4 Coaching – Becoming the Change: Leadership Behavior Strategies for Continuous Improvement in Healthcare


The primary job of a performance improvement officer is to coach the leaders of an organization, from the C-suite to frontline managers, to embrace and display the principle-driven behaviors that will guide improvement. And yes, that is a little different than the job description at most places.

In the two decades since the methods of the Toyota Production System first began migrating into healthcare, the job of the performance improvement (PI) office has been that of resident expert. Staffed with former manufacturing engineers who were learning healthcare and caregivers who were learning lean thinking (or any of its variations), the PI office was there to plan and lead rapid improvement events, teach people how to use the tools, and get results.

Teaching the tools and running events is still an important part of the job. It’s just not nearly as important as coaching people on how they get those results. We know this now, after seeing and studying more than a hundred transformation attempts, because we have consistently seen that deployment of tools without the right intentions leads to marginalized and then dead-end efforts.

The initiatives that have real legs, on the other hand, consistently have leaders who focus on the cultural and behavioral aspects of transformation. These leaders know that how they communicate with others is just as important as what they say. These are traits that rarely come naturally, however.

People who pursue leadership positions usually value boldness, quickness, and having the right answer. Thinking about how their behavior affects people—and results—is not often in that profile. That’s why they need coaching.

For many readers, this idea of PI officers as behavioral coaches to CEOs and CMOs will seem a far stretch. Will the PI director be welcomed into executive offices to talk about personal behaviors?

Fortunately, we have two terrific examples of organizations that have adopted the PI coaching model. There are many more health systems out there doing this kind of work—including some that have already been highlighted in this book—but these examples show two solid variations on the PI coaching structure that can serve as a template.

Carlos Scholz-Moreno, director of strategic initiatives for a large multispecialty physician group in Northern California (which prefers not to have its name published), has led the PI effort in two major health systems. He will give an inside look at setting up a coaching PI office from the ground up.

Didier Rabino, who helped lead the transformation of HealthEast in Minnesota and then stayed on as vice president and lean sensei when that health system was merged with the far larger Fairview Health Services, was trained in the classic Toyota Production System methods and brings that sensibility to the work.

Both leaders keep their PI offices small and focused on coaching and teaching. Their jobs are to spread the culture and knowledge of continuous improvement, they say, not to run events and update metrics.

Years ago, we advocated a larger PI office, based on the recommendations of John’s first sensei. It was a common rule of thumb in manufacturing that the PI staff should equal approximately 1 percent of the organization’s total FTEs1 and be used as a training ground for the health system’s future leaders. (Some consultants argued strenuously for an office equaling 3 percent of FTEs.)

In Management on the Mend, John highlighted organizations that had a lot of PI engineers, plus a steady stream of interns from the rest of the organization. These interns were rising leaders who took a 12- or 24-month rotation through the PI office to deeply learn the principles, tools, and behaviors.

These were valiant efforts but, in the end, cumbersome. If a health system had 700 managers, it would take decades to give all those managers a rotation through the PI office and provide the kind of in-depth training needed to transform the culture. Also, those larger PI offices led to the assumption that all those people on staff should be doing everyone’s PI work: updating A3s, tracking various metrics, creating huddle boards. This was detrimental to spreading knowledge deeply through the organization.

In that last book, John also showed Didier’s method at HealthEast. The PI staff was small and focused on coaching leaders to conduct their own improvements. In the end, this was the model that proved most effective at spreading improvement knowledge and the right behaviors.


We might not have met Didier had it not been for Kathryn Correia. The former president of hospitals at ThedaCare during its transformation, Kathryn arrived as CEO at HealthEast in 2012 determined to bring change to the culture as well as to the care processes. In her experience, the two could not be separated.

Early on, Kathryn took a couple of her new colleagues on a field trip to see what good looked like. At Andersen Windows in Menomonie, Wisconsin, they found it. On a walk through the plant’s morning routine—with their huddles, their clear process for escalating problems, and safety checks—Kathryn saw a frontline operator uncover a problem and signal for help. Two local leaders immediately met the operator, asked respectful questions, and used PDSA thinking to contain the problem and begin root cause analysis to find corrective action.

“You can’t fake that stuff,” Kathryn said. “That was the culture we wanted.”

So, Kathryn recruited the plant manager to HealthEast, and Didier—a former cabinet maker and engineer who studied with lean masters at Steelcase in Europe and Michigan before undertaking the transformation of Andersen Windows—began to learn healthcare.

