1 What Change Looks Like – Becoming the Change: Leadership Behavior Strategies for Continuous Improvement in Healthcare

What Change Looks Like

“I know I need to change. But how?”

The voice of the chief medical officer over the telephone was tight and unhappy. After 15 successful years as a medical specialist and another 3 years in leadership at this 500-bed American hospital, she was not used to stumbling on the job. She had gone to all the best schools and, in her experience, solutions were almost always within easy reach.

Now we were asking her to look at her hospital units in another way. Specifically, we were asking her to look without talking, without offering solutions. It was so much harder than she thought.

Like many doctors of her era and training, she was taught to take control of situations. If someone had a question, she liked to say, she had the answer. And because she always delivered that answer with a smile and a sense of camaraderie, people loved her.

Yet every day, this CMO left the hospital not knowing what she did not know. She knew there were problems of unknown origin and that she talked more than she listened. She wanted to change, she told Kim. The idea of learning from frontline medical staff about their problems made perfect sense. But how do you change decades of training?


So, a few weeks later, Kim arrived for a visit and went to gemba with the CMO to observe her in action. The CMO’s first job in this personal coaching session was to observe a nurse going through the process of discharging a patient from the hospital. The CMO notified a nurse in advance that they would be observing the process in order to learn—not judge—and then listened as Kim listed the rules.

The CMO should ask no questions unless she saw something that was clearly putting the health and safety of a person at risk. If that happened, the gemba visit would be suspended while the situation was corrected. Otherwise, no questions. If the CMO could not help herself and absolutely had to speak, she was given a list of four acceptable questions that were focused on the purpose of the work. The CMO nodded; she was ready.

On the medical-surgical unit, a patient’s lead nurse handled all the administrative and quality checks to be completed before a patient left the hospital, such as ensuring necessary testing had been performed, follow-up appointments had been scheduled, and prescription medications were dispensed.

In this case, the lead nurse was off shift and so the role fell to a nurse unfamiliar with the patient. The fact that he was being watched by the CMO and Kim probably made the nurse even more careful. He seemed determined to check every bit of information twice. But just as he confirmed that one test was done, a colleague alerted the nurse that he was getting a new patient. He left to complete that admission process, then returned and was interrupted again with questions about another case.

Still, this nurse diligently stepped through the discharge process, trying to get a patient safely home and free up a bed.

“Are you going to call the pharmacy and see if his prescriptions are ready?” the CMO asked.

The nurse nodded, interrupted his current task, placed that call, and continued on. Placing that call to the pharmacy was on the nurse’s list; he just had not arrived there yet.

Back in the CMO’s office, Kim asked how she thought the observation had gone.

“Why do you think the discharge process took so long?” the CMO asked in return. “We really need to streamline that process.”

“Well, part of the problem may have been that the nurse was interrupted 13 times,” Kim said. “And 5 of those times were by you.”


We begin with the story of this observation because moments like this are the raw material that culture is made of. We can make all the announcements that we like about transformation and organizational excellence, but none of it is real until leaders exhibit the behaviors that align with their words.

If a leader says she is there to observe and ends up directing the action, we know there is another agenda.

No matter how warm and personable this CMO was, every time she reminded the nurse to do a bit of work, she was saying that she knew the process better than he and that she was watching for compliance. And so they remained locked in their old roles: the know-it-all boss, the subservient underling.

This is a dynamic we have seen play out dozens of times, and it has led us to a simple lesson: motivations matter, and at gemba, people can see yours.

If nurses believe that a CMO wants to see the established process working well, they will work to make it go smoothly—glossing over the issues they struggle with every day.

If nurses believe that an executive is observing a process in order to see what is and find ways to help solve problems, they will be more comfortable pulling problems to the surface where everyone can examine them. Only then do we have the right environment for transformation.

The next question, of course, is, what will motivate that CMO to behave in a way that will help her see problems? To overcome those decades of training, she needed more than a few instructions on how to act. She—and every other executive and leader in that hospital—needed internal drivers. And for that, they needed common principles to guide them.

Creating a common code of fundamental truths that serve as the foundation for reasoning and action—in other words, principles—is what successful human organizations have always done. Whether people were banding together to learn or worship or create security, the organizations that lasted all had a common code to guide behavior. (Even if people did not always follow it perfectly.)

It turns out that the principles that will drive the right leadership behavior are pretty common. While every health system needs to collectively decide on its own principles, we will continue to cite the Shingo Institute principles (shown in Figure 1.1 on next page) throughout this book because they are the clearest example we know and perfectly align with the traits and behaviors required in organizational excellence:

FIGURE 1.1 Shingo Principles

   Create value for the patient.

   Create constancy of purpose.

   Think systemically.

   Assure quality at the source.

   Improve flow and pull.

   Seek perfection.

   Embrace scientific thinking.

   Focus on process (not people).

    Respect every individual.

   Lead with humility.

