http://informahealthcare.com/jic ISSN: 1356-1820 (print), 1469-9567 (electronic) J Interprof Care, 2014; 28(6): 501–506 ! 2014 Informa UK Ltd. DOI: 10.3109/13561820.2014.939744

ORIGINAL ARTICLE

Using realist synthesis to understand the mechanisms of interprofessional teamwork in health and social care Gillian Hewitt1, Sarah Sims2 and Ruth Harris2 1

Cardiff School of Social Sciences, Cardiff University, Cardiff, UK and 2Faculty of Health, Social Care and Education, Kingston University and St George’s, University of London, Kingston Upon Thames, Surrey, UK Abstract

Keywords

Realist synthesis offers a novel and innovative way to interrogate the large literature on interprofessional teamwork in health and social care teams. This article introduces realist synthesis and its approach to identifying and testing the underpinning processes (or ‘‘mechanisms’’) that make an intervention work, the contexts that trigger those mechanisms and their subsequent outcomes. A realist synthesis of the evidence on interprofessional teamwork is described. Thirteen mechanisms were identified in the synthesis and findings for one mechanism, called ‘‘Support and value’’ are presented in this paper. The evidence for the other twelve mechanisms (‘‘collaboration and coordination’’, ‘‘pooling of resources’’, ‘‘individual learning’’, ‘‘role blurring’’, ‘‘efficient, open and equitable communication’’, ‘‘tactical communication’’, ‘‘shared responsibility and influence’’, ‘‘team behavioural norms’’, ‘‘shared responsibility and influence’’, ‘‘critically reviewing performance and decisions’’, ‘‘generating and implementing new ideas’’ and ‘‘leadership’’) are reported in a further three papers in this series. The ‘‘support and value’’ mechanism referred to the ways in which team members supported one another, respected other’s skills and abilities and valued each other’s contributions. ‘‘Support and value’’ was present in some, but far from all, teams and a number of contexts that explained this variation were identified. The article concludes with a discussion of the challenges and benefits of undertaking this realist synthesis.

Interprofessional practice, realist synthesis, teamwork

Introduction Interprofessional teams are a common feature of modern health and social care, where they are perceived as a means to enhance care quality and efficiency and patient safety and therefore strongly advocated within the healthcare policy of many countries (Reeves, Lewin, Espin, & Zwarenstein, 2010). Increasingly, patients have complex and long term conditions that require treatments from a range of health professionals and good quality care is dependent upon those professionals collaborating together in teams. Although there are numerous definitions of ‘‘teams’’, a general consensus exists that they are ‘‘comprised of a small, manageable number of members with an appropriate mix of skills and expertise, who are all committed to a meaningful purpose and have collective responsibility to achieve performance objectives and outcomes’’ (Harris et al., 2013, p. 22). A large body of research on interprofessional teamwork exists, much of which is descriptive and unempirical (Reeves et al., 2010). Furthermore, the rapidly changing health and social care landscape in the UK means the need persists for innovative research that will inform the development and management of increasingly complex interprofessional teams. The authors and colleagues therefore adopted the realist approach (Pawson & Tilley, 1997) in a

Correspondence: Professor Ruth Harris, Faculty of Health, Social Care and Education, Kingston University and St George’s, University of London, Kingston Hill Campus, Kingston Upon Thames, KT2 7LB, Surrey, UK. E-mail: [email protected]

History Received 5 November 2013 Revised 20 February 2014 Accepted 25 June 2014 Published online 21 July 2014

multi-method study of interprofessional teamwork along the stroke care pathway (the Teams Study) (Harris et al., 2013). The realist approach was developed to evaluate complex social interventions, namely programmes that offer one or more resources, but depend upon people’s responses to the resource to generate the anticipated outcomes (Pawson, Greenhalgh, Harvey, & Walshe, 2005). Interprofessional teamwork was considered to be a complex social intervention because it provides individual professionals with a resource (the team and its members), but the impacts of teamwork depend on the ways in which individuals respond to their team membership. For example, therapists and nurses might respond to being in a team together by sharing their knowledge of a patient more frequently or in greater depth. Such a response is referred to as a ‘‘mechanism’’ and realist researchers seek to identify the mechanisms that underpin complex social interventions. Contexts that determine whether or not mechanisms are ‘‘triggered’’ for particular groups of people or in particular situations are also identified, along with context-dependent outcomes. Context-mechanism-outcome (CMO) configurations can then be generated and used to address the realist question of, ‘‘What is it about teamwork that works for whom, in what circumstances and why?’’ (Pawson et al., 2005). This novel way of interrogating interprofessional teamwork was used to identify the mechanisms of teamwork, thereby creating a conceptual framework for use in this study and others. As far as we are aware, this is the first use of the realist synthesis method in the teamwork literature and builds on earlier work with the method undertaken by Hammick,