Meanwhile, he was showing Kathryn a new way to think about the work of a performance improvement office. With a staff of less than 10 people, he began conducting workshops on A3 thinking, daily management, and strategy deployment for staff throughout the system. But instead of doing rapid improvement events for specific areas, he and his team coached leaders to run their own.

“When we had improvement workshops, our goal was to achieve business results, of course,” Didier says. “But it was also to transfer knowledge. We had assessments after every workshop to find out what they learned, but we also were watching to see who was holding the marker at the end of 81the workshop. Were we still leading the group or did the local owner take over? That was our goal.”

The PI office was also training apprentices—rising leaders in the organization who spent about 24 months in PI, learning to coach, use the tools, and lead workshops.2 The apprentices would arrive in pairs and be assigned to one of the staff advisors—mostly engineers who came from manufacturing and had deep expertise in lean. In general, the apprentices would coach frontline leaders while Didier and the advisors would coach executives.

Executive coaching took on two forms. Didier coached the executive team as a group as they defined their common executive standard work, developed strategy deployment, and did the crucial work of deselecting initiatives. Strategy deployment is the system for ensuring that top breakthrough strategies, as defined by senior leaders, are understood and acted upon at all levels of the organization and that frontline staff know how their actions effect strategies. This can be accomplished through the daily management system, where people in frontline huddles address issues such as safety (patient falls), quality (infection rates), and resource uses (staffing) identified by senior leaders. Strategy deployment is often tracked using an x matrix, described in more detail in Chapter 10.

Also, when executives were willing and interested, Didier and the PI advisors entered into more personal coaching relationships. The one-on-one coaching typically included 30- to 60-minute sessions featuring a few simple questions:

   What did you learn since we last met?

   Where do we stand now?

   Where do we want to be?

   How will we know we are heading in the right direction?

Executives used personal improvement A3s to focus their efforts on behavioral change. But most of Didier’s coaching time with senior executives took place in the work.

“Our job, as I saw it, was to allow the leaders to be in specific conditions, experiencing real things,” Didier says. “All of our beliefs are based on what we experience. So, we wanted them to have experiences while we were there to talk about it.”


To illustrate his methods, Didier recalled an executive named Dave who led a very large chunk of the health system’s business, revenue-wise, but who sometimes felt disconnected from the rest of the leadership team. Dave did not want to be left behind on this new direction for the organization, so he asked for personal coaching early.

Didier guided Dave to a self-assessment on 10 ideal behaviors. Dave identified two or three gaps between where he was and where he wanted to be and then set himself a goal to pursue excellence in visual management and self-discipline. Dave said he wanted to lead his team by example.

Knowing that leading with questions instead of answers was a core priority, Dave and Didier created a simple PDSA form that would help Dave ask for feedback from his team. Dave regularly asked his team members to rate his behaviors on a scale from one to five—rare, event-based, common, consistent, and uniform—and then used that information to guide his personal work.

“It was easy to see Dave’s behaviors changing from one month to the next. He became one of our best leaders,” Didier says. “And then, it was time to go further. He needed to lead from the gemba. So I asked him, what if we took his office away? Could he gather all the information he needed at gemba to make high-quality decisions?

“Dave turned around and asked his team that question. How could they make all necessary information visible? Dave’s division was like a medium-sized business. He had a print shop, a legal office, sales. With his team, he identified where the information was located, decided how they would make it visible, and defined the routes to connect the locations.”

Involving the frontline team was a revelation. Once they learned what was important to the senior leadership team, people became highly engaged in making that information transparent. The relationship between the front line and senior executives became more collaborative. A new type of teamwork emerged.

Not all coaching attempts were as successful, of course. Some leaders cannot give up the dream of power and control, no matter what words they say about humility and respect. This usually shows up in meetings or gemba walks when the leaders in question interrupt conversation about a problem or current condition with their own can’t-fail solutions.

In the best-case scenario, the power brokers end up being surrounded by colleagues who remind them of the process preferred by the group. And if a coach is present in that meeting, he can respectfully ask, later on, how the power broker felt that his suggestions were received. Still, for a personal coaching relationship to work, a leader must be open and willing to change.


Group coaching, especially in a structured setting like strategy deployment sessions, is a different matter. Because executives are accustomed to having a leader present options and guide them through material in strategy sessions, having a coach take that role is not so personally intrusive.

