Editor's note: J. Bryan Sexton, PhD, is associate professor and director of the Patient Safety Center for the Duke University Health System. He is generally considered one of the world's foremost experts on the relationship between culture and patient safety. We interviewed him in 2006 about safety culture for AHRQ WebM&M and thought it was time to check in on how his thinking and research has evolved.

Dr. Robert Wachter, Editor, AHRQ WebM&M: One thing that you pioneered was a focus on how people interact with each other, listen to each other, and work in teams. How do you think that has changed over the last 5 or 6 years?

Dr. J. Bryan Sexton: Safety culture as a term 10 years ago was not well understood, but there is a much broader awareness of it now. In particular, aspects of safety culture like teamwork: Do you feel comfortable speaking up? Can we resolve conflicts? Can we ask questions to clarify ambiguities? Those concepts are much better understood so the awareness is up, and with that senior leaders, clinicians, and also increasingly patient advocacy groups and patients themselves are expecting their teams to be functional. At the same time, our resources are more limited than ever in doing teamwork "the right way." There are TeamSTEPPS training programs, which can be amazing, but take time and energy and you need the right people involved at the right levels. With increasing expectations, you realize that you're not anywhere near where you should be yet—that dissonance makes a lot of people uncomfortable.

RW: You noted more and deeper understanding of the elements of safety culture and teamwork—people think and talk about it more than they once did—but that is overlaid with the burden and even burnout. In general, if you swooped in and looked at a surgical team or a medical team somewhere around the country, are things better than they were 7 or 8 years ago?

BS: In pockets. It's not consistent. We learned along the way that teamwork is important. Whether your stance toward patient safety and quality is reactive or proactive is very important. Having trust with leadership is important. But there was a key element that we weren't examining at the beginning: innovation readiness versus innovation fatigue. You can measure that reliably with the standard burnout tools. The research on burnout in health care has shown pretty consistently that about one in three physicians meet the criteria for emotional exhaustion or burnout. Linda Aiken has shown that about 34% of nurses meet the criteria for severe emotional exhaustion or burnout. It is highly variable depending upon the unit or medical office building in which you work. Some of these clinics have a 0% burnout rate. In others, 70% to 80% of them meet the criteria for burnout. When you assess with a good metric that a unit that has no burnout, well then they are ready to work on teamwork. They have the wherewithal needed, they are bringing their whole selves to work, and they have what it takes to implement some process improvements and changes. Now that we are collecting the right data, we can identify with more precision who's ready for quality improvement (QI) and who's not. I think we're entering an era of Safety Culture 2.0. A big part of that is understanding the innovation readiness of our workforce.

RW: That sounds daunting, in that you'd think that the group that needs it the most is the group where there is some burnout. Yet your argument is that they are simply not ready for it. You have to address the burnout before you can address improvement in other domains.

BS: It's almost like Maslow's hierarchy of needs—where do you start? Sometimes where you start is to get people back to their fighting strength so that they can work on process improvements. For example, the leadership at Duke University Health System has resilience in the workforce as one of our strategic priorities in our 5-year plan for patient safety and quality. So we have to measure and know the resilience rates among all of our different units and labs, but we also have to enhance it in places where it's not where it needs to be. Part of that research is: what's the threshold below which you can do QI and above which you should probably work on the burnout first.

We have a whole course that we offer at the Duke Patient Safety Center called the Enhancing Caregiver Resilience Course. (It used to be called Bouncing Back from Burnout, but that had a negative ring to it.) When we put work settings through this we found that the resilience is enhanced, the work–life balance improves, sleep quality improves, days of work missed due to illness go down, meals skipped go down, they get to work on time more often, they get home on time more often, and they feel like they're more engaged in what they're doing at work. By working on resilience, the byproduct interestingly is: delays go down, teamwork goes up, and disruptive behavior goes down.

