Turning Doctors into Leaders

Artwork: David Maisel, History’s Shadow GM8, 2010, Archival pigment print, 40″ x 30″

The problem with health care is people like me—doctors (mostly men) in our fifties and beyond, who learned medicine when it was more art and less finance. We were taught to go to the hospital before dawn, stay until our patients were stable, focus on the needs of each patient before us, and not worry about costs. We were taught to review every test result with our own eyes—to depend on no one. The only way to ensure quality was to adopt high personal standards for ourselves and then meet them. Now, at many health care institutions and practices, we are in charge. And that’s a problem, because health care today needs a fundamentally different approach—and a new breed of leaders.

Listen to an interview with Dr. Thomas Lee.
Download this podcast

Most recent discussions of health care have focused on its rising costs, but these financial challenges are really just a symptom. What is the real “disease”? The usual suspects have surprisingly small roles. Greed and incompetence surely exist, but economists agree that they don’t account for double-digit annual cost increases on their own.

The good and the bad news is that the biggest driver of rising costs is medical progress: new drugs, new tests, new devices, and new ways of using them. These tools are frequently marvelous and complex, and their use requires increasing numbers of personnel trained in narrow fields. Patients with complicated conditions end up seeing a variety of physicians who are often spread across several institutions.

Of course this progress is welcome, and at times it seems miraculous. The Red Sox pitcher Jon Lester was diagnosed with lymphoma in September 2006, but he reported to spring training in 2007 and pitched a no-hitter in 2008. Steve Jobs is still on the job. Many patients diagnosed with heart failure can now go back to work after receiving a new type of high-tech pacemaker.

But this explosion of knowledge is going off within a system too fragmented and disorganized to absorb it. The result is chaos. In my own organization, Partners HealthCare, a poignant example involves the widow of a young man who died of cancer. In the last days of his final six-week stay in the intensive care unit, she demanded that all his doctors have a meeting with the family. The family didn’t really need the meeting, she said—the doctors did. She wanted to be sure that the various physicians were actually talking to one another, because she so often received inconsistent or even contradictory messages from them. The confusion she described does more than distress families, of course. It leads to redundant care and errors that raise costs and threaten quality.

Tough Medicine

To effectively attack this chaos we need a new kind of leadership at every level of the health care system, from large integrated delivery systems like Partners to hospitals to physician practices. The specific kinds of work and performance measures may differ from one setting to another, but the key responsibilities of leadership are the same. To understand what they are, leaders must first absorb three painful messages:

Performance matters.

Most clinicians are hard workers, but the quality of their work should not be measured by how many patients they manage to see or tests and procedures they call for. What matters is their results. This is controversial, because comparing outcomes is notoriously complicated. After all, how well patients eventually do depends heavily on how sick they were to start with. Nonetheless, the bottom line is how patients fare. How often do they survive their illnesses and recover from their disabilities? How frequently do they get infections and other complications? Are their informational and emotional needs met?

“Value” is not a bad word.

When employers and insurance companies use the term, many providers suspect that it’s code for cost reduction. But Michael Porter, of Harvard Business School, and others have been pointing out for years that in health care, “value” means something else: achieving good outcomes as efficiently as possible. It may never be expressible as a numerical ratio (quality divided by costs) that allows meaningful comparisons among providers. But measuring outcomes and costs does allow providers to push for improvement—and to learn from their competitors.

Improvements in performance require teamwork.

Individual clinicians and hospitals have only limited control over the fate of their patients. At any organization that provides health care, superior coordination, information sharing, and teamwork across disciplines are required if value and outcomes are to improve.

Many leaders of providers can pinpoint the moment when they realized that their world was changing; often it came when someone outside the organization started measuring its performance. Although few providers welcome this development, it provides context for a new breed of leaders. Traditional health care leaders try to buy time, fend off change, and maximize revenue under the existing payment system while they can. The new leaders focus on outcomes and use performance measurement as a motivating tool to organize their colleagues and drive improvements.

The challenges are similar whether these leaders are working in a large integrated delivery system, a hospital, a large multispecialty physician group, or a small physician practice. Although their tactics will vary from one setting to the next, the broad roles that leaders need to assume will not.

