Dr. Robert Wachter: Helping Health Care Get Well

Robert M. Wachter, M.D., is professor and associate chairman of the Department of Medicine at the University of California, San Francisco, where he holds the nation’s first endowed chair in Hospital Medicine. He is also chief of the Division of Hospital Medicine, and chief of the Medical Service at UCSF Medical Center. He has published 200 articles and six books in the fields of quality, safety, and health policy. He coined the term “hospitalist” in a 1996 New England Journal of Medicine article and is past president of the Society of Hospital Medicine. He is generally regarded as the academic leader of the hospitalist movement, the fastest growing specialty in the history of modern medicine.

He is also a national leader in the fields of patient safety and health care quality. He is editor of AHRQ WebM&M, a case-based patient safety journal on the Web, and AHRQ Patient Safety Network, the leading federal patient safety portal. Together, the sites receive two million visitors a year. He has written two best-selling books on patient safety: Internal Bleeding and Understanding Patient Safety. Dr. Wachter has discussed patient safety and quality on Good Morning America, PBS’s NewsHour, CBS Sunday Morning, and NPR’s Talk of the Nation, and been quoted in virtually every major newspaper and newsmagazine. He received one of the 2004 John M. Eisenberg Awards, the nation’s top honor in patient safety. In 2010, Modern Healthcare magazine named him the 10th most influential physician-executive in the U.S. (the third year in a row in which he was the most highly ranked academic physician on the list) and one of the 100 most powerful people in health care. He is a member of the board of directors of the American Board of Internal Medicine and has served on the health care advisory boards of several companies, including Google. His blog, www.wachtersworld.org, is one of the nation’s most popular health care blogs.

This Conversation took place on October 29, 2010, between Bob Wachter, Mark Neuenschwander (see bio with his essay in this issue) and Al Erisman in Dr. Wachter’s office at the UCSF Medical Center.

◊ ◊ ◊ ◊ ◊

Ethix: You are a hospitalist and coined the term. What does it mean and why is it important?

Robert Wachter: A hospitalist is a generalist physician who coordinates the care of hospitalized patients. The best analogy is emergency medicine. It used to be that when you came into the emergency room, a nurse would see you, take your vital signs and then call your regular doctor, who would come into the emergency room to take over your care. About 20 or 30 years ago, we realized that this model did not work very well. And so the emergency medicine physician was born – a generalist physician whose specialty was the site of care, in this case the emergency room. Hospital care had become sufficiently complicated that you need a generalist physician – a physician comfortable with kidney, heart, and GI problems – to assume the role of primary doctor and coordinator of care. Hospital care had become sufficiently complicated that you needed a generalist physician to assume the role of primary doctor and coordinator of care.

About 15 years ago, I noticed that something like this was beginning to happen, and wrote an article for the New England Journal of Medicine that identified the emerging role of the hospitalist. In it, I articulated the advantages of having a doctor separate from your primary care doctor coordinating your overall care in the hospital. I called these doctors “hospitalists.”

There were a couple of hundred hospitalists when I first coined the term, and about 30,000 today, making it the fastest growing specialty in the history of medicine. Some worry about having someone other than the patient’s regular doctor in charge of hospital care. But the hospital has become so complex, the pace so fast, and technology so complicated that having a hospitalist who acts as the orchestra conductor adds a lot of value.

In addition to their focus on patient care, hospitalists have also taken the lead on patient safety. We have recognized that good physicians not only take care of their individual patients well, but they also make the systems they work in work better. When we started the hospitalist field, one of my mantras was you have two really sick patients. One is this person in front of you and one is this building you are working in. Both are desperately ill and require your help. We’ve always been pretty good at the individual patient part because that is what we learn in medical school and residency. But physicians often did not know what I was talking about when we talked about the system part. So, we have had to gain a whole new set of competencies in systems thinking and systems engineering and culture. The things that nobody taught us during our training.

We learned a great deal about systems thinking at Boeing. It is easy to think about “my little piece” without being aware of how that piece affects so many other things.

Many people from other fields cannot believe how backwards we were when it came to this idea of systems thinking. But it was a natural outgrowth of a set of social values that said the individual physician-patient relationship is the main lens through which we see patient care. There is something about that that is quite noble. But it led to a mom-and-pop operation kind of philosophy – everything was one doctor and one patient – and then trying to make that work in a hospital like this one: a breathtakingly complex billion-and-a-half dollar organization with 10,000 employees. You cannot do that well if the players think only about their own tasks, and do not think about how these tasks fit with the whole system.

How do you measure improvement that hospitalists may have brought to the system?