502

G. Hewitt et al.

Freeth, Koppel, Reeves, and Barr (2007) on interprofessional education. This article introduces the realist approach to evidence synthesis and describes the realist synthesis undertaken for the Teams Study. It presents findings for one mechanism (‘‘support and value’’) as an example and concludes with reflections on the process of undertaking a realist synthesis. This is the first in a series of four articles reporting the findings of the realist synthesis. The second article reports the findings for the ‘‘collaboration and coordination’’, ‘‘pooling of resources’’, Individual learning’’ and ‘‘role blurring’’ mechanisms (Sims, Hewitt, & Harris, in press a). The third article reports the ‘‘efficient, open and equitable communication’’, ‘‘tactical communication’’, ‘‘shared responsibility and influence’’ and ‘‘team behavioural norms’’ mechanisms (Hewitt, Sims, & Harris, in press). The fourth and final article reports ‘‘shared responsibility and influence’’, ‘‘critically reviewing performance and decisions’’, ‘‘generating and implementing new ideas’’ and ‘‘leadership’’ mechanisms drawing overall conclusions from the findings of the synthesis and their implications for healthcare delivery and further research (Sims, Hewitt, & Harris, in press b). Realist synthesis Realist synthesis identifies and tests CMO configurations using evidence from the literature. For a comprehensive description readers are referred to Pawson, Greenhalgh, Harvey, and Walshe (2004). Briefly, realist synthesis:  Identifies the mechanisms that programme designers thought would underpin the intervention.  Tests those mechanisms using empirical evidence from the literature.  Identifies and tests other, unforeseen mechanisms that might underpin the intervention once it is implemented.  Explores which contexts ‘‘trigger’’ the mechanisms for which people and in which circumstances.  Identifies positive and negative outcomes of the intervention, depending on which contexts and mechanisms are present (CMO configurations).  Synthesises the evidence in order to refine the theory the intervention rests on. Realist synthesis is a more flexible and iterative process than conventional systematic review. Its focus, for example, is directed by the emerging evidence rather than being tightly defined at the outset and the reviewer iteratively develops the search strategy as the review progresses, meaning multiple, responsive searches are conducted. There are, however, two main stages to the search process. The first identifies the purported mechanisms that underpin the intervention, using diverse sources such as policy documents, editorials, other reviews and interviews with key informants. The second looks for empirical evidence that supports or refutes the mechanisms. The reviewer now looks for contexts that trigger the purported mechanisms and the outcomes they generate and also looks for unforeseen mechanisms and their associated contexts and outcomes. Another area in which realist synthesis differs significantly from conventional systematic review is in its approach to quality appraisal of potentially relevant studies. Instead of assessing methodological quality and judging a study’s acceptability for inclusion on that basis, the realist reviewer looks at the quality of the inference the author is making from their data and asks if their inference makes a credible contribution to the mechanism being tested (see Pawson, 2006 for further detail). Realist synthesis therefore draws on a much wider range of evidence than conventional systematic review and makes critical and cautious use of ‘‘methodologically weak’’ studies.

J Interprof Care, 2014; 28(6): 501–506

Table I. First stage search strategy: free text terms and operators. theor* OR feature* OR trait* OR characteristic* OR attribute* AND interprofessional OR multiprofessional OR interdisciplinary OR multidisciplinary OR transdisciplinary OR multi agency OR collaborat* OR integrate* AND teamwork* OR team work* OR team-work* AND health*

