The importance of good leadership in producing what is required of an organisation is accepted unquestionably, from football teams to global enterprises. Its key role within health care has been a rather slower development but has recently been acknowledged as vital,¹ particularly now in the development of quality care through clinical governance. However, rhetoric does not ensure good care, so this paper looks at the evidence for why leadership matters, what various theories of leadership might tell us, and what the implications of this are for policy and development.

Leadership exists at every level throughout an organisation,² and usually includes management tasks. Although there is often a distinction made between leadership and management³—management being seen as the seeking of order and stability while leadership is about seeking adaptive and constructive change—leaders are also likely to want to be able to produce and manage periods of stability, often at the same time as planning future changes. For this reason, the two inevitably overlap.

In the case of health care there is also a longstanding distinction between leadership in the person of the chief executive (or his or her directors) and team or clinical leadership in the persons of uniprofessional groups such as doctors or nurses and multidisciplinary teams which are the face of care of which the patient is most aware. Most of this discussion is applicable to both types of leaders, but it should also be borne in mind that not all of leadership theory sits comfortably in clinical settings in particular, nor always in health care as a whole.

Research into leadership

Leadership research is wide ranging, covering the personality or behaviour of the leader, the context in which leadership takes place, and the people who are led. It is only through the combination of all three that we can begin to appreciate the true complexity of what it means to lead any group of people successfully. Even these traditional areas neglect the emotional experience of the leaders themselves.⁴ As Sarason explained: “When leaders talk of the experience of leadership they talk more about the role of that person, far more about duties and responsibilities than about the maelstrom of feelings, fantasies, ambitions, conflicts, guilt and joys that are always in the picture . . . they give us a job description and not a personal experience.”⁵ No research design can possibly capture such complexity and any advances in leadership theory must involve a consideration of the leader's characteristics, behaviours, and the situation simultaneously.⁶

Personality, leadership and quality

While some leaders emerge through their own influence and the support or acknowledgment of staff around them, others are assigned leadership roles. In a study of male college students⁷ emergent leaders were shown to be more dominant, intelligent, and confident about their own performance and were also identified more often as leaders by others in the group. Short term leaders such as these may produce short term goals, but not longer term ones which involve carrying their staff with them along the way. Nevertheless, in terms of actual outcomes from different types of personality, these findings show a core of agreement with others which look at the performance that emerges from different leadership characteristics. For example, research with airline crews found that error levels were lowest where the leaders were warm, friendly, self-confident, and able to stand up to pressure,⁸ what proponents of the Big Five personality constructs used for job selection would label as “agreeableness” and “emotional stability”.⁹ Other studies which have looked at “great leaders” have come up with a similar collection of characteristics including intelligence, self-confidence, determination, integrity, and sociability.¹⁰

These findings coincide closely with the man in the street's picture of the strong, extrovert, confident leader. However, a recent UK study of health service leadership, in which staff and leaders were themselves asked what makes a good leader, concluded that staff wanted leaders who could do the best for them, what they called “the model of leader as servant”.¹¹ The other important abilities which emerged in this study were about engaging others as partners in the way forward and forming an environment for creative thinking—all very much about the ability to provide staff with discretion and control wherever possible, characteristics which incidentally have also been shown to be very important in terms of lowering staff stress¹² and in creating a sense of justice in staff which leads to higher satisfaction with their leaders.¹³ Within health services there is another type of leader who is thought to be important in terms of quality. These people are the “opinion leaders” who have been shown to be influential in bringing about change in terms of evidence based health care.¹⁴ As in the theory of charismatic leadership, they provide strong role models for the beliefs and values they wish others to adopt, appear competent to those being led, and articulate ideological goals.¹⁵ Within the clinical setting, such leadership skills are very valuable so long as they are appropriately in line with the aims of the organisation as a whole.

