Dr. Robert Wachter, Editor, AHRQ WebM&M: Knowing a lot about teamwork in other industries including the military, what did you think you'd find when you entered the world of health care?
Dr. Eduardo Salas: Well, I thought I would find that health care and medicine in general understand the science of team training and simulation. Much to my surprise, I spent the first few years essentially educating health care professionals on what we know about teamwork, team training, and the like. The lack of awareness about this very robust science that could help them with their problems surprised me.
RW: As you reflected on that surprise, did you have any theories why this was such a foreign science to health care back then?
ES: It's like any other science. We are so invested in our own journals, societies, and conferences that we don't look to other disciplines that might be studying this. Fortunately, I think in the last 10 years, most fields have become more multidisciplinary. There's a little more cross-fertilization. But, I found over the years that disciplines are so insular and in their own box that they forget to reach out. So, I spent a lot of my time writing and giving talks in schools of medicine and hospitals about this science and what it could do for them.
RW: Did you discover that part of the reason also was that people in medicine didn't recognize that teamwork was an important part of what they were doing for a living?
ES: That came second. Even today, there's pushback in some sectors, and some physicians still reject this notion despite all of the evidence we have that this seems to work. I found there was (not to be disrespectful) some arrogance that said, "I'm not so sure this is relevant to us, and if it is can you just give me a book to read, then we'll figure it out." That was the usual response I received when I first started this. Now, it's a little different; but back then there was a lot more resistance, a lot more pushback, a lot more thinking that this is just a very fuzzy concept and that we don't need it.
RW: What did you come to believe were the reasons for that?
ES: The first thing is how physicians are trained and educated. I read recently about the cowboy mentality: "I know everything; therefore, I'm the one that has all the solutions." Some of the resistance stems from the nature of the training and the education physicians are receiving. The second thing is that organizations or health care providers are not organized around teams and do not send any signals that teamwork is what is needed. The third thing is the difference in status, which was very prevalent in the military 20 years ago. That has changed, but you have this hierarchy—there's the nurse, the technician, and so forth—and everybody seems to think the best way to communicate is among themselves and not across each other. There's also the lack of recognition that health care is a team sport. Every patient is receives care from a lot of people, no matter what condition they come in, and there's lack of recognition of that. I found leadership in hospitals or in schools of medicine didn't know what to do with this concept at that time. "What does this mean? Do I need to select people differently now?" Obviously you need to train them differently. "How do I reinforce what they do?" The way health care pays clinical care providers is also a problem. A whole host of these little things were, in a sense, barriers to adopting the concept that teams do help with patient safety and that teams do many things that individuals by themselves cannot do.
RW: As you entered health care, what concepts or lessons did you find worked perfectly well in health care, and which did you find weren't a perfect fit because health care is so different?
ES: I get asked that a lot. I think that just about 80% of what we learned in the military about team performance and teamwork transfers to health care in America and in many other places for that matter. Every discipline has that mentality: "We are different." Cardiologists think they're different than pediatrics—as a hypothetical example—but I found a similar mentality in the military. "The A14 is different than the F18; therefore, we do things differently." I learned that what is different is equipment or the setting, but the behaviors, conditions, and attitudes needed in order to function were the same. There was this pushback in health care at the beginning; everybody wanted to look at things through the eyes of their own discipline. It took me many years to convince the Pentagon that teams matter. The way we convinced them was with data. We went out and collected data with real military teams doing real things in the real world. We found differences between the effective teams and the ineffective teams. The same thing has happened in health care. I decided that the best way to convince CEOs, CFOs, deans, or whomever is with data, and the data has come in: Teams who are trained on teamwork principles do better than those who were not trained on them.
RW: Sometimes people bring up the idea that, for example, a cockpit really only has two or three people at fairly similar levels of the organization and hierarchy that have to be able to speak up to each other. But in health care a team might have 20 or 30 people, very fluid, and it might be a clerk who has to speak up to the Chief of Neurosurgery—huge differences in training, language, and status.
ES: Of course, there are differences. But at the end of the day, what is your mission and what is it you need to be prepared for? In health care, some professionals reject the teamwork notion from aviation because they think pilots go from A to B and that's it. Therefore, sometimes I use an analogy to respond to that thinking. That response is like saying, "Well here comes a patient, I'll give him a pill, then he'll just go away and come back if he feels bad again." It's not simplistic like that. Pilots go from A to B, but a lot of things can happen in between: unexpected events, crises that they have to deal with or be prepared for. It's the same thing in health care, so I frankly don't buy the argument that one is so different than the other one. Are there differences? Of course there are. But at the end, we're talking about task interdependence. So, if a pilot and co-pilot are as task interdependent as a surgeon and nurse, they will need the same knowledge, skills, and attitudes to function effectively.