Methods The realist synthesis, conducted in 2008–2010, aimed to identify and explore the mechanisms of interprofessional teamwork. The initial review question was: ‘‘Through what mechanisms does interprofessional team working affect clinical outcomes and patient experience, and how does context influence those mechanisms?’’ The first stage search strategy (Table I), to identify the purported mechanisms of teamwork, was run through the electronic databases AMED, CINAHL, MEDLINE and IBSS with the English language limiter. The resulting 301 records were screened by the authors, who read any potentially relevant article in full. This included any type of article (theoretical, opinion, research, etc) that focused on the functioning or benefits of interprofessional teams in health and social care. Recent health policy documents were also read. Each author then independently identified provisional mechanisms of interprofessional teamwork. Discussion with the research team and study advisory group, which included senior academics and clinicians from a range of disciplines including social science, psychology, physiotherapy and nursing, on the meaning of ‘‘mechanism’’, led to the review question being modified slightly to ‘‘Through what mechanisms does team working affect outcomes and experience (patient, carer, staff and service), and how does context influence those mechanisms and outcomes?’’ The authors then pooled their provisional mechanisms and agreed and defined nine (Table II). These were circulated to the advisory group who, on the basis of their research expertise in teamwork, suggested a tenth mechanism of ‘‘Leadership’’. The second stage search aimed to identify empirical evidence that could be used to test the ten provisional mechanisms. Additional mechanisms were also sought. A new search strategy was developed, using free text terms and subject headings appropriate for each database. The search combined terms for inter/multi/trans-disciplinary or -professional with terms for team and teamwork and with health-related terms such as rehabilitation and community care. English language and study type limiters were used. A further four electronic databases were included: HMIC, Psychinfo, ASSIA and Scopus. Searches retrieved 1865 records, which were screened for reports of original research on teams that cared for adult client groups. Records where this information was ambiguous were retained. Studies of pre-qualification interprofessional education were excluded. A total of 762 records remained (Figure 1). Original research studies from the first stage search were retrieved, along with any relevant reviews. Reference lists of the latter were screened and potentially relevant titles followed up. Further, snowball sampling was undertaken throughout the review whereby the reference list of every article read in full was screened for further relevant articles. Two of the authors began reading the collected articles in full. Detailed inclusion criteria were not applied, but articles had to

A realist synthesis of interprofessional teamwork

DOI: 10.3109/13561820.2014.939744

503

Table II. Mechanisms of interprofessional teamwork. Mechanism Shared sense of purpose

Description a,d,g

Pooling of resourcesa,d,e Collaboration and coordinationa,e Efficient, open and equitable communicationa,f Shared responsibility and influencea,d,f Support and valuea Critically reviewing performance and decisionsa,d,g Generating and implementing new ideasa,g Individual learninga,d,e Leadershipb,d,g Tactical communicationc,f Role blurringc,d,e

Team behavioural normsc,d,f

Teams have an explicit and shared understanding of their objectives, values and vision which leads to a consistent approach to care and enhances team member motivation and commitment. Team working enables team members to pool their diverse knowledge, skills, experience, influence, resources and networks. This leads to a more comprehensive understanding of the patient and more efficient use of resources. Role clarity within the team allows members to collaborate and coordinate their work. This reduces duplication or omission of services, ensures team members’ skills are fully utilized and increases continuity of care. Team members can offer their opinion and challenge one another and all contributions are given due consideration. Free and efficient flow of information helps avoid error, builds trust and allows decisions to be rigorously debated. Team members can influence team decisions and share responsibility for them. This makes members more committed to implementing team decisions, gives junior staff a greater sense of responsibility and provides members with support. Members of a team feel supported and valued and trust and respect each other’s skills, knowledge and perspectives. This fosters motivation and commitment to the team and its goals and improves staff relationships. Team working provides opportunity for group reflection, giving the team space to critically review its performance and decisions and to share feedback. This can result in better quality decision-making, the resolution of problems and greater self-regulation. Team working provides opportunity for collective learning and development, including sharing ideas and identifying and implementing of new ways of working. This can improve team effectiveness and improve patient care. Team working provides individual learning opportunities and relationships which promote professional development and help foster mutual respect for other professions, resulting in a higher quality of patient care. An identifiable leader establishes the team’s culture, engages and motivates the team, ensures communication flows and creates a safe climate for constructive debate. Through this they elicit commitment to the team and its objectives. Team members control the amount or type of information they share with other team members for their own or what they perceive to be their patient’s advantage. This can avert open disagreements and help to negotiate team hierarchies or conflicting treatment models. A shared body of knowledge and skills between team members means that some elements of a professional’s roles can be taken on by others if needed. This overlapping of roles helps ensure the carryover of tasks when a professional is absent. It can also aid professional development and lead to greater continuity of patient care. Teams develop and share behavioural rules (explicit or implicit) which govern acceptable ways of working together, e.g. expected standards of practice, behaviour and attitude. Failure to follow the team’s norms can have negative consequences, including unpopularity or exclusion from the team.

