Editor's Note: James L. Reinertsen, MD, heads the Reinertsen Group, a prominent health care consulting firm based in Wyoming. Prior to that, he was CEO at Beth Israel Deaconess Medical Center, where he developed a reputation for his unwavering focus on safety and quality. He is also a senior faculty member at the Institute for Healthcare Improvement (IHI), where he has taken a role in teaching leadership skills and promoting the engagement of health care boards and "C-suites" in patient safety efforts. He was a prime driver behind the IHI's decision to include the "Boards on Board" initiative as part of its recent 5 Million Lives Campaign. We asked him to speak with us about the role of boards in improving patient safety.

Dr. Robert Wachter, Editor, AHRQ WebM&M: How did you get interested in leadership and medical administration?

Dr. James L. Reinertsen: As I started into practice, I kept noticing things that didn't work, which would affect the way my day went. For example, seeing a patient when the chart wasn't there just ruined my day. I kept saying somebody ought to fix these things. And it turns out, "somebody" was me. I eventually worked my way up from fixing things in my little hallway to fixing things for the whole organization. Luckily, Paul Batalden was the director of quality for our group, and he pointed me toward Deming's work, and then other seminars and readings. I learned that, rather than just fixing things for our practice or the guy down the hall, with the right methods I could fix things for 100, then 300 doctors. And it just kind of snowballed.

RW: So fast forward to being CEO of a major academic medical center, what did you learn about trying to improve quality and safety that you didn't know until you found yourself in that situation?

JR: Well, Deming said that the job of a leader is to choose the boundaries of the system that he wishes to improve. And he said that the smaller the boundaries of the system you choose, the more likely you will make an improvement. And the greater the boundaries of the system you choose, the greater the impact of your improvement will be, but the more difficult improvement will be. What I discovered when I took on the job of trying to improve a very large, very complex system was that some systems are so large and complex that it's difficult to achieve improvement at that system level. You have to find a "sweet spot" between improving things at a very small level, a hallway, let's say, and at the other end of the scale, boiling the ocean. Academic medical centers are someplace close to boiling the ocean. They're very large, very complicated, very difficult systems to work on. And what I learned was how difficult it is to make big changes in such systems.

RW: Is that an insoluble problem?

JR: It may be that we have to parse the efforts more cleverly and bundle the bite-size efforts in a way that achieves system-level change. It's what really pushed me to the question of choosing leverage points—places where there's a high likelihood that a small change would produce a big change at a higher order of the system. Choosing and acting on leverage points was probably the main personal learning that came out of my experience in that large complicated system. In other words, those systems are so big and difficult that it's impossible to imagine a plan from the top, to take the system from one place in performance to another. You cannot imagine all the combinations and permutations and second-order effects. It's unpredictable: You might get surprising levels of change in a positive sense, but you also might get complete failure. But the point is you're focusing on a workable number of leverage points. You're making a bet that changing those items will change the larger system.

RW: And what makes a good leverage point?

JR: We get hunches about what leverage points might be from observing the experience of people trying to change these kinds of systems. For example, very few of these systems make and achieve great changes in measured levels of patient safety without the board being on the case, by being interested, watching, caring, asking questions, and holding to account. Where boards' attention is largely elsewhere, systems have a lower likelihood of achieving major improvements in safety. So I would identify the board as a potential leverage point.

RW: Why is it such an uphill battle to get the board engaged?

JR: The typical board composition—community leaders, business leaders, bankers, executives, community activists—is an eclectic group of people who have only a general idea of what they're there to do. And when you look at what boards and hospitals do accomplish, you find that they tend to mirror the historical activities of the administration. Administrators of hospitals have tended to take care of the finances and facilities and to a large extent have delegated the quality and safety of care to the physicians, and possibly the nurses. So a division of responsibility has historically grown up where the administration and the board look at finances and facilities, and the doctors and nurses take care of the patients and the quality and safety. Unfortunately, that has never been the way it should be. That is to say, the board has always been responsible for everything. Boards—no matter the type of organization or industry—represent the owner. In the case of hospitals, they represent the community and the people who live in it—and they're there to look out for the interest of those people.

RW: What was the sea change that all of a sudden got boards engaged in this?