I’d like to begin with an analogy. Many hospitals have specialized critical care physicians, called intensivists, who are available to the ICU patients around the clock. There is good data that says that if you have them, patients do better. The emergence of intensivists is a recognition that someone who is sick enough to be in the intensive care unit not only needs all the technologies and nursing care, but also needs a physician who “lives” in that place.

But in the rest of the hospital, we had a huge gap. Under the old model of hospital care, we expected the patient’s primary care doctor to serve as the coordinator and overseer of hospital care. Since the patient has a pre-existing relationship with this doctor, this sounds sensible enough. The problem is that it does not work, when you place these expectations on a primary care doctor seeing patients every 14 minutes in the office and who now has a really sick patient in the hospital. You cannot be in two places simultaneously.

Can you give me an example of how this is helpful through the eyes of a patient?

Sure. A patient is admitted, desperately ill with severe infection, and we know that the patient needs rapidly to be assessed, to receive the correct antibiotics and fluid promptly, to have his or her blood tests monitored quickly, to get different specialists involved … and then someone to be there as the different specialists weigh in, each seeing his or her own piece of the “elephant” and saying, “I think you should do this for the kidneys, and this for the heart, and this for the liver.” In the old days, there would have been nobody around to coordinate all of that information.

Today in a good hospitalist system, that just happens. I am there by the bedside, I am talking to the family, to the consultants, to the primary care doctor. And another part of what I do as a hospitalist, when I get a break in my clinical work, is to be part of a group that ensures that we have bought the right computer system and implemented it in a sensible way, and that we have the right protocols and checklists in place. That is behind the scenes but it is just as important. The patient does better today than five or 10 years ago because of this focus on the system.

This is not just a money question. We have come to recognize this is good for the patient. Hospitals are dangerous places. The bugs that we have in the hospital are worse than the bugs that you have in your living room.

Learning From Other Industries

You have suggested that other businesses and other industries were thinking about the systems-level ideas before the hospital. Are there other things that the health care community can borrow from other industries?

Answering this question has been one of the healthiest things that has happened within the patient safety and quality of care movements. We have gotten past our usual myopia and begun to ask questions like: How is aviation so remarkably safe? How does FedEx get its package to you with 100 percent reliability the next day when we cannot get a discharge summary to a primary care doctor within two weeks? We have learned about manufacturing principles like lean production. These learnings allow us to look at our work through a different set of lenses.How does FedEx get its package to you with 100 percent reliability the next day when we cannot get a discharge summary to a primary care doctor within two weeks?

For example, it didn’t cross our minds until a few years ago that checklists would be a valuable addition to our work. Why? Because we are really smart, we have fancy degrees, so why do we need a checklist in the operating room? But no pilot would consider flying without a checklist. Further, it would be inconceivable for a pilot to walk into a cockpit and say to Boeing, “I like the cockpit of my 747 designed in a special way because I am a unique person.” Pilots have given up that level of customization because they know that standard procedures are safer. In health care, one surgeon will want the operating room set up his own special way for the same procedure that the next surgeon will want a different way. That is lunacy, but it was an accepted value in health care. So standardization, checklists, double checks, and team training have become very important.

A study recently came out in the Journal of American Medical Association. Of the 80 VA hospitals that implemented team training and some checklists, the mortality rate plummeted, when compared to other VA hospitals that had not yet implemented these things. It turns out you can change culture the way aviation did.

After the Tenerife accident [i.e., the1977 collision of two Boeing 747 passenger aircraft, the worst accident in aviation history, on the runway of Los Rodeos Airport on the Spanish island of Tenerife], aviation had to learn the importance of teamwork and good communication. I am told by my colleagues in aviation that such an accident could not happen today. Somebody would suspect that there might be 747 in the way and speak up vigorously to the pilot. But we also have to recognize that changing the culture of aviation is easier than doing so in medicine. To prevent another Tenerife, you needed to change the relationship between two people – the pilot and co-pilot – who have similar training and expertise and values, who communicate in a common language, and who are sitting in a sealed cockpit with 20 minutes to get ready before the flight takes off.

The Culture Challenge

In health care, the issue might be that the high school trained clerk sees something that seems amiss and she has to speak up to the 50-year-old chief of cardiac surgery, who has 17 years of post-college education and residency training, speaks a different language and has a different social status. It is just so much messier.

For this reason, analogies to other industries often irritate people in health care, particularly physicians, because they imply we are the only ones who are this careless, that we somehow have our head in the sands and just cannot figure out how to do our work safely. There is a long tradition of people coming to health care from other industries knowing how to “fix us.” Invariably they leave two years later with their tail between their legs saying, “I have never seen anything this complex.” Part of that is our resistance to change, but part of it is that flying a patient through hospitalization is much more complex than flying a 767.

The real epiphanies have come from bringing together people who say, “There is a different way of thinking about the world, let us try it in health care,” with those who recognize that these changes need to be carefully adapted to health care’s unique structure and culture.