RW: What is the state of the evidence that better teamwork is associated with better outcomes and safety than worse teamwork, versus the state of the evidence that teamwork can be improved and that makes a difference in outcomes?
ES: There have been a number of studies in team training. In 2008, I published a meta-analysis of team training in the journal, Human Factors. Essentially, we found a very robust effect that team training accounts for 20% of the team performance variance. That's good news and bad news. Team training does have an effect. It gives you better outcomes across safety, fewer errors, and those kinds of things. But this also says that 80% is left to other interventions, other things you need to do in order to account for performance. That's where I think organization and leadership come into play, with all the other things. The Journal of the American Medical Association just published a huge study from the VA on team training and found that those who were trained in teamwork had better clinical outcomes. The study is not perfect, but the data are certainly encouraging. There is a lot of medical team training evaluation going on, and I've been reviewing some of them. They all show that those who know about teamwork do better. They feel better, they know the concepts, they see the behaviors on the job, and there are better outcomes. Again, the outcomes are not perfect and it's very difficult to get access to clinical outcomes, but I think that the data is very encouraging. I think we're getting to a point where we can say that team training does work. Now the question is: What do you do with the other 80% of the variance that is remaining? And that's what I've been writing lately. In the military, we also did the same thing with real experts doing real Navy tasks. For example, we found that those who were trained on a simple strategy of self-correction and team correction outperformed the teams that did not. They had better communication, decision-making, accuracy, and so forth. There is now mounting evidence that teamwork does matter and that team performance can be improved by simple interventions.
RW: What does a great team training program look like?
ES: Even before I recommend a team training program or a team training system, I usually tell a leader, "The first thing you need to do is know your own organization. Where are the problems?" Most organizations will tell you, "We have good communication in our organization." Communication is a big bucket, so you need to peel the onion. What are the team problems you are having? Tell me where your needs are. Tell me where the gaps are. Where are the issues or concerns? The second thing is you need to do an organization readiness assessment. Is the organization ready to deliver the team training? Because if it doesn't have the resources, if it doesn't have the leadership aligned, then at the end of training people have all these great skills but they go back to the organization and the organization has not changed, it has not done anything to support the new skills, and of course with time, people revert back to old methods and don't apply the newly learned skills. Those two things are critical to understand before you implement any team training. What are your needs? Where are your gaps? Peel back the onion of communication so that you can match the training to the skills you need, and secondly, make sure you have the resources and incentive, as much as you can, to facilitate these behaviors on the job, as well as a culture that will support all of this.
Then you select a team approach. Any training has four components at the end of the day: information, demonstration, practice, and feedback. I'm not married to any team training program, as there are a few out there that are well-designed. But, I tell them that the science of learning tells us that people learn more when they practice and get feedback on the skills that they've been trained on, and not from a lot of information and demonstration. So, if you can afford simulation, if you can afford some sort of practice, then get a program that allows you to do these four things: information, demonstration, practice, and feedback and spend more on the practice and feedback. Most team training programs spend a lot of time on information and demonstration. These are 4- to 6-hour workshops where there might be a little role play but not a lot. They spend a lot of time talking about skills; they show videos of good and bad performance, but allow very little time for practice and feedback. So I always recommend practice—that's how you get this ingrained in your peoples' minds.
RW: Let's say I've done that and I feel like I've gotten a pretty good team training program and I've addressed that 20%, and I still have that 80% black box. I call you into my organization and ask you to help me. What do you tell us?
ES: I would ask, "How are you reinforcing that the trainees can now apply the skills on the job? What are your policies? What are your procedures? What types of signals is top leadership sending? Is the CEO sending memos or speeches to say, 'this matters, and this is why it matters'?" The rest of the 80% will come from leadership within the organization, at least as we know it now. If I just went through the best team training program in the world, and am now trained, and then I go to work and my boss tells me, "That training you just went to, I think it's useless," that 3-second interaction just killed my motivation to apply the new skills. So you must have supervisory support. In my view, at the end of the day organizations get the behaviors that they measure and reinforce. If you want teamwork, you'd better measure it and you'd better reinforce it. But a lot of hospitals call me for what you just asked me. They say, "We just invested in this team training program, and we're not getting the behaviors that we want." I respond by saying, "Let's look at what you are doing as an organization. What are the policies and the procedures, etc.?" And that's where they say, "Oh, we didn't know. Oh, we needed to do this?" So in the end, it's the culture that sustains the training. It is what the organization values and what they do in order to preserve those values. The good organizations do that; they create the climate for implementing newly trained skills. Some of it is very subtle and some very obvious, but you have to work at both.
RW: Let me shift gears slightly away from teamwork and more toward individual cognition. You've written quite a bit about situation awareness and how people can be aware of their environment and perhaps catch themselves before they make errors. What do we know about that and how do you teach people to think differently?