Articulating Vision and Values

The reorganization process starts with articulating the rationale and goals for change. Change is hard in any field, and medicine’s altruistic core values actually reinforce practitioners’ resistance to disturbing the status quo. My generation’s traditionalists know that they are good people who work hard, and they have the courage of their convictions as they point out the risks of change. So the vision expressed by leaders in health care must convey both understanding and resolve. It should acknowledge the importance of what clinicians currently do, but make explicit that they have to work differently in the future. It should be direct about the measures by which they must succeed. And it should be both optimistic and realistic, expressing the beliefs that care can get better and that delivering superior care is the best business strategy.

An effective vision helps people accept inevitable changes and put information and events into context. For example, many physicians and hospital leaders have a viscerally negative reaction to public reporting on the quality of care they provide. They know the limitations of the data and are appalled that providers might be ranked numerically on the basis of inadequate, easily misinterpreted information. Their typical reaction to a decision to release data on provider quality: Civilization is coming to an end.

In contrast, consider how the cardiac surgeon Delos M. Cosgrove, who became CEO of the Cleveland Clinic in 2004, folded the imperative for performance measurement into a broad vision. If the clinic was committed to the idea of “patients first,” he argued, it had to not only make a serious commitment to measuring patient outcomes but also demonstrate that commitment to the world. Cosgrove immediately took the measurement systems that had evolved in one part of the organization and disseminated them throughout the clinic. At first the new data were available only to insiders; now they are published, warts and all, on the clinic’s website. Physicians were indeed uncomfortable with these changes, but seeing performance measurement as a tool to help (and attract) patients, rather than as just a carrot or a stick, brought them along.

Leaders at Seattle’s Virginia Mason Medical Center made a similar commitment to the notion of patients first, but they took it a step further by making explicit its clear corollary: Physicians and everyone else come second. Whereas patients in most cancer centers do the walking—to the laboratory, to doctors’ offices, to chemotherapy infusion rooms—patients visiting Virginia Mason’s new cancer center are ushered into well-appointed rooms where doctors, nurses, and lab technicians come to them. These rooms are filled with natural light from large windows; the physicians work in windowless cubicles in the floor’s interior.

Virginia Mason’s cancer center embraced its patients-first vision so zealously that some of the doctors on the staff left. But those who remained, despite some grumbling, have helped engineer the center’s financial turnaround and rise to national prominence.

Organizing for Performance

Focusing on performance in health care is more radical than it sounds. In the era now waning, the conventional wisdom has been that true quality can’t be measured. Thus performance has generally been gauged by the volume and profitability of services delivered.

In the traditional world, medicine is organized around what doctors do rather than what patients need. For example, hospitals often have separate units for cardiology, cardiac surgery, cardiac anesthesiology, and radiology, each of which includes doctors and other clinicians who contribute to the care of patients with heart disease. Every unit has a physician leader and an administrative staff. At many hospitals the various units independently submit their bills (“claims”) to insurance companies and patients. That’s why patients are so often confused by multiple bills.

These clinicians may actually work well together in caring for individual patients, but increased costs and dysfunction are inherent in separated administrative structures. The units are staffed by people with good intentions, but they all have turf to defend—and in the mainstream of American medicine, threatening someone’s turf is a quick path to destructive conflict. In the absence of compelling reasons to change this arrangement, inefficient structures remain stubbornly in place. And for clinicians to embrace a radical redesign of care delivery—well, that would be an unnatural act when they are organized according to their specialties and contented to remain so.

This fragmentation often goes deeper than the organizational division of physicians. At many hospitals relationships between doctors and administrators are downright antagonistic, and financial interests are poorly aligned or even in direct conflict. For instance, hospitals want to shorten lengths of stay because they receive a lump sum for a patient’s entire admission, but doctors are paid for each visit on each hospital day, so the sooner patients go home, the less they make. Under most insurance plans, neither is rewarded for doing the extra work that might prevent a readmission to the hospital.

