Article Text

Download PDFPDF

Improvement leaders: what do they and should they do? A summary of a review of research
  1. John Ovretveit
  1. Health Innovation and Evaluation, Medical Management Centre, The Karolinska Institutet, Stockholm, Sweden
  1. Correspondence to Dr John Ovretveit, Medical Management Centre (MMC) Floor 5, Berzelius väg 3, Karolinska Institutet, Stockholm SE-171 77, Sweden; jovret{at}


Background There is research evidence to support the growing recognition that leadership influences successful improvement; also that it is one factor explaining slow, partial or failed improvement. However, the evidence is not strong, especially about how important one or more leaders actions were compared with other situational factors.

Methods This paper summarises the evidence found in a review of research into different leader's roles in quality improvement.

Results Actions suited to the situation and type of improvement appear to be the most successful, and this has implications for developing leaders to ‘fit’ their actions to their situation and the improvement, and for research to help them do this better.

Conclusions The full review lists other practical implications for leaders where there is good evidence, and notes other literature which could provide guidance for leaders in the absence of research. It also considers the limitations of the research, and specific subjects for future research where knowledge and practical guidance for leaders are especially needed.

  • Leader
  • quality
  • safety
  • improvement
  • healthcare
  • healthcare quality
  • leadership
  • management
  • patient safety
  • research

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


How to lead improvement? This is a question many health leaders and management teams are asking. Does research give an answer? A review of research was carried out in 2009 for The UK Health Foundation, which drew on an earlier review for the Swedish association of county councils.1–3


The method for searching and synthesising the research was one used for previous management and public health research reviews.3 4 This differs in some respects from the method for systematic reviews of clinical research in involving an initial scan to refine the terms and questions, and in the criteria for strength of evidence for selecting studies for review. An initial scan was made of studies shown on the Social Sciences Citation Database databases, CINAHL, Embase, Medline, Pubmed and some quality improvement journals. The aim was to identify how leader, leadership, quality improvement and related terms were defined, as well as previous research reviews or comprehensive reports on the subject. It also aimed to discover how the links between leadership, improvement actions and outcomes had been conceptualised and studied.

This initial scan found little empirical research, that the terms were used in different ways, and no other systematic reviews of research into leadership for improvement. The initial questions were then revised slightly to six questions, three of which are addressed in summaries in this paper: (1) What is improvement leadership? (2) Can leaders influence improvement? (3) Which individual leader actions are associated with successful and unsuccessful improvement?

Definitions of terms were created which reflected their use in the literature. Abstracts and papers were selected according to whether empirical research about the leader's actions in relation to improvement was reported. They were assessed in terms of strength of evidence using a grading scale based on those used in previous reviews of management research, where randomised controlled or even experimental trials are rare, or disputed as valid for certain questions.4 5 The assessments of strength of evidence also related to whether leadership for improvement was the focus of the study, or whether the study referred to this as one finding, or whether there were conclusions for improvement leaders which could be derived from the study. After assessment, some studies which did not meet the ‘strength of evidence’ criteria were retained because no other evidence was available and/or they were highly relevant to the questions. The methods are described more fully in the reports.1–3

What is improvement leadership?

Research studies use differing definitions, or do not define the terms precisely. Many definitions of ‘leadership’ and ‘leader’ were found, some which overlap with or differentiate the concept from ‘management.’6 Sometimes ‘leading’ is used to refer to actions by a leader or a group, and ‘leadership’ to attributes of a leader.

Most empirical research has been carried out into senior leaders or physician leaders, and most studies focus on the individual leader's competencies or actions. However, a small number of studies conceptualise leadership as a relationship or a collective activity. The origins of the term ‘shared leadership’ are unclear, but this and related terms became more prevalent in the leadership research literature in the late 1990s. The concept is also sometimes referred to as ‘collective,’ ‘distributed’ or ‘blended’ leadership.7–9

The search also found the concept of ‘improvement’ was defined in different ways. Originally emphasising more limited quality and safety improvement, it has been extended to include reducing waits and increasing access, and, broader still, to cover cost savings and productivity improvement.10 ‘Improvement’ in the literature refers both to an activity and to an outcome: to using methods and systems to make care better and less wasteful, as well as to the end result of a service which is better, as judged by a particular stakeholder. The concept overlaps with ‘change’ and raises the question as to whether leading all or certain types of improvement is different to leading other changes, especially with the expansion of the changes covered by the term ‘improvement.’

Reducing waiting times or infection rates are examples of a direct ‘first-level improvement.’ Research suggests that, for such changes to be made and sustained, other more general changes to systems and culture need to be made to support care providers to make change, termed here generic support ‘second level improvements’ made to the organisation.10 Different leader actions may be needed for a ‘direct’ improvement to those for a ‘support’ improvement, although there is little evidence showing this.

Little research has studied leadership for improvement, especially in healthcare. A definition of ‘leading improvement’ was formulated, and a model is given in the report which focuses on this link:

The ability appropriately to apply leadership improvement principles and to adapt to a specific situation the actions found by other leaders to be successful for improvement…

Starting and enabling improvement: influencing yourself and others to do something you and they would not otherwise do to improve the care of patients and reduce waste.3

Can leaders influence improvement?

There are different implicit or explicit conceptions in the research of what would be evidence of ‘success’ in leading improvement, sometimes reflecting different stakeholder perceptions. The most common is the same as the common criteria of ‘success in improvement’: better outcomes—for patients and/or lower costs; or intermediate achievements such as process or organisational changes which are thought to result in better outcomes.