Although we have to wait 2 full years to get the turnover data, we do collect intention-to-leave data and that is also improving. So we can actually keep our good people who we would otherwise lose through burnout and attrition. Utilizing the resilience focus before we even move over to TeamSTEPPS, executive walk rounds, learning from defects, or one of these other kind of quality improvement standards, the teamwork norms and the safety norms improve as well. It's like there's a glass ceiling, and when you have so much burnout you won't be able to take on anything new and sustain it. By getting that shot in the arm of resilience, we're able to show some additional improvements in teamwork and safety just as a byproduct of focusing on bringing your whole self to work.

RW: In safety you usually hear the term "resilience" to refer to the ability to recover from a glitch or a curve ball. It sounds like that's part of what's going on here, but it's broader than that—the opposite of burnout.

BS: I think of it as the polar opposite of burnout. There's more to it than that, but if you want to just boil it down, burnout is: I feel frustrated by my job, I'm working too hard at my job, I feel burned out from my work. And resilience is just the opposite. If you're a resilient unit you're going to bounce and not break. If a colleague committed suicide, if you went live with Epic, if you got a new manager, if you moved into a new building—you bounce, you do not break. But if you're lacking in resilience it just takes little things to push people over the edge and you see disruptive behavior flare-ups. When you're burned out you're not engaged, so you do things that are more expensive, you make the same mistakes again and again, and tragically, you don't even notice that you're making the mistakes.

RW: You've said that it's very hard to train people on new things and innovation and get them engaged when they're burned out. I imagine it's hard to train them on getting unburned out when they're burned out. How do you approach that?

BS: One of the most frustrating things about working in burnout is you have to have a lot of compassion. Because you'll schedule a meeting and nobody will show up. You'll agree to something and nobody does it. So it takes really small baby steps with very, very minimal amounts of commitment on their part to get these things up and running. Unlike a unit where everybody is showing up and they're firing on all pistons, on these burned out units you have to start really small. The first couple of meetings we have to help them find a way to talk about what's going on. Because they usually think it's something else. They oftentimes will say, "Well, the problem around here is that Mark's an idiot. If somebody would just fire Mark we'd be fine." And you unpack that, you say, "If you take Mark out of the situation and replace him with some other person, are you going to see this problem happen again?" They said, "Well yeah, maybe it's not Mark. Maybe there are system issues going on here." Perhaps what we need to do is instead of focusing on readmission rates, instead of focusing on bloodstream infection rates, or patients who left without being seen, maybe we can do a little project and collect a little data and make sure that you're skipping fewer meals and getting home on time more often. And now everybody's leaning forward and wanting to work on that. They don't want to work on a clinical outcome when they're not feeling connected to their work. But if you can show that by doing some little things to help them take care of themselves better, then ultimately they can take better care of their patients and that's even more rewarding for them. It starts with those little steps.

RW: I would suspect that in addition to "If Mark was gone life would be better," you also hear, "if we had more money," or "if we worked fewer hours, life would be better." But institutions can only do that so much.

BS: At the Patient Safety Center, we work with a lot of different health care systems and many of them think that they know what the problem is with their lab or ICU or pharmacy. When you do these structured interviews or spend a little time doing a debriefing with staff you'll find out that they've got a problem with a particular issue, maybe a technology, maybe a manager, maybe a process that they do, and they've been saying it for many years and no one's listening. What ends up happening is these people who feel powerless, they start punching sideways. And what a lot of organizations characterize as a disruptive behavior environment or an environment where there's a lot of "bad apples," from my research perspective, it looks like these guys have been neglected for a long time. So they do what anybody does on a desert island—they take it out on each other. We find that by working on the burnout, conflict goes down and sleep quality improves. We haven't actually demonstrated an increase in sleep quantity, but sleep quality improving is a consistent finding across our studies. And that is directly related to disruptive behaviors, delays, and handoffs. Ultimately, I think that's a big part of what we're doing, relabeling what needs to be addressed in some of these "problematic" areas.