So when you consider something from another industry, you find there are some things that look interesting, but the solution must be adapted to make it work.

Exactly. An example is the aviation safety reporting system, which has been spectacularly successful. The system involves collecting data on anything that seems “glitchy,” even near misses, and reporting this to an organization run by NASA. There it gets analyzed and is connected back to the FAA, leading to changes by everyone from Boeing to American Airlines. When the patient safety field began, we tried to emulate this model. And so we started admonishing caregivers to report everything, near misses, unsafe conditions, etc. What we didn’t understand is that the aviation safety reporting system, gets 35,000-40,000 reports in a year, across the entire U.S. commercial aviation fleet. In my 600-bed hospital, we alone received about 15,000 incident reports last year. And we are one of 6,000 hospitals in the country! The volume of glitchiness in health care is staggering. Figuring out what to report and how to deal with the data is something we are still resolving.

The Use of Checklists

A good illustration of this adaptation is the book The Checklist Manifesto (issue 68), which I found fascinating.

Me too.

The big challenge in that book seems to be the cultural change. Gawande shows impressive data that following a careful checklist, you can reduce post-operative infection. But it might mean a nurse needs to direct a surgeon to do something different, and this seems like a big change to the culture.

Checklists have taken on this “magic bullet” gloss, like they’re a panacea. As you well know, it is much more complex than that. The checklist is a tool, but the question is whether it is embedded in a safe culture. If it is embedded in a culture where people know why they are using the checklist and respect it for what it can do as a tool, then it brings success. If not, then it just becomes a perfunctory, “All right, whatever, check the boxes,” and you have not enhanced safety very much. I think the point that Gawande made in the book, was that checklists can be culture-changing prompts.

For example, one step in the checklist was to have everybody introduce themselves. In the OR before you start the surgery, this has a subtle effect of humanizing the environment and dampening down the hierarchy, making it more likely that a young nurse will speak up to the chief of surgery. So it is more than the checklist. You are trying to change the set of relationships that have built up over a century, where someone lower on a totem pole of power and authority feels uncomfortable speaking up to power.

In the Tenerife crash, it wasn’t that the flight engineer was sure there was a 747 in the way. It was that he was not sure that there was not one there. The interesting question is: What is the default setting of our caregivers? Is the default setting one in which I am going to assume it is right unless I am sure it is wrong, because it probably is right? If I “stop the presses” and it turns out there was nothing wrong, is there a subtle pressure on me to not to do that again?

I think that is an accurate description of what people feel in most health care organizations. The young nurse or clerk who saw something that seemed to be little bit out of whack thinks, “This is weird.” Her next instinctive response was to think of the four most dangerous words in health care, or any other industry: “It must be right.” Safe industries purge those four words from their vocabulary. They say, “Unless I am sure it is right, I will assume it is wrong. I will do whatever it takes to make sure it is right and the organization will always support me.”“It must be right.” Safe industries purge those four words from their vocabulary.

We have come to realize that UCSF, the organization that we are sitting in now, does not have a culture. Johns Hopkins does not have a culture. Even the organizations with very strong leadership like Gary Kaplan’s [Virginia Mason Medical Center] do not really have a culture. Instead, the UCSF intensive care unit has a culture. And the step-down unit 30 feet away has a different culture. The splay between those individual clinical units on culture surveys is as large as the differences one sees looking from one institution to another. So part of the question is what my organization can learn from the Mayo Clinic or from Boeing. But another important question is, “What can the intensive care unit on the 11th floor learn from the intensive care unit on the 12th floor, where the culture is really terrific?”

I can’t emphasize how much we are at our infancy in this whole domain in medicine. I went to a reputable medical school, residencies, and was on the faculty for 15 years and learned precisely nothing about any of this until about 10 years ago.

Is that starting to change in the medical schools today?

Yes, it is happening. The medical students are very interested in this and, in fact, they will probably drive the more senior folks to change. They believe instinctively in collaboration. This idea of systems-thinking is more natural to them, and obviously the role of information technology is a no-brainer. Even when they look at good hospital computer systems, all they can see is how incredibly clunky they are compared with Facebook!

When I started medical school, many of my colleagues aspired to be a solo practitioner, hanging up a shingle, running our own small business. Very few medical students think that way now. Traditionally for physicians, autonomy has been one of the core values of the profession. I think it was misplaced. Today’s trainees do not want to be cogs in big bureaucratic machines. They think, I am a highly trained professional, spent a lot of years becoming that, and so I won’t work in an organization where my opinion isn’t valued. But I also want to work in an organization where there is a system, and it works well. I don’t need to run my own shop.

I read something recently about faculty members that bears on this. What most faculty members value about their positions is not the teaching, not the research, but the autonomy. This, of course, flies in the face of systems thinking.