Some evidence reported is negative: inaction or ‘the wrong actions’ is a common factor associated with quality failure. Some is positive, where certain actions were found to be associated with ‘success.’ But in both, certainty of attribution and proof of causality was weak in all the studies largely because causal models were not theorised. Investigations into organisational failure indicate that lack of action or certain types of action is a factor in tragedies or poor performance.11–15 The search found two studies reporting that CEOs had limited influence in large hospitals.16 17 It is possible that leadership may be less important for some types of improvement, such as re-engineering or larger more complex changes which may depend more on factors other than leaders actions in the service in question.18

Popular literature and some leaders attribute successful improvement to the leader's actions, often to a leader's ability to motivate and influence others to work on improvements using effective methods and/or implement changes proven elsewhere. More research is finding that the leader's influence is less important for achieving improvement than a number of other conditions which are needed to support those working on the improvements.19 20 These ‘context factors’ in implementation are beginning to be studied,21 but there is little evidence of their influence or how much the leader's actions influence improvement compared with other factors. Factors reported as likely to be important are good data support, expertise in using different methods, physicians attitude towards improvement, regulatory requirements and resources to ‘back fill’ to cover the personnel time taken from clinical practice. Many of these ‘supporting conditions’ at the local level are the results of leaders actions at higher levels. The leader's ability to lead appropriate adaption of the improvement content and implementation to the context may be one critical factor in successful improvement.

Overall, the review found many publications stressing the importance of leadership, but that only a few studies provide observational evidence to support this view. There are limitations to the scientific quality of the few studies which have examined leaders' influence on improvement: the associations are weak, even to intermediate effects such as completion of projects, and success is often defined subjectively and over the short term. No studies have rigorously tested the proposition that leaders are the main influence on improvement in healthcare. None provide causal models or theories which are research evidence-based rather than experience-based (consultant) of exactly how leaders influence improvement (the ‘mechanisms’ or ‘pathways’ of influence).

Which specific leader actions have been found to be associated with successful and unsuccessful improvement?

The cumulative weight of evidence of similar findings from observational studies is that senior leaders do have an important role in quality and safety improvement. However, the evidence from individual studies is weak, with some disconfirming evidence from some studies.16 17 There are different views about which types of research methods are best for answering this question, and about the extent of attribution or generalisation which are possible with different methods.5 21 One view is not only that the impact of the leader and what they should do depends on the context, but also that there is a complex interaction between the leader and their context which should not be separated and that leader and context influences affect the improvement actions of others and ultimately a change in patient outcomes.

There is evidence that leaders are more likely to be successful if they choose strategically significant improvements which are amenable to improvement interventions, skilfully adapt the methods for the situation, and persistently follow and revise the programme.22 23 There is far less evidence about operational managers actions in improvement, and none which gives good evidence of the effects of their actions.

It is possible that, in some situations, actions described as ‘directive’ may be needed, where the leader decides which improvements to make, and oversees the work. This can be modified with consultation and empowerment. In other situations, a more effective approach may be for a leader to start a participatory and shared process suited to the situation where different people can lead on different subjects at different times. This is one version of a more collective approach to leadership, which can be more challenging for the leader, not least because it requires both clarity about who does what, and flexibility in tasks and roles. The question of which actions are more effective in different organisational or regional cultures has not been systematically examined, but anecdotal evidence suggests considerable differences between the USA, Eastern Europe, the Middle East and Asia.24

Empirical research shows that ‘engaging’ doctors is essential to quality improvement,25 26 as well as some evidence of ‘successful involvement’ and the role of senior leaders in ‘winning involvement.’27 Identifying and influencing ‘opinion leaders’ to promote quality improvement appears to be one successful way to gain involvement,28 but other actions are needed such as providing time, resources, data, evidence of results and incentives.

The search did not find any systematic empirical studies into nurse leadership for improvement with evidence strong enough to formulate evidence-based actions for different nurse leaders. Indeed, few detailed research descriptions of nurse leader actions for improvement have been published. The differences, if any, between professions in the leader actions which are successful for leading improvement have not been systematically studied.

Some findings from the review3

  • Improving a service is now an essential part of running it.

  • Leaders at all levels ‘lead improvement’ in some way but could be more effective if they used recent knowledge about the subject.

  • What a leader can achieve depends in part on the context created by higher-level leaders, such as resources delegated for improvement or computer systems supporting improvement.

  • Evidence from research can help a leader to enable people to contribute to planning and implementing improvements, and to work through differences to find creative solutions.

  • The findings of one study about effective leadership may not apply in another setting.

  • Research shows that many start an improvement, fewer finish successfully, and even fewer develop a shared and sustained capacity in their service to make and lead improvement.

  • A focus on implementation—of proven effective changes and of organisational capacity to improve—appears to be important in leading improvement.


One finding from this review of the research is not a surprise: the absence of leadership is related to poor quality and safety performance, and the widespread view that leadership is important for making improvement. But how important compared with other factors, and what exactly a leader should do is less certain. There are no evidence-based models of exactly which leaders' actions are effective for enabling others to carry out improvement activities in one or more situations, and how this relates to clinical and cost outcomes. There is little evidence about which situational factors do or should influence a leader's actions, or about how leader actions differ by level and type of leader role.

However, research has found that people in many different positions lead improvement, not only formal leaders. Also, that ‘heroic’ individualistic leadership approaches are not associated with improvement in healthcare in the UK,29 and possibly in some other cultures and health systems. What may be successful are leaders' actions which galvanise the sustained effort of other leaders and personnel to work on improvement (‘alignment’) or a ‘collective leadership’ approach. Which actions these are appears to depend on the situation and whether the leader is ‘selling, starting, spreading or sustaining’ one improvement or an improvement strategy.



  • Linked articles 043745.

  • Funding The Health Foundation, UK.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles