HEALTH CARE

FEBRUARY 2009

When clinicians lead

Health care systems that are serious about transforming themselves must
harness the energies of their clinicians as organizational leaders.

James Mountford and Caroline Webb

 2

The health care industry faces daunting challenges. Across developed countries, cost
inflation continues unchecked; the average US household, for example, spends more on
health insurance than on mortgage repayments. Profound quality and safety problems
persist—there are about 90,000 avoidable deaths a year in the United States alone.1 Many
health systems face recruitment challenges despite large pay raises for doctors, and an
increasing number of clinicians say they would advise young people against choosing careers
in medicine.2

So further change is still needed, despite years of progress in the quality of health care
around the world. This transformation will require leadership—and that leadership must
come substantially from doctors and other clinicians, whether or not they play formal
management roles. Clinicians not only make the frontline decisions that determine the
quality and efficiency of care but also have the technical knowledge to help make sound
strategic choices about longer-term patterns of service delivery.

Unfortunately, the conventional view of health care management divides treatment from
administration—doctors and nurses look after patients, while administrators look after
the organizations that treat them. But we can learn from a number of pioneering health
care institutions that have achieved outstanding performance by radically challenging this
assumption.

Our research also highlights the powerful barriers that hold back the development of
effective clinical leadership. Understanding these barriers offers pointers toward the best
ways to build clinical leadership across health systems.

Why clinical leadership matters

Consider the case of Kaiser Permanente, a large US payer and provider operating in several
states. In the late 1990s, Kaiser Permanente Colorado was struggling with worsening clinical
and financial performance and losing top doctors to private practice and rival organizations.
A new executive medical director—Jack Cochran, a pediatric plastic surgeon—made clinical
leadership an explicit force for improving outcomes for patients. Defining the role of the
clinician as “healer, leader, and partner,” he revamped Kaiser’s leadership-development
programs for doctors. Within five years, Colorado had become Kaiser’s highest-performing
affiliate on quality of care and a beacon of quality within US health care. Patients were
significantly more satisfied, staff turnover fell dramatically, and net income rose from zero
to $87 million.

The Veterans Health Administration, within the US Department of Veterans Affairs

(VA), provides another example. Established as a public-sector health system for retired
military personnel, it was performing so poorly in the mid-1990s that some prominent
voices suggested closing it down. A new CEO—again a doctor—sponsored an improvement
program in which clinical leadership played a central part. Ken Kizer reorganized the
Veterans’ Health Administration into 21 networks, each with accountable clinical leadership,
across the United States. The program also introduced clinically relevant performance

1
See Crossing the Quality Chasm, Institute of Medicine, 2001.
2
See Janice Hopkins Tanne, “US GPs are unhappy, underpaid, deluged by paperwork, and want to retire, study says,” BMJ, 2008,
Volume 337, Number 2711.
 3

measures, with corresponding rewards, and new information systems, including one for
electronic medical records. The VA soon became a leader in clinical quality: for example, the
risk of death for men over 65 in the VA’s care is 40 percent lower than the US average. The
satisfaction level of patients rose to 83 percent, 12 percent above the national average, even
as the VA’s patient numbers doubled over the following decade.

What do these and similar examples tell us about clinical leadership? Improvements
happened because clinicians (most notably doctors) played an integral part in shaping
clinical services. This expanded role did not come about through one-off projects; nor
were changes in formal job descriptions the primary driving force. Rather, what changed
for clinicians was their professional identity and sense of accountability. All staff, whether
clinicians or not, came to share a common aim: delivering excellent care efficiently. Doctors
collaborated with administrators on important clinical decisions—such as how to expand or
reconfigure services—in full knowledge of the trade-offs and resource implications.

Even more thought was given to patients and their needs—for example, clinicians not only
paid attention to clinical outcomes for patients but also further emphasized the overall
quality of the patient experience. The performance of clinical units was tracked in real
time. A lapse in safety, rather than being explained away, triggered a multidisciplinary
conversation to help learn lessons for the future. There was a sense that clinicians were,
more broadly, extending the responsibility they feel for their patients to the organization
itself.