I think that is accurate. When you come out of medical school and begin to do clinical work, you immediately begin working shoulder to shoulder – in intense and often chaotic circumstances — with nurses, pharmacists, and respiratory therapists. But in most medical schools, we have never put you in the same room with those people to learn to work together effectively. As we train people in teamwork, our instinct is to get the medical students in a room and teach them about teamwork. But we need to get the medical students and nursing students in the same room and teach them about teamwork. That is harder, but I think more important.

The Role of Technology

Sometimes technology is also a fundamental culture-changing factor. Tell me about how technology has affected the practice of medicine, whether it is MRIs or information technology or robotic surgery.

I would distinguish the technology that we use to do things to and for patients from the ubiquitous technology that we use to move information around and communicate more effectively.

People frequently ask me, “Why are you people, meaning physicians, such Luddites? You are in a billion-dollar hospital, but I still see doctors writing in chicken-scratch on pieces of dead trees. Don’t you people like technology?” I say, “Are you kidding me? Come to the cardiology suite or come to the radiology department.” We love technology, and we are not bad at it. If you look at the history of technology and medicine, you generally see a very rapid implementation curve.

We love it, though, when we can stick it in a corner of the building and only have 10 people who have to really know how to use it. And we love it when we can bill for it. We face the challenge as we try to implement ubiquitous technologies that transform the nature of everybody’s work. We cannot bill for them, their goal is improving the quality of care, the safety of care, and the efficiency of care. This category includes computerized order entry, electronic health records, bar coding, and smart pumps. It is laughable how long it has taken to get these useful technologies out to the front line, but there are a lot of reasons for this.

The technologies have not been very good, very user friendly. That’s partly because adoption rates have been slow, so the IT companies have not had the resources to do it right nor the amount of user pushback needed to get these very complex technologies to sing. At our hospital, we recently pulled the plug on an IT system from a reputable company to switch to a different system just because the initial one was not performing the way that we had hoped. Unfortunately, that is not a rare case.

We need to cross that tipping point, going from, “This is too hard, the technology is too primitive, what if I make the wrong choice?” to “I cannot hire and keep good nurses and doctors without effective technology, patients are not going to come here unless they see this technology, we are not going to be able to perform well without this technology.” I believe we have finally passed that point and that is very exciting.

That means we are going to go from 10 percent or 15 percent adoption in some of the ubiquitous technologies like computerized order entry and electronic records to 70 percent or 80 percent adoption in the next five to seven years. That’s a lot of change, and there will be a lot of shaking out in the health care IT industry.

Diagnostic Systems

You wrote recently that technology-based diagnostic systems might become a reality in a few years. Why do you think that?

You have to be a little circumspect because people have said that for about 30 years.

Actually about 50 years. There were promises of such systems in the early 1960s.

You are right. This capability was “Just around the corner.” It was way over-promised and it way under-delivered and because of that I think there was a backlash, certainly in medicine. After a lot of effort in the ’70s and ’80s to develop artificial intelligence systems that can replace doctor’s thinking, we learned two things. The first one is the computers weren’t as good as we thought they were. The comforting part is it turns out that what is going on between the ears of physicians is actually pretty complex.

This led people to begin studying how doctors think. It turns out to be really very, very interesting and very nuanced. It is algorithmic, but it is algorithmic with all sorts of twists. Part of the challenge is that when I see a patient and I ask her to tell me about her complaints — her chest pain, where it starts, where it radiates to, what brings it on, what doesn’t — and then I also want to know about her past history, what medicines she’s on, what it was like growing up in Seattle or being a teacher. At the end I have several hundred facts. Maybe three to five of them will be truly relevant in helping me think through what is going on. But these crucial facts do not come with flashing lights telling me which ones they are.

And so the early AI systems stumbled. While they could develop a list of potential diagnoses that might have a few clever and relevant possibilities, several others were plain wacky. It was clear that the systems were nowhere near ready for prime time, and so field became dormant for about 20 years.

But now we are in the age of Google and extraordinarily powerful machines, and there is renewed interest in human-language recognition and sifting through huge databases. It is too labor intensive for the doctor or programmer to load into the computer’s memory the 10 causes of this and the 20 causes of that. Instead, today’s computers can sift through terrabytes of data and make connections between words that tend to be associated with each other in the same journal article or in the same textbook.

There is now an IBM supercomputer that is competing against humans in “Jeopardy.” (See video.) I am told there will likely be a contest between this IBM computer and the all-time Jeopardy champion, Ken Jennings. If I were a betting person, I would bet on the computer. Given how challenging Jeopardy questions tend to be, this gives me hope that a truly useful diagnostic computer might soon be on the horizon.