a

Mechanism identified in first stage search. Mechanism suggested by advisory group. c Mechanism identified during second stage search. d Focussed search conducted. e Mechanism reported in second paper of series. f Mechanism reported in third paper of series. g Mechanism reported in fourth paper of series. b

report empirical studies from the health literature that addressed interprofessional teamwork and were relevant to one of the provisional mechanisms or suggested a new mechanism. Bespoke data extraction forms were designed to summarise articles and record CMO configurations identified in the study findings. Forms were also completed for excluded articles, with the reason for exclusion recorded. Most often this was because the article was not relevant, i.e. did not address any mechanisms; other reasons included lack of useful detail about mechanisms, unclear methods and not reporting original research. Early on, nine articles were read and discussed by the authors to check consistency of data extraction. Discussion of all articles read up to that point raised three more potential mechanisms (Table II) and these were defined and incorporated into the data extraction form (Appendix). As reading and data extraction progressed it became clear that following the realist synthesis method exactly as described by Pawson et al. (2004) was too ambitious with the resources available as numerous CMO configurations were being identified.

Instead, main findings pertaining to each mechanism were summarised. Information on contexts and on outcomes for patients, teams, staff and organisations was noted, but individual CMO configurations (as in Box 1) were not recorded. A narrative approach to synthesising the findings on mechanisms, contexts and outcomes could then be adopted. The data extraction form was amended accordingly. After reviewing 232 articles, of which 109 were included (Figure 1), evidence for each mechanism was synthesised by drawing together the information on contexts, mechanisms and outcomes from the data extraction forms. The aim was to test and develop the provisional definitions of the mechanisms (Table II). Synthesis of the evidence started at this point rather than after reviewing all 762 records because the realist synthesis method is an iterative and cyclical process, where reading, searching and synthesis occur together and inform each other. For each mechanism, relevant sections of articles were re-read and similarities and differences in their findings sought in order to build a comprehensive description of the mechanism, its role in

504

G. Hewitt et al.

J Interprof Care, 2014; 28(6): 501–506

Box 1. CMO configuration of ‘‘Support and value’’ mechanism.

1865 records identified

Records screened for original research and interprofessional education

87 records identified from reviews and snowball sampling

35 records retrieved from first stage search

1103 excluded 762 records remaining; 110 papers read in full

A study of cooperation and autonomy in two Dutch hospital teams (geriatrics and oncology) found that over the period of their existence (three years) trust had grown, so members knew each other better and trusted one another much more than they had in the teams’ early days. Consequently, initial reluctance to involve other members in decisionmaking had largely disappeared and the teams openly reflected on individuals’ decisions during patient discussions, where voicing criticisms of other members had become much more acceptable (Molleman, Broekhuis, Stoffels, & Jaspers, 2008). Long tenure of team (C) ! Trust between team members (M) ! Shared decision-making and critical reflection (O)

232 papers read 123 excluded 109 included

3 new mechanisms identified; evidence for mechanisms synthesised

Focused searches; 19 additional papers included

Figure 1. Flow chart of second stage search.

interprofessional teamwork and the contexts that triggered or inhibited it. Data extraction forms contained a section for notes on each study’s methodological quality and the robustness of the conclusions drawn. This information was drawn on during the synthesis, not to include or exclude studies, but to inform decisions regarding the strength of the contribution the evidence from that article should make to the synthesis. Further focussed searches were undertaken for mechanisms with weak evidence or only a small number of articles (Table II). Appropriate key words were identified and used to search the remaining records from the second stage search, the Journal of Interprofessional Care, MEDLINE and CINAHL. Relevant papers identified were then fed into the synthesis of the mechanism. Realist synthesis draws on the qualitative research principle of saturation, meaning that searching ceases when no new evidence for a mechanism is emerging. However, the review was halted following the focussed searches for pragmatic reasons, despite a large number of articles from the second stage search remaining unread. Whilst some evidence may therefore have been missed, sufficient had been reviewed to begin to explore and test the mechanisms identified.

Findings Support and value mechanism This section reports the evidence pertaining to the ‘‘Support and value’’ mechanism within the realist framework of context, mechanism and outcome. Evidence for the other twelve mechanisms are reported in later papers in the series (Hewitt et al., in press; Sims et al., in press a, in press b). Sixty-one articles reporting 59 studies were included in the synthesis of this mechanism (Harris et al., 2013). Most used qualitative research methods and approximately half were conducted in the UK. Studies were undertaken in a variety of inpatient and community settings and with many types of interprofessional team, strengthening the relevance of the findings across health. Box 1 illustrates a CMO configuration of the ‘‘support and value’’ mechanism.

