Evaluating the evidence on employee engagement and its potential benefits to NHS staff: a narrative synthesis of the literature

Article history

The research reported in this issue of the journal was funded by the HS&DR programme or one of its preceding programmes as project number 12/5004/01. The contractual start date was in June 2013. The final report began editorial review in March 2014 and was accepted for publication in November 2014. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HS&DR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

none

Copyright statement

© Queen’s Printer and Controller of HMSO 2015. This work was produced by Bailey et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Context for the evidence synthesis

Employee engagement has been a topic of growing significance in recent years, bolstered in the UK by the work of the ‘Engage for Success’ movement, which has asserted that there is evidence of a link between high levels of staff engagement, organisational performance and individual well-being, as well as lowered rates of absenteeism and intent to quit. 1,2 This association was also underlined by Dame Carol Black in her 2008 report to the UK government, Working for a Healthier Tomorrow, in which she argues that features of job design, management and leadership are linked to the health of the workforce. 3

Academics have similarly argued that a range of positive organisational outcomes are associated with high engagement levels, such as improved performance,4 productivity,5 customer service6 and organisational citizenship behaviour (OCB),7 as well as positive individual outcomes such as well-being,8 reduced sickness absence9 and reduced intent to quit. 10

Engagement has grown in significance to the extent that it has been identified by the UK’s Chartered Institute of Personnel and Development as one of the core professional competencies for human resource management (HRM) practitioners, and is frequently cited as being one of the key challenges facing the HRM profession.

Within the NHS, engagement has come increasingly to the fore, with the establishment of a ‘Staff Engagement Policy Group’ at the Department of Health in 2008, the creation of a staff engagement indicator within the annual NHS Staff Survey in 2011 and the development of a range of resources on engagement by NHS Employers. 11 Sir David Nicholson, Chief Executive of the NHS in England, has been a member of the Sponsor Group supporting the work of the current Engage for Success Taskforce. The 2013 Francis Report12 indicated the potential risks of low engagement levels within the NHS and concluded that the NHS needs to foster a culture where the patient is put first, and staff are fully engaged.

However, the 2012 NHS Staff Survey results suggest that, although the staff experience is very positive in some respects, there is also cause for concern. 11 For example, only 26% said senior managers acted on staff feedback, 35% felt that communication between senior managers and staff was effective and 40% felt that their trust valued their work, while 38% reported feeling unwell as a result of work-related stress. 11 All of these factors have been found in academic research to be linked with levels of engagement. 8,10,13 Furthermore, 55% would recommend their organisation as a place to work; although an improvement on 2011 and comparable with findings elsewhere,10 this means that a large proportion of employees still do not feel positive enough about their employers to recommend them. Despite a growing demand for resources and advice on engagement within the NHS, there has hitherto been no systematic evidence synthesis that summarises the findings of research on engagement and shows how these may be relevant for developing and embedding engagement strategies in a NHS context. The purpose of this report is to address this overarching question and to provide a synthesis of the evidence relating to engagement, both within the workforce as a whole and within health contexts in particular.

This task is by no means clear-cut. There is a great deal of uncertainty over what engagement means, and its theoretical underpinnings. For instance, MacLeod and Clarke1 found over 50 different definitions of engagement while preparing their Engaging for Success report, and academics frequently refer to the definitional complexity of the field. 14–16 Definitions drawn from the practitioner domain tend to focus on engagement as an active verb, ‘engaging’, and highlight the notion that employee engagement is something done to employees to ensure they ‘buy in’ to the organisation’s overarching goals and values, often with the expectation that, if employees are engaged, then they will want to give something back to their employer. 1 This conceptualisation is closely linked to the more established constructs of involvement and participation: ‘doing engagement’. 17

However, this conceptualisation of engagement is not necessarily aligned with the development of the field within the academic literature. 18–20 Here, the construct of employee engagement was first introduced by Kahn21 to signify the authentic expression of self in role, involving physical, cognitive and emotional dimensions, and Kahn’s work has heavily influenced subsequent writings. 4,8,10,22 Engagement is thus considered within the organisational psychology field to be a multifactorial behavioural, attitudinal and affective individual differences variable,23–25 ‘being engaged’. 17 More recently, attention has turned to the topic of engagement from a critical HRM and organisational sociological perspective,26 raising new and as yet unanswered questions about the ontological status of engagement.

Linked to this, there is also considerable debate over the factors deemed to drive up levels of engagement, and the evidence is not so clear-cut as advice in the management literature would suggest. Academic research has suggested that a very wide range of factors at the levels of the individual, the job, the line manager and the employer may all be relevant. 27 These include, for instance, aspects of job design such as autonomy, meaningfulness and person–job fit4,21 and aspects of organisational climate such as voice and value congruence. 4,10 Specifically within the context of health-care workers, experiences of negative affect within the context of the job demands–resources (JD-R) model have been shown in one study to impact on engagement outcomes,28 while research by the Institute for Employment Studies found that the key drivers of engagement were staff perceptions of feeling valued by and involved with the organisation. 13

Equally important is an understanding of the underlying process by which engagement is thought to operate, and the theoretical frameworks that may be especially relevant. A number of theories have been proposed that might ‘explain’ how engagement works. For example, psychological traits such as perceived self-efficacy and a proactive approach to work, together with positive affect, are argued to generate an energetic, enthusiastic and engaged state. 29 Job design theory has also been found to be relevant, since for instance Kahn’s21 theory of engagement is rooted in Hackman and Oldham’s30 proposal that job characteristics drive attitudes and behaviour. Bakker and Demerouti31 also argue that the JD-R model demonstrates how job design can generate engaged states. However, there is as yet no agreed theoretical framework that may be of particular relevance in explaining engagement within the NHS context.

Bearing in mind these gaps in knowledge, the purpose of this evidence synthesis is to systematically bring together the research and evidence on engagement that is relevant in the health sector, in order to provide a thorough grounding for the development of a set of practice guides and materials that will be of direct, practical benefit to NHS managers and organisations. As Briner et al. 32 argue, ‘a synthesis of evidence from multiple studies is better than evidence from a single study . . . it is the collective body of evidence we need to understand’ (p. 24). It is therefore hoped that assembling evidence from a wide range of studies into engagement will bring about a more nuanced understanding of what engagement is, and how it works.

Review aim, scope and questions

The aim of this report is to present the results of a systematic evidence synthesis on engagement. Specifically, there are four research questions:

1. How has employee engagement been defined, modelled and operationalised within the academic literature?

2. What evidence is there that engagement is relevant for staff morale and performance?

3. What approaches and interventions have the greatest potential to create and embed high levels of engagement within the NHS?

4. What tools and resources would be most useful to NHS managers in order to improve engagement?

Thus, the first aim is to examine the ways in which engagement is defined and measured within the academic literature. The second is to examine the nature and quality of the evidence available that links engagement with morale and performance outcomes through a systematic review of the literature. The third is to examine the research findings that purport to demonstrate the antecedent factors to engagement. Based on the first three questions, the final research question concerns identifying other resources and evidence (‘grey literature’) that are of practical relevance to practitioners in the NHS. The results of this question are addressed through the production of a series of practitioner outputs provided in the appendices to this report. The main part of the report provides evidence from a systematic evidence synthesis on engagement. A core aspect of the evidence synthesis is to critically evaluate the quality of evidence currently available from a variety of sources in order to ensure that the report and other outputs from the study are based on best evidence. A problem that we have faced in the preparation of this report has been the wide variety of terms used to refer to employee engagement. These include ‘work engagement’, ‘personal engagement’, ‘job engagement’, ‘task engagement’, ‘organisational engagement’ and ‘employee engagement’. For simplicity, we have tended to use the term ‘engagement’ throughout.

Structure of the report

Following this introduction, Chapter 2 describes the rationale underpinning the methodology for the evidence synthesis, and details the stages of the process of piloting and refining search terms, searching for studies, sifting studies against inclusion and exclusion criteria, and extracting and synthesising data.

Chapter 3 addresses the question: ‘What is engagement?’ Engagement is a contested term that has been defined and operationalised in many different ways. 16 In this chapter, we provide an overview of definitions and measures used within the academic literature, and evaluate the areas of both strength and concern. We also present the major theoretical frameworks used to explain the engagement process, and report on the occurrence of both measures and theories within the selected studies. The chapter concludes with some consideration of how engagement as a construct relates to the wider field, and an evaluation of its construct and discriminant validity.

In Chapter 4, we examine the results of the evidence synthesis relating to the link between engagement and morale, and in Chapter 5 we examine the results relating to the association between engagement and performance outcomes. Chapter 6 focuses on the antecedents of engagement, and evaluates the strength of the available evidence concerning approaches within the workplace that can create and embed high levels of engagement.

In Chapter 7, we bring together the evidence presented in the earlier chapters and synthesise the overarching themes emerging from the review of the literature. We highlight areas of strength within the extant literature, as well as areas where further development is required. We present the overall conclusions based on our evidence synthesis, indicate the implications for policy and practice, and make recommendations for future research.

Introduction

This chapter outlines the methodological approach to the evidence synthesis. This commences with an examination of the engagement literature from a methodological perspective and is followed by an explanation of the rationale for the use of a narrative approach to evidence synthesis. The remainder of the chapter then details the specific methods used at each stage of the synthesis, explaining how the search terms and strategy were developed and the data were produced. The chapter also explains the methods used to review the grey literature sources in conjunction with the main data collection and analysis process. The grey literature was evaluated for its relevance to the evidence synthesis report, and for inclusion in the practitioner outputs arising from this project that are detailed in this chapter.

The engagement literature and evidence synthesis

Evidence review methodology

Briner46 sets out the process whereby a systematic evidence review is conducted according to these core principles within the field of management. He suggests it is a process that should be moulded around the issues and review questions, but it is not expected to proceed in a linear fashion. Systematic review is a method of choice because it can be ‘applied or modified depending on the questions being asked’ (p. 21). 45 Nonetheless, the first principle of organisation means that a systematic approach must be taken in which the basis of all decisions about quality, relevance and credibility is clearly defined, alongside the outcomes of those decisions. To achieve this, Briner46 sets out five stages to the review process:

1. planning, which includes developing the research questions

2. locating studies through a structured search

3. evaluating identified material against eligibility criteria for inclusion/exclusion as evidence

4. analysis and thematic coding (data extraction)

5. reporting.

We have set out below how these stages were applied in this project.

Evaluating material against eligibility criteria for inclusion/exclusion

The quality of any evidence review depends almost entirely on the quality of included studies. 55 Before any data can be extracted from the studies, it is therefore crucial to assess each one using clear and explicit inclusion and exclusion criteria in order to evaluate the relevance and quality of each contribution. 46,49 This process should be free from bias and as replicable and systematic as possible. 45 Korhonen et al. 55 state that this evaluation should be carried out as transparently as possible, as this is ‘a key requirement for the reliability of the synthesis and transferability of the results, as well as for the identification of theoretical possibilities’ (p. 1030). We critically evaluated all the studies in two phases: (1) sifting the abstracts of all identified material against a series of inclusion criteria and (2) extracting data from included material as the basis of the synthesis.