A growing body of research supports the assertion that effective clinical leadership
lifts the performance of health care organizations. A recent study by McKinsey and the
London School of Economics,3 for example, found that hospitals with the greatest clinician
participation in management scored about 50 percent higher on important drivers of
performance than hospitals with low levels of clinical leadership did. In the United States
and elsewhere, academic studies report that high-performing medical groups typically
emphasize clinical quality, build deep relationships between clinicians and nonclinicians,
and are quick to learn new ways of working. 4 A recent study by the UK National Health
Service (NHS) found that in 11 cases of attempted improvement in services, organizations
with stronger clinical leadership were more successful,5 while another UK study found that
CEOs in the highest-performing organizations engaged clinicians in dialogue and in joint
problem-solving efforts.6

In many ways, this evidence is unsurprising. Large health care systems and providers rely
on complex and rapid decision making from thousands of people hundreds of times a day,
often with life-or-death consequences. A command-and-control approach to leadership is

3
Pedro J. Castro, Stephen J. Dorgan, and Ben Richardson, “A healthier health care system for the United Kingdom,”
mckinseyquarterly.com, February 2008. This research involved interviews with more than 170 general managers and clinical heads
of departments in the UK National Health Service. Responses covered the effectiveness of overall management and
of performance management, as well as the level and effectiveness of clinical leadership.
4
For example, Stephen M. Shortell et al., “An empirical assessment of high-performing medical groups: Results from a national
study,” Medical Care Research and Review, 2005, Volume 62, Number 4, pp. 407–34; Lawrence Casalino et al., “External
incentives, information technology, and organized processes to improve health care quality for patients with chronic diseases,”
Journal of the American Medical Association, 2003, Volume 289, Number 4, pp. 434–41.
5
Managing Change and Role Enactment in the Professionalised Organisation, National Co-ordinating Centre for NHS Service
Delivery and Organisation, 2006.
6
Enhancing Engagement in Clinical Leadership, Academy of Royal Medical Colleges and NHS Institute, 2007.
 4

The virtues of distributed leadership

Many people associate clinical leadership with highly Third, frontline leaders are great clinicians who focus
visible, formal leadership roles. These are certainly squarely on the direct delivery of patient care but
part of the equation, but our research suggests that also see continuous improvement in the way the
there are at least three distinct types of clinical leaders organization delivers care as their responsibility. If, for
(exhibit). example, clinical records repeatedly go missing from
consultations, or if patients frequently fail to show up
The most obvious are the institutional leaders: for appointments, frontline leaders take ownership
sophisticated clinical leaders who often occupy in solving the problem. To do so, they need some
formal, executive-level roles. They can communicate a awareness of systems- and quality-improvement
powerful, clinically based vision and have deep, broad techniques and must know the basics of leadership,
skills in both leadership and administration. These such as an awareness of their personal style and how
skills are both “hard,” such as strategic thinking and to work well in teams.
planning, and “soft,” such as negotiation and influence.
A typical institutional leader might be a medical director These three categories (and the format of the exhibit)
who manages services across a multisite organization, might suggest that leadership is organized into a
earning the support of colleagues by demonstrating hierarchy, but all levels are needed and none has
how change will improve quality of care. greater worth than the others; the last thing clinical
leaders should do is devalue a clinician’s core
Service leaders are the second type: passionate activity of direct patient care. Indeed, clinicians on
advocates of their own units or teams, who are different “levels” are likely to be peers, with similar
also aware of the context and requirements of the remuneration and professional status but varying
whole organization. They have detailed knowledge degrees of leadership focus and specialization. Although
of the
Webrelevant
2009 clinical evidence base and constantly institutional and service leaders have greater overall
innovate to improve
Clinical patient
leadership care. Service leaders are
sidebar responsibility, the far more numerous frontline leaders
accountable
Exhibit 1forof the
1 overall performance of the service, ultimately hold the key to realizing the organization’s
bothGlance:
clinically and financially.
Research vision
suggests that at least three distinct types by using
of clinical their day-to-day
leaders exist. experience to inform
the constant improvement of services.
Exhibit title: Three ways to lead