RW: As we think back on the early years of the quality and safety movement, let's say, 8 or 10 years ago, do you think we made an error approaching quality and safety with great enthusiasm and building these programs and teamwork training and measurement programs without addressing the burnout, or do you think this is an inevitable stage? You begin with early enthusiasm and early adopters and some success stories, and then you always realize that you've left some people behind and you've got to slow down a little bit and make sure everybody's on board?

BS: I remember these very explicit conversations with industry leaders where their concerns were: "I just don't think we have the bandwidth to take all of this on all at once, did we make a mistake?" If you see that something works, you have a moral obligation to get it to as many people as you can. At the same time, I don't think that people who were tweaking and improving processes in health care had any idea that what started off in one project making life easier for patients and for health care workers, a couple of years later it becomes this incredible burden. I don't think that that was a misstep, but going forward, patient-centered care requires a high level of resilience in our health care workers (people-centered care). As patient-centered care is our goal, we have to do much more deliberate work for health care workers. The reason that's complicated in terms of building resilience is because health care workers are just not good at taking care of themselves. It's so easy to come in early, stay late, skip the meal, finish that e-mail, see that last patient, squeeze one more in, and not go to the bathroom. I think in 1995 you could do that and get away with it. But our resources and our time are so lean right now. The complexity of what we're doing has increased, our volumes are up, and our acuity is up. Squeezing more stuff in means that eventually our health care workforce is burning out, and our burnout rates are higher than they've ever been. What's making me sad is that the health care providers who are most committed to getting the best possible outcomes for their patients are most vulnerable to burnout. It's not the lazy ones; it's not the ones who don't care. We need to support our health care workers more directly. We need to give protective time to health care workers so that they can take better care of themselves if we're going to continue to demand higher degrees of excellence from them.

RW: Aside from the issue of resilience and readiness of the providers, what else have we learned about teamwork training programs and other ways of trying to improve more traditionally defined safety culture?

BS: Parts of it were oversold and parts of it, unfortunately, were undersold. Some people would say, "Let's shut down the whole hospital, send everybody through TeamSTEPPS to learn how to do this teamwork stuff and then we're all going to be in this, holding hands in the hot tub, and everything's going to be great." That actually didn't happen. There was no evidence that shutting down the whole hospital results in people being pleasant with each other and everything getting better. What we found is that if you're using data to pinpoint which units specifically are struggling with teamwork—and within those units that are struggling, what aspects of teamwork are they struggling with—TeamSTEPPS has a whole litany of tools for conflict resolution, speaking up, handoffs, briefings, and debriefings. Maybe you don't need briefings and debriefings but you do need something for speaking up. Or maybe you don't need something for speaking up but you do need something for conflict resolution. If you tailor the tools to address the problem based on a cultural assessment that's reflective of the underlying norms at the work setting or clinical area, then there's a tremendous amount of support. That does work. If you're doing a generic approach—let's try to solve that with a generic tool or by throwing everything at folks—that's generally not a good idea because you won't find support for that in the literature or by talking to TeamSTEPPS training centers around the country.

RW: Interesting. Because I remember in the early days people talked about mandatory team training for everybody.

BS: There were places where they would shut down the OR or the ED, and I think that that lesson has been learned. It's been reframed in the most recent versions that just came out in the past 9 months of TeamSTEPPS. They're carefully selecting a tool and matching it to a need that you have in a specific clinical area. So I think that has improved somewhat. But at the same time, teamwork and burnout rates do go hand in hand.

RW: Johns Hopkins surgeon Martin Makary wrote a book in which, in part based on your work, he argued that care would get safer if we published results of safety culture surveys, because they tend to correlate to some degree with safety outcomes. What do you think about that idea?