Though we are not there yet, it is really getting exciting as these systems sift through the world’s literature of every article, every PubMed listing or the contents of every textbook, and also through the 20 million records of patients being managed through large systems like the VA or Kaiser Permanente. We need the computer drawing data from my actual workflow as well. Every time I take care of a patient, I enter a whole bunch of data about that patient. The computer could then follow the patient to see what disorder the patient ultimately had. If I initially said “I think this patient has pneumonia,” but that particular patient turned out to have Wegener’s granulomatosis, the computer can learn from that. The diagnosis for the next patient can become more likely to be accurate over time.

It’s like Amazon.com’s function that says, “Customers like you who brought this book also seem to like this book.” But instead it would work like this: “Patients like this one, who you think have disease X turned out to have disease Y, and it was often mistakenly diagnosed as X by morons like you. You should think about Y.” I am guessing that level of diagnostic decision support is 10 years away.

Even today, with systems not as good as the one you describe, can these tools give you useful information?

Oh, sure. Even today in my hospital, every computer screen has direct access to an electronic medical textbook and the entire world’s literature on medicine. In the old days, 20 years ago, my team might admit a bunch of patients, they would be running around crazy all night just taking care of them, putting fingers in dikes. I would give them a call at 11 o’clock at night and they tell me a little bit about what is going on. I would do a little supervision, but then go to my bookshelf at home, pull out the textbook and read the chapters so the next morning I could give them a lecture on the topic at hand.

I cannot get away with that anymore because by the next morning they have gone online and concluded, “I think this might be a weird case of lupus.” Even if they do not know much about lupus, with one click of the button they can access the world’s literature on lupus. It is not much of a stretch to build in enough of an AI system that says to them, this looks like lupus, but here are four or five other diagnoses that might fit this set of facts that you have given me.

There is a book on AI that discusses this, When Things Start to Think by Neil A. Gershenfeld. We keep saying AI is a long ways away, but then we go to an elevator, press the button, and the best elevator is selected to come. Traffic flows are projected. These are decided by AI systems. This book has a whole collection of things that computers will never do (e.g., play soccer, play world-class chess, …) but all of these have now been solved. These systems can do a lot, but there is always more we think or hope they will do.

One of the challenges of a diagnostic system is the “hidden data” that a human would naturally gather. Suppose a patient walks in and it is 95 degrees outside and he is wearing a heavy coat. Most diagnostic systems would not pick up the outside temperature and note the clothing of the patient, but you as a doctor know instantly that this is helpful.

The question is what part of our work can be algorithmic and managed perfectly adequately by computers and self-managed by patients? I suspect you know that most people are not coming into the hospital for bone marrow transplants, but are seeing doctors to manage their blood pressure and their cholesterol and get their Pap smears and things that strike me as being relatively algorithmic. There are context issues that have to be paid attention to, but these are not so complex that the patients couldn’t handle themselves if prompted by a good computer system.

We have, for both myopic and financial reasons, organized care around the patient coming into my workplace. The future will be much more integrated, centered around the patient’s life and their home. A lot of that will be facilitated by IT. People like me, who tend to work in complex hospitals with really, really sick people, will be the last ones to go. But much of day-to-day office care will be entirely transformed.

Let’s go back to first principles and look at the goal of the health care system. We have to answer the question: How do we create the most value by providing the best and safest care at the lowest cost? That question immediately confronts you with the kinds of questions that every business asks every day. Have we organized this with the right people? Do we have a person earning $300,000 a year doing a task that a $75,000 person could do? Do we have a person earning $75,000 doing a task that a computer can do? When we ask those questions in health care, we will find that we need fewer doctors than we have now and more and more teams of people. Physicians will be taking on more management roles, there will be more use of other kinds of providers, and more work that is being done by patients and their IT systems.

Managing Technology Change

Why is technological change so difficult in health care?

I am assuming that you did not need federal intervention to get Boeing to computerize, nor Amazon.com, nor Wal-Mart. The business pressures are such that those organizations simply had to do it. Further, if companies put out IT systems that did not serve the needs of these businesses, they rapidly went out of business, and better ones replaced them. The market worked pretty well.

Technology in health care demonstrates a huge market failure. Of our 6,000 American hospitals, you find highly developed and functional technology — electronic health records, computerized order entry, and bar coding — in perhaps 300 or 500 them, a shockingly small number.

So I believe we needed government investment to kick start the health care IT field – and there are about $20 billion being invested to do just that. They are currently working out the grand rules, because you could see how this would be very wasteful and ineffective if not done carefully. You can’t buy a 30-year-old computer and stick it in the corner of the office, turning it into a planter, and receive a government subsidy. Computerization has to promote the overall goal of improving safety, quality, and efficiency. There is so much social good having our health care system wired. It allows more transparency, rapid measurement of quality, rapid implementation of new standards, improvements in safety, improvements in efficiency. We need to provide the resources to get to a certain level and then to ratchet up these standards over time.