At HealthEast, and subsequently at Fairview, Didier employed a kata3 type series of questions, commonly used at Toyota, to be sure that everyone was clear about the goals.4 Using a flipchart and markers to make notes, Didier would ask the group:

1.   What is our quantified target condition?

2.   What is our current condition?

3.   What are the obstacles that prevent us from achieving the target condition today? How do we know these are the real obstacles and that removing them will take us to the target condition?

4.   What are the potential countermeasures to remove each obstacle? How do you know they will be effective? What are the potential adverse effects or risks associated with them?

5.   Which one should we try first?

This was usually accomplished in several sessions with the senior executive and her team. A major goal of these sessions was always to deselect initiatives that were extraneous to the target state. After all, Didier says, it does not matter how well we work on the wrong problems.

Deselecting initiatives is, of course, one of the most difficult things a senior team does. Healthcare leaders are notoriously additive. There are so many worthy ideas for improving patient care and adding to our knowledge, it is hard to say that one idea is not worthy. But this is an issue of respect. Keeping the number of initiatives to a critical few shows respect for the time and stress levels of the people in an organization who will inevitably shoulder the burden of this extra work. And it shows respect for the initiatives that are selected. Keeping the number small means we want these efforts to succeed.

In California, where Carlos Scholz-Moreno was leading the transformation of a very large multispecialty physician group and coaching the executive team in strategy deployment, the path to deselection was in visual management. That means he used an x matrix with senior leadership. (See Chapter 10, on the x matrix, for an example.)

The x matrix offers at-a-glance information about the strategic initiatives in play, ranked by importance to goals, and identifies each initiative by its impact on True North. Initiatives are color-coded by the time and resources required to complete the work:

   A red dot indicates the work needs five or more hours per week.

   A blue dot means two to four hours.

   A green dot shows less than an hour of work required per initiative, per week.

It is a time-consuming document to put together and requires regular updating. But the x matrix is powerful because it clearly identifies the limitations of an organization’s resources and serves as a constant reminder of these boundaries.

This helped Carlos guide his executive team—24 people with a lot of strong opinions—to create a manageable list of 10 mission-critical initiatives in active play. Waiting in the wings were another 11 initiatives deemed important and 7 more on a wait list. This is a big improvement over the 135 strategic initiatives that people were juggling when Carlos first

In the system Carlos helped set up, senior leaders had a steady, standardized cadence for managing strategic initiatives during their regular weekly meetings. Each initiative was reviewed monthly. Everything on the wait list was reconsidered quarterly. Leaders adopted strategic filters—a series of yes-or-no questions and observed conditions that created a decision tree, allowing leaders to prioritize initiatives—in order to deselect initiatives entirely or move wait-listed items up to active status.

Having a clearly defined process and visual management of strategy deployment helped make strategy meetings more productive, less rancorous. But the greater change lever, Carlos says, was the leadership-wide commitment to adopting and displaying the Shingo principles and behaviors and receiving regular coaching from Carlos and his group.


Before Carlos arrived in California, he had already spent a number of years thinking about the role of a continuous improvement office. After years in manufacturing leadership—with companies that made underwear and then guitar strings—Carlos joined the largest municipal health system in the United States, New York City Health + Hospitals. Over the course of six years, he rose to become assistant vice president of process improvement.

Those were six years of sometimes-turbulent learning, and while there, Carlos led his group away from depending on rapid improvement events to focus instead on a daily management system and frontline problem solving. He had lived through the sugar high of big events and had landed in a very different place.

“The default way of thinking—tools get results—might get you there initially, but this only gives you the illusion of gaining control,” Carlos says. “Really, you’re creating a culture of compliance. Tools only answer the question how, but never answer why.

Arriving in California in late 2017, Carlos met the new health system’s senior executive team, and what they really wanted to know was how? How will you get our continuous improvement efforts back on track? How will you fix our system? Carlos wanted everyone to take a step back, however, and ask some more foundational questions.

“We did an A3 on what our (PI) role should be in the organization. We started with very open questions to senior leaders like, ‘What do you think we should do? How can we best help?’ The responses varied wildly,” Carlos says. “Some people thought we should be updating their A3s. You know, doing all that lean stuff for them.”

Senior leaders of this three-hospital, 10-clinic system had been working with lean consultants for about eight years. That work had been project based. The improvements they implemented stayed in pockets that shrank over time. The executive team recognized this and did not want to return to their old methods.