BS: First and foremost health care providers need to be aware of what their strengths and weaknesses are, and I don't think that's the case right now. People might say, "Teamwork's no good in this unit." What does that mean? What part of teamwork isn't good? And do you really know what to do about that if that's the problem? Until there's a lot more clarity on the part of the health care providers as to what it means, what to do about it, where to go to get help, what this stuff is called, and to know that there are resources available to help us through that process, we have to be careful about making everything transparent immediately. Because it starts to shut people down and it drives problems underground. People start to game the system. They don't give honest responses on surveys. If you get threatened with termination as a manager if your employee engagement scores are low, you very quickly learn how to do things that get your employee engagement scores up. It doesn't mean that you have an engaged workforce, it just means that you're "teaching to the test," so to speak. We have to be careful of those types of things happening with our safety culture because it's not far away from happening. People are starting to look at safety culture surveys, employee engagement surveys, NDNQI, physician satisfaction surveys, and HCAHPS, and they're starting to merge together. I think we need to be careful. That's not a very clear answer, but I would rather see more of an emphasis on helping people understand what they have in their hands than moving toward making the data public.

RW: If you asked people what the largest victory in the last 10 years in the safety field was, people would probably say checklists. How do you interpret the work on checklists? Obviously the point has been made that it's not just a checklist but relates deeply to culture and communication and probably burnout as well. So can you make some sense of the enthusiasm about checklists in light of the way you see the health care world?

BS: The idea of the checklist became the symbol of how it happened. It's actually much more complicated than that, but it sure felt good to be able to point to something and say: This is the thing that we take out of the ICU and try to push into the OR and take out of the ICU and push into the ED, or some other clinical area. The concept of simplifying work is very appealing and it clearly links back to what we need to do for innovation readiness. That is, if you're ready for innovation, we need to package it and make it stick. When you read the word checklist, the fine print is getting this done within our local resource restrictions, with our local champions and our local curmudgeons, and despite inertia of other things like new technology and a new building or some other big distracting things. At its core the checklist represented something that people could point to and say there's science here, there are steps to take, there's a process that we have to go through. That felt reassuring because in the very early days a lot of this felt squishy. Let's tick this box and that box, and before you know it we've improved safety.

Ten years from now, people will probably still say that the projects that led to checklists being the big change, they're valid. As people take additional looks at why those projects worked and why they have been relatively hard to replicate, I think ultimately we're going to come back to the bandwidth of our workforce to take on these changes and then be able to sustain them. Ten years from now, I bet we'll be talking about how the new focus on health care worker wherewithal made all the difference in the big strides that were made.

RW: Yeah, it's seductively easy; if you just put in a checklist you'll fix this really complicated problem. And certainly competing against teamwork training, or working on resilience, or installing Epic, checklists seem like, "Boy, we can get that done in a day."

BS: If I could give you just one very brief anecdote, we replicated the Michigan Keystone ICU project and a Robert Wood Johnson project with a bunch of Adventist Health ICUs. In the replication we showed it wasn't the 66% reduction of bloodstream infections, it was 81.5%, which was amazing. We were all very excited. As we were coming together to congratulate each other and celebrate the success in a big face-to-face meeting with all the executives, the sponsors of the units, physician champions, the nurse managers, and the research coordinators—everybody across all these ICUs—we talked about the success of the project, the funders were happy, everybody was happy. As I was in the hallway getting a cup of coffee between some of the sessions, I was literally attacked by a nurse manager who said, "Dr. Sexton, I came in early, stayed late. I haven't seen my kids in months. I'm now going through a divorce. I got this project done and our bloodstream infection rates are down, they're at zero, they've been at zero for 7 months and we're going to keep them there. But I've got to tell you: Don't give us anything else to do. The next time you want to do a research project you do it on me. You do it on me because I'm broken. This project killed me. I need you to understand that." More and more people saw her talking to me, and they all kind of ganged up. It was in that moment I realized that we had used quality in a way that created harm. I wanted to look at how we do quality a little differently. And that was when Duke recruited me to make resilience a big part of their health system, and I couldn't say no. I'll go back to that hallway interaction with that nurse as something that really changed the course of my career. Who knew that health care quality improvement would go through a phase where the path to better patient care would be through better self-care among health care workers? This is powerful and meaningful work, and it is a genuine privilege to be a part of it at Duke.