Sifting the results

Academic literature

All the identified titles, abstracts and referencing information from the structured search were downloaded onto RefWorks. Patterson et al. 45 recommend that each item be ‘sifted’ by two members of the research team independently and evaluated against a pro forma which sets out clearly the quality and relevance thresholds for inclusion. Using a checklist of agreed criteria in this way helps to address the potential impact of reviewer bias. Where there is some dispute or doubt over inclusion, the item should be referred to a third reviewer. The agreed inclusion/exclusion criteria and categories for the sifting process are:

1. include

2. exclude – dated before 1990

3. exclude – not in English language

4. exclude – empirical but study design does not include employees

5. exclude – opinion piece only/no evidence

6. exclude – item not related to research questions

7. exclude – other (specify).

Given the volume of literature to be sifted as well as the dispersed nature of the project team, it was important to develop a systematic and co-ordinated way of sifting the material. Thus, a bespoke database was developed using Excel Professional Plus 2010 (Microsoft Corporation, Redmond, WA, USA), into which all items of literature were imported and assigned a unique reference number. Each member of the team was then randomly assigned an equal share of the 5771 items identified in the structured search as either first or second reviewer for assessment against the stated criteria.

The database included a series of user-friendly worksheet-based interfaces, the first of which allowed project team members to call up each individual item from the 5771 results (title, author, source, abstract and referencing information) using the allocated reference numbers. A second enabled the reviewer to evaluate relevance and quality according to the agreed criteria. Given that the item abstracts (or, in a minority of cases, titles only) were the initial basis for assessment, the criteria as shown above were weighted more towards relevance (e.g. ‘dated before 1990’, ‘not in English language’, ‘empirical but study design does not include employees’, ‘opinion piece only/no evidence’ and ‘item not related to research questions’), with the view that quality would be better evaluated at the second stage once full items were obtained. Items were included in the next stage where they appeared to be of direct relevance to the research questions, and appeared to include either empirical evidence from employees or a theoretical contribution to the field.

A third interface of the bespoke database was designed to systematically record the outcome of the sifting process by logging the following information: item reference number, reviewer’s name, fields within each record that had been checked, type of engagement discussed, whether peer-reviewed or not, specific relevance to the research questions and, if excluded, the exact reason why. From these records, it was possible to identify disputed items easily and reallocate them to a third reviewer (sample screenshots from these interfaces are illustrated in Appendix 4).

In order to develop inter-rater reliability and further minimise the potential impact of reviewer bias, prior to starting the sift process the project team undertook a number of pilot ‘sifts’ followed by tele-meetings to identify areas and causes of uncertainty, and to build critical reflection and consensus into the evaluation process. 42 A kappa rating was calculated from the results of pilot sifts using all six reviewers from the team, and only when a score of 0.75 was achieved [generally interpreted as ‘substantial agreement’ (p. 361)56] was it agreed to proceed with the sift.

However, as the project team sifted the results of the structured search, it was clear that, while a great many of the results met the relevance criterion, they would not be included at the data extraction stage because of the quality criterion. As with the grey literature material, the search identified a large quantity of material that simply did not contain any substantive evidence or duplicates. Thus, after consulting with the project adviser, it was agreed that only items from peer-reviewed academic sources should be put forward to the next stage. The project protocol was amended to reflect this change. Of the original 5771 items identified in the full search, 5168 were excluded on grounds of relevance at this stage (i.e. not peer-reviewed, duplicated or not in English). This left a total of 603 items to be potentially considered for data extraction. These 603 items are included in the References section of this report.

Each of these 603 items was then reviewed in greater depth by two members of the project team, and 389 of them were excluded on grounds of quality (e.g. rigour), relevance (e.g. conflation of engagement with other concepts such as job satisfaction) or other reasons (Table 8). This left a total of 214 items to be included for full data extraction.

TABLE 8 - Review of potentially included materials prior to data extraction

UWES, Utrecht Work Engagement Scale.

Basis for exclusion		Total (%)
Lacking empirical data	For example, opinion piece/normative	35 (6)
Quality	Poor quality of item (e.g. improper scales; missing values or values not reported; measures not stated; sample issues; data not analysed)	95 (16)
	Measuring engagement using one dimension only of UWES	7 (1)
	Measuring engagement using two dimensions only of UWES	46 (7.5)
	Not peer-reviewed	6 (1)
Relevance	Study measures only individual/demographic factors as antecedents	31 (5)
	Not focused on concept of engagement, employees or work context	124 (20)
Other	Duplicated items evident only on close scrutiny	14 (2.5)
	Validation study only (of existing scale/not testing variables)	6 (1)
	Item unobtainable via usual sources	25 (4)
Total excluded		389 (64.5)
Total included for full data extraction		214 (35.5)

Grey literature

To assess the quality and identify materials suitable for data extraction from the grey literature identified in the structured search, a series of ‘sift’ questions were applied to each of the materials. These ‘sift’ questions were devised within the project team with particular reference to the more explicitly practical emphasis within research question 4 and the production of practitioner outputs. These were:

• Is the material relevant or useful to an NHS practitioner (in the context of staff engagement)?

• Does the material contain evidence?

• Does the material include a described methodology?

• Is the research original to this source?

• If the material forms part of a series, is this the most recent?

By applying these quality criteria to the results of the structured search of grey literature revealed, the team deemed only six grey literature sources to be of sufficient quality for inclusion in the data extraction, including one referred from the academic literature search (Table 9). It enabled a greater focus on a small number of high-quality materials from these sources in the production of practitioner outputs.

TABLE 9 - Number of items for data extraction from grey literature sources

GSR, Government Social Research Service.

Source	Number of suitable items
Chartered Institute of Personnel and Development	5
Institute for Employment Studies	3
Kenexa	3
The King’s Fund	1
GSR	1
Referred from main literature search: Strategic Human Resources Review	1
Total number of items	14

Data extraction

Academic literature

The second stage involved obtaining complete versions of the 214 included items in order to evaluate and extract data from them. To facilitate analysis, a data extraction form (see Appendix 5) was devised to record the evaluation of items against a range of quality criteria including methodology (robustness of design and analysis), relevance to health-care contexts and relevance to the research questions (see Appendix 6). This approach was agreed with the advisory group.

Of the 214 items included for full extraction in the synthesis, five were qualitative studies and four were meta-analyses. These were organised according to their specific relevance to the research questions (Table 10). A total of 67 out of the 172 empirical papers (39%) within this evidence review were included in at least one of the four meta-analyses, while nearly half of these (n = 32) had been included in all four meta-analyses. However, to avoid distorted effect, none of these meta-analyses was included in the data extraction tables detailed in Chapters 4–6 but they are discussed separately within each chapter. Throughout the extraction process, additional studies were being added to the search results and sifted as a result of the citation and reference tracking strategy, along with others identified by ‘alert’ services from journals and databases using keywords.

TABLE 10 - Number of studies in the academic literature relevant to each research question

Because items were relevant to more than one research question, these do not add up to 214.

These 38 items did not contain empirical data and were relevant to research question 1 only. However, the underpinning models and theories of the 172 empirical studies were also considered in addressing research question 1, as reported in Table 13.

Research question	Number of relevant studiesa
1. Models and theories	38b
2.1. Morale	47
General	35
Health	12
2.2. Performance	42
General	36
Health	6
3. Antecedents	155
General	113
Health	42

Originally, it had been proposed that each full item would be evaluated by two researchers and coded to identify its primary contribution to knowledge. Given the volume of included studies, it was decided that each item would be reviewed in full initially by one researcher, who completed the data extraction form. However, in practice the vast majority of items, about 75%, were evaluated twice anyway as the report authors reassessed the items included for each of their respective chapters. To describe stages 1–4 of the search and data extraction process, a Preferred Reporting Items for Systematic Reviews and Meta-Analyses-style flow chart57 (Figure 1) was prepared according to the format proposed by Liberati et al. 57 The flow chart summarises the process of evidence synthesis from the planning to the data extraction stage of the project.

FIGURE 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses-style flow of information through stages 1–4 of the evidence synthesis.

Grey literature

The 14 materials identified in Table 9 were taken forward for data extraction for use in the production of the practitioner materials arising from this project (see Appendices 7–16). The dates of materials from which data were extracted ranged from 2004 to 2013. They included two single-organisation case studies, while the other papers discuss data from more than one organisation. Studies were based in the UK, the USA or mixed-country settings, and four were based either uniquely or partially in a health-care setting. A copy of the data extraction form is provided at Appendix 17. Although none of the practitioner (‘grey’) literature was able to satisfy the peer-reviewed criterion for inclusion in the main evidence review, a review of good-quality practitioner materials was conducted in order to inform the practitioner outputs and address research question 4. This material was therefore separated from the main evidence review and is the subject of a separate review of practitioner material (see Appendix 7).

Summary

In this chapter, we have described the methodological approach underpinning this evidence synthesis. Following the recommendations of Briner,46 we adopted a narrative approach in five stages (planning, locating studies, evaluating material, analysis and coding, reporting).

In collaboration with the project advisory group, we refined the project protocol, detailed research questions and search terms, and conducted a series of pilot searches in order to help refine and focus our search strategy. The full search of academic literature was conducted using five databases and a wide range of grey sources. A total of 5771 studies were included in the preliminary sifting exercise whereby the abstract and/or title for each item was reviewed by two or in some cases three members of the research team.

The application of quality and relevance criteria along with the removal of non-peer-reviewed items led to the inclusion of 172 empirical articles, four meta-analyses and 38 theoretical papers in the final data extraction exercise. Items that were published in the English language after 1990, and that met the appropriate quality and relevance thresholds for the type of study, were included in the evidence synthesis. Items identified from six sources through searching the grey literature are included in the practitioner-oriented materials arising from this project.

In the next chapter, we examine the results of the evidence synthesis in relation to research question 1: what is engagement?

Introduction

In this chapter, we address the first research question, namely:

 How has employee engagement been defined, modelled and operationalised within the academic literature?

This overarching question can be broken down into the following three subquestions:

1. How is employee engagement defined within the academic literature?

2. How has engagement been measured and modelled within the academic literature?

3. What theories have been used to underpin models of engagement within the academic literature?

In order to address these, we undertook the following analysis:

1. Extraction of information relating to the definition, measurement and theorisation of engagement from all the studies included in the evidence synthesis for research questions 2 and 3 (see Chapter 2): a total of 172 papers.

2. Review of relevant information from a number of literature reviews and conceptual papers focused on defining engagement that were identified in the second stage of the data extraction process but that either did not contain empirical data or contained empirical data that did not meet the quality threshold and so were excluded from the data extraction for research questions 2 and 3: a total of 38 papers.

3. Consultation of three recent academic books focusing on engagement. 16,59,60 The research team identified these as being the only academic books with an exclusive focus on engagement.

4. Consultation of further conceptual articles focusing on defining engagement that were known among the research team or that were identified through a snowballing approach.

5. Consultation of a number of conceptual articles or literature reviews that critiqued or questioned the engagement construct.

The chapter is organised as follows. First, we present an overview of the broad history and development of engagement, and outline the definitions and measures of engagement used within the literature. Next, the findings relating to question 1, the extraction of definitions and measures used in the empirical papers that formed the substance of our data extraction, are presented. This delineates the principal approaches that have been used within the empirical literature. Next, we outline the theoretical frameworks that have been used to explain the processes of engagement, before presenting an analysis of the critiques that have been proposed of the engagement construct. We conclude by highlighting the principal areas of agreement and disagreement with regard to engagement at a theoretical level, a topic that is explored further in Chapter 7, in the light of the evidence presented in Chapters 4–6.

