Dr. Robert Wachter, Editor, AHRQ WebM&M: What got you interested in health care?
Steven Spear: My link into health care came out of concern with a problem many organizations have: how to deliver great value despite the extraordinary complexity of the work. The commonality that seemed to span across sectors was that the very best in each sector was able to outperform its rivals because it was able to generate and sustain rates of improvement and innovation that were unmatchable by their rivals. The background on this was that I was a student at Harvard Business School, in the 1990s, trying to understand how Toyota—which was doing exactly the same work as GM, Ford, and Chrysler—had been able to race ahead of them. What resulted was a realization that Toyota had developed a fundamentally different approach to managing the very complex interdisciplinary collaborative work demanded of designing great cars and designing the great production systems that could make those cars on a repeated, reliable basis. While rivals were in the mindset of planning, command, and control, Toyota had learned that no amount of planning would ever get you to a perfect answer. What you're trying to accomplish is too complex. However, they realized that if you manage your work to achieve constant discovery—doing your work with a close eye toward what you don't understand, seeing problems as they occur, solving those problems, converting what you don't understand into something that you do understand, and then putting to use what you now understand and what others don't understand—you could improve and invent more quickly, for longer, and in more areas than anyone else. As we've seen over the last 20 or 30 years, Toyota has been the innovative leader in its industry.
As I was finishing my doctorate, a surgeon appeared in my office and said, "You may not know it, but health care is a real mess. My colleagues and I have tried everything we possibly can to fix that mess, and we cannot seem to crack the problem." He said, "Is there anything we can learn from Toyota?" The answer I had to give was, "I really don't know. I don't know much about health care, but I'm perfectly willing to go have a look."
Around 1999, we had a very hospitable invitation from the senior leadership at CareGroup, which is the Beth Israel Deaconess Medical Center and its affiliated hospitals. They gave us access to a 42-bed community hospital in Needham, a suburb of Boston. They asked us to have a crack at something that looks like manufacturing, which is medication administration. Over the course of a couple of months, we worked with the staff—physicians, nurses, administrators—around how to help them get the right drug to the right person at the right time at the right dose without asking the nurses to do endless workarounds to make sure that will all happen. We recorded some nice successes.
Coincident to this effort, Paul O'Neill, who had up to that point been CEO of Alcoa, invited us to Pittsburgh to explain what we were up to in terms of managing systems of care delivery, improving the processes by which care is delivered, and developing a more responsive self-correcting, self-improving system. Similar to what happened in Boston, a bunch of people stood up and said, "Well, we've tried everything else so there's nothing to be lost by trying this, too."
This brings us to about 2000. Starting then, we got very involved in the Pittsburgh community creating what they came to call the Perfecting Patient Care System. They started tackling some really tough problems: patient falls, surgical site infections, central line infections, and so forth. And along the way, some of these hospitals scored fantastic improvements.
In my new book, Chasing the Rabbit, I write about Allegheny General Hospital, which had a terrible problem with central line infections. When they really dug into the data, they found that of about 1750 patients seen in 2003, about 40 patients had acquired a central-line–associated bloodstream infection. Of those patients, 19 died. So of those who got a central line, 1 in 100 were getting killed by it. Over the course of the next several months, they started tackling this problem, these constant breakdowns in care, in the ways we learned from Toyota: see a problem and know that the problem you see is something you don't understand, solve the problem, convert what you don't understand to something you do understand, and then put it to good use. And they went from 19 deaths out of 1750 patients with central lines, down to a handful in just a few months, and over the course of the next few years the mortality went to zero. And they sustained zero. And Allegheny General wasn't alone. The region as a whole cut the infection rate 70%, and other hospitals eliminated other "never events" like falls and surgical site infections. It was much better care with less human suffering and less financial cost by avoiding complications. The truth is, there are standout hospitals, clinics, and practices around the country that have done much the same. Always getting better, safer, more effective, and more efficient.
RW: There's a long history of people coming into health care with industrial backgrounds who say, "I know how to fix this because I've seen it work in industry." These people often find that there are so many disanalogies that the lessons aren't transferable. Which ones do you think really are relevant and which ones don't translate well to health care?