ES: We know that there are limits to how people process information. Essentially, every experience that an individual has, builds a repertoire in their head and it becomes a template—a template for how to act, how to think, and so forth. The way to train this is through simulation. Simulation accelerates expertise by providing experiences through scenarios that help to connect and then create a template that an individual puts in his or her head. At the same time, in order for individuals to monitor their own thinking and behavior, we teach them what you might call self-regulation. It is kind of like stepping out of where you are and saying, "Am I following the right path? Do I have all the information I need? Have I received all the communications that I need to have in order to execute?" There are opportunities where the simulation can be stopped (even though it changes the dynamic of the task a little; but this is for training purposes, so you have license to do that), and questions can be addressed. You continually tell the trainee that he or she needs to take a different path because there's a crystal ball which says that you're going to crash. These types of experiences create self-regulation. Always thinking about what you're thinking. Again, in the military we did that with simulations. I'm not talking about high-fidelity simulation necessarily; we can also do this with low fidelity.
RW: It sounds like it begins with thinking about your own thinking. But you can't do that constantly or your brain hurts—so you must have to learn to read your own internal thermostat and recognize this is a situation where I might make an error because there's so much stuff going on that I need to stop and think about it. How do you get people to recognize that moment, because paradoxically that's the moment maybe when you're least likely to do that?
ES: It comes with experience. I'd say it's expertise. That's why experts, while they do make errors, make fewer errors than novices. The reason experts make fewer errors is that they build this template in their head that creates a set of patterns—pattern recognition is critical to all of it. Usually what happens is the experts say, "I've been here before and this is the thing I need to do." Or they say, "I don't think I've been here; I haven't been in this situation before, but I think if I go this way I'll be successful." Now all of this happens in seconds in their head, but the ability to be able for a moment to step back and say, "What is this? Have I been here before or not?" is crucial to anticipating when a mistake might be made. With experience employing the strategy I just explained, I think people do evolve a mindset where they can self-regulate. Experts are good at self-regulation but the other thing experts do is imagine probable outcomes. Every individual engages in this type of mental simulation. You project very quickly what's going to happen after you make that mistake—when golfers kneel down and look at the ball in relation to the hole, they're doing a mental simulation of where and how the ball might go. Basketball players do it, and I'm sure surgeons do it as well. It's a way for us to process information and that's how we can check ourselves.
RW: I hadn't really thought about this notion that the signal for me to think carefully about this and hit the pause button really is that this violates a pattern that I've either learned about or done many, many times before.
ES: Chess players do it all the time and most of the studies on pattern recognition come from master chess players. They project the game—the way they move their pieces—way, way ahead.
RW: As you bring up pattern recognition and chess players, earlier this year a computer beat the Jeopardy! champions. What does the future hold as we begin to distribute some of our intelligence through information technology systems rather than confine it to the human brain?
ES: One area that I'm following more and more is neuroscience. I think we're finding a lot more about the brain and how firing A will lead to B. To me the future is uncovering more about how the brain works. I think we have technology to do that. Other things that are happening—I think we're going to use more and more technology no matter what. We have to start thinking about human–system integration. When the computer, your iPad, your tablet, or whatever is your teammate, and is giving you information—how do we create mental models with that? How do we create a trust in these technologies? Those are the issues that I think are confronting us now. Technology is going to evolve, and the technology is doing more and more of the thinking for us. So the question is: How do we use technologies as teammates? How do we process that information? How do we trust that information? This whole human–system integration, I think, is the future. How do we integrate the social/technical system such that you reduce errors?
Simulations are also evolving. We can now simulate just about everything and mimic the real world with tremendous realism. But I still believe that it's not the realism that matters, but what you do with it. The instructional features you're training and so forth. We cannot get in people's heads, but how do we get proxies that have a sense of how much individuals and teams are processing, what knowledge they have, and what knowledge they're missing. Current research is trying to capture in real time and diagnose in real time what teams and individuals are thinking. Hopefully in 10 years we'll know more. Those are different buckets that I see evolving and maturing and helping us.
RW: Anything else that you wanted to talk about?
ES: Just how we started; there is a science of teams and team training. There's a lot of information out there that is useful, practical, and can be applied. I think in 5 years we're going to stop talking about team training. And it's all about the culture of the organization, what leadership does. What the CEOs, CFOs, what they value and reinforce. Hopefully the next generation of health care providers will think differently. As you know, the way you were trained makes a difference in how you perform in your job. Similar to the way the culture of the aviation industry has evolved, which has taken 20 years to change, changing the culture of the health care industry will require a new generation of doctors that are used to collaborating, that are used to communicating a lot, and other things. We have a new School of Medicine here at UCF. The way we're teaching is very different than what I've seen in other schools. The next generation will embrace a lot more teamwork and collaboration, which will accelerate the future of the health care industry.