Not all systems of personality classification have clear implications for the quality of care produced. For example, in terms of Myers-Brigg personality types, any type is said to be able to lead well, if differently. The Myers-Brigg type inventory (MBTI)¹⁶ is based upon Jung's personality types, dispositions which predispose people in various ways in terms of where they get their strength (Es or extroverts get it from others, Is or introverts from within themselves); how they gather their data about the world (Ss or sensors from the here and now, Ns or intuiters from what they can create from it for the future); how they make their decisions (Fs or feelers from person-based values, Ts or thinkers from rational logical processes); and whether they prefer the process of gathering the data (Ps or perceivers) to that of making the decisions (Js or judgers)* It is important that leaders should recognise their own strengths and weaknesses that result from their particular type, develop their own “least preferred areas”, and recognise and reward the different skills that are offered by others in the team or by different sections of the organisation. Although chief executives are commonly ENTJs,¹⁸ perhaps because it conforms best to the Big Five characteristics described above, some of our most famous leaders of British industry have, in fact, been very different types. A preponderance of different types in different sections of an organisation—for example, chief executives who are predominantly extrovert, consultants who are mainly introvert,¹⁹ nurses who are mainly Fs, making their decisions by their value systems rather than by the logic and rationality used by their T colleagues, the doctors and managers—will affect communication, persuasion, and decision making and this must be taken into account. It means perhaps that a principal attribute for good leadership is to use the skills around you well.

Different types of leaders can certainly affect decisions reached. One recent study tested the ways various leaders affected the level of groupthink in a team.²⁰ Groupthink occurs where teams close up against outside messages and strive prematurely for unanimous agreement on a course of action.²¹ The study showed that those groups with “promotional leaders”—who promoted their own preferred solutions—produced more groupthink, discussed few facts, and reached a decision more quickly than groups with non-promotional leaders. Again, this has important implications for both clinical and management teams in health care where multidisciplinary team working and decision making are so important.

Other key factors in the new quality agenda²² are the recognition and reporting of errors and learning from these as individuals, teams, and organisations.²³ There are some important studies which show how the personality of the leader might affect this process. As described above, there is evidence from research with airline crews to demonstrate the effects of a leader's personality on the errors made by the team. We have discussed the warm, friendly, and self-confident leaders whose teams produced fewer errors, but the research also showed that error levels are higher where the captains are characterised by arrogance, hostility, boastfulness, or being dictatorial.⁸ It therefore seems that leaders are able directly to affect the safety of their teams' actions and outcomes—an extremely important finding for patient care. However, this research was carried out in the automated environment of the flight deck where each error was recorded unchangeably. In health care such situations are rare and we depend much more upon the honesty, accuracy, and memories of staff to record what goes wrong and so learn from it. Edmondson²⁴ has shown that the personality of the leader plays a part in the number of errors recorded. In a study of medication errors and quality of teamwork in nursing she was surprised to find that good teams recorded more errors than poor teams. To explore this further she interviewed the team leaders themselves and found that the authoritarian dictatorial leaders led the poorer teams who reported, perhaps not surprisingly, fewer errors. This shows the importance of not taking healthcare data at face value: repressive dictatorial regimes are almost bound to produce data which are less than accurate. Team leadership should perhaps be judged more by demonstrating how good their teams are at detecting errors and learning from them, rather than by simply recording and reporting the number of errors produced.

Leadership style and quality

One of the most enduring distinctions in leadership research is that between the two styles of transactional and transformational leadership. Transactional leadership stems from a traditional view of the leader having power and authority over followers, and the use of power to achieve goals and objectives.²⁵ In MBTI terms, transactional leaders are more likely to be Ss, focusing primarily on the details of the “here and now”.²⁶ The principal components of this style are contingent rewards for staff and management by exception, which involves a focus on problems and mistakes. Some might say that the UK government's management of the health service very much fits this mode of leadership, and that it inevitably flows down through commissioners to most healthcare organisations and even clinical teams. It is perhaps least common within the primary care culture.