The provisional definition of ‘‘support and value’’ (Table II) posited that individuals respond to team membership by supporting one another and showing respect to and valuing each other’s skills and knowledge. Evidence to support this was found, but there were also reports of teams where the mechanism was clearly absent. The most common manifestation of the mechanism was where all team members’ contributions were seen as important and valued equally. This meant that everyone was trusted, their knowledge and skills were recognised, interest was shown in everyone’s opinions and members encouraged one another to participate (Craigie Jr & Hobbs, 2004; Pethybridge, 2004; Shaw, Walker, & Hogue, 2008). Social bonds were evident in some teams where members were friends as well as colleagues and interactions included humour and chat (Kvarnstrom & Cedersund, 2006; Reeves & Lewin, 2004). Other manifestations included praise between team members, backing one another up and sharing emotions when stressed (Apker, Propp, & Ford, 2005; Wilson, McCormack, & Ives, 2005; Wittenberg-Lyles & Parker-Oliver, 2007). When the mechanism was absent, individuals reported feeling their contributions were not valued and a lack of trust, obstructive behaviour and bullying were evident (Lingard, Espin, Evans, & Hawryluck, 2004; Rice Simpson, James, & Knox, 2006; SilenLipponen, Tossavainen, Turunen, Smith, & Burdett, 2004). A number of contexts suggested why ‘‘support and value’’ was present in some teams but absent in others. Another mechanism, ‘‘Efficient, open and equitable communication’’ was tightly linked to ‘‘Support and value’’, being both a context and an outcome of it. As a context, open communication meant all team members could contribute and get to know one another, helping build cohesive teams and respect across professions (Craigie Jr & Hobbs, 2004; Fear & de Renzie-Brett, 2007). Trust and respect were also enhanced by familiarity and this context was most prominent in teams with long-standing members or where members worked in close proximity (Reeves & Lewin, 2004; Rice Simpson et al., 2006). Role clarity and ambiguity were also important contexts, whereby the former promoted and the latter undermined trust within teams (Atwal, 2002; Gantert & McWilliam, 2004). Professional hierarchy was also influential as individuals in ‘‘lower status’’ roles sometimes felt less trusted and valued, whilst others did not necessarily believe that equality between the professions existed (Coombs & Ersser, 2004; Morris, Payne, & Lambert, 2007; Sargeant, Loney, & Murphy, 2008). The ‘‘Efficient, open and equitable communication’’ mechanism was an outcome for ‘‘Support and value’’ as well as a context. When elements of ‘‘Support and value’’ were present, team communication improved and when absent, communication was inhibited as individuals felt unable or unwilling to voice their opinions or challenge others (Field & West, 1995; Piquette, Reeves, & Leblanc, 2009). ‘‘Support and value’’ impacted on the ‘‘Collaboration and coordination’’ mechanism too, as supportive teams showed improved collaboration and unsupportive teams collaborated less well (Lingard et al., 2004; Salhani & Coulter,

DOI: 10.3109/13561820.2014.939744

2009). Teams where the mechanism was absent also exhibited tension and conflict and individuals reported feeling marginalised, unappreciated or frustrated (Coombs, 2003; Salhani & Coulter, 2009). When individuals did feel supported and valued, however, they reported improved self-esteem, a sense of belonging and an ability to cope with stress (Apker, Propp, Ford, & Hofmeister, 2006; Fear & de Renzie-Brett, 2007; Silen-Lipponen et al., 2004).