Why doesn’t that happen on its own? We have a system where there is no pressure to computerize. A hospital can say, “We are a terrific hospital, everybody loves us, the Chamber of Commerce thinks we are great, we are a pillar of the community.” But then we find the hospital is profoundly unsafe, low quality, and has poor efficiency.

Over time, we need an environment in which hospitals and physicians think, I will not be able to succeed – to produce care of the highest quality and the lowest cost – if I don’t have a highly functional IT scaffolding.

The Cost Issue

You have a unique situation where the customers do not really care what it costs. They care whether their care is reimbursed.

Completely. That is a huge issue in health care. The insurance system does something really pernicious: It disconnects the patient and the provider from that value equation. If the patient knows that his care is being paid for by somebody else, he logically wants more care and more technology. If the provider is just going to get paid per widget, then the provider also wants to do more.

There’s a story I like about this, called “The expensive lunch club.” A person moves to a new town and goes out to lunch in a restaurant where there are big tables with lots of people at them. There is a spot at one of the tables and he sits down. The waiter comes over and asks, “What you would like, sir?” He isn’t very wealthy, and so he says he just wants the salad. The waiter looks at this person a little cockeyed but says, “OK.” He moves to the next customer and asks the same question. She said, “I am in the mood for the filet mignon.” This person seems to be shabbily dressed and you are wondering how she can afford the filet mignon. And the next person says, “Filet mignon sounds good, and I will also have the lobster.” So the original customer turns to the woman next to him and ask, “What the hell is going on here?” And she says, “Perhaps nobody told you the way lunch works here. What we do is we add up the bill and divide it by the number of people per table.” Of course, the customer feels like a moron, call the waiter back immediately and says, “I want to add on the lobster.” If this is the kind of restaurant where the waiter’s tip is 15 percent of the whole bill, what do you think happens if you ask the waiter for his recommendation? You think he is recommending salad or lobster? And if this town has a lot of these lunch clubs, do you think you’ll find more restaurants specializing in salad or lobster?

American health care has largely been an expensive lunch club and it has to change. A lot of the battles around health reform were about how to change it in a way that makes sense, that ultimately drives all of the players to producing higher value. You don’t want the cheapest care (because we are a rich enough country that we can afford good care) but neither do you want the most expensive. That is incredibly tricky to do because people that were used to the old expensive lunch-club system and they like their lobster and steak. Changing this turns out, not surprisingly, to be incredibly difficult politically.

New Health Care Legislation

We have just passed new health care reform legislation in this country. What is your view of this legislation?

I think it is about the best thing that could have come out of this political environment. Watching the process, though, was like watching sausage getting made. You really had to cover your nose and eyes. It would make you a political vegetarian.

But if I had just gone to sleep for a year when the process was beginning and awakened a year later and looked at the outcome, I would have said “That is not bad.” It moves us very sharply toward universal coverage, and it moves us in the right direction regarding cost containment, although not nearly enough. It has a lot of my wish list around quality and safety, including more transparency, funding for a Medicare innovations center, and an easier pathway to take an innovation and make it Medicare policy. And, of course, though it wasn’t part of the health reform legislation but rather part of the stimulus package, there is the $20 billion to promote IT implementation.

As a student of politics, I find it remarkable to see how the legislation has become a piñata in the political debate. I don’t understand the anger. I don’t understand the pushback. At its core, the legislation seems like a fairly rational way of dealing with some of the flaws in our present health care system without scrapping our employer-based system of health insurance. I personally believe that a single-payer system would be better, but in America we will not trust our government to run something as large as our entire health insurance system. So I can’t envision a bill that could move us toward universal coverage, cost containment, and quality improvement that would be much better than what we got, without defying the political laws of gravity.

The death-panel discussions were the most disgusting hypocritical caricaturization of something I could imagine. Some don’t think people are going to die.

In 2008, a governor came out and declared that day (April 16, 2008) to be health care decision day. This particular state promoted more open discussions between patients and caregivers about end-of-life care through a governor’s proclamation. The state was Alaska and the proclamation was signed by governor Sarah Palin. And a year later, a single line in the health care bill said we will reimburse doctors to have discussions about advanced directives and end-of-life care with patients. This line was transformed into rationing death panels where you are killing grandma, by the same governor.

That said, the quality of the debate makes me fearful for the future. Take the “death panel” lie – a hypocritical caricature of what was actually in the bill – funding to cover the time it takes for doctors to discuss end-of-life issues and advance directives with their patients. But now, because of the turbocharged media and blogosphere, even cynical caricatures can be rapidly amplified, and the wackiest falsehoods can drown out the facts. I find this intensely scary, because it is likely to be a pattern for how our country grapples with complex issues going forward.