JR: Well, first was a public awareness of quality and safety problems in hospitals. The IOM report that came out in 2000 made a big splash and caught a lot of board members unaware. So when they read in the paper that 50,000 to 100,000 people were being killed in hospitals each year, their immediate response was pretty similar to that of the hospital industry—"Well, that may be true but not in our hospital." The second thing that happened in many hospitals was that they had a shocking incident. That's a very common activator of boards. If somebody cuts off the wrong leg or there's a major disaster, it's a sentinel event that comes to the board. When that kind of event occurs, there's this incredible wailing and gnashing of teeth. "How could that have happened here?" And it illustrates why boards haven't been so good at this: boards have always thought that quality is better than it really is. There's interesting data on this subject. Surveys have consistently shown that when you ask the board chair, "What do you think of quality in your hospital?" they generally think it's several times better than either the CEO, the nursing staff, the physicians, or anybody else you can poll inside the system does. I'll tell a story that I think illustrates how that came about. To the extent that board members knew anything about quality, they were always hearing the good news. "Hey, just look at this wonderful stuff we just did," they'd hear—"A patient lost their hand in a farm accident and we put it back on and it's working again, and aren't we marvelous." But the routine stories of medication errors with harm, serious post-op and surgical infections, and other complications, the board wasn't hearing about those. As an example, Jim Conway, the former chief operating officer of Dana Farber, went to talk to a particular hospital (which I won't name) about what Dana Farber learned when Betsy Lehman was killed with an overdose of chemotherapy. The board listened with horror to this story, and then the board chair turned to the CEO of this institution and said, well, at least it's nice to know that things like that don't happen here. At which point, the CEO said, oh, but they do. And there was a stunned silence. And the board chair said, but you never told us. And the CEO said, well, yes, we did tell you. And then the board chair said something very telling. He said, but you never told us in a way we understood. That's the point—when the board realizes that harm is happening in their institution, that's when we can engage boards to become better at this.

RW: If the CEO has a hard time figuring out how much to bite off when he or she is trying to change the direction of the battleship, what's the magic that happens when the board is engaged?

JR: Well, it's a matter of channeling attention in the organization. One study from 350 hospitals showed that if a board spends more than 25% of its time discussing, working on, and hearing data about quality and safety issues, that is very strongly correlated with better quality and safety performance. Because when the board is paying attention, listening, and asking questions, the natural pressure that then comes to the management, administration, and medical establishment leadership from that tends to have a great trickle-down effect in the organization. I can tell you as a chief executive officer, and I can speak for many of my colleagues on this subject, board attention to quality helps when staff says, we cannot do that, we've got too many other things to do. If I, as CEO, can look them in the eye and say, the board is expecting us to come back in 3 months with a plan and within a year with real improvement, then they can't push this one aside. That creates what I call will or backbone. Boards have that effect. They don't get in and fix things, obviously. That's not their job. But they can expect them to get fixed. And when boards do get this bit in their teeth, it changes the agenda within the organization in very important ways.

RW: What is the role of measurable data versus the energizing effect of horrible incidents?

JR: Well, I'm glad you asked this question, because it's something I think is one of the real important learnings in the last few years about how to work with boards on improving their capability. The answer is that it's a "both...and"—not an "either...or." Measurement is critical. Basically, if you can't measure it, it's really hard to work on improving it. When boards ask for measures, what they generally are asking for are measures of the performance of the whole system. So, to make it too simple, boards ask for things like the customer service ratings of the entire institution, not things like the average time to answer a call light on the fifth floor. When boards look at finance, they look at the operating margin, they look at days of cash on hand, they look at the balance sheet, and so forth. They don't look at every little dot, jot, and tiddle of financial activity down inside the organization, inside departments. What we've done in quality with boards historically has been a travesty. We've sent them all the data from down inside the departments. It's no wonder they've been confused. They've received measures on who got a beta blocker on cardiology versus medicine, and on and on. It's not uncommon for me to go to a board in a typical hospital and find that when they get their quality reports, there are sometimes more than 300 data elements on it. In other words, everything we're measuring about quality is just dumped on the board. And it's not surprising that they throw their hands up. They can't sort the wheat from the chaff. On the measurement question, what's critical is for boards to, along with medical leadership support, choose a few key system level measures and track those, asking two key improvement questions: are we getting better, and how do we compare to the theoretical ideal? There's a third question that boards are very interested in—how do we compare to others? That third question introduces all kinds of other measurement difficulties. The risk adjustment questions are challenging. There are measurement and reporting delays because you have to send it out to get all the data from whoever has to vet the stuff and get it back. So you end up getting data that's 2, 3, 5, 7, even 9 months old when you do the "compared to whom" measurements, and you very often lose something in the translations. The best boards right now are actually asking for their own data on things like hospital-acquired infections and adverse drug events, close to real time. They're hearing about the rates of central line infections, but they're also hearing stories—stories about the 30-year-old computer programmer who is a father of three, admitted with pancreatitis, who developed a central line infection and wound up in a nursing home with a tracheotomy. The best boards are hearing both the stories and the data, and when they're doing it really well, they're hearing the stories that illustrate the data points in a matched time set.