Discussion Evidence supported the provisional definition of the ‘‘support and value’’ mechanism (Table II) and suggested why its presence varied across teams (contexts). Although commitment and motivation specifically were not identified as outcomes, those described above underline the importance of the mechanism for both staff quality of working life and team functioning through its impact on communication and collaboration. The clear effect of the mechanism on staff, however, was not mirrored by any notable effect on patients, who were rarely mentioned, whether in terms of clinical outcomes or experience. Therefore, to answer the realist question of ‘‘What is it about teamwork that works for whom, in what circumstances and why?’’, the evidence synthesised in this paper demonstrates that supporting one another, showing respect and valuing each other’s skills and knowledge works well for interprofessional staff where all members of the team can communicate openly and get to know each other. This familiarity occurs in stable teams where members have clearly defined roles and work in close proximity with an absence of hierarchy enabling team members to trust each other and fully contribute to the work of the team. Realist synthesis is a relatively new approach to literature reviewing and as such it is useful to consider how it adds to the reviewer’s toolkit. Using realist synthesis to explore the health and social care literature distilled interprofessional teamwork into an evidence-based framework of its candidate social processes (Table II) and illuminated the contexts and outcomes associated with those processes. Unlike conventional review methods, its purpose was not to pass judgement on interprofessional teamwork as a successful or failing policy by focusing on outcomes, but to provide a means of conceptualising and understanding the processes of teamwork in a way that carries potential for use in team development and service improvement. Pawson et al. (2004) caution, however, that realist review ‘‘requires sustained thinking and imagination to track and trace’’ mechanisms through the literature and we would concur with this. We found it an intellectually challenging process, particularly as there are few published syntheses to provide guidance. Time was needed in the early stages of the review to debate and clarify our understanding of ‘‘mechanisms’’ and how they related to teamwork. We returned to this frequently throughout the review as our understanding was often tested, particularly by the way in which mechanisms interlinked: a mechanism could trigger one or more others, so the first mechanism was a context for the second and the second an outcome for the first. Box 1 illustrates this as the outcomes, shared decision-making and critical reflection, are also mechanisms themselves (‘‘Shared responsibility and influence’’ and ‘‘Critically reviewing performance and decisions’’). This issue has been noted by Byng, Norman, and Redfern (2005) after undertaking an empirical realist evaluation, where they had difficulties deciding whether a factor was a context or a mechanism, or indeed both. A related issue we faced was the difficulty of clearly differentiating the mechanisms of teamwork. We sometimes felt a degree of overlap existed between them, which meant that a process described in an article could be ascribed to more than one

A realist synthesis of interprofessional teamwork

505

mechanism. For example, if a team were discussing a patient and sharing what they knew of the patient’s home circumstances, was that ‘‘Efficient, open and equitable communication’’, ‘‘Pooling of resources’’ or both? This, coupled with how tightly interlinked the mechanisms appeared to be, meant that at times it seemed somewhat artificial to separate out teamwork into such discrete processes. A brief reading of the interprofessional teamwork literature highlights how complex and context-dependent it is, so to impose the degree of clarity and order on it that realist synthesis does may not accurately reflect the realities of teamwork and produce a misleadingly simple framework. We also found it difficult to make sense of the multiple CMO configurations we identified in the papers and ultimately found it helpful to step back from that level of detail and instead summarise information on each mechanism and its contexts and outcomes separately. This made the process of synthesising each mechanism more straightforward as the information on each element (context, mechanism or outcome) could be easily drawn together to look for patterns, similarities and differences. Examples of some of the CMO configurations highlighted in the synthesis are included in the three subsequent papers in this series (Sims et al., in press a, in press b; Hewitt et al., in press). The inclusive and wide-ranging nature of realist synthesis make it a substantial undertaking, particularly in a field where the literature is profuse, and Pawson et al. (2005) acknowledge that in many cases the limits of time and funding will determine when searching ceases. The volume of literature we faced was problematic and we found it difficult to decide when to stop searching and reading. This synthesis, however, was not a discrete research project, but the first stage of a wider study and as such was limited by the time available. Two researchers worked fulltime on the review for a year, supported by other research team members, but even at this level of human resource it was not possible to follow the realist synthesis method exactly as described by Pawson et al. (2004). They recommend, for example, searching across disciplinary boundaries to learn how mechanisms work in other fields, but we restricted our searching to health and even then could not read all the records we identified (Figure 1). We have therefore potentially weakened the synthesis by omitting evidence from other fields and failing to ‘‘maximize learning across policy, disciplinary and organizational boundaries’’ (Pawson et al., 2005). One strategy to narrow the focus of a realist synthesis is to select some of the provisional mechanisms identified in the first stage search to explore in depth and not pursue the others. The mechanisms we identified, however, were to form an analytical framework for the rest of the study, so such a strategy was not appropriate for this synthesis. Realist synthesis offers researchers a logic through which to explore policies and interventions and gains strength from its pluralism and flexibility (Pawson et al., 2005), but it is not an approach to be undertaken lightly. Pawson et al. (2004) explicitly caution that it is not for novices as it requires broad knowledge of disciplines and methodologies and skills in searching for and assessing evidence. Despite the intellectual and practical challenges encountered, however, we found realist synthesis a useful means of interrogating the large literature on interprofessional teamwork. It helped us articulate the social processes and actions that potentially constitute teamwork and provided an evidencebased analytical framework for the subsequent empirical study.