Transformational leadership looks for ways to motivate followers with a view to engaging them more intimately in the process of work—it is “performance beyond expectations”.²⁷ Transformational leaders can initiate and cope with change, create something new from something old. In Myers-Brigg terms they are usually the Ns—keen to go beyond the evidence to build something new. They are entrepreneurial, take risks, and are often informal in their relationships, always seeking to develop individuals and respond to their needs and interests. Such a style—a complex mix of personality and behaviour and changing context—sounds quite similar to what NHS staff said they wanted in a leader.¹¹ However, Alimo-Metcalfe and Alban-Metcalfe found that staff needs were for a combination of the transformational and transactional—perhaps for both the ability to lead change but also to hold things stable in ways such as the use of reward and the focus on error that are recognisable as part of our early socialisation. Certainly, research has failed to show conclusively that transformational leaders are always best,²⁸ and perhaps leaders should recognise their own style and partner it as much as possible with those with other skills.

Stress, leadership and quality

There are theoretical and even common sense reasons why and how leaders might affect the final product that is delivered by their organisations or their teams—in this case, patient care. There is also a small but growing area of research findings which confirm some of these reasons but also provide evidence for a pathway of care from the leader through the staff to the patients themselves.

Working back from the professional-patient interface, we have evidence first that stressed staff produce inferior care.²⁹ Evidence has come, for example, from cognitive testing, looking at the effects of fatigue on decision making,³⁰ or from looking at the relationships between symptoms of stress, insomnia, and errors and how this strengthens over time as junior doctors begin a new post.³¹ In one study of the effects of work stress on care, young doctors described events from general carelessness through to errors contributing to patient deaths which they attributed primarily to their exhaustion, overwork, lack of support, or the symptoms of depression³²; 40% described becoming irritable or even abusive to patients and to colleagues, showing that the effects spread into care but also around the team itself. In addition, the largest patient satisfaction survey ever conducted showed that the highest correlations were with the cheerfulness, friendliness, and sensitivity of staff.³³ Clearly, the psychological wellbeing of staff intimately affects the quality of patient care in a variety of ways. Moreover, stress in health service staff, at least in the National Health Service, is considerably higher than other members of the workforce, with about 28% showing levels above the threshold for symptoms.³⁴

Tackling job stress is clearly a crucial step in producing better care, so we need to study its causes carefully. In this regard there is growing evidence that team functioning is an important predictor of stress levels.³⁵³⁶ In a study of NHS health staff in 19 organisations, those who reported being in no team had the highest stress levels, followed by those in teams with inferior functioning, with those in well functioning teams having the least symptoms.³⁶ A well functioning team will be able to provide good support to each other and step in to help when this is seen to be necessary.²³³⁷ However, meta-analytical studies have also shown that the principal cause of stress in the work place is “the boss”,³⁸ and so it is reasonable to suppose that good leadership produces good teams with low stress and better patient care.

One way that this might happen is through the perceptiveness of the leader to the needs and views of the staff. In a recent study,³⁵ the authors measured attitudes to work and stress levels in house officers and also the consultants' views of their house officers' attitudes to work. The “gap” between the consultants' views and the actual views of their staff was highly related to staff stress levels. In addition, it was a large predictor of whether the team was functioning well or not. In other words, the team leader's skill in accurately recognising the views of his or her staff members was an important factor in their wellbeing and in the general working of the team. Similarly, in a large study of non-health organisations³⁹ the “gap” between perceptions of risk in the chief executive and perceptions of risk on the shop floor again demonstrated that the larger the gap between the perceptions of leaders and their staff, the more errors were being made. These studies show a clear path back through patient care to staff stress to team leadership.