RW: What can boards do to help engage the doctors? Are there things that you've seen that have really worked, from the standpoint of board engagement, to bridge that gulf?

JR: Actually there's some interesting data on this subject, and the evidence from the field suggests that where boards have really established a meaningful conversation with the medical leadership, you wind up seeing a higher likelihood of better quality and safety results. The case study experience out of things like the Pursuing Perfection program offers examples like the McLeod Regional in Florence, South Carolina, where the medical staff is seriously engaged in quality. And the board is in the middle of it. So what do they do? Well, it turns out that one thing they do is promote a conversation between members of the board and members of the medical executive committee. It can be in different formats, but it goes something like this: "We delegated this quality and safety responsibility to you. Tell us what you're doing about it. We've seen the results and they aren't as good as we'd like them to be. What's the medical staff's plan to make them better?" So there's this sense of accountability and responsibility that is a real partnership. It's not a browbeating, but it's a serious question.

RW: Does that involve the board asking, how can we help you, the medical staff, in these efforts?

JR: Here's one example. I know a hospital where the biggest trauma surgeon in the region had his privileges to do surgery and to take care of patients suspended for 3 weeks. It was a huge disruption to the trauma service and had serious economic consequences to the hospital. But it was done because he wasn't completing his dictations of his op notes on time in any way that was useful for the care of the patients. And the medical staff and the board said together—not one or the other, but together—now this is serious enough we're going to send a signal. A few months later, when the medical staff executive committee adopted a policy of full barrier precautions on central line insertions, everybody knew they were serious. This isn't a power play or anything, but sometimes the medical executive committee needs the support of the board, before they take difficult but appropriate actions against their colleagues. Far too often in our hospitals we have just said, wink, wink to these things and allowed them to continue. And that has had a toxic effect on what I call the culture of discipline around serious rules.

RW: Talk a little bit about the new IHI campaign. The "Boards on Board" piece was a departure from the other types of initiatives. How did that happen? What was your role in it, and what do you think its inclusion will mean for this effort?

JR: Well, I think it's a reflection that boards have been an underutilized leverage point in our work in quality and safety, and our recognition that they can be very powerful in driving a different level of performance. I was delighted to see it evolve into a plank, if you will, of the new campaign. Yes, I was involved to some extent in the creation of the plank itself. Jim Conway, who has now moved to IHI from Dana Farber, along with Jamie Orlikoff, Diane Miller, a variety of the IHI Fellows, got together and we put the board plank together. One of the reasons that we felt comfortable doing this is that this was no longer just a set of opinions of a bunch of people who had worked in health care for a long time that this might be important. In the last year or two, we've seen three fairly large studies of boards and how they work and how they impacted quality and safety. These studies all came out with consistent findings: when the board is engaged, when a quality committee exists, when the board is spending 25% more time on quality, when the board is holding the Chief Executives accountable for achieving quality and safety aims, things get better. These findings gave more weight to this board plank than would have been the case without that evidence. In terms of the evidence base for this, we ought to recognize something very important. The evidence is at the level of descriptive theory, i.e., these are associations. There is an association between the board spending 25% of its time on quality and better quality and safety results. It is not yet at the level of a normative theory—being able to say that if you spend 25% or more of your time as a board on quality you will get better quality results. We don't know that yet. Nevertheless, it's better than what we used to have, which was just a bunch of hunches. So we felt comfortable advancing it as a serious plank in the campaign. Frankly, I think it will have far more widespread impact than just on the harm issue. When the board gets the harm bit in its teeth and really gets good at watching over the data on harm in their hospitals and expecting performance to improve and holding to account and channeling attention and having serious conversations with the medical staff, my guess is it's going to extend over to other aspects of quality as well, not just the safety agenda, which frankly would be a great thing.