Organizing to deliver high performance (rather than units of service) can help break down all these barriers. As performance starts to matter, for example, some providers are moving toward structures for the delivery of care that are defined by patients’ needs. In many cases, the first step is colocation—putting the various types of physicians who provide most of the care for a patient population in one place. Sometimes an opportunity for colocation is created by the construction of a new facility dedicated to patients with specific conditions, such as cardiac disease or cancer. More often, institutional leaders must move groups around in an elaborate multiyear effort to bring physicians from different disciplines but the same patient population closer to one another.

But colocation alone can’t guarantee a well-coordinated effort to improve patient outcomes. That’s why Delos Cosgrove abolished the Cleveland Clinic’s traditional departments and replaced them with “institutes” defined by patients’ conditions. He realized that as a cardiac surgeon, he needed to collaborate more with cardiologists than with surgeons who operated on other parts of the body. So he brought together the clinic’s cardiologists, cardiac surgeons, and vascular surgeons in the new Heart and Vascular Institute, and started capturing and publishing information on how its patients have fared.

In similar facilities, such as the Head and Neck Center at Houston’s M.D. Anderson Cancer Center, physicians remain members of their various departments, but they’re in close proximity on two adjacent floors. Over time they have come to identify more with their cancer-center roles than with their departmental affiliations.

The work of organizing care around the needs of patients instead of physicians’ turf and politics plays out on smaller scales as well. Within many leading hospitals today, physicians from the departments of surgery, medicine, and radiology work together to perform hybrid procedures (such as simultaneous open-heart surgery and abdominal aneurysm repair) that previously would have been performed separately.

Large-scale organizational changes like these require strong leaders and a cultural context in which they can lead. For obvious reasons, such leaders gain additional leverage if they are physicians and their organization employs its doctors. At the Cleveland Clinic all physicians are on one-year renewable contracts, which sends a powerful message about the importance of team spirit.

Rule of Thumb

Generally speaking, the number of people an organization needs to train in process improvement is the square root of the total number of personnel. Thus, if you have 100 people,

…

Not every institution will have the leadership wherewithal to undertake such transformative change. But even when integrating departments of clinicians seems unrealistic, strategically chosen performance measures can spur progress.

Developing a Measurement System

The first challenge in creating a performance measurement system is getting everyone across an organization to use the same “language”—that is, to measure the same things in the same way. Otherwise it’s easy, and understandable, for resisters to challenge the validity of apparent differences. But once providers believe that apples are being accurately compared with apples, peer pressure and other incentives will help spread best practices. (See the sidebar “Using Peer Pressure to Improve Performance.”)

Using Peer Pressure to Improve Performance by: Kelly W. Hall

Financial incentives in health care, as in any other industry, are necessary but not sufficient to optimize people’s performance. If, for instance, they are linked to too many or

…

Consider this example of the importance of a common language, which comes from my own organization. Bloodstream infections are a serious and frequent problem in patients who have indwelling catheters in their arteries and veins, and in January 2008 the Massachusetts state government announced plans to begin public reporting of each hospital’s rate of such infections. Knowing that reporting was coming, my colleagues began including data on bloodstream infections in the internal quality report cards that are shared with our board and other senior leaders. These reports revealed apparent differences in the rates at two of our hospitals, but meaningful comparisons were hindered by their differing methods of detecting the infections.

Hospital A monitored for infections by drawing blood samples through the indwelling catheters—an approach that was painless and easy but more likely to lead to contamination or false positives. Hospital B checked for infections by drawing blood through fresh needle sticks. Hospital A usually had higher rates of infection, but its physicians always argued that this was simply a false finding resulting from their particular detection method. After they finally adopted Hospital B’s method, however, they found that they still had a higher infection rate. With both hospitals measuring in the same way, claims about the source of the difference ended. Infection-control leaders became very interested in what else Hospital B did differently.

When data are uniform and reliable, leaders can push for the standardization of best practices throughout an organization. For example, clinicians at all the hospitals in our system have agreed to attach colored tape to catheters inserted under less-than-ideal conditions in the emergency department. The tape tells doctors and nurses in the intensive care unit to change those lines as soon as possible once the patient is stable—a practice that we expect will further reduce our infection rates.

In this case, standards were not dictated from the top of the organization. What did come from the top was pressure to collect data in the same way at all hospitals and use it to improve care. Innovation occurs at the front lines of health care; our senior managers would never have thought of using colored tape on catheters. But they could and did create the environment in which such ideas spread.