Exhibit Overall identity Sources of power Selected leadership

skills and knowledge
Three ways required
to lead Institutional t Clinician executive t Highly credible to t Corporate-level strategic
leader acting as steward of colleagues as clinician thinking, talent manage-
Few whole organization and leader; able to ment, succession planning
t Little direct contact communicate vision t Political savvy; strong
with patients skills in negotiation
and influence
Service t Passionate advocate t Highly credible to t Fluent service-management
leader for own service, colleagues, primarily as skills—eg, strategy/people
feels responsible for clinician; well connected, development, budgeting
its clinical and financial can tap into centers t Detailed knowledge of
performance of excellence evidence-based medicine in
t Moderate level of direct t Innovative, willing to own clinical area
contact with patients take risks
Frontline t Great frontline clinician t Passionate about t Understanding of systems-
leader who focuses on clinical work, credible and quality-improvement
delivering and improving to colleagues techniques—eg, process
Many excellent patient care t Close to patients and mapping, operational
t High level of direct frontline realities; can see improvement
contact with patients opportunities for t Self-starter, able to work
improvement well in teams
 5

untenable in such a complex and uncertain environment: it is impossible to determine, from

the top, the right decision in any given situation. Distributed leadership models reap benefits
by enabling people to make effective decisions locally, guided by the organization’s overall
aims and norms, without the need for excessive bureaucracy and top-down intervention (see
sidebar, “The virtues of distributed leadership”). In essence, the most successful health care
organizations treat all employees as potential leaders in their own spheres—and not least the
clinicians.

What stands in the way

Despite accumulating evidence of the positive impact of clinical involvement in the delivery
and improvement of service, health care organizations often struggle to achieve this kind of
participation. To understand the barriers to clinical leadership, we conducted interviews and
workshops involving nearly 100 clinical professionals. Our research highlighted three main
issues.

First, we found an ingrained skepticism among clinicians about the value of spending
time on leadership, as opposed to the evident and immediate value of treating patients.
Participants explained that playing an organizational-leadership role wasn’t seen as vital
either for patient care or their own professional success and therefore seemed irrelevant to
the self-esteem and careers of clinicians.

Moreover, many participants expressed discomfort with knowing that the impact of
clinical leadership is often hard to prove. Clinicians develop a skeptical mind-set about
changes to treatment approaches—a mind-set that is rooted in the precept, “first, do no
harm.” They also have a clear view of what constitutes robust evidence—one that is rooted
in evidence-based medicine for clinical interventions. As compared with biomedical
standards (particularly randomized controlled trials), clinicians see the study of leadership
as fundamentally ambiguous, even weak. This attitude becomes entrenched early in people’s
careers (in medical school, typically, for doctors), and there is no concerted effort to broaden
it later on.

Second, it became clear there were weak or even negative incentives for clinicians—
especially doctors—to take on service leadership roles. Leadership potential generally isn’t
a criterion for entry into the clinical professions and often isn’t a major factor in promotion.
Nor is there a well-defined and respected career path for those with an appetite for formal
leadership roles—in stark contrast with well-trodden clinical and academic tracks. Peer
recognition is low or nonexistent: those who reduce their clinical practice to take on formal
leadership roles are often described by colleagues as having “gone over to the dark side.” The
difference between leadership and research is instructive: the latter is well systematized, its
importance in clinicians’ careers is widely recognized, and the incentives to undertake it
are clear: research publications are crucial to securing the top jobs or professorships, which
carry great prestige and influence and (frequently) financial rewards.