So, they agreed with Carlos—in theory—that they needed to change their leadership in order to change their processes. They decided to try coaching, but some remained wary—especially when Carlos laid out his plans to train his small PI staff to be the coaches.

“Some people said, ‘No way. I’m a senior leader. I need someone super experienced.’ We said, ‘Look, we’re always learning. We are learning to be good coaches. You are learning to be learners.’ Finally, everyone agreed to a standard coaching schedule, once or twice a week, minimum,” Carlos says.

This was a pivotal moment for Carlos. One of the hardest jobs facing any PI leader is building credibility with executive leaders. Most PI leaders did not go to medical school or earn advanced degrees in hospital administration. They are outsiders in the system, and it can be difficult for a CMO to consent to take instruction from a former manufacturing engineer. This is where an external coach—brought in specifically to help senior leaders grasp the idea of behavioral change—can be very useful.


Also important here is a solid plan and process for conducting coaching in a respectful and effective way. Knowing this, Carlos started a new A3 with the following problem statement: We don’t have a consistent coaching method.

As part of the countermeasures for that A3, Carlos and his PI staff created standardized coaching questions and schedules—a baseline that they could measure and judge as they went along.5 He shared the schedule and coaching concept with executives and told them that he would ask for feedback every three months.

A group of six leaders in three model cells were assigned to coaches from Carlos’s staff, under his oversight. Every coaching relationship began with clearly stated, standardized ground rules.

“We begin by asking, ‘Do I have permission to be candid with you?’ ” Carlos says. “We need to develop trust. And it must be personalized, so we ask people how they like to learn—reading alone, for instance, or maybe by talking while being in the work—and how honest they really want to be. Because, if I’m completely honest, you might end up angry sometimes or frustrated. A leader might need to be told that their team members are afraid to be candid with them. That can be a hard conversation, so we need to talk about how to have those revelations come out.”

Meanwhile, Carlos began personally coaching four C-suite executives. The team created a curriculum so that the executives would know, in advance, what their hour of coaching would look like, what they would learn and discuss. Everyone began with Carlos’s idea of square one: True North and the principle constancy of purpose.

“After talking a little about what we mean by constancy of purpose and its importance, we have an initial reflection. We ask, ‘How do you, as a leader, create constancy of purpose? How do you communicate with your team, your direct reports, about what matters? When do you talk about performance? Show me in your calendar, where you practice constancy of purpose,’ ” Carlos says. “Normally, they are firefighting all over the place, so they can’t really show constancy. That’s OK. We can then start thinking about how to consistently communicate what matters.”

At the end of the in-person coaching session, they also agreed on an activity to be observed. If a leader needed to better understand their role at gemba, it would be a gemba walk. If team dynamics were an issue, the coach sat in on meetings or status exchanges to observe. After the observations, they met again to discuss what the coach saw.

If the meeting being observed was a staff member’s performance review, for instance, the coach would quietly observe, and then afterward, the coach would ask questions such as, How did you create a safe environment? How do you know that it’s working? How do you collect information on performance?

In every case, the coaches would tell the leader in advance what questions they might ask in the follow-up interview. This became a very important point, Carlos says. “You don’t want to catch leaders by surprise. They don’t like that.”

In addition to one-on-one coaching, the executive team also took up this work as a group. During the weekly executive team meeting, executives agreed to set aside 30 minutes for a group reflection, led by Carlos. He might ask how they understood constancy of purpose. How did they communicate it to others? Not everyone was willing to share at every meeting, but soon, conversations started.

Individually and as a group, for instance, executives looked at their calendars and began rating the meetings they attended each week. They found that not all of them were value added. Some meetings got eliminated; some were changed in content or scope.

And as executives created standardized work, they shared that, too. For every item of work, they showed themselves as green or red, for having succeeded or not. One of them might have a line item to check up on capital projects every week, for instance, or to practice asking open-ended questions, and there would be animated discussions about how they were earning a green dot for that work.

After 18 months of regular group reflections in the weekly executive meeting, senior leaders did not always need Carlos present to explain or referee. Sometimes, he stayed away on purpose to allow their reflection habit to be self-directed.


In the PI office, meanwhile, Carlos and his team created a process map of the coaching relationship. They met weekly to discuss who was deliberately practicing certain behaviors and how they could address gaps. They shared the questions they were asking, the problems their learners were encountering, and helped each other improve.