Finally, an effective measurement system requires clear metrics that detail costs and outcomes for episodes of care or even entire patient populations. These data can be wielded in ways of varying impact. Although workers in any environment, medical or otherwise, will respond to negative motivators (“Reduce the infection rate or you will be humiliated”), positive ones (“Reduce the rate because you want to provide the best possible care”) can be more effective. Such value-oriented performance measurement should become the focus of internal improvement efforts—before measurement is forced upon leaders from the outside.

Building Effective Teams

Working in teams does not come easily to physicians, who still often see themselves as heroic lone healers. Nonetheless, developing teams is a key leadership function for health care providers of all types.

Working in teams does not come easily to physicians, who still often see themselves as heroic lone healers.

Consider how teams at Pennsylvania’s Geisinger Health System (where I sit on the board) have helped cut hospital readmissions by half. In the United States about 20% of Medicare patients discharged from the hospital are readmitted within 30 days. These “bouncebacks” should be seen for what they are—failures of the delivery system to meet patients’ needs. Even in retrospect many readmissions seem unpreventable, but others result from confusion about what medicines the patient should be taking, what signs might suggest that a complication is brewing, when the patient should go back to which doctor, and so on.

In that confusion lies an opportunity for well-organized providers. The obvious ingredient in Geisinger’s recipe for success is placing nurses (“care coordinators”) in the offices of patients’ primary-care physicians. The care coordinators stay in close touch with patients whose cases are complex, particularly when they are about to be discharged from the hospital or have recently gone home. They figure out which patients need to see which physicians and when.

The more subtle ingredient in this model’s success—the secret sauce, as it were—is a culture in which care coordinators can actually coordinate care. It requires that physicians be both team leaders and team players. Not long ago, in the strict hierarchy of medicine, nurses were largely regarded as technicians whose job was to follow orders. No decision was made without a physician’s knowledge and consent. The notion of a nurse as a critical contributor and independent decision maker on a clinical team would have seemed absurd.

That’s changing, because providers that deliver care in the traditional way simply can’t match the performance of Geisinger and other organizations where physicians work in teams with care coordinators. In these organizations the coordinator’s role is something like that of a point guard in basketball, with the physician acting as a combined general manager and player-coach. Leading these teams requires physicians to hand off considerable responsibility to nurses. The payoff is improved performance on the metrics that matter most to them and their patients.

Team building is a critical competency for leaders of physician groups, particularly the increasingly common groups of 25 or more that include doctors from a range of specialties. Most of these doctors joined their groups not because they wanted to work collaboratively with others to improve performance but because they wanted to be in a big organization where someone else would worry about administrative hassles, they’d find some safety in numbers from market forces, and it would be possible to make additional revenue from ancillary services such as radiology and laboratory testing.

But the fortunes of these groups and others in all areas of medicine will depend on leaders who can improve performance by inspiring (or requiring) teamwork. In many markets insurers are incorporating costs and quality in insurance product design, so that patients pay more—or aren’t covered at all—if they want to see physicians in more-expensive, less-efficient, or lower-quality groups. If patients don’t come, the groups will fail. The ability to build high-performance teams confers competitive advantage.

Improving Processes

Health care teams can’t view their purpose as time-limited or focused on one project. The day will never come when readmission rates are low enough, heart-attack treatment is fast enough, or all the processes of care delivery are efficient and reliable enough. Thus leaders must work relentlessly to reduce errors and waste and improve outcomes—for example, by preventing bouncebacks or reducing the time between a heart-attack patient’s arrival in the emergency room and the opening of his or her blocked artery. To do so they need a culture of process improvement and the disciplined use of its methods, such as lean management, data collection, brainstorming, intervention, and impact analysis—and a long-term commitment to applying them. That culture and experience can be ingrained in a variety of ways.

The now classic health care example comes from Virginia Mason Medical Center, which a decade ago was in danger of losing market share and its best physicians to the numerous outstanding hospitals in the region. In 2001 the center’s president, J. Michael Rona, happened to sit next to John Black, then the director of lean management at Boeing, on a plane. Black had sent hundreds of Boeing managers to Japan to study the Toyota Production System. By the end of his flight, Rona was convinced that Virginia Mason needed to do the same thing.