The financial disincentives for doctors to take on organizational leadership roles is

illustrated by the NHS, where salary scales are lower for managers than for doctors and
where devoting time to leadership activities may reduce the scope for income from private
clinical practice and even jeopardize research funding. And because measurement of the
quality of care has been rudimentary, it has been impossible to reward those who build the
best services.
Third, we found little provision for the nurturing of clinical-leadership capabilities.
Organizations generally lack meaningful processes for finding, inspiring, and stretching
those clinicians who possess the greatest potential as leaders. Leadership and management
training is frequently absent from core curricula for undergraduate or postgraduate trainees
and for the continuing professional development of clinicians.

The programs that are available to clinicians are often run externally rather than in house,
making it harder to focus the development experience on the real day-to-day challenges
participants and their services face, reducing relevance, and hindering the translation
of learning into action—especially important given the lack of follow-up support in the
workplace. The biases of clinicians are at play as well: having had years of training to
perform their clinical role, many assume that months or even years of formal training are
needed before anyone can safely be let loose as a leader.

The path to leadership

Despite an explicit focus on clinical leadership across many health systems, traditional
ways of working and mind-sets become so entrenched that many health care organizations
struggle to develop clinical leaders successfully. But some straightforward measures can
yield significant results.

Shifting beliefs
Perhaps the highest barrier to the greater involvement of clinicians in shaping the future
of patient care lies in the historical beliefs of clinicians themselves about the value of
leadership and management. One way to address this problem is to be far more systematic
about gathering stories, told authentically and compellingly by those who participated or
observed, that highlight the value of great clinical leadership. By “making heroes” of clinical
leaders of all types, both in formal management and in frontline roles, organizations can
create a stronger bank of role models and also spark a sense of possibility. These stories
should highlight the benefits both to patients and to the teams delivering care—benefits
such as greater autonomy or simply the sense of pride in achievement. In Boston, for
example, Partners HealthCare celebrates distinctive clinical leaders not only at annual
award ceremonies but also day to day, through e-mail, in-house journals, and informal
conversations.

Health care organizations need to build a solid, credible evidence base to show the
importance of clinical leadership. While approaching the topic as though it were a clinical
trial is difficult, organizations should track measures of clinical-leadership development
and correlate them with their impact on quality and costs. Regional health care systems or
authorities have an influential role to play here, given their scope for pooling analysis across
a number of organizations.

Creating the right environment

To create the kind of evidence base described above, health care organizations need, at
a minimum, basic performance data from which meaningful comparisons can be made.
Sound, transparent performance information will also encourage clinicians to play a wider
role in making decisions about the best ways to care for patients and to manage resources.
Many organizations have found that circulating clinician-level performance data—whether
made anonymous or not—prompts competition among clinicians, which in turn encourages
them to become involved in improvement efforts. Generally, the health care sector lags
behind others in implementing the infrastructure and processes that could provide such
basic information. There are exceptions, however: a substantial investment in systems for
generating reliable, timely performance data underpinned the transformations described
above at both the VA and Kaiser Permanente.

Policy makers and organizations must also retune incentives—above all, by removing
egregious disincentives for clinicians to become service and system leaders; these
disincentives include paying clinicians significantly less in such roles than they would make
by remaining in full-time clinical practice. Correcting these problems is important not only
for direct financial reasons but also because of the wider signals that incentives send about
the value and prestige attached to clinical leadership. Where it flourishes, in organizations
such as Health Partners, in Minnesota, clinicians in formal leadership roles typically
receive a small premium over colleagues who focus solely on direct patient care. Too great a
financial premium, however, would make patient care less attractive and damage what ought
to be the peer-to-peer relationship between leaders and other clinicians.

As people come to appreciate the link between performance and enhanced clinical
leadership, health systems can also encourage it indirectly by finding appropriate ways
to reward organizations that perform well and by creating meaningful consequences for
those that don’t. The VA, for instance, operates on the principle of earned autonomy: high-
performing regions and organizations receive substantial freedom to operate with less
central oversight, whereas those that underperform are scrutinized closely.