Let’s return to that medical-surgical unit for a moment and see how the action—going to gemba to observe—would be different when driven by these principles. If our CMO thought of her task as being inseparable from principles such as respect for people and focus on process, it would have been difficult to interrupt the nurse as he worked through the discharge procedures. Instead of assuming the nurse would forget a step, the CMO would have been more curious about how well the process worked and whether the nurse had everything he needed. She definitely would have waited to ask questions until the patient’s discharge was complete. That simple shift in focus creates an atmosphere where people are confident enough to point out problems.

Throughout this book, we will be examining moments like this through the lens of different principles to get at the root of the behaviors that create a culture. At the same time, we will be looking at the actions people must take at every level of leadership to get to organizational excellence. Both types of change—internal (reflection on principles) and external (behaviors)—must occur at all levels of the organization.

Because there are so many moving parts to a transformation like this—with specific actions taking place at each level of leadership, tools being taught, and common principles being learned and adopted by all—it is easiest if we visualize how it all works from ground zero, the model cell.


First, however, we need to define a model cell. In the past decade, the term has suffered a reduced reputation. In many organizations, what’s called a model cell is actually just the experimental redesign of an area, separate from the rest of the organization and not of strategic importance. Improvement goals are often as low as 5 to 10 percent, and when it comes time to spread the work to other areas, it’s just one pretty neat idea that is introduced around.

What this creates is an island of pretty good, surrounded by an organization that is unchanged. It is rarely worth the trouble.

A true model cell is the first act of a revolution. It strikes at the heart of an organization in that it begins with a business problem that is central to the organization’s future. The model cell has an executive sponsor, not just the attention of an improvement team, and the goal is 50 to 100 percent improvement on the model cell’s critical measures within six to nine months.

Does 50 to 100 percent improvement sound out of reach? In 2017, a 450-bed acute care medical center on the West Coast of the United States chose to focus on the ED to ensure that patients were seen faster and moved into the hospital more efficiently when required and that costs were controlled. Over the course of six months, workflows and job roles were redesigned, and a new management system was put in place. At the end of that six months, the number of people who gave up and left the ED without treatment dropped by 98 percent, even while volume in the ED rose by 7 percent. The number of patients being boarded in the ER decreased by 29 percent, and the average cost for ED cases fell 13 percent.

The model cell is where the most tangible change begins. But that does not mean it is the first thing to happen in the organizational excellence transformation.


First, senior leadership needs to assess the organization’s current conditions, its goals, and the gap between the two. As a group, they identify the few critical measures that become their True North (more on that in Chapter 4) and agree upon their principles. Only then does the leadership team have the necessary information to select the site for the model cell and choose the executive sponsor.

Now, the real learning begins. Leaders can choose one executive sponsor for the model cell, but all executives will need to lead various aspects of this transformation. The question is, what kind of people should lead the new world? What traits and behaviors are most desirable?

The leadership traits the team identifies should naturally flow from the principles being adopted. As a starting point, we advise groups to adopt these five fundamental leadership traits:

1.    Willingness to change

2.    Leading with humility

3.    Curiosity

4.    Perseverance

5.    Self-discipline

These are not lofty goals. There are many studies that show leaders who display these traits are more effective at creating positive results.1 More important, there are behaviors that can be learned and practiced to strengthen these traits. People are capable of change.

In order to change, of course, leaders must assess their current conditions and identify their goals and gaps. This means asking questions such as, “Am I glad to see a problem come to light? Do I really listen to my direct reports, or am I already thinking of a solution?” We highly recommend using a personal A32 for this work.

Once a team has identified their desired traits and assessed their own behavioral goals and gaps, we advise leaders to create a system of responsibility to one another. Maybe it will look like the mini 360-degree reviews created by UMass Memorial and detailed in Chapter 6, or the group reflections used by the California executive team profiled in Chapter 7. There are many ways to create responsibility, and all of them are legitimate as long as they provide respectful and timely feedback to leaders.

For most humans, there are gaps between how we think we are behaving and how others perceive us. Feedback and responsibility are our mirrors.

As leaders are learning new behaviors that support the work of the model cell, they will also be working with staff and medical personnel to redesign workflows in the target area and develop the routine of a daily management system.

Prior to the redesigned system being launched, executives and the improvement team will be preparing to spread the new way of designing workflow—not just the coolest idea from the model cell—to other areas. The management system will spread too. This means that senior leaders must become coaches, helping others to embrace principle-driven behaviors while solving their own problems.

This is the component of model cell work that most teams miss: behavior change. One thing will not happen without the other.

Over the years of doing this work, one of the most consistent truths we have seen is that new methods rarely take root without new leadership behavior, and so any benefit to the patient is fleeting. Yet new leadership behavior alone, of the kind we have outlined here, can create measurable improvements to patient outcome.