SS: Fair question. A common pitfall is when outsiders try to impose "solutions" that worked in one circumstance on a circumstance that is wholly different. The key is not importing the answer: it's in importing the robust process by which the great answers are discovered. So, the important similarity between health care and other industries is not the specific science and technology—you can get lost focusing on that. It is in the challenge of successfully engaging people across a broad range of specialties in a common purpose. Individuals have to perform well within their discipline, but the disciplines and the specialties also have to integrate harmoniously. Health care often fails to do that well, ironically, with good reason. For a very long time, we as a society understood very little about medicine and invested very heavily in deepening our scientific and technological knowledge of how the human body works, how it suffers failure, and what sort of interventions will help to reverse those failures. As science and technology knowledge advances, we start fragmenting disciplines into subdisciplines in order to deepen the knowledge in those fragments. But of course to deliver on the value of all those discoveries, the pieces have to come back together.
And this brings us to the "failure mode." Academic medical centers and major teaching hospitals are organized around the teaching and research mission, which is around discipline, function, and specialty. They're not necessarily organized around the processes by which care is delivered.
True, there are big differences between health care and other sectors. The most perplexing is that, like in other sectors of the economy, there are some who are very good at providing care and there are some who are not nearly as good, but it's very hard to get rewarded for excellence. Part of that problem is that it's very hard to measure the performance of providers. But even with measurement, if patients and payers can't make informed choices based on those measures, then it's very hard to direct resources toward those who are using those resources well, and it's very hard to direct resources away from those who aren't putting those resources to good use.
Compounding this is that people who manage health care delivery organizations have often risen through a function or a specialty. They've never been trained to think about and to manage the integration of the pieces into a well-functioning process or bring the pieces together into a well-functioning system. So even if you had measurement and even if you had choice on the part of the patients and the payers, there's still this issue on the provider side. Are the providers prepared to up their game in response to the signals that the market is giving them?
RW: In many health organizations, the organizational arrangement is such that the doctors are individual entrepreneurs and have to be herded. Is that an important difference between health care and industry in trying to create a culture that will improve quality, safety, and value?
SS: This whole notion that because someone is in control they can exercise influence and compel compliance makes me chuckle—I think that anyone who says that doesn't have children. This morning, the organization that we call The Spear Household had to get three kids out the door with all the stuff that they needed to make it through an elementary school day, which in terms of complexity is not real high. This isn't treating a disease or building a car. Yet it was hard and demanded an awful lot of persuasion and encouragement and refocus. To move from the jocular to the serious, this notion that because you employ people you can tell them what to do is a profound misunderstanding. I have friends, colleagues, and students who serve in the military who, when asked about this issue of leadership control, say, you've got to be kidding. Even on board a ship where you have clear ranks and hierarchies, there's not a single officer who succeeds as captain of a ship because what he commands gets done. There are a lot of other ways in which things can get done and a lot of other ways of exercising influence and authority, but it's not strictly through command.
Another example is Toyota compared with its competitors. If you look at GM, Ford, and Chrysler, they had very strong command–control approaches, and the reality is that as much as they commanded, they had very little control. There are some screaming examples of extraordinary social dysfunction inside their plants because people were being asked to work on products for which they knew there was no market. They were being asked to work in ways that were nonsensical. They were being asked to adhere to standards that they had no part in developing, and even if they had a part in developing them, there was no way to adhere to those standards. They were operating not under the promise of shared gain and reward but under the threat of punishment and unemployment.
It's not a question of union workers. There's a great example, close to where you are in the Bay Area, which I discuss in Chasing the Rabbit, in which GM was running a plant in Fremont, California, near Oakland. That plant had an extraordinary reputation. It had record levels of defects, poor quality, and inefficiency. Within the GM system, it was the leader in union grievances against management. It's been documented that frontline supervisors would carry weapons into the plant for personal protection. Absenteeism was 25%. There was documented alcohol and drug abuse in the shop and prostitution in the parking lot. It was unbelievable. And GM, in some frustration, shut the plant down. Soon after the plant was reopened, it became a joint venture with Toyota. GM contributed the physical facility; Toyota would manage the facility and bring in their products and cars. Half the cars came out as Toyotas, the other half as Geo Prisms, but it was the same car. Within a very short period, productivity in that plant was great. Quality was super. Soon after the start of the joint venture, that plant started winning awards from JD Power, and it continues to win those awards today. It was a unionized plant, still UAW, but the union grievances against management just disappeared. Absenteeism went from 25% to 3%. There was no longer a need to bring knives and guns into the plant for personal protection. The drug and alcohol abuse disappeared. The punch line is that 98% of the hourly work force who worked for the joint venture had been GM employees beforehand. These were the same "drug-addicted slackers"—the only difference was that they were being managed differently.
