, when I was conducting rounds on the intensive cardiac care unit, I noticed that our savvy interns, residents, and fellows were becoming increasingly frustrated. Ever-growing clerical work was one culprit; another was the proliferation of obstacles to trying obvious clinical strategies for improving a patient’s likelihood of survival, speeding recovery, or achieving better function. It was clear that our healthcare system was structured in a way that led us to spend our time and efforts on activities that were not aligned with the patient’s best interest. This led to me start every day of rounds by making them answer a rhetorical question: “What do doctors do?” 

The reply: “They do what they are paid to do.” 

This was a blunt way of encouraging our brightest young people to think about how the system needed to change. In fact, as I revisit elegant work by Michael PorterDon Berwick, and Mark McClellan, it seems obvious that the importance of moving from a “fee-for-service” system to one in which payment is based on the value of the care received has been clear for some time. In short, our healthcare system has been profoundly shaped (some might say distorted) by the fee-for-service model. We devote much more effort to high-margin procedures than to relatively simple practices that have been shown to be effective in preventing disease and enabling people with chronic disease to live longer and with better function and quality of life....Read more

What Is Our Dependent Variable?

June 30, 2017

Robert M. Califf, MD
Director, Duke Forge

We are entering an era in which the full stream of biomedical data—from genes to genome, to traditional medical data gathered from EHRs, to social interactions and environmental context—will enable us to better understand the drivers of health and healthcare delivery. But despite the amazing potential of this “information revolution,” my many visits to academic centers and their associated integrated health systems have shown me that a tremendous tension exists as well. This is because faculty and staff clearly perceive the trap created by our reimbursement system, which channels so much energy into optimizing billing and reimbursement, bending practice to approaches that are not aligned with the best patient care.  I’m convinced that this tension is at the root of much clinician dissatisfaction and burnout.

As we embark on our new Center for Health Data Science, we should take this tension into account.  We can think about this in terms of the dependent variable in the actionable analyses we conduct. One option is to focus on clinical efficiency: how do we develop analytical solutions that lead to implementation approaches that optimize the current system to deliver services as efficiently as possible, assuming that the reimbursement system channels healthcare to produce better health outcomes? Alternatively, we could focus purely on health outcomes: how do we develop analytical solutions that lead to implementation approaches that lead directly to optimal health regardless of the reimbursement system? An ideal health policy is one in which payment and clinical care that produces better outcomes are perfectly aligned....Read more

A New Center at Duke

Forge sigil-only.png

June 19, 2017

Robert M. Califf, MD
Director, Duke Forge

Following the completion of a truly inspiring experience at FDA, I chose to return home to Duke while also devoting half of my time to Verily Life Sciences, a member of Google’s Alphabet family.  Many people have asked me why I made this choice. The answer centers on an infectious excitement surrounding the topic of health data science and what it can do for health and healthcare. Discussions that began at Duke before I left for FDA had borne fruit in the form of a campus-wide center designed to work across a spectrum spanning discovery science, social science, clinical research, and clinical delivery of healthcare, with the goal of developing powerful approaches to turning data into useful information and knowledge and acting on that knowledge to improve health and healthcare. 

The confluence of major increases in computing power and analytical capability is creating a potential so enormous as to be almost beyond comprehension—a true information revolution. But at the same time, our country is experiencing an unprecedented divergence of health outcomes that mirrors gaps in wealth and education. In other words, wealthy, highly educated people are benefiting from information that allows them to lead longer, more functional lives, while others are suffering. Ominously, for the first time since 1918, national life expectancy is declining instead of increasing....Read more