The origins and definitions of employee engagement

Interest in engagement first arose as part of the wider development of the positive psychology movement that has burgeoned in recent decades as a counterbalance to the predominant focus on negative psychological states. As Youssef-Morgan and Bockorny61 note (p. 36), the earlier emphasis on factors such as stress, burnout and poor performance offered limited opportunity to understand strengths, optimal functioning and fulfilment at work.

William Kahn is widely acknowledged as being the first academic to research and write about engagement, which he referred to as ‘personal engagement’. In his seminal article,21 Kahn claimed that personal engagement or disengagement arises when ‘people bring in or leave out their personal selves during work-role performances’ (p. 702). Thus, personally engaged workers are those who express themselves authentically at work in three ways: cognitively, emotionally and physically. This authentic expression of self-in-role is contrasted with disengagement, whereby the individual ‘uncouples’ his or her true self from his or her work role, and suppresses his or her involvement. Since Kahn’s original research, interest in engagement has mushroomed, leading to the publication of significant numbers of publications, especially since 2005. 62

Kahn’s original notion that engagement is the investment of the self into work roles has been developed further into the concept of ‘work engagement’, or the ‘relationship of the employee with his or her work’ (p. 15). 62

However, along with this burgeoning interest has been considerable confusion and uncertainty about what engagement means, leading Christian et al. 27 to conclude: ‘engagement research has been plagued by inconsistent construct definitions and operationalizations’ (pp. 89–90). A range of different terms has been used, including ‘work engagement’, ‘job engagement’, ‘role engagement’, ‘organisational engagement’ and ‘self-engagement’, with associated variations in the measures and theoretical underpinnings used. 63 Some have gone so far as to argue that engagement may be no more than old wine in new bottles. 18,19,64,65

There has been uncertainty over whether engagement is a relatively stable personality trait or a state that is susceptible to fluctuation over time, as well as whether it is a one-, two- or three-dimensional construct. However, the emerging consensus is that engagement is a psychological state, as summarised by Christian et al. :27 engagement is ‘a relatively enduring state of mind referring to the simultaneous investment of personal energies in the experience or performance of work’ (p. 90). Parker and Griffin29 extend this by arguing that engagement is an active rather than a passive psychological state, and therefore is associated with energetic states of mind. There is, additionally, broad agreement that engagement is not a one-dimensional construct but rather comprises several facets. 66

Below, we explore the most widely used definitions and conceptualisations of engagement found through our data extraction process. Drawing on and extending previous typologies such as that of Shuck63 and Simpson,36 we categorise the definitions and operationalisations of engagement within the literature under six headings, and review each in turn (Table 12):

• Personal role engagement, including the work of Kahn21 and researchers who have sought to operationalise his theoretical framework.

• Work task or job engagement, including the work of the Utrecht Group,71 which has focused specifically on the notion of engagement with work tasks.

• Multidimensional engagement, drawing on the work of Saks,72 who distinguishes between engagement with work and engagement with the organisation as a whole.

• Engagement as a composite attitudinal and behavioural construct, drawing on the work of various consultancy firms and researchers who regard engagement as a broadly defined positive attitudinal state; this approach is what is commonly referred to as ‘employee engagement’.

• Engagement as practice: scholars within the HRM field have recently begun to focus on engagement, and there is a small emergent literature on engagement as an employment relations practice. 16,17

• Self-engagement with performance: one measure has been developed that regards engagement as the extent to which high levels of performance are salient to the individual.

While some meta-analyses of engagement such as that of Christian et al. 27 restrict themselves to a narrow view of engagement purely in terms of an individual’s engagement with his or her work tasks, we seek here to be inclusive in terms of encompassing the range of various definitions within the literature, while remaining rigorous in terms of the application of a quality threshold.

Several measures that have been published in peer-reviewed journals were excluded on grounds of quality. Most significantly, following Christian et al. 27 we excluded several papers that drew on the Gallup Q12 engagement measure (e.g. Jones and Harter81). This was because of concerns that the Q12 is a composite, catch-all measure that lacks construct and discriminant validity. 20,36,62 Little and Little in their critique of the measure82 note that ‘merely attaching a name to a collection of survey items does not make it a construct’ (p. 112), while Guest18 argues that the consultancy perspective on engagement is akin to a management fad.

We also excluded scales developed to capture engagement specifically within the context of health, such as that of Mark et al. 83 This scale included nursing expertise, commitment to care and tenure, aggregated into one overarching engagement construct. Again, this measure lacked validity. The Shirom–Melamed vigour scale, which has also been used to measure engagement,84 was excluded on the grounds that the scale more accurately captures the single notion of ‘vigour’ rather than engagement.

The included measures are reviewed in the following paragraphs.

The discriminant validity of engagement

A fundamental question is whether engagement can be considered a distinct construct in its own right or the overlaps between engagement and other, similar, constructs are so great that it adds little or nothing. 19 The notion that engagement might be little more than an assemblage of other constructs has been referred to as the ‘jangle fallacy’ (p. 97). 27 Several studies have addressed this point, but the findings generally suggest that at least some of the conceptualisations of engagement can be classed as a distinct construct.

Christian et al. 27 examined whether or not engagement as defined by the Utrecht Group demonstrates discriminant validity as compared with more established constructs such as job satisfaction, job involvement and organisational commitment. They found evidence to support the notion that engagement is different from these other constructs and showed that engagement has incremental validity over other job attitudes in predicting performance. However, they also concluded that there is sufficient overlap, of around r = 0.50, between engagement and these other constructs to conclude that they share ‘conceptual space’ (p. 120). Hallberg and Schaufeli99 similarly found that engagement measured by the UWES has discriminant validity compared with job involvement and organisational commitment.

However, although Viljevac et al. 100 found that both the UWES and the scale proposed by May et al. 22 showed discriminant validity compared with affective commitment, and that the UWES showed some evidence of differentiation compared with job involvement and intent to stay, they also found that neither measure showed discriminant validity compared with job satisfaction. Partial convergent relationships were found between the UWES and the May et al. 22 scale, suggesting that they are similar but not overlapping constructs. Viljevac et al. 100 conclude: ‘our results suggest that neither measure should be considered an adequate measure of work engagement’ (p. 3706). Wefald et al. 101 could not confirm the three-factor structure of the UWES, and found, further, that the measure was not able to predict a range of outcomes when satisfaction and affective commitment were controlled for.

With regard to the UWES and the Britt102 scale, Wefald et al. 101 conclude: ‘these findings potentially suggest that the way engagement is typically measured may be inherently flawed, and that engagement, as measured by both Schaufeli and Britt, may be redundant with the more established constructs of job satisfaction and affective organisational commitment’ (p. 87). These findings suggest that some degree of caution should be exercised in interpreting the findings of quantitative engagement studies. More research is needed that explores the construct, discriminant and predictive validity of the engagement scales currently in use.

No research to date has examined the evidence emerging from the body of qualitative studies of engagement in the organisational sociology/industrial relations field, in which engagement is defined in a very different way from the psychology literature, as practice. This emergent field holds some promise as an antidote to what Godard103 has termed the ‘psychologisation of employment relations’ (p. 1).

Taken together, these findings suggest that further research is required to demonstrate engagement’s properties as a distinct psychological state that can explain more of the variance in other attitudes and behaviours than other states, and to explain and synthesise the wide range of definitions of engagement that have arisen.

Definitions and measures of engagement: conclusions

As Schaufeli62 notes, ‘probably the most important issue in defining engagement is “where to draw the line?” ’ (p. 19). As we have seen, the variety of ways in which engagement has been defined suggests that the construct has been subject to ‘fixing, shrinking, stretching and bending’ (p. 19),17 creating a bewildering array of definitions and meanings, and ways of measuring and evaluating levels of engagement. Although Macey and Schneider25 suggest that engagement is a synthesis of trait engagement (including personality traits such as conscientiousness), state engagement (including satisfaction and involvement) and behavioural engagement (including proactivity and role crafting), the emerging consensus appears to be that engagement should be considered as a relatively enduring psychological state experienced by an individual in relation to his or her work activities that is affected by various antecedent factors and leads to a range of outcomes. The most widely used definitions of engagement, those of the Utrecht Group71 and Kahn,21 share a number of commonalities, in that both regard engagement as multifaceted and as comprising three dimensions: an affective, a cognitive and an energetic dimension.

A number of studies have been conducted to examine the construct validity and discriminant validity of engagement in the wake of accusations that engagement has conceptual overlaps with constructs such as commitment, involvement and satisfaction. Evaluating these has not been a primary focus of this investigation; however, meta-analysis conducted by Christian et al. ,27 as well as primary research,4,99 suggests the cautious conclusion that engagement is sufficiently different from these other attitudes to be considered as an attitude in its own right. However, more research is needed to address the concerns that have been raised and that examines the emerging view of engagement as employment relations practice.

Theoretical frameworks: introduction

Engagement scholars have used a wide range of theoretical frameworks to ‘explain’ engagement, either in the sense of explaining how engagement fits within wider positive psychological paradigms, or in order to explain the relationship between engagement, its antecedents and its outcomes. To identify what theories have been used to underpin models of engagement within the academic literature, we undertook an analysis of the principal theoretical frameworks used in empirical studies of engagement to ‘explain’ the engagement process. This involved reading the papers to extract the principal explicit or implicit theory/ies underpinning the analysis presented. The findings are presented in Table 13.

However, these findings should be interpreted with some caution; in some cases, the theories were not made explicit in the paper. We have coded 21 papers as ‘unspecified’ where no theory was mentioned and it was not clear on reading the paper what the author’s intention was. In other cases where no theory was mentioned, we have inferred based on available information within the paper what the author’s intention was. In many instances, authors referred to a range of different theories; for most of these, we have reported on the main theoretical frameworks only, but in some cases it was not clear that there was one overriding theory, and so we have included several theories for one item. The results presented in Table 13 should therefore be regarded as indicative of general trends within the literature.

Engagement definitions, models and theories: synthesis

The roots of engagement within the academic literature can be traced back to the work of William Kahn. 21 Although Kahn saw engagement as the expression of the preferred self in relation to work roles, and some subsequent researchers have sought to operationalise Kahn’s framework, the academic field has come to be dominated by the work of the Utrecht Group, which has argued that engagement is a positive, activated state of mind experienced by individuals in relation to their work, and has three facets: vigour, dedication and absorption. The UWES has been used as the means to evaluate engagement levels in 83% of the studies of engagement deemed to have met the quality threshold for inclusion. The JD-R framework was mentioned as a major theoretical framework in 39% of the empirical papers included. The majority of studies using the UWES are based on the JD-R framework.

Some other scales have been developed based on a conceptualisation of engagement as something that employees can experience in relation to their employer overall, or towards their work group; however, these studies are in the minority.

More recently, studies are starting to take place that consider engagement as a strategy for people management78 or form of communication. 79 This perspective is more aligned with the practitioner perspective, whereby engagement tends to be viewed either as an employment relations strategy or as a composite attitudinal and behavioural construct. 18,148 However, no quantitative measures that met the quality threshold were identified, so this body of work comprises qualitative studies.

The evidence from prior studies concerning the validity and reliability of the engagement construct has been equivocal, although there is some preliminary evidence that engagement has both construct and discriminant validity. This issue is explored further in subsequent chapters, particularly Chapter 7.

Summary

In this chapter, we have reviewed how engagement has been defined within the academic literature, discussed the main measures used to evaluate engagement and examined the main theories used to explain the engagement process. Our analysis showed the predominance of the UWES as the means of evaluating engagement levels, and the JD-R analytical framework. In the next chapter, we present the results of our data extraction in relation to the association between engagement and morale.

Introduction

The focus of this chapter is on research question 2.1:

 What evidence is there that engagement is relevant for staff morale?

In order to address this we have developed two subquestions:

1. What is the evidence that engagement is relevant for staff morale within the workforce in general?

2. What is the evidence that engagement is relevant for staff morale within the context of health?

The purpose of addressing this question is to evaluate the evidence concerning the link between engagement and morale outcomes. In particular, this chapter aims to identify which morale outcomes have been demonstrated empirically to be most significantly affected by, or at least associated with, high levels of engagement within the general workforce and within the health context specifically. To address these questions, we undertook the data extraction process described in detail in Chapter 2.

First, we review the general background and context for the research questions (next section). We then present the evidence we have assembled from our data extraction exercise in relation to the general workforce, followed by an analysis of the health context. We have included in this latter section any study that includes a sample of health-care workers, even if part of a wider sample involving a range of occupations. Finally, we bring together these findings to suggest which morale outcomes are most likely to be associated with high levels of engagement (Conclusions).

Background to morale outcomes relevant to engagement

A main concern within the prior research has been to identify which morale outcomes are associated with high levels of engagement. We have interpreted ‘morale’ in a wide sense and these morale outcome factors can be grouped under two headings:

• well-being and health perceptions, includes measures of life satisfaction,149 general and psychological health (e.g. the 12-item General Health Questionnaire;150 Beck’s Depression Inventory151), stress/burnout (e.g. MBI152) and various other aspects such as affect,153 work ability154 and recovery155

• work-related attitudes, including measures of organisational commitment,156,157 job or career satisfaction,158,159 occupational or career success160 and turnover intentions. 161,162

In total, 35 studies examined the relationship between engagement and morale outcomes within the general workforce, and 12 studies examined the relationship between engagement and morale outcomes within the health context. Table 14 shows the breakdown of studies that satisfied the inclusion criteria, first across the two main categories, and second across the subcategories. It should also be noted that many studies examined more than one outcome (although not usually more than two or three), and a minority included outcome variables from both well-being/health perception and work-related attitude categories. Therefore the numbers in the subcategories do not add up to the numbers in the overall categories.

Furthermore, many studies examined the link between engagement and morale outcomes as part of a larger hypothesised model. These models tended to include a number of antecedents theorised to be related to engagement (see Chapter 6), engagement as a mediator (and, for some, other mediators such as burnout) and potential morale and/or performance outcomes (see Chapter 5). The focus for this chapter is on the specific relationship between engagement and morale outcomes; therefore, no particular details regarding any larger hypothesised model tested have been included here.

Almost all included studies have examined engagement as a psychological state experienced in relation to work in general terms (see Chapter 3, The origins and definitions of employee engagement), and have used quantitative, survey-based methods to examine how engagement relates to self-reported morale indicators. Therefore, most studies reviewed in this chapter reflect a specific and narrow research focus. Moreover, the vast majority of the studies utilised a cross-sectional, between-persons design, so causality is difficult to demonstrate.

The workforce in general

The data on morale outcomes in relation to the workforce in general are reported in Table 15.

The health context

The data relating to morale outcomes in the health context are reported in Table 16.

Conclusions

A total of 47 studies had examined the relationship between engagement and at least one morale outcome (35 related to the general workforce; 12 to the health context).

Of these, 21 tested the associations between engagement and at least one well-being/health indicator. The most robust finding identified was the positive association between engagement (as a holistic factor) and life satisfaction (four out of four studies). Of these studies, two were longitudinal: one in the health context202 and one in the general workforce. 181 Having longitudinal evidence was a key element for identifying ‘robust’ findings.

The other consistent finding identified was that engagement (as a holistic factor) was negatively correlated with burnout measures (five out of five studies). However, these studies relied solely on correlations and cross-sectional designs, and many measured burnout with only one or two dimensions (e.g. emotional exhaustion and/or cynicism). There is still debate regarding whether burnout and engagement are independent or overlapping constructs. The meta-analysis by Cole et al. 92 suggests that ‘employee engagement, as gauged by the UWES, overlaps to such an extent with job burnout, as gauged by the MBI, that it effectively taps an existing construct under a new label’ (p. 1574).

Other well-being/health perceptions, such as depressive symptoms, were consistently related to engagement, as a holistic factor (eight out of nine studies), yet this covered a range of different aspects of well-being/health where many were only examined by one study, such as psychological distress, and did control for burnout. Although the meta-analysis by Halbesleben195 found that engagement is positively associated with health outcomes (yet little detail is given about which these are), the meta-analysis by Cole et al. 92 indicates that engagement may not explain any unique variance in health complaints above that of burnout. Indeed, the study by Buys and Rothmann168 also supports this finding. Taken together, it could be suggested that engagement may be specifically related to positive well-being/health experiences rather than negative well-being/health experiences.

Out of the 46 studies, 31 tested the associations between engagement and at least one work-related attitude. The most robust finding identified was the positive association between engagement (as a holistic factor) and organisational commitment (nine out of nine studies). However, all but one utilised a cross-sectional design, so causality cannot be established. Despite this, the findings are supported by two meta-analyses: Halbesleben195 found that engagement was positively associated with organisational commitment and Cole et al. 92 found that the dimensions of engagement accounted for a small to moderate amount of unique variance (beyond that of the burnout dimensions) in organisational commitment.

Engagement (as a holistic factor) was consistently found to be positively associated with job satisfaction when job satisfaction was deemed the outcome (nine out of nine studies). However, all of these studies were cross-sectional. Nevertheless, this association is supported by Cole et al. ’s92 meta-analysis, which found that the dimensions of engagement (as measured by the UWES) accounted for a small to moderate amount of unique variance (beyond that of the burnout dimensions) in job satisfaction. Yet, in the only longitudinal study, Yalabik et al. 191 found that job satisfaction may, in fact, be an antecedent rather than an outcome of engagement. Therefore, further longitudinal research is needed to confirm this finding.

Although there is consistent evidence to show that engagement (as a holistic factor) is negatively associated with turnover intentions (22 out of 22 studies), which is supported by the meta-analytic findings of Halbesleben,195 four of these studies also found that this relationship may be mediated by other work-related attitudes, specifically organisational commitment and job embeddedness. 101,132,197,198 However, 19 studies utilised cross-sectional designs and only three utilised time-lagged designs, so causality cannot be fully established.

A particular issue arose during the course of this evidence review. It was found that when engagement was examined not as a holistic factor, but as three subfactors (i.e. representing vigour, dedication and absorption dimensions of work engagement), findings became less consistent and more complex. However, dedication seems to be consistent in its (significant) association with morale indicators (9 out of 10 studies). This is supported by Cole et al. ’s92 meta-analysis, which found that dedication was the most significantly related to job satisfaction and organisational commitment. The issue of examining individual dimensions is particularly important, as engagement has been widely conceptualised and defined as a holistic yet multidimensional construct. The finding that subdimensions of engagement are more inconsistently related to morale outcomes than is a composite single factor of engagement is important, as it indicates that the latter is more appropriate than the former when examining the relationship between engagement and morale outcomes.

In summary, although the findings are far from conclusive, four consistent links between engagement and specific morale outcomes were identified. The first, and most conclusive, is that high levels of engagement are related, and may lead to higher levels of life satisfaction, as demonstrated by consistent findings across three cross-sectional and two longitudinal studies. The second is that high levels of engagement are associated with low levels of burnout, as demonstrated by consistent findings across five cross-sectional studies. However, concern is still warranted over whether or not the two are completely independent and linked in a causal rather than intercorrelated way. The third and fourth are that high levels of engagement are associated with high levels of organisational commitment and weak intentions to leave the organisation, as demonstrated by consistent findings across 9 and 22 studies, respectively, of which four were longitudinal. A tentative proposition is that the relationship between engagement and turnover intentions is mediated by organisational commitment and/or by job embeddedness, as demonstrated by three cross-sectional studies and one that was time-lagged. Another tentative proposition is that, although engagement and job satisfaction are positively related, job satisfaction may act as an antecedent rather than an outcome of engagement, as demonstrated by Yalabik et al. ’s191 longitudinal study.

In the next chapter, we consider the evidence relating to the association between engagement and performance outcomes.

Introduction

The focus of this chapter is on research question 2.2:

 What evidence is there that engagement is relevant for performance?

In order to answer this question, we have developed two subquestions:

1. What evidence is there that engagement is relevant for performance within the workforce in general?

2. What evidence is there that engagement is relevant for performance within the context of health?

The purpose of addressing this question is to find evidence that sheds light on the link between engagement and performance outcomes at the individual, work group and organisational levels. In particular, this chapter aims to explore the empirical evidence concerning the association between employees’ levels of engagement and performance levels within the general workforce and within the health context specifically. To address these questions, we undertook the data extraction process described in detail in Chapter 2.

First, we review the general background and context for the research questions (see Background to performance outcomes associated with engagement). We then present the evidence we have assembled from our data extraction exercise in relation to the general workforce, followed by an analysis of the health context. We have included in this latter section any study that includes a sample of health-care workers, even if part of a wider sample involving a range of occupations. Finally, we bring together these findings to identify the performance outcomes that most likely result from high levels of engagement (Conclusions).

Background to performance outcomes associated with engagement

One central question in engagement research relates to the extent to which employees’ levels of engagement are related to higher performance outcomes. We have categorised performance outcomes into individual versus higher-level (team, organisation) outcomes. Individual performance can be further grouped into three headings:

1. In-role performance is related to behaviours that are generally specified by the job description and contribute to the organisation’s technical core. 211 We have included constructs such as in-role performance, quality of care and service quality.

2. Extra-role performance is related to behaviours that support task performance by enhancing and maintaining the social and psychological environment. 211 We have included constructs such as citizenship behaviour, adaptability and innovative work behaviour.

3. Counterproductive performance (or deviance behaviour) is related to behaviours that harm the organisation and are an indication of an employee’s withdrawal behaviour. 212

In total, 42 studies examined the performance outcomes of engagement. Of those studies, six were carried out in a health-care context. Table 17 shows a breakdown of performance outcomes that were examined in these studies. It is important to note that many studies examined more than one outcome (although not usually more than two or three), and that the total number of outcomes measured across all studies therefore exceeds the number of studies included in this review.

It is noteworthy that the majority of studies included in this section examined the relationship between engagement and performance outcomes as part of a larger model. In general, these models explored how antecedents are related to engagement (see Chapter 6) and how engagement is related to other outcomes aside from performance (see Chapter 4). The models also explored mediating, moderating and reciprocal relationships between engagement and its correlates. We have included some information about these more complex relationships in this chapter, when they were relevant in explaining the mechanism through which engagement is related to performance outcomes.

The workforce in general

The health context

The data relating to performance outcomes in the health context are reported in Table 19.

Conclusions

Overall, 42 studies examined the relationship between engagement and at least one performance outcome. The majority of the studies (36) were carried out in the general workforce, while six studies were carried out in a health-care context.

Whether or not higher levels of engagement within the workforce were positively related to higher-level (organisation, unit, team) performance outcomes was analysed 13 times. These studies provide some, but inconclusive, support for a positive association between engagement and performance. The majority of the studies relied on employee perceptions of organisational performance variables, rather than using objective performance data, such as financial outcomes or employee turnover data, and only a few studies used third-party data, such as customer ratings, as a measure of performance.

At the individual level, the relationship between engagement and in-role performance was the focus in the majority of studies. These studies unanimously provided support of a positive relationship between both variables. Eleven studies used third-party rating or objective performance indicators to assess employees’ performance. This lends weight to the argument that employees who are engaged with their jobs perform better on the tasks that are assigned to them.

A substantial number of studies also analysed the relationship between engagement and extra-role as well as counterproductive performance outcomes. While these studies support the notion that engagement is positively related to extra-role performance and negatively related to counterproductive performance, it is important to note that the majority of studies used a cross-sectional, self-report design. Hence, common methods bias might have influenced the results in these studies. The evidence on the relationship between engagement and extra-role and counterproductive performance is therefore less convincing than the evidence on the link between engagement and in-role performance.

Overall, the studies suggest that engagement has positive performance outcomes. This is supported in the meta-analysis by Halbesleben. 195 Based on seven studies and a total sample size of 6131, Halbesleben found that engagement, and specifically the vigour component, was positively related to an overall composite of performance. However, as the number of studies included in this meta-analysis was relatively low, future studies need to validate the relationships analysed in the present chapter, using longitudinal research design and third-party ratings of performance, specifically, to measure extra-role and counterproductive performance. In the next chapter, we consider the evidence relating to the antecedents of engagement.

Introduction

The focus of this chapter is on research question 3:

 What approaches and interventions have the greatest potential to create and embed high levels of engagement within the NHS?

In order to address this, we have developed two subquestions:

1. What evidence is there concerning approaches and interventions within an organisational setting at (1) the individual, (2) the team or (3) the organisational level that create and embed high levels of engagement within the general workforce?

2. What evidence is there concerning approaches and interventions within an organisational setting at (1) the individual, (2) the team or (3) the organisational level that create and embed high levels of engagement within the health context?

The purpose of addressing this question is to find evidence that sheds light on the approaches and interventions that have been demonstrated empirically to have the most significant effect on, or at least association with, high levels of engagement within the general workforce and within the health context specifically.

We undertook the data extraction process described in detail in Chapter 2. We have specifically excluded from this analysis of antecedents the following factors:

• Demographic variables such as age or gender, except where these are relevant to understanding and interpreting study findings. This is because demographic factors alone do not constitute an approach or intervention.

• Personality variables such as conscientiousness, neuroticism and extraversion, except where these are relevant to understanding and interpreting study findings. This is again because personality variables do not constitute an approach or intervention.

This is to say not that either demographic and personality factors may not be salient for engagement in different contexts and circumstances, but rather that these factors lie beyond the scope of this review.

First, we provide an overview of the general background and context for the research questions. We then present the evidence we have assembled from our data extraction exercise in relation to the general workforce, followed by an analysis of the health context. We bring together these findings to suggest which interventions or approaches are supported by the strongest evidence.

Background to approaches and interventions to foster engagement

A fundamental concern within the body of research on engagement has been to identify the factors associated with, or antecedents of, high levels of engagement. It should be noted that the bulk of the literature on employee engagement focuses not on the evaluation of specific approaches or interventions, but rather on the psychological antecedents of engagement. This is because the research is based within the organisational psychology literature, which does not have a history or tradition of evaluating organisational interventions. Instead, the focus is very much on psychological factors at the level of the individual. Several of those studies that did focus on organisational interventions did not pass the quality threshold for inclusion. It is disappointing that so much of this literature does not enable an evaluation of specific interventions, which would have been of most interest and relevance to practitioners, and, as the field develops further into industrial sociology and organisational behaviour, it is probable that there will be a significant development of studies that examine this aspect in more detail.

Despite this, there are a number of themes that have emerged from the engagement literature that point towards promising approaches to enhancing engagement. These factors can be grouped under five headings:

• individual psychological states such as experienced psychological safety or availability, which are included in our report, since such states can be influenced by organisational factors

• experienced job design-related factors such as task significance, variety, meaningfulness and autonomy, job demands and job resources

• perceived leadership and management factors such as leadership style, authentic leadership, perceived supervisor support

• individual perceptions of organisational-level factors such as perceived organisational support, organisational mission, climate or culture, and perceptions of colleagues and team

• organisational interventions or activities such as specific training and development courses or communication activities.

It should be noted that many studies examined a range of antecedents rather than just one and so may be referred to under several headings.

Almost all included studies have examined engagement as a psychological state experienced in relation to work in general terms, and have used quantitative, survey-based methods to examine how engagement relates to other attitudes held by employees. Therefore, most studies reviewed in this chapter do not examine engagement in association with an actual organisational intervention or activity (although a small number do, and these are described below). It is therefore necessary to infer what interventions might support or encourage the development of high levels of experienced engagement among employees through examining these attitudinal associations. A very small number of studies that met the quality threshold have been published recently that have examined engagement in the sense of ‘doing engagement’ as a way of managing the employment relationship (see Chapter 3), and these are similarly outlined below.

In the next section we review the evidence relating to the association between antecedents and engagement for the workforce in general, and in the section after it we review the evidence from studies specifically related to health. We have included in this latter section any study that includes a sample of health-care workers, even if part of a wider study involving a range of occupations.

Antecedents: the workforce in general

A total of 113 studies met the inclusion criteria. These are reported in Table 20. It should be noted that some studies are relevant in more than one category.

Fifty six of the 113 studies were based in Europe, of which eight were based in the UK, 16 were conducted in the USA/Canada, seven in Australasia, 21 in Asia, six in South Africa and one in Cameroon (see Appendix 18). The preponderance of studies emanating from the Netherlands reflects the concentration of studies conducted by the Utrecht Group.

As explained in Chapter 3, the vast majority of studies used variants of the UWES. Only four studies were qualitative and drew on interviews. The majority of studies used self-report surveys at one time point (71), and hence drawing conclusions over causality from these is problematic.

However, 41 studies that examined antecedents involved more complex methods, for example diary studies, time-lagged surveys or data from multiple informants, as listed in Table 21. There was a significant variation in the size of the sample used in the studies, and in the sectors surveyed. However, most studies reported on data from the service sector, and the majority included data from one or two organisations only. Table 22 summarises the data from this analysis.

Antecedents: the health context

In this section, we report on the findings relating to antecedents of engagement within studies in the health context. A total of 42 studies met the inclusion criteria (Table 23); most focused exclusively on health, while others reported on samples involving health-care workers and those in other occupational groups.

Most studies used self-report surveys; however, a minority also used complex methods as shown in Table 24.

Twenty-three studies took place in Europe, of which two were in the UK and four were conducted in multiple European countries; seven in the USA/Canada; four in Australia; four in Asia; one in South Africa; one in Uganda; one in multiple continents; and one in Israel (see Appendix 18). All the studies in this category used variants of the UWES to measure engagement.

Table 25 reports the data relating to antecedents of engagement within a health context.

Conclusions

A total of 113 studies within the general workforce and 44 studies within the context of health that focused on the antecedents of engagement met the inclusion criteria. The majority of these studies used self-report cross-sectional surveys, and these studies can demonstrate correlation or association but not necessarily a causal relationship. A very small number of studies used qualitative or ethnographic methods. Forty-one studies within the general workforce and 9 in health used more complex methods such as diary studies, time-lagged surveys, pre-/post-intervention analysis or mixed methods. These studies are better able to demonstrate causal relationships between the variables. Studies were included from all continents but most research had taken place in Europe, the USA and Canada (64%). Only two studies had taken place in the health-care sector in the UK, neither of which was conducted in England (see Appendix 18).

The majority of studies in the general workforce and all those within the health-care sector used the UWES and in consequence the majority of studies examined engagement as a psychological state of mind rather than as an intervention or management approach. Only five studies in total, three from the general workforce and two from the health-care sector, reported on the findings of an intervention aimed at enhancing engagement and only two studies examined engagement as a management style.

Antecedents were grouped under five headings: individual psychological states, experienced job design factors, perceived leadership and management factors, individual perceptions of organisational factors, and organisational interventions. While many studies examined the interaction of a complex range of different variables, job design factors had received most attention (64 studies), followed by perception of the organisation/team (53 studies), psychological states (52 studies), leadership and management (36 studies) and specific interventions (nine studies). Within health, features of job design had been studied to a much greater extent than variables in other categories.

A very wide range of variables had been studied under each heading, with many factors having been examined in just one study. A number of factors were used in different studies as antecedents, mediators and moderators. Given that findings were also often mixed or contradictory, coupled with the cross-sectional nature of many of the studies, it is difficult to discern any clear-cut emerging evidence in support of any one or set of specific antecedents or interventions that support engagement.

With regard to individual psychological states, the weight of evidence appears to suggest that positive states such as positive affect, optimism, personal resources, self-efficacy and resilience are more likely to be associated with high levels of engagement than are negative states such as pessimism, anger or negative affect; several studies in this area used complex methods such as diary studies, time-lagged surveys or dyad surveys. However, the affective shift model121 proposes that the dynamic interplay between positive and negative states may have a role to play within wider models of engagement. Within the health-care sector, the 11 studies under this heading mainly used cross-sectional, self-report approaches, although Sonnentag et al. ’s96 two-wave study showed a link between low levels of negative affectivity and engagement, and Mauno et al. 140 in a longitudinal study in a hospital showed that job satisfaction and also paradoxically job insecurity were associated with high levels of engagement.

A large number of studies have investigated the role of job design factors such as job resources in relation to engagement both within the wider workforce and in the context of health specifically. Generally, the emerging consensus within the general literature and within health care is that job resources can promote engagement, while job demands may be neutral, negative or positive in relation to engagement levels. Three studies using complex methods found an association between job crafting and engagement (see Experienced job design-related factors). However, researchers have included a very wide range of factors within models of both resources and demands, and have measured these in different ways, and as antecedent, mediating and moderating variables, so evidence that resources promote engagement should be seen as tentative. Nevertheless, these findings corroborate those of the meta-analysis conducted by Crawford et al. ,301 which included the results of 55 largely cross-sectional surveys and found that job resources and demands predicted engagement, with resources being somewhat more predictive than demands: job resources and challenge-type demands were positively related to engagement, while hindrance demands were negatively related; together these three aspects explained 19% of the variance in engagement levels. Similarly, the meta-analysis conducted by Halbesleben195 involving 53 studies found a positive association between job resources and engagement and a negative association between demands and engagement.

Studies of leadership and management have broadly concluded that positive experiences of leader and manager behaviour promote engagement while negative experiences, such as abusive supervision and supervisor incivility, deplete engagement levels. Most evidence, within both the general workforce and health care, linked supervisory support with engagement. Studies using complex methods within health found that empowering leadership and transformational leadership were associated with engagement. 133,134

Similarly, studies of perceptions of the organisation/team have coalesced around the notion that positive experiences, such as value congruence, identification, perceived organisational and coworker support, all promote engagement, while negative experiences, such as intergroup conflict or coworker incivility, lower engagement levels. Many studies under this heading used cross-sectional methods. Within health, very few studies under this heading used complex methods, although the findings of cross-sectional surveys generally also showed that supportive environments fostered engagement. Some interesting findings are starting to emerge from studies of the general workforce concerning the spill-over effects of engagement; associations were found in one study127 that team engagement and individual engagement levels are correlated, and another study showed that colleagues’ engagement was associated with individual engagement. 126

The limited number of studies that have evaluated the effects of a specific intervention or approach have yielded mixed results, although there does appear to be positive evidence from some of these,217,263 notably from studies that involved surveys at multiple time points. Two studies have been conducted within the health-care sector; while Bishop282 found that participation in an offsite programme focused on the true meaning of caring boosted engagement among 17 nurses, Rickard et al. 293 did not find a significant change in engagement levels among a large group of nurses and midwives following participation in a workload exercise. Overall, the number of studies conducted to date is too small, and the overall conclusions too mixed, to reach any definitive conclusions about the salience of workplace interventions for raising engagement levels.