So the issue that you can compel compliance in one setting and not in another is misunderstood. I'm not sure you can compel compliance in any situation. What you can do is engage people in such a way that they recognize that if they work together everyone wins. That comes back to the basic premise of the research I've been doing over many years, which is that some organizations—Toyota is an example in the auto industry, Southwest in commercial aviation, and some of the hospitals we've been referring to—are able to generate far more value with much less effort than anybody else. Now if you're in a situation in which you can generate more value with less effort, it means the people who are paying you benefit because they get more value for their dollar. The people who work for you benefit because they're more productive and enjoy the fruits of their labor. The people who depend on you—whether it's customers in the industrial setting or patients in the health care setting—benefit because they're getting better care and better service than they would elsewhere. So it really is a win-win-win across the board.
RW: How does "stop the line" work in industry and how does it translate to health care? Also, what have we come to understand about workarounds and how they get in the way of improving organizational performance?
SS: The issues of stopping the line and working around are two related behaviors. Fundamentally, it comes down to this basic issue: If you're trying to engage many people in a collaborative effort to create value for somebody, something is going to go wrong. It's inevitable. And it's going to go wrong because what you're trying to do is so complex that your understanding has to be imperfect. And because your understanding is imperfect, at some point that imperfection of understanding—or, said differently, that ignorance—is going to pop up and present itself as a problem. In health care, it might be that the report from one shift to another is incomplete or inaccurate. In the more common cases, it's the nurse who goes to pour meds at 9:00 in the morning and discovers that she doesn't have all the meds that she needs for a particular patient. Those problems are manifestations of systems that are imperfect and have to be imperfect because they were designed by people.
So if a nurse goes to pour meds and there's a med missing, what does she do? She calls the pharmacy. Or if she cannot call the pharmacy, she faxes the pharmacy, pages the pharmacy, emails the pharmacy, or asks a colleague to do the same. The problem with that approach is that all the factors that created the problem in the first place—breakdown in communication, misplaced order, whatever it happens to be—all those factors are still sitting there. They haven't been identified, they haven't been addressed, and they haven't been resolved. And because they're still there, they're going to cause a problem for somebody else.
So let's think about that same moment in which the nurse goes to pour meds and the med is missing. There's a slight misunderstanding when people look at Toyota. They think that, in that moment, the nurse pulls a cord and all the work shuts down. That's actually not what happens at Toyota. What happens at Toyota is that if someone goes to do his or her work and they find that they're at a moment of struggle, they call attention to the problem. In the moment they have the problem, the problem gets swarmed. When they call attention to the problem, whether it's a first level or a second level, the supervisor comes running. The first words out of the supervisor's mouth are, "What's the problem? Why do you think the problem happened? What can we do to help?" The follow-up questions, which continue during the shift and after the shift if it really proves to be a difficult problem, are part of the constant experimentation to discover why the problem happened and what you can do to make the problem go away.
You can swing it back into the health care setting. So you have a nurse who goes to pour meds and she doesn't have the medication she needs; the first thing she does is say, "Hey, charge nurse, I've got a problem. I was supposed to have this medication here. The medication is not here and the patient needs it." Now without stopping the work, what could happen at that point is that the nurse does go procure the medication. The patient needs it. But as she's doing that momentary workaround, the charge nurse is now investigating the problem—almost like the investigators on CSI investigating the crime scene—and saying, "We were supposed to have medication? We don't have medication. How could that have happened?" So the charge nurse or someone else designated as the real-time problem solver figures out why the medication is not there and then engages people from the pharmacy, physicians, nurses, and so forth to figure out how you can prevent that problem from recurring. So it's this very dynamic process of detecting abnormality, containing the abnormality, and dealing with the abnormality, so the abnormality doesn't recur. Now from what our experience has been in the industry, organizations that identify the problem, investigate the problem, and solve the problem are very robust. Likewise in hospitals where folks have tried this approach: when you have a problem, don't avoid meeting the patient's needs but call attention to the problem so we can understand why it happened. Organizations that have taken that approach of very dynamic rapid discovery of what they don't know and how to do better have made vast improvements.
RW: Does this not happen in health care because the nurse doesn't think about her work that way?
SS: The easy answer is that the nurse doesn't understand what to do. The real answer is that leadership doesn't know what it should want nurses to do. CEOs and presidents don't understand what is supposed to be happening in their organization. People rise to the rank of CEO, president, chief medical officer, chief nursing officer, and so forth for a variety of different reasons. Very few whom I know have risen to those levels based on their excellence in managing a process in general and in managing the delivery of care specifically. In any organization, people in the organization will follow the behavior modeled by their leaders. If you have it at the department level and the department head is willing to do this, the department is great. But the rest of the hospital is not. And if you have it at the level that the president and CEO are willing to do this, then the hospital itself can be great.
It is very funny that you asked this question because I was on the phone last night with a software company and they kept saying, "How do we engage the front line?" I said the way you engage the front line is you do this. You the CEO, you the president, you the business unit president. You do this and the front line will be engaged. You don't do it and they will know you don't care. No matter how much you espouse that this is important, unless you do it yourself, no one else will believe that it's important.
RW: Anything else you'd like to talk about in the safety and quality world?
SS: There are three things on the front page. There's bank failures, corporate failures in the auto industry, and health care. What's the link between the three? One is that they all cost a lot of money. But there's the other link, which is cause. And the reason that the Big Three failed is that they, despite literally decades of opportunity, failed to learn a different approach to managing the complex work of creating value through design and production from Toyota and other great companies like Toyota. If you look at the financial market, what will happen is that for all the finger pointing that's going to go on, we're going to find very few people who broke the law. What we are going to find is that people actually followed the rules and that the rules were bad rules and led to bad outcomes. If people had been sensitive when the bad outcomes were small, we could have changed the rules before we had this economic calamity.
The third issue of course is health care. Now the health care debate—people have talked about health care rationing and increased spending. But the fundamental premise for most of the discussion around health care is that the resources we put into health care today are spent as effectively and efficiently as possible. People who are inside health care know that premise is fundamentally flawed. In fact, the resources that we put in—whether you measure those resources in terms of dollars, hours, effort, or just the best of intentions by well-trained, well-meaning human beings—if we get half of our potential out of the investment we make, that would be a long shot. The way to phrase the reform question is: How do you provide better care to more people in less time with less effort and less cost? If we start tackling it that way in terms of how do we improve, innovate, and invent on the delivery of care, this whole issue of rationing and taxing is off the boards, because we're really talking about providing twice the care at half the cost. We're so far from our potential these days that to even have a discussion based on the premise that we're at our potential is just mind bending.
RW: So from a policy perspective, what should an administration do to promote the kinds of changes that you're talking about?
SS: If I had to draft a one-page policy advice paper to the president and his chief advisors, it would be that there are people currently in the health care sector who are figuring out or have figured out how to radically increase the amount of care they provide at no increased cost. The problem is that they cannot get rewarded for doing so, and because of that they cannot grow their own organization and they're severely constrained in influencing other organizations to do what they're doing. So if you want us to uncork that genie, first, demand measurement. Since the federal government is the single largest payer, whether it's through Medicare, Medicaid, or the VA, they're in a position not only to demand measurement but also to work toward developing measures. Some people say, "You can't measure care because every patient is different." We've made huge advances in measuring lots of different things, in many sectors and disciplines. I don't believe that we can't measure quality in health care. I just don't think we have figured out how. We haven't put the same discipline in as we've put in elsewhere. That's the first thing. Demand that people start measuring quality. Maybe the first measures are very simple: yes or no, did a never event happen? But let's start with the simple and then move to the sophisticated. Let's not just write off the whole project.
Second, once we start getting measurements of quality coupled with cost, start making decisions in terms of where the federal government is going to start putting its resources. Is it going to send patients to places that suffer a very high incidence of never events or is it going to send them to places that don't? Then you've created the opportunity for everyone to piggyback—the insurance companies can piggyback, the states can piggyback, employers can piggyback, patients can piggyback—on the combination of measurement and choice. Once people start piggybacking, then we uncork the tremendous capacity for invention and innovation that is in American society and let people start figuring out how to work toward those measures and those incentives.
At the end of the day—and I think in relatively little time, not decades—we end up with a situation where we're getting much better care at much less cost. And this whole issue of nationalized health care, and rationed health care, disappears from the policy discussion.