Dealing With End-of-Life Issues

Let’s turn our attention to the moral and ethical issues in health care? What do you see as the big issues?

The end-of-life issues are extraordinarily profound. I might be caring for a patient – both she and her family are grappling with how hard to push at the end of life. They are trying to weigh the value of days or months remaining and how likely it is that some new chemotherapy or technology might make a difference. Looming over these intense and challenging discussions is the cost issue, with Medicare slated to go bankrupt in eight years partly because we are not very good at saying no.

This is part of what is so unbelievably gratifying and challenging about being a doctor. There is lot of stuff we have talked about that really is pretty algorithmic but the important discussions are in a catastrophic situation where there is a knowledge imbalance. I know lot more about it than they do, yet we as a society have decided, appropriately, they are in charge. There is a context of people hoping against hope that something might work. The chances of it working are very low and then there is this price tag in the background. That to me is the dominant ethical issue that we are facing in health care.

We have had the luxury, up to now, of essentially putting on context blinders. The doctor, the patient, and the family are told that they need not think about how much their clinical decisions cost. That is someone else’s responsibility. That is really nice – who wants to think about resources when someone is dying – but it turns out to be bankrupting our society. So, how do you have this conversation – both between individual doctors and patients and with society as a whole – knowing that the consequences are stark. If we don’t learn to say no, not only will be bankrupt our health care system but we will bankrupt our entire society. There will simply not be enough money to have a decent school system, or run competitive businesses. So I feel that this is the number-one ethical issue, by a long shot.

Access to Health Care

What about the people who don’t have insurance and don’t have access to health care. Is that an issue?

I think it is wrong, immoral, and dysfunctional, and we have to fix it. It turns out they do have some access because, when they really sick, those people come to the emergency room, we admit them, and we don’t let them die on the street. They get the same quality of care as the billionaire in the private room next door. The problem is, of course, that somebody has to pay for it. So the hospital pays for it. But the hospital has to stay in business and so it then charges the insurance company and ultimately every business providing insurance a higher rate than businesses want to pay because they are in a competitive industry. This shell game has gotten us through so far, but it’s now falling apart.

Globalization is part of that, correct?

Globalization is a big part of that. If I were GM and was paying a lot of money in health insurance premiums but only competing against Chrysler and Ford, then I don’t care that much. If I am competing against Toyota and Nissan, then I care a lot, because they don’t share the same health care cost burden.

While it is true that with today’s system anyone can get access to health care in the emergency room, people can’t necessarily get chemotherapy or prenatal care or all kinds of things in the emergency room.

Yes. I don’t want to portray it as a good system. It is terrible system, because if you don’t have insurance, your kid can’t get in to see his pediatrician for his asthma, and you can’t get your screening for early detection of your cancer, or treatment of your blood pressure, or preventing a stroke. It is terrible for the patients and it also is terrible for the whole system. The patients get very expensive, specialized care when they get sick, but it is often sickness that we could have prevented.

Other developed countries don’t have this problem. They have decided that access to insurance for health care is a social good. In America, we had this unique carve-out where we believe that certain things are public, that they should be made available to everyone, like the fire department and public schools. There is not much debate about those. But we have decided that health care is different. If you are of the view that health care should be like the other social goods, you are characterized as being somewhere between a socialist and a communist. It is very bizarre.

Access as a Moral Issue

T. R. Reid, in his book The Healing of America (issue 70) makes the case that health care is a moral issue and tries to put the argument in this moral framework.

Our politics don’t allow for that. Wherever you stand on the political spectrum you have to say that we are a remarkably selfish society. People in America give to charities, care deeply about their families, and actually care about their communities. When we can put a circle around a group of people that we know, and who we can identify with, we are actually quite charitable. But we are terrible at going beyond those circles and helping others who who we can’t see, can’t touch, who may not look or think like us.

In America, I don’t think the moral argument will have political traction. So instead, universal health care insurance gets framed as a pragmatic argument. In other words, we don’t make a moral argument that we need to care for our fellow man. Rather, we say these uninsured people are gumming up your emergency room or these people are going to spread tuberculosis to you on the bus. Or we say that this could be you … you think you are OK today because you have decent insurance from your company, but you are one lay-off or one cancer away from not being able to buy insurance. Those are all reasonable ways of selling the concept of universal insurance, but they don’t really address the moral argument that this is simply the right thing to do.

Effectiveness of Health Care

In America, we spend more per person for health care than any other country and yet our outcomes are not as good. The outcome measures include longevity, quality of life after 60, infant mortality, and so forth. How do you explain that?