Introduction

In this report, we have presented the findings of an evidence synthesis focusing on three aspects of employee engagement: what engagement is; what antecedents are associated with high levels of engagement; and what performance and personal outcomes are associated with engagement. We have examined evidence concerning all three factors in relation to the workforce in general, and in the specific context of health care.

This report is timely; engagement is enjoying significant levels of popularity as a concept, notably in the UK, where the ‘Engage for Success’ movement has raised awareness of the potential for engagement to affect individual well-being, corporate performance and national productivity, and where the NHS has come under pressure to consider raising levels of engagement as a potential solution to some of the major challenges of staff morale, retention and performance. The question underpinning this report is: is this focus on engagement justified? Is there, in fact, any evidence that engagement levels make a difference and, if so, what does the research tell us are the factors most likely to yield high levels of engagement?

In this final chapter, we first summarise the methodology used in our study, and then outline the main findings arising in relation to each of our research questions. We then bring these together into an overarching synthesis, set within the context of some broader and more fundamentally challenging questions about the nature and meaning of engagement.

Methods for the evidence synthesis

We founded the approach to our evidence synthesis on the recommendations of Briner and Denyer39 for the conduct of systematic reviews using a narrative approach, and adhered to the principles of quality, relevance, transparency, replicability and credibility. We aimed to produce a critical narrative around the evidence in order to generate new insights into engagement, drawing on evidence obtained systematically from a review of relevant literature.

We started with four overarching research questions:

1. How has employee engagement been defined, modelled and operationalised within the academic literature?

2. What evidence is there that engagement is relevant for staff morale and performance?

3. What approaches and interventions have the greatest potential to create and embed high levels of engagement within the NHS?

4. What tools and resources would be most useful to NHS managers in order to improve engagement?

Research question 4 was addressed outwith the scope of this evidence synthesis report, but based on the background research outlined in Chapter 2 and on the results of the evidence synthesis. In addition to this report, and to address question 4, the project has yielded a set of outputs for practitioners including a conference, a workshop, a webinar, four podcasts, a set of practitioner guides, a report on the practitioner literature and a research paper on measuring engagement. These documents, or links to appropriate websites, are attached as appendices to this report.

The other three questions were further refined into detailed and specific questions that could be directly addressed from the literature, as outlined in Chapter 2, Planning. The searching and sifting process yielded a final total of 214 items comprising four meta-analyses, 172 empirical articles and a further 38 theoretical/conceptual pieces. Data were extracted from these 214 items and three books using a data extraction form designed by the project team (see Appendix 5) and approved by the advisory group to enable evaluation for quality and relevance.

By following the recommendations of Briner and Denyer39 for the conduct of evidence syntheses, we found far more items of potential evidence relating to engagement than we had envisaged at the outset, even by restricting publications to the English language and to the period after 1990. This vast body of work reflects the burgeoning interest in engagement that has emerged particularly in the past 10 years on the part of both academics and practitioners. It reflects the wide range of meanings attached to engagement; as a ‘fashionable’ label,18 it has been used in many writings about all manner of aspects of the employment relationship (e.g. ‘involvement’), the experience of work, and leadership and management styles/approaches (e.g. ‘medical engagement’). It also highlights the extent to which some terms have become conflated. Disentangling what is distinctive and significant about engagement has been a major task of this evidence synthesis.

Following the data extraction stage, each research question was addressed in a separate chapter of this report. Evidence obtained from each item was tabulated under the appropriate headings, with summary results presented together with an overall synthesis of the findings for each research question. Research question 1 was addressed in Chapter 3; research question 2 was addressed in Chapters 4 and 5; research question 3 was addressed in Chapter 6. In parallel with the main data search and extraction exercise, the grey literature was systematically searched to generate material aligned with the main search and relevant to the practitioner outputs (see Appendices 7–16).

It is perhaps appropriate at this stage to reflect on the narrative evidence synthesis methodology and its application in the context of this particular study. We noted earlier the almost overwhelming body of literature that was returned via our initial searches, necessitating the narrowing of our search terms in order to render the review manageable. This process led to a very specific focus on the topic of employee/work engagement. In applying the type of stringent quality thresholds necessary for this type of review, we also had to take the decision to focus our review on peer-reviewed sources only, and exclude conference papers and practitioner publications. In consequence, we eliminated from the search a range of subjects initially identified by the project team and the advisory group as potentially relevant, including areas particularly important for practitioners such as voice, involvement and participation, which are closer to the ‘lay’ definition and understanding of engagement. In consequence, the main review was unable to shed much light on ‘engagement as practice’ and instead focused on the body of work on engagement as a psychological state.

Alongside the main evidence synthesis, we undertook a parallel review of the grey literature with particular emphasis on the work of widely cited and reputable professional bodies and consultancy houses. It was initially hoped that this secondary review would bolster the report by yielding robust insights relevant to a practitioner audience. This review also led to an initially vast number of potential items for inclusion. However, through the application of the same stringent quality criteria as were applied in the main review, none of this grey literature met the threshold for inclusion in the evidence synthesis report. This situation highlights the challenges of applying robust evidence synthesis review methods in the context of a topic that enjoys multiple and occasionally competing definitions and interests across academic and practitioner spheres.

Despite the failure of the grey literature review to meet the same quality thresholds as the main review, we felt that some items from this body of work not only were of high quality but also had some relevance and ‘real-world’ application that rendered the findings of interest to practitioners. It was in this spirit that we developed the outputs for practitioners listed in Appendices 7–16, which are based on the review of the grey literature. As highlighted elsewhere, these outputs should be read in the light of the fact that they are not based on evidence sifted to the same quality thresholds as the academic studies. As we argue later in this chapter, it is hoped that future research will seek to bring together these disparate bodies of literature into a more coherent whole that addresses the needs of both academics and practitioners alike.

Synthesis of findings for research question 1: how has engagement been defined, modelled and operationalised within the academic literature?

Synthesis of findings for research question 2.1: what evidence is there that engagement is relevant for staff morale?

It has been widely argued that engagement is associated with higher levels of individual morale. But what evidence is there that this is in fact the case? We considered this question under two headings:

1. well-being and health perceptions, including measures of life satisfaction149 general and psychological health (e.g. the 12-item General Health Questionnaire, Beck’s Depression Inventory),150,151 stress/burnout (e.g. MBI),152 and various other aspects such as affect,153 work ability154 and recovery155

2. work-related attitudes, including measures of organisational commitment,156,157 job or career satisfaction,158,159 occupational or career success160 and turnover intentions. 161,162

We further considered this question in relation to both the general workforce and health care specifically. A total of 35 studies relating to the general workforce and 12 in relation to health care met the quality threshold and were included in the evidence synthesis. Several of these studies addressed more than one outcome criterion and a number also examined the association between various antecedents and engagement. Most studies were cross-sectional, so imputing causality should be treated with caution. From the studies focusing on well-being and health perceptions, the most consistent finding was a positive association between engagement and life satisfaction; five studies examined this link and two used complex methods. Engagement was also consistently found to be negatively associated with burnout (10 studies), although these studies were cross-sectional.

Thirty-one studies examined the link between engagement and work-related attitudes. The most consistent finding to emerge from these (nine studies) was that engagement was positively associated with organisational commitment; however, almost all these studies were cross-sectional. In a further 15 studies, engagement was found to be positively linked with job satisfaction (where this was treated as an outcome measure), although only one of these studies was longitudinal191 and this study further suggested that job satisfaction may act as an antecedent rather than an outcome of engagement (see also Synthesis of findings for research question 3). Twenty-four studies found engagement to be negatively associated with turnover intentions, and four of these studies showed this association to be a mediated relationship (e.g. by commitment). Most studies were cross-sectional, although some used complex methods. Within the context of health, one 2-year study by Mache et al. 204 involving surgeons found a positive association between engagement and work-related ability. It was noted that in those studies where engagement was broken down into different facets, rather than treated as a higher-order factor, the associations became more complex and tenuous.

If just those studies that took place within a health-care context are considered, all 12 included studies took place outside the UK, and only the UWES was used to evaluate engagement levels. Only one of the 12 studies was longitudinal,202 and this study found a positive association between engagement and life satisfaction, and a negative association between engagement and depressive symptoms for dentists in Finland over 3- and 4-year intervals. The other studies found associations between engagement and positive psychological outcomes such as psychological and general health and well-being, and between engagement and job satisfaction, career satisfaction and commitment. Negative links between engagement and intent to quit were found in seven cross-sectional studies. Generally, it is difficult to draw definitive conclusions in relation to the link between engagement and morale-related outcomes for health-care professionals on the basis of this body of evidence. Nevertheless, the weight of evidence tends to support the notion that engagement is linked with positive outcomes for health-care workers, and the evidence is strongest in relation to the link between engagement, life satisfaction and absence of depressive symptoms. Since this study involved just dentists and took place in Finland, it is uncertain if the results would be replicated in different health-care contexts or in the UK.

Synthesis of findings for research question 2.2: what evidence is there that engagement is relevant for performance?

To answer this question, we developed two subquestions:

1. What evidence is there that engagement is relevant for performance within the workforce in general?

2. What evidence is there that engagement is relevant for performance within the context of health?

We classified outcomes as individual or higher level (e.g. team, unit, organisational). Individual outcomes were considered under the following headings: in-role performance; extrarole performance (e.g. citizenship behaviour); and counterproductive performance (e.g. deviant behaviours). A total of 42 studies focused on these areas, of which just six were in a health-care context. Again, many studies examined more than one outcome variable, and many also examined wider models of engagement that included antecedents.

The relationship between engagement and higher-level performance at the unit, team or organisational level was explored eight times. These showed support for the notion that engagement is associated with performance; however, most studies used individual perceptions of performance outcomes rather than objective performance data, and only a small number of studies used third-party data such as customer ratings. Five studies took place within a health-care context and examined quality of care at the team/unit level. The results of these studies were inconclusive.

At the individual level, 22 studies examined the link between engagement and individual task-related performance outcomes within the general workforce; of these, 12 used self-reported performance data, which can be subject to error. Ten studies used third-party performance ratings, such as by coworkers, supervisors or customers. These studies using multiple informants, and often also other complex methods such as longitudinal analysis or diary studies, all showed a consistent association between engagement and performance outcomes either directly or as part of a mediated relationship. Thus, we can conclude that there is substantial support for the association between engagement and individual task performance outcomes. Within the health-care context, two studies examined the link between engagement and individual performance and both showed a positive link.