Rona and the center’s CEO, a physician named Gary Kaplan, began taking groups of their colleagues to Japan for two-week immersion courses in TPS. Rona and Kaplan let it be known across the organization that leadership roles would most likely be reserved for people (including physicians) who took the training and adopted the lessons. Some of their star physicians left as a result. But Virginia Mason has used its version of TPS to reduce costs, improve quality and service, and strengthen its financial health.

Of course, not all hospitals or practices can or should fly their staff to Japan. But they can find (and increasingly are finding) alternative ways to bring the process-improvement culture inside. Many organizations have sent their midlevel and senior leaders for training at Intermountain Healthcare in Utah, where Brent James, the chief quality officer, runs a highly respected process-improvement course. And James has generously helped some of these organizations, including mine, start their own courses to spread this expertise.

Dismantling Cultural Barriers

Why is collaboration so hard in a field that attracts idealistic people who want to do good? Why are performance measurement and improvement so problematic for some of the smartest, hardest-working, and most competitive people in society? Why is the concept of value rejected by providers that would benefit if they improved their care?

In a word, autonomy. The cultural barriers to change in health care—doctors’ resistance to being measured, their need be “perfect,” their reluctance to criticize colleagues, their resistance to teamwork—reflect a deep-seated belief that physician autonomy is crucial to quality in health care. Doctors have historically seen themselves as their patients’ sole advocates, with the rest of the world divided into those who are helping and those who are in the way. A temper tantrum in the pursuit of patients’ interests was acceptable behavior. Some of my most respected colleagues have confessed a wish that no one would even talk to their patients except through them.

Precious as this passion for patients’ interests might be, physician autonomy is not synonymous with quality. For the needed structural and operational changes—performance measurement, process improvement, teamwork—to become mainstream, doctors must accept that to be all-caring is different from being all-knowing or all-controlling. To foster such acceptance, leaders can use three approaches:

Appeal to the better angels.

People who are drawn to health care want to focus their life’s work on something good: helping patients. Altruism is core to the identity of physicians and virtually everyone else in medicine. Health care leaders cannot succeed without making it explicit that they share and will act on the same aspiration.

Twice during the past several years I watched James J. Mongan, the recently retired CEO of Partners HealthCare, sit quietly while colleagues discussed whether to make a given practice standard across our network of physicians. The first time was January 6, 2005, when we considered whether to mandate the use of electronic medical records. The second was March 6, 2007, when we weighed automatically enrolling patients with heart failure in a disease management program, as opposed to waiting for physicians to refer them.

We believed that both steps would improve patient care, but we hesitated because we knew we’d anger some physicians by curtailing their autonomy. They might leave our network and take their referrals with them. In both instances, though, Mongan brought the discussion to an end by saying, “I really think this is the right thing to do.” No one could argue with him, and no one did. In medicine, more than in most other fields, a senior leader’s appeal to doing “the right thing” can serve as a trump card.

A senior leader’s appeal to doing “the right thing” can overcome physicians’ resistance.

Show them the data.

Physicians are quick to challenge performance data and to identify methodological problems with them. But the fact is that they are mesmerized by data and cannot look away. Brent James may be the ultimate health care data swami; for decades he has put the collection and sharing of data on quality and efficiency at the core of Intermountain Healthcare’s culture. Rather than make a frontal attack on physicians’ autonomy, he wears down their resistance to change by showing them how their practice varies from the norm.

Define strategy around patients’ needs.

What doctors know and do is constantly changing, but the needs of patients remain the same. They get diseases, they worry, and they hope to be cured or relieved of suffering. Meeting their needs is what health care is about. As we’ve seen, the leaders of organizations like the Cleveland Clinic, Intermountain Healthcare, and Virginia Mason Medical Center have taken the concept of patients first from an abstraction to a robust organizational strategy.

A shift to value-oriented, performance-driven health care requires doctors to adapt or even reject some ways of working that are embedded in medicine’s past. Difficult as this change will be, I am optimistic that the new generation of leaders will achieve it. In truth, they have no choice. Defending the status quo is no longer a viable strategy, even in the near term.