I think we have a dysfunctional [health care] system that does not produce the outcomes that we want, particularly for the level of investment we make. There is no question that there is something to the argument that there are social determinants of health that are more important than doctoring. Some of those countries that do better than we do are smaller, more homogeneous, and have fewer social burdens, whether it is alcohol, drug use, or obesity. But these factors would argue we should put more resources into addressing those social determinants of health.

But even when you adjust for those other determinants of health, you still see outcomes that are not commensurate with the expenditures and the effort. Why? Our system is not organized very well. We have good doctors and nurses, people trying to do the best that they can. But to produce the best care at the lowest cost, you need a system. You need certain management structures, you need information technology, you need ways for people to communicate with each other that work, you need ways of moving evidence into practice that are seamless. We have very little of that.

A big part of that is structural. In most American hospitals, the organization employs the nurses and pharmacists, but the doctors work independently, often in tiny, mom-and-pop practice models. If you were going to design a system that produced the best care at the lowest cost, this system would not be the one you’d choose. Instead, you’d construct a system like Kaiser Permanente, the VA, or the Mayo Clinic — organized care delivery models in which the doctors are part of the system.

My hope is that health care reform will drive us in the right direction. There are incentives built in to create accountable care organizations, where the doctors and the hospitals and the nursing homes become part of the unified system of care. It is tricky to do, and it will only happen when the alternatives are worse. Doctors need to realize that it is only through working collaboratively with a larger system of care that we can achieve the outcomes we want. Hospitals need to realize that if they manage the doctors with too heavy a hand, the doctors will feel like they are cogs in a machine, and they won’t perform as well.

We also have a payment system that essentially promotes the expensive lunch-club model. If at the end of the day I make far more money doing a hip replacement or a coronary bypass surgery than not doing it, in very subtle but important ways, my inclination is to do it. You have to change the incentive system.

Did you see Atul Gawande’s article in the New Yorker, “The Cost Conundrum?”

Sure.

He discusses the role of financial incentives in health care costs between two Texas cities.

Yes. It’s a great article. He takes El Paso as an example of a place that is doing reasonably well and McAllen, Texas, as an icon of overuse and runaway expenses. But El Paso is not very good. There are very few beacons in American health care.

The payment system that just pays people to do more units of work will get what you would expect, more units of work without a whole lot of consciousness about the quality of their output. The alternative system, paying people a fixed salary or putting in a certain amount of money at the beginning of the year, risks the opposite, which is under-spending and too much rationing.

My guess is we need some combination of fixed payments and operating within a budget. We also need reporting and transparency, with incentives built around quality and safety, as well as the patient’s experience.

Professionalism

There is an old statement that there is no system that can’t be “gamed.” It seems to me that no matter how you construct this system, someone is going to try and find an advantage somewhere.

Absolutely. Speaking for the physician component of this, we are professionals. The patient’s interest trumps self-interest. Most doctors and nurses take that seriously. But we have learned that you can’t suspend the economic laws of gravity. Most of the “doing more” is not from fraud or abuse or people just willingly doing procedures on patients that don’t need them. It is when it is a close call, and a lot of medicine is close calls. The incentive system will drive you to say, “Let’s go ahead and get that extra CAT scan, it might tell us something useful.” On the other hand, the incentive system could say, “We are dealing with a fixed budget here. I don’t think that CAT scan is going to give me much useful information.” It is at the margin – these uncertain decisions – where those incentives matter. A lot of medicine is decision-making under uncertainty.

There is a very prominent author in the quality world named Avedis Donabedian, who developed the “structure, process, and outcome” framework for measuring quality. He was a very serious health care scientist. Donabedian was interviewed toward the end of his life and he said something really profound: “The secret of quality is love.” All of this other stuff is very important (measurement, transparency, and incentives), but at the end of the day this is about people caring about other people. We have to be absolutely sure that we don’t lose that love, that passion, while we are developing structures and incentives that make sense.

Personal Motivation

You are obviously passionate about these things. Where does that passion come from? What drives you?

I would love to give you a pat answer about some sort of deep wellspring of caring for humanity, but that does not feel like the driving force. I was a political science major in college. I have just always been intensely interested in the way things worked. I hate silliness and irrationality. When I entered into medicine, I did so not just because I cared about people and wanted to do the right thing, but I also because I found it endlessly interesting. And so a lot of my career has been devoted to trying to understand complex health care issues and articulate them to people – caregivers, patients, and policymakers – in a way that makes a difference.

Every now and then, I will get a letter from a nurse or a doctor, who read something I wrote or heard something I said and that tell me that it changed the way they think about their work. Or a patient tells me that the doctors and nurses in the hospital saved his life or made him feel better because of something I had done. At times like these, I feel like I’ve made a real contribution, and it feels awfully good.