Seventeen studies examined the link between engagement and extrarole performance within the general workforce. All these studies found a link between engagement and various aspects of extrarole performance including citizenship behaviour, innovative work behaviour, personal initiative, knowledge sharing and creativity. However, the majority of these studies were based on cross-sectional self-report data. Two studies examined the link within the health-care context and both similarly found a positive link; notably, one study by Hakanen et al. 235 conducted over 3 years and involving 2555 dentists found a positive link between engagement at time 1 and personal initiative at time 2. Three studies among the general workforce (but none within the health-care sector) found a negative link between engagement and counterproductive behaviour.

Overall, the strongest support was found for the link between engagement and individual in-role performance, as all studies showed a positive association between the two. Support was also found for a negative link between engagement and counterproductive performance outcomes, albeit most studies relied on cross-sectional self-report data. A link between engagement and higher-level performance outcomes was also found, but most studies relied on subjective performance evaluations.

Only six studies took place within the health-care context, and none of these was conducted in the UK. All of them used the UWES and all were cross-sectional self-report questionnaire surveys, except for one study which used a longitudinal design235 and one that used a combination of methods: structured observations and survey data. 234 The inference of causality from most of the studies is therefore problematic. The most persuasive study, that of Abdelhadi and Drach-Zahavy234 in Israel, showed that, among a sample of 158 nurses, engagement was positively associated with nurses’ patient-centred care measured by structured observations, and that engagement mediated the relationship between service climate and patient-centred care. Hakanen et al. ’s235 longitudinal study of 2555 Finnish dentists showed that engagement and the use of personal initiative were positively and reciprocally related at two time points.

Synthesis of findings for research question 3: what approaches and interventions have the greatest potential to create and embed high levels of engagement within the NHS?

Our third research question concerned the link between approaches and interventions and engagement within the general workforce and within the health-care context. We excluded demographic variables and personality variables from this analysis, as neither constitute an approach or intervention, although we did consider these when they were relevant for understanding a wider model.

Our analysis showed that the antecedent factors fell under five broad headings:

• Individual psychological states, such as experienced psychological safety or availability. Forty-one studies within the general workforce and 11 studies within the health-care context fell under this heading. A very wide range of psychological states was investigated. A number of these studies used complex methods such as diaries, time-lagged surveys or supervisor/employee dyads, lending weight to the overall finding that the states of self-efficacy, resilience and personal resources generally are positively associated with engagement. These psychological states were examined in terms of their direct association with engagement and as part of wider mediated/moderated models. There was also evidence that positive affect, psychological empowerment, experienced meaningfulness, safety and availability, job satisfaction, recovery/relaxation experiences and in fact a wide range of other psychological states may be relevant for engagement. Broadly, these indicated that positive states are more likely to be linked with engagement than are negative states, and that relaxation and recovery are likely to raise engagement levels. However, one study using the affective shift model121 suggested that it is in fact the interplay between negative and positive affect that is relevant to engagement.

• Experienced job design-related factors, such as task significance, variety, meaningfulness, autonomy, job demands and job resources. Forty-three studies within the general workforce and 22 within the health-care sector came under this heading. A large number of studies, including those using complex methods, have found evidence that a range of job resources are associated positively with engagement. However, mixed results were obtained with regard to job demands; it may be that challenge demands raise engagement levels and hindrance demands lower engagement, but further research is needed to assert this with confidence.

• Perceived leadership and management factors, such as leadership style, authentic leadership and perceived supervisor support. Twenty-eight studies within the general workforce and eight within health care were included. Generally, studies falling under this heading found a positive link between what might be regarded as positive or enabling approaches to leadership, such as supervisory support, ethical leadership, authentic leadership, charismatic leadership and trust in leaders, and follower engagement. This included several studies using complex methods such as diary studies or involving multiple respondents. Conversely, two cross-sectional studies found negative associations between more negative aspects of leadership, such as uncivil or abusive supervision, and engagement.

• Individual perceptions of organisational and team-level factors, such as perceived organisational support, organisational mission, climate or culture and perceptions of colleagues and team. Forty-one studies within the general workforce and 12 within health care were included. A number of studies, including one using complex methods, found a link between perceived organisational support and engagement, although the findings relating to the link between the psychological contract and engagement were inconclusive. Researchers have studied a very wide array of variables under this heading, with many being the focus of just one study, so drawing firm conclusions from these is difficult. A number of studies found links between team-level and individual-level engagement, several using complex methods. It would seem that there is provisional evidence that the engagement levels of teams and individuals within them are associated. Further support is lent to this conclusion by the small number of studies that have examined negative aspects of the organisational environment, such as bullying and harassment, coworker incivility and interpersonal conflict, which have all found the experience of these to be negatively linked to engagement.

• Organisational interventions or activities, such as specific training and development courses or communication activities. Seven studies within the general workforce and two within health care were included. A very small number of studies published in peer-reviewed journals focus on evaluating interventions, which, arguably, is the aspect of engagement that is of most interest to practitioners. Given the scarcity of studies, their individualistic nature, their methodological limitations and the range of interventions studied, it is difficult to draw any robust conclusions from this body of work. However, some studies using complex methods have yielded some potentially interesting findings. One study by Bishop282 found that participation in an offsite programme focused on the true meaning of caring raised engagement levels among 17 nurses and another study among the general workforce also found that mindfulness training raised engagement levels. 263 Brummelhuis et al. 244 found that new ways of working (choosing where and when to work) appeared to raise engagement levels, and Carter et al. 217 found that participation in a forum theatre training intervention buffered the drop in engagement levels following the announcement of a merger.

With regard to health-care specifically, we found that 42 studies met the inclusion criteria; two of these took place in the UK and all used the UWES to measure engagement. Several of these used complex methods such as diary studies, longitudinal surveys or mixed methods. Considering the findings that might be of most help and relevance to health-care practitioners, a number of complex studies which considered the association between job design factors and engagement are of potential interest. 96,290 These suggest designing jobs in such a way that health-care workers who perceive their employer to be providing them with the resources they need to do their work are more likely to be engaged. Weigl et al. ’s300 time-lagged survey of 416 hospital physicians in Germany also found a link between job control, working relationships and active coping with engagement. Most of the eight studies looking at aspects of leadership and management were cross-sectional. However, Gillet et al. ’s133 study involving nurse/supervisor dyads in France found a link between transformational leadership and follower engagement, and Hornung et al. ’s134 two-wave study of 142 doctors in Germany found that leader consideration towards employees led to higher engagement levels.

A small number of studies used complex methods to examine individual perceptions of aspects of organisational and team-level factors and engagement. Notable among these are the study by Abdelhadi and Drach-Zahavy,234 whose mixed methods study of 158 nurses in Israel found a link between service climate and engagement, and that by Weigl et al. ,300 whose time-lagged survey showed that work relationships were linked with engagement.

Disappointingly, only two studies in the health-care context examined specific interventions. Although Bishop’s282 study found that participation in an offsite programme focused on the true meaning of caring raised engagement levels among 17 nurses in the USA, Rickard et al. 293 did not find that nurses’ and midwives’ engagement levels in Australia were influenced by participation in a workload intervention exercise.

In conclusion, studies of antecedent factors have lent most support to the potential relevance of the following factors for raising engagement levels among individuals, although the limitations with regard to health-care workers outlined above should be borne in mind:

1. positive psychological states, notably self-efficacy, resilience and personal resources

2. job-related resources and jobs enabling individuals to experience meaningfulness, safety and availability

3. positive and supportive leadership approaches, including supervisory support, ethical leadership, authentic leadership, charismatic leadership and trustworthy leaders

4. perceived organisational support

5. team-level engagement

6. participation in training or development interventions designed to enhance personal coping strategies, resilience, or interventions allowing individuals choice and discretion in ways of working.

Summary of findings for research question 4: what tools and resources would be most useful to NHS managers in order to improve engagement?

The purpose of the grey literature review was to try to achieve inclusivity of any relevant materials to this evidence synthesis in order to enhance rigour and overcome bias and, specifically, to address research question 4, to consider what materials and tools from this wider resource might of relevance to practitioners in the health-care context. Disappointingly, very little evidence from the review of grey literature was helpful in answering this research question. From six sources of practitioner materials we identified 14 items which covered a range of elements associated with engagement drawn from a number of countries, sectors and organisation sizes. Given the nature of the study methods, none of them was able to establish causal links between the particular interventions, attitudes or behaviours being analysed and subsequent improvements in engagement. The findings were instead correlations or associations. None of these items was considered to be of a sufficiently high quality to include in the main synthesis report.

In contrast to the most common approach in the academic literature, where the concept of engagement is perceived as a positive psychological state, the review of these practitioner materials suggested the majority of definitions used in practice consider engagement as a general positive attitude towards the organisation, rather than an experience related to work activities or the job role. In analysing the materials, seven key approaches emerged suggesting how engagement can be improved. These were:

• Senior leadership. Practitioner studies showed an association between positive perceptions of or trust in leaders and increased engagement, although one study found that higher levels of employee engagement were associated with lower ratings of senior manager effectiveness.

• Role of the line manager. The role and behaviour of the line manager was one of the key factors associated with employee engagement. Types of behaviours shown to be correlated with engagement were clear and respectful communication, recognising and involving team members, and being supportive and approachable.

• Appraisal, performance management and training. Good-quality appraisals, having performance development plans and being able to undertake training and development opportunities were shown to be linked with higher levels of engagement. One important caveat was that a poor appraisal may be linked to lower levels of engagement than having no appraisal at all.

• Meaningfulness. In one study, meaningfulness, defined as ‘the extent to which employees find meaning in their work . . . where people can see the impact of their work on other people or society in general’ (p. 23),193 was shown to be a relevant factor associated with high levels of engagement.

• Employee voice. This refers to the opportunities employees have to give input into decisions affecting their work and to be properly consulted. It was identified as a strong driver of engagement in a number of practitioner studies.

• Team working. Being part of a well-structured team that has shared and clear objectives was associated with increased levels of engagement. Other related factors such as perceived organisational support and coworker quality were also found to be associated with high levels of engagement.

• Work–life balance. There was evidence that people on flexible contracts, those satisfied with their work-life balance and those feeling that their work–life balance was supported by their employer demonstrated higher levels of engagement.

These approaches are explored in much greater depth in a series of practitioner outputs provided in Appendices 7–16 of this report. Although of interest and relevance to a practitioner audience, the fact that the items identified in the grey literature search did not meet the quality threshold means that these findings should be treated with a degree of caution.

Overall synthesis of findings

In this synthesis, in order to address our research questions, we have examined the antecedents of engagement separately from the outcomes, and we have further broken down and analysed separately the findings relating to each potential outcome and antecedent. However, it should be noted that the majority of the studies reviewed in this report have in fact examined a range of antecedents and outcomes, and that in many instances both antecedents and outcomes have been examined within the same study. Engagement itself has been treated as an antecedent, mediator, moderator or outcome, depending on the focus of the study. It is beyond the scope of this rapid review to examine these holistic models in any detail, given their range and complexity. Nevertheless, it should be borne in mind when considering the overall findings presented in this report that a main focus within the literature has been on examining engagement as embedded within a broad network of factors.

In summary, we have found that engagement first appeared in the academic literature as a discrete construct in 1990. Over the past 10 years in particular, there has been a significant increase in interest in engagement, with the publication of a very rapidly growing volume of research findings. Uncertainties and disagreements over the meaning, nature and measurement of engagement have led to a splintering of the literature and the proliferation of different understandings and interpretations. Engagement has been viewed as a state, as a composite and as practice. Despite this, the academic field has come to be dominated by the work of the Utrecht Group and its associates worldwide, with ‘state engagement’ seen as a higher-order construct comprising vigour, dedication and absorption within the JD-R framework.