One recent study has looked at these links directly. Corrigan et al⁴⁰ measured leadership style in 31 mental health teams and asked their clients to rate their satisfaction with the treatment programmes and their quality of life. Both these factors were inversely associated with laissez-faire leadership styles and positively associated with both transformational and transactional leadership. The ratings by the leaders and the subordinates of their team leadership independently accounted for 40% of the total variance in client satisfaction. In addition, leadership style in terms of personality can have real consequences for patient safety, as mentioned above with research in airline pilots,⁸ and with differences in staff accurately reporting error.²⁴

This section has shown that one important way in which leaders affect patient care and satisfaction is through their management of teams and its effect upon the levels of stress experienced by the team members. Although most evidence comes from the team level, it is likely that this is also true for whole organisations and their chief executives. It implies again that there are important ways of assessing leaders other than by meeting assigned objectives. Rather, we can judge by the wellbeing of their staff in terms of absence, turnover, and disruptive behaviours.⁴¹

Conflicts of leadership for quality

Transformational leadership continues to be the style that is presumed best for health services, primarily because it has a focus on change. However, it is not always easy to achieve when, in reality, most change is imposed upon leaders and its expected outcomes are detailed. Alongside this are the very transactional methods of performance monitoring, clinical audit, re-accreditation, controls assurance, central error reporting, league tables, and so on—all aspects of quality which are laudable and probably valuable in themselves but which are likely to influence leaders to take a much more transactional style than they might otherwise have done. In addition, the more centralised such controls become, the less a leader is able to give staff the discretion and participation which they desire.¹¹ This is a central conflict in the agenda for better health services—to make staff accountable while allowing their creativity and participation to flourish. Leaders will best achieve it, we suggest, by using local participation to develop systems of strong accountability⁴² but, like all these conflicts, tension is always likely to remain.

The second clear conflict for health service leaders is between quality and efficiency. One characteristic of good leadership which we might note from other industries is the consistency of the message which is essential for building trust. Without it a leader is swiftly paid instead with cynicism⁴³ and quality is bound to suffer. Binney and Williams⁴⁴ quote a senior manager from Nissan: “To achieve real quality everyone in the organisation has to genuinely believe it and act on the belief. Management must mean what it says. As soon as a senior manager lets a car go through which is not the right quality level `because I have to meet the schedule' the battle is lost.” Nevertheless, the reality for many healthcare leaders is one of often desperate choices—for example, of being an inner city general practitioner with a vital need of a partner, faced only with the application of a known but not entirely trusted locum.⁴⁵ It is not always a question of money, but of whom you can get to do a difficult job. We are not suggesting that quality should be compromised, but in situations like this it is likely to be so, whatever choice is made. This reality of inadequate resources needs facing as part of health leadership training, rather than avoiding.

The final conflict is again more fundamental to health services than to other organisations. Because of the dark side which is always present in health care (disease, distress, disability, and death), there will always be a natural tendency for non-clinical leaders to avoid an appreciation of what is actually happening on the front line: that care is sometimes not good enough, mistakes do take place, that people—staff and patients alike—suffer as a result. As we have seen from the “gap” studies, the difference in perceptions caused by this avoidance can affect both staff morale and probably also care. Leaders need to get close to patients' and staff's experiences—to do what has been called “the walk of shame”⁴⁷—but they need training and support to face the unacceptable in health care in ways which are different from other types of organisations.

These conflicts are unlikely to fade or to be resolved. It is important that leadership training and development takes them into account and helps people to tackle them daily rather than using models from a culture less complex than that faced by leaders in health care.

Conclusions

We have shown that leadership skills can have real benefits to patient care, and have described some of the behaviours and characteristics that might underpin them. It seems clear that certain traits such as arrogance, authoritarianism, and strong competitiveness may be prejudicial to good leadership, and that sociable, confident people who work well under stress have a head start in making good leaders. In terms of whether their style should be transformational or transactional, it appears from current evidence that both are going to be necessary for health services. If leaders feel more in tune with one approach than the other, it may be more important that they ensure that others are playing the role they find more difficult rather than presume they must do both themselves.

Just as important is the suggestion that any assessment of good leadership needs to go beyond performance monitoring and to look at the effects on staff wellbeing, the ways in which staff are used and developed to enhance their strengths, and the ways that leaders can show they are able to recognise and learn from the errors and inadequacies which will always be a part of health care.