For each relationship, they set up a kind of PDSA for the coach to fill out, including notes on what the coach intended to achieve in the upcoming session, what the leader was learning, what the obstacles might be, what would be the focus of the next session, and how they intended to follow up. So, there were always two parallel lines of standard work being tracked: one for the learner, one for the coach.

In addition, Carlos was getting feedback from the learners every three months; he used that to tweak processes or redirect his coaches, as needed.

As the first year of this work was completed, Carlos had four coaches on staff, each of whom had four learners. They spent about 25 percent of the time coaching—about 10 hours of the week shadowing their learners or in one-on-one sessions. The rest of their time was spent in project management, training workshops, and facilitation of rapid improvement events.

“Rapid improvement events are necessary and will always be part of our system,” Carlos says. “That’s the way people learn the tools. But without a system to support the RIE, it won’t last. It’s not the tool that creates change; it’s the behaviors.”

As the first set of coaching relationships matured, there came greater demand. Executives identified new leaders that needed to deeply understand the work and, as usual, new executives joined the organization. The PI office needed to expand capacity.

So, Carlos set about hiring and training new coaches. He knew he needed five more people but started with two in order to experiment with training. He found people who 93had good technical knowledge and a few other traits he had identified.

“I knew we needed people with real self-awareness and strength in interpersonal relationships and problem solving,” Carlos says. “They also need to be assertive, which is tricky because, as a coach, they can’t be perceived as being pushy. They need to be unafraid to ask questions. So, not an accommodator and good with conflict resolution. We need collaborators, not competitors.”

He figured his newly hired consultants would need six months of shadowing and learning. Carlos guided them through personal A3s and self-assessments. They created personal development plans; Carlos and the PI team leader did weekly status-exchange meetings with them on their progress. In those meetings, he would ask: What was your last experiment? What did you expect to have happen? What happened? And then they would show Carlos how they were applying the principles to their work.

With the rest of his PI staff, Carlos identified a series of markers that would show when the consultants—who were also conducting training workshops and helping to run improvement events—would be ready to take on solo coaching. He is still surprised that the new consultants passed their markers in three months instead of six and were ready to take on coaching relationships of their own. Within five months, Carlos was beginning to interview new candidates.

Another surprise: within about three months of beginning work with one of those new coaches, an executive learner sent back a note in his quarterly feedback form saying that his new coach was more useful to him than the super-experienced outside expert who was shadowing and supporting the coach.


One final note on this matter of a new role for the PI office that we have learned from long and sometimes painful experience: the PI office must move forward in partnership with either HR or organizational development (OD). Whichever office has taken on the mantle of training within the organization must be consulted. This should be a collaboration, although we understand that this can run up against executives who will treat change like a turf war.

In the best-case scenario, the HR or OD leader who oversees systemwide training will be one of the first executives to enter into the coaching relationship so that the leader can experience it firsthand. And if there are disagreements, they must be brought to the surface and given full consideration.

Too many times, both of us have allowed powerful executives to harbor silent disagreement. We have thought, “That person will come along, once they see the amazing results our teams can produce.”

Leaders who remain on the outside of this work for too long can become resentful, however. And resentment is a hard spot to recover from. It is better to invite chief executives into a coaching relationship at the beginning—even if it is less formal—so that all leaders have the opportunity to see and understand the nature of continuous improvement.

One good method is to set clear timelines for each part of the organization to begin working in the steady cadences of a daily management system. To prepare each area, have executives receive coaching at the same time that frontline leaders undergo intensive training—often referred to as belt training—before the area makes the transition. In this way, each level of leadership can be receiving training and practice appropriate to their positions.

In Chapter 5, you will meet two organizations that found success with this rollout of the work. And in Chapter 6, on leadership in the support organizations such as HR, finance, and information technology, we will examine how good partnerships are formed across these boundaries.

First, let’s meet some managers.

1. FTE = full-time equivalent employee. An office with 10 full-time staffers and 6 part-time employees has 13 FTEs.

2. Because the PI office was small, it did not have capacity for a large number of leader/apprentices. It was understood that these were a few cases and not the necessary path to advancement in the system.

3. Kata is a Japanese word meaning form, most commonly used to identify the choreographed training exercises for martial arts practitioners.

4. See Mike Rother’s book Toyota Kata for a more complete illustration of these methods.

5. Along with our colleague Tom Hartman, we assisted Carlos in this work.