Some evidence emerged from our review of the literature to suggest that engagement is associated with beneficial outcomes for the individual, notably life satisfaction, organisational commitment and job satisfaction, while being negatively associated with turnover intentions and burnout. There was also some consistent evidence, from a range of studies using different methodologies, that engagement is associated with higher levels of performance at the individual level, and with extra-role performance. The literature on antecedents of engagement was diffuse; however, evidence emerged which showed that positive psychological states, job-related resources, positive leadership, perceived organisational support, team-level engagement and some interventions designed to foster engagement were associated with higher levels of individual engagement.

Implications for practice

The findings of this review have a number of implications for practice, notably within a health-care context. These implications should, though, be considered alongside a number of caveats. First, many studies are cross-sectional and based on self-reports, and therefore it is difficult to be sure of the direction of causality, or to determine whether or not common method bias is a factor. Second, because of publishing norms within the social sciences, replication studies are almost non-existent; consequently, many relationships between antecedents, engagement and outcomes are examined in single studies, so a cumulative body of evidence has not been assembled to support or refute particular propositions. Third, the majority of research within the engagement field has focused on engagement as a psychological state and has not examined issues of most interest and relevance to practitioners, such as the impact of initiatives aimed at raising engagement levels. Fourth, the amount of variance in engagement levels that has been found in research studies is in the majority of cases very small, even when the variance is statistically significant. Whether or not these differences in fact make a practical difference in an organisational setting is often unclear. Finally, the variety of ways in which engagement has been defined and measured means that there is a lack of comparability across the body of research on engagement that makes generalisation difficult.

Bearing these points in mind, this synthesis nevertheless sheds light on some aspects of engagement that are of relevance to practitioners. Specifically, the synthesis has found evidence that high levels of engagement are beneficial for both individuals and organisations, and therefore it is desirable for organisations to consider finding ways of raising levels of staff engagement.

The evidence synthesis suggests that six factors may be especially salient for raised engagement levels:

1. The positive psychological states of self-efficacy and resilience, and perceived personal resources. Organisations that find ways to foster these states among the workforce may help to bolster engagement levels. States such as these can be fostered through a range of initiatives including personal development and training that strengthen employees’ self-belief, or through mentoring or coaching on the job.

2. Job design features, including the provision of job-related resources and the design of jobs allowing individuals to experience the psychological states of meaningfulness, safety and availability through their work. It has been suggested that employees are more likely to experience these last three states when they work in jobs that enable them to express their authentic selves. This can be achieved through ensuring that there is a good ‘fit’ between the individual’s role preferences and his or her job. Line and senior managers can articulate the meaning and significance of the work that is being undertaken, and provide employees with the physical and intangible resources needed to perform the job successfully.

3. Positive and supportive leadership and line management. Studies have found a link between various types of leadership, including authentic, transformational, empowering, charismatic and trustworthy leadership behaviours, and engagement, and between supervisory support and engagement. These findings suggest that leadership behaviours have an important role to play in the experience of engagement on the part of followers, and that organisations that invest in leadership and supervisory/management training and development may enjoy higher levels of engagement among the workforce.

4. Organisational support. Studies have found that, where employees perceive high levels of organisational support, they are more likely to be engaged. Thus, organisations that demonstrate a genuine concern for their staff, provide staff with support to perform their jobs and support employees at a personal and individual level may enjoy higher levels of engagement.

5. Team-level engagement. Some studies have shown that there may be a link between the level of engagement enjoyed at team level and that of the individual employee. This is an emergent area of interest, but nonetheless there is some evidence that engagement spills over between teams and individuals. Therefore it would appear important for organisations to consider and seek to manage not only individual workers’ engagement levels but also those of teams and work units.

6. Participation in training or development interventions designed to enhance personal coping strategies, resilience or interventions allowing individuals choice and discretion in ways of working. Although this is another new and emergent area of research, there is some evidence from intervention studies that participation in programmes or initiatives aimed at fostering high levels of personal engagement, or that enable individuals to feel involved and empowered in aspects of their working arrangements that affect them, may help raise engagement levels.

Recommendations for future research

Despite the growing volume of research on engagement, our evidence synthesis has highlighted an urgent need for further research on a range of topics. Out of 5771 items identified in our search, only 172 empirical studies met the quality threshold, suggesting that a great deal of what has been written about engagement is at best incomplete or undertheorised.

1. There is a general need for further longitudinal research on both the antecedents and the outcomes of engagement within the health-care context specifically. Only a small proportion of the studies that were included in the evidence synthesis were based in this sector, and, in particular, only six studies have focused on the performance outcomes of engagement within a health-care context. Only two studies had been conducted within the UK focusing on the antecedents of engagement. There is little evidence drawn from longitudinal studies to date.

2. More research is needed that focuses on engagement as practice and, in particular, there is a need for more longitudinal studies that examine the impact of initiatives aimed at enhancing engagement levels within the context of health care. It would be useful to gain further insights into what interventions have the most impact and under what conditions.

3. Most of the high-quality evidence that is available lies outside the health-care context. The highly specific and professionalised nature of the health-care context, along with its external performance demands, means that some of the frameworks and findings identified in the wider literature may be more – or less – significant. More robust research is needed which applies and contextualises the more generic frameworks around employee engagement to the health-care context.

4. Further research could usefully examine the link between engagement and patient safety, and the quality of care. It was disappointing to note the dearth of high-quality research in this area.

5. More multimethod, qualitative or ethnographic research on engagement within health care would be welcome, allowing a greater sensitivity to the relevance of context.

6. Very little research within the engagement field has considered issues of diversity and equality. For instance, more research that investigates the antecedents and outcomes of engagement, as well as the experience of engagement, from the perspectives of employees from various backgrounds would be welcome.

7. Further studies that investigate the interaction of engagement at different levels – individual, work group/team and organisational – would shed light on the experience of engagement.

8. Research that evaluates the comparative salience of a range of different antecedents to engagement would be welcome; hitherto, studies have focused on a relatively limited range of antecedents, so there is a dearth of research that compares and contrasts the potential importance of a range of antecedents for engagement levels.

9. It would be useful to know more about the focus of individuals’ engagement – for instance, whether people are engaged with their job, their work team, their organisation or their profession – and what are the implications of this.

10. All research on the antecedents of engagement with a health-care context included in this review used the UWES to measure engagement. Quantitative studies that use other measures and conceptualisations of engagement to test out alternative perspectives on engagement that may be relevant within a health-care context would be welcome.

Conclusions

In this synthesis, we have reported on the evidence accumulated in relation to the meaning, antecedents and consequences of engagement as published in the English language since 1990. In conclusion, despite the enormous amount that has been written, there is in fact still very little about engagement that can be asserted with any degree of certainty; we do not really know what engagement means, how to measure it, what its outcomes are or what drives up levels of engagement.

The literature is fractured, with so many different meanings attached to the engagement that it does not make sense to talk of engagement as one single construct. Schaufeli62 captures some of the tensions within the literature on engagement when he notes that the prevailing academic definitions of engagement as a psychological state are very narrow, but that, by extending these to incorporate behaviours, the distinctiveness of engagement becomes lost. Jenkins and Delbridge78 bemoan the acontextualised writing of much of the engagement literature, and Guest18 notes that engagement as a composite construct risks being dismissed as a management fad. While Keenoy26 argues that the notion of engagement is inherently managerialist, others have commented on engagement’s potential ‘dark sides’: Bakker et al. 306 suggest that engagement may tip over into workaholism, and Schaufeli and Salanova93 note that high levels of engagement risk leading to burnout when the balance of give and take between employer and employee is disturbed. The sceptics’ view, that engagement adds little or nothing to our understanding of workplace attitudes over and above more established constructs such as commitment and satisfaction, has not yet been fully disproved. 19,27

What can be said now is that there is a body of evidence which lends some support to the view that high levels of engagement are beneficial for individuals and employers, and that aspects of what might be considered good management and leadership practice may serve to raise engagement levels. However, even in studies where statistically significant relationships have been found between a range of antecedents, a range of outcomes, and engagement, the degree of substantive difference that is ‘explained’ in relation to engagement is often negligible and may be of relatively little practical concern in the workplace.

In order to be sure that advice given to practitioners is founded on best evidence, there is a significant need for further research on the topic. Of particular relevance here is the fact that there is a dearth of research on engagement set within health-care organisations in the UK published in academic journals. Given that much of the extant research on engagement does not take account of context, it is difficult to be sure of the relevance and applicability of current findings for this setting. Despite this somewhat pessimistic conclusion, the topic of employee engagement continues to show significant promise as an area for research and practice. There is much scope for further research that seeks to develop and extend current conceptualisations and theorisations of engagement through investigations that take greater account of the organisational and political contexts within which engagement is enacted and experienced.

Contributions of authors

Professor Catherine Bailey (née Truss, principal investigator and corresponding author) oversaw the project, contributed to the development of the search strategy, undertook the sifting of studies for inclusion and data extraction and prepared the final data for publication. She wrote Chapters 1, 3, 6 and 7 and produced the full report. Her areas of expertise include employee engagement, HRM and meaningful work.

Dr Adrian Madden (research fellow) contributed to the development of the search strategy, conducted the preliminary trials and main searches, managed the project databases, undertook the sifting of studies for inclusion and data extraction and prepared the final data for publication. He wrote Chapter 2 and assisted with the production of the full report. His areas of expertise include research and review methodologies, employment relations and meaningful work.

Dr Kerstin Alfes (co-investigator) contributed to the development of the search strategy, undertook the sifting of studies for inclusion and data extraction, advised on the interpretation of statistical data and prepared the final data for publication. She wrote Chapter 5. Her areas of expertise include employee engagement, strategic HRM and work intensification.

Dr Luke Fletcher (research assistant) undertook data extraction and prepared the final data for publication. He wrote Chapter 4. His areas of expertise include strategic HRM, employee engagement and well-being.

Dilys Robinson (co-investigator) contributed to the development of the search strategy, undertook the sifting of studies for inclusion, managed the process of search and extraction of the grey literature, oversaw the team members based at the Institute for Employment Studies and led on the development of practitioner materials arising from the project. Her areas of expertise include employee engagement in the health context, workforce planning and evaluation.

Jenny Holmes (research assistant) contributed to the development of the search strategy and undertook the sifting of studies for inclusion and data extraction. She wrote sections of Chapter 2 and contributed to the development of practitioner materials arising from the project. Her areas of expertise include employee engagement, HRM and employment policy.

Jonathan Buzzeo (research assistant) contributed to the development of the search strategy, and undertook the sifting of studies for inclusion and data extraction. He wrote sections of Chapter 2 and contributed to the development of practitioner materials arising from the project. His areas of expertise include labour market inequalities and research methodologies.

Professor Graeme Currie (co-investigator) provided academic advice on the development of research terms, the conduct of the sifting process and data extraction, and read the final report. His areas of expertise include HRM, leadership, identity and organisational change.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HS&DR programme or the Department of Health. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HS&DR programme or the Department of Health.

Terms associated with ‘engagement’ in the literature

“Absorption”

“Active engagement”

“Affective engagement”

“Agency AND engagement”

“Authentic engagement”

“Behavioural engagement”

“Cognitive engagement”