Supporting real learning

Any effort to encourage clinical leadership has to include support for professional
development. But (perhaps surprisingly) the best starting point is not to create or
commission a training course. Health care organizations must first define what they want
from their clinical leaders—what skills and attitudes they hope to encourage, whether there
are differences across professions or roles, and where the need to develop leadership is
greatest. They can then target their efforts wisely and help clinicians identify and overcome
any shortcomings.

The US Army’s West Point Leadership Academy, for example, recruits, trains, and develops
leaders in accordance with the explicitly defined leadership model of the army and its
threefold “be, know, do” philosophy. From the moment new trainees arrive at West Point,
this model is emphasized, along with the need for trainees to demonstrate that it has an
ongoing influence on their development. Some health care organizations with a development
focus have made their expectations similarly explicit: Heart of England NHS Foundation
Trust, in Birmingham, UK, and New York Presbyterian Hospital have worked hard to
define their expectations of clinical leaders at different levels. This enables them to target
their development programs very precisely and to create enough leaders to meet their
organizational needs.

As with all training efforts for adults, it will be necessary to address the practical issues
clinical leaders face. There are obvious benefits to programs that are truly centered on real
work: the power of learning by doing, the importance of immediate feedback, the integration
of concept and application, and the encouragement that comes from seeing a tangible impact.
 And for clinicians, development programs with real work at their heart can help enormously
in demonstrating how patients benefit when clinicians lead the improvement of services. A
leadership program involving a dozen UK hospitals and both clinical and nonclinical staff,
for example, focused on redesigning pathways (strict treatment steps) for patients with
stroke and hip fractures. The program, positioned as a quality-improvement effort rather
than a training or development course, had a remarkable impact on lengths of stay, mortality
rates, and costs—all of which fell by up to 30 percent. It also created enthusiasm for leading
service-improvement efforts more generally, with enduring benefits after the formal
program had ended.

Related articles The most powerful clinical-leadership initiatives go even further, with integrated
development journeys tailored to the evolving needs of individuals. At Kaiser Permanente,
“A healthier health
the choice of technical skills covered in leadership programs matches the participants’
care system for
self-identified needs: for example, the head of a primary-care clinic might be trained in
the United Kingdom”
scheduling, multidisciplinary teamwork, and group visits. Physicians with particular
“Service-line strengths, such as interpersonal effectiveness, are asked to share their expertise by teaching
strategies for US colleagues. Leaders don’t receive just a single boost; a series of interventions reinforces their
hospitals” development over time, creating groups that learn together and make the link to real work.

“US hospitals For more formal leadership-development programs, health care organizations should
for the
consider introducing processes to select participants in order to underline the value of the
21st century”
programs and, more broadly, the prestige associated with being on the organizational-
leadership track. For example, Singapore’s National Institute of Education (NIE) sifts
through the whole teaching workforce to identify high-potential candidates to be future
head teachers. Entry into the head-teacher track is highly competitive, and a series of gates
determines a candidate’s subsequent progression. This approach helps signal the value the
NIE attaches to teachers who step up to become leaders.

Starting from isolated pockets of excellence and innovation, clinical leadership still has a
long road to travel. But it is an essential road for both clinicians and their patients. A deep
commitment to patient care and to traditional clinical skills will always remain the core of
a clinician’s identity. To achieve the best and most sustainable quality of care, however, a
commitment to building high-performing organizations must complement these traditional
values. All the evidence suggests that patients will see the benefit. Q

The authors would like to thank the many clinicians and others in various health systems whose
insights have contributed to this work. Many colleagues have also participated, especially Penny Dash, John Drew, Nicolaus
Henke, Chris Llewellyn, and Ben Richardson.
•

James Mountford is a consultant in McKinsey’s London office, where Caroline Webb is a principal.
Copyright © 2009 McKinsey & Company. All rights reserved.