In the fall of 2016, Dr. Susan P. Ehrlich, CEO of Zuckerberg San Francisco General Hospital (ZSFGH), did not set out to prove this point about leadership behavior. She just happened to have two major initiatives happening at the same time. One was her attempt to bring organizational excellence to the hospital, beginning with identifying True North and focusing on leadership behaviors. The second was the flu.

Flu season is pretty similar all over the country. It escalates in fall, peaks in late winter, and dissipates in late spring. And 2017 looked like it was going to be a bad year for flu. Still, Susan was determined to focus the 55 top leaders at ZSFGH on aligning their behaviors with the goals of organizational excellence.

They all began by defining their principles and five desirable leadership traits—willingness to change, humility, curiosity, perseverance, and self-discipline—and then worked with their peers in developing a personal A33.

Part of the leaders’ personal A3 countermeasures plan was to create leadership standard work, including daily, weekly, and monthly activities that kept them on track toward achieving the organization’s strategic goals, as well as their personal goals. They also used radar charts4 and scheduled time for self-reflection so that they could gauge their personal growth toward the five desirable leadership traits.

The executives met in small groups to share their personal A3s with one another, and Susan convinced some of the leaders with more developed plans to go first. These peer-to-peer learning sessions proved to establish a camaraderie about the changes they were all going through. To create responsibility, they began with simple self-reflection but soon changed to 360-degree reviews, hoping to better see how others viewed their behavior.

This was no small task to take on during flu season. But the truth is, there is always an emergency or some budding crisis in a hospital. Perfect opportunities to take on extra work are rare.

This year, however, the flu kept getting worse. By late spring, patient volumes in the ED were not decreasing, but continuing to rise. In July 2016, the hospital’s ED saw about 175 patients a day. In January 2017, the average patient population was 215 per day and rising. Volume did not peak until January 2018, at 240 patients per day.

In most hospitals, the reaction to this crisis would be the same: rising costs, patients turned away, and staff training sidelined while all hands responded to the emergency.

But Susan and her team doubled down on their plans. They established weekly standup rounds for leaders to review patient flow through the ED—which had just undergone a major overhaul as the hospital’s model cell—and hospital inpatient services.

Leaders listened to frontline staff and one another as they focused on staffing, supplies, and patient transportation. They recruited colleagues to help them understand issues and spent time at the front line practicing their desired leadership traits, asking questions, and removing barriers.

In the thick of it, the leaders revisited their True North metrics and selected a few new performance metrics: percentage of time on ambulance diversion, mean ED length of stay, and number of patient days attributable to nonacute patients. Using A3s, they told each other the stories of multiple rapid experiments focused on areas where patient flow was getting stuck. They included staff ideas to address the problems and presented results to one another and to Susan, who made it a point to spend time at gemba with her team.

Patient volume in the ED remained high all summer, and as fall approached, and with it the promise of a new flu season, every executive at ZSFGH became engaged in this kind of problem solving. By the end of the year, 100 percent of leaders were trained in A3 thinking—an improvement of 23 percent. And 87 percent of leaders had adopted standard work for themselves—an improvement of 58 percent.

These behavioral changes had direct impact on quality scores that ZSFGH leaders were tracking and, by extension, on patient outcomes.5 During 2017, while patient volume in the ED increased by 13 percent, important measures fell:

   Mean length of stay in the ED—decreased by 9 percent

   Time on ambulance diversion—decreased by 25 percent

   Days attributed to nonacute patients—decreased by 35 percent

This type of research linking leadership behavior to patient outcomes is still nascent. In fact, reliable scientific studies on the use of organizational excellence or continuous improvement tools—independent of leadership behaviors—is not robust, either. There were interesting studies at UC Berkeley in 2019, but to this point, healthcare leaders have not had the benefit of clear scientific conclusions regarding improvement methods.

What we do know is that organizations focused on continuous improvement provide better results for their patients. And leaders that consciously adopt principle-driven behaviors—and then coach others to do the same—ensure that continuous improvement work can thrive.

In this book, we focus on the behaviors required of leaders at all levels of a health system. Like most important initiatives, this change must begin at the top of an organization. But what we consider the top might be surprising to some.

1. Jim Collins, “Level 5 Leadership: The Triumph of Humility and Fierce Resolve,” Harvard Business Review, July/August 2005.

2. The A3 is a more detailed version of the plan-do-study-act cycle. It is the scientific method for problem solving on a single sheet of A3-sized paper, expanded to include the business context and possible root causes for the problem. A personal A3 focuses on a leader’s plan to practice desirable behaviors. Chapter 8 describes the personal A3 in much more detail and shows a sample.

3. See Chapter 8 for a complete description.

4. See Chapter 9 for details.

5. These results were originally published in Susan P. Ehrlich, MD, MPP, and John S. Toussaint, MD, “Changing Leadership Behavior Gets Real Results,” NEJM Catalyst: Innovations in Care Delivery, October 10, 2018.