Acknowledgements We are grateful to Sally Brearley, Vari Drennan, Geoff Cloud, Nan Greenwood, Fiona Jones, Mark Joy, Lalit Kalra, Ann Mackenzie, Fiona Ross, Sally Redfern and Scott Reeves for their support throughout this work.

506

G. Hewitt et al.

Declaration of interest This project was funded by the National Institute for Health Research (NIHR) Health Services and Delivery Research (HS&DR) programme (project number 08/1819/219). The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the HS&DR programme, NIHR, NHS or the Department of Health. The authors report no conflicts of interest. The authors are responsible for the writing and content of this paper.

References Apker, J., Propp, K.M., & Ford, W.S.Z. (2005). Negotiating status and identity tensions in healthcare team interactions: An exploration of nurse role dialectics. Journal of Applied Communication Research, 33, 93–115. Apker, J., Propp, K.M., Ford, W.S.Z., & Hofmeister, N. (2006). Collaboration, credibility, compassion, and coordination: Professional nurse communication skill sets in health care team interactions. Journal of Professional Nursing, 22, 180–189. Atwal, A. (2002). A world apart: How occupational therapists, nurses and care managers perceive each other in acute health care. The British Journal of Occupational Therapy, 65, 446–452. Byng, R., Norman, I., & Redfern, S. (2005). Using realistic evaluation to evaluate a practice-level intervention to improve primary healthcare for patients with long-term mental illness. Evaluation, 11, 69–94. Coombs, M. (2003). Power and conflict in intensive care clinical decision making. Intensive & Critical Care Nursing, 19, 125–135. Coombs, M., & Ersser, S.J. (2004). Medical hegemony in decisionmaking – A barrier to interdisciplinary working in intensive care? Journal of Advanced Nursing, 46, 245–252. Craigie Jr F.C., & Hobbs, R.F. (2004). Exploring the organizational culture of exemplary community health center practices. Family Medicine, 36, 733–738. Fear, T., & de Renzie-Brett, H. (2007). Developing interprofessional working in primary care. Practice Development in Health Care, 6, 107–118. Field, R., & West, M. (1995). Teamwork in primary health care. 2. Perspectives from practices. Journal of Interprofessional Care, 9, 123–30. Gantert, T.W., & McWilliam, C.L. (2004). Interdisciplinary team processes within an in-home service delivery organization. Home Health Care Services Quarterly, 23, 1–17. Hammick, M., Freeth, D., Koppel, I., Reeves, S., & Barr, H. (2007). A best evidence systematic review of interprofessional education: BEME guide no. 9. Medical Teacher, 29, 735–751. Harris, R., Sims, S., Hewitt, G., Joy, M., Brearley, S., Cloud, G., Drennan, V., et al. (2013). Interprofessional teamwork across stroke care pathways: Outcomes and patient and carer experience. Final report, NIHR Service Delivery and Organisation Programme. Hewitt, G., Sims, S., & Harris, R. (in press). Evidence of communication, influence and behavioural norms in interprofessional teams: A realist synthesis. Journal of Interprofessional Care. Kvarnstrom, S., & Cedersund, E. (2006). Discursive patterns in multiprofessional healthcare teams. Journal of Advanced Nursing, 53, 244–253. Lingard, L., Espin, S., Evans, C., & Hawryluck, L. (2004). The rules of the game: Interprofessional collaboration on the intensive care unit team. Critical Care, 8, R403–R408. Molleman, E., Broekhuis, M., Stoffels, R., & Jaspers, F. (2008). How health care complexity leads to cooperation and affects the autonomy of health care professionals. Health Care Analysis, 16, 329–341. Morris, R., Payne, O., & Lambert, A. (2007). Patient, carer and staff experience of a hospital-based stroke service. International Journal for Quality in Health Care, 19, 105–112. Pawson, R. (2006). Digging for nuggets: How ‘bad’ research can yield ‘good’ evidence. International Journal of Social Research Methodology, 9, 127–142. Pawson, R., Greenhalgh, T., Harvey, G., & Walshe, K. (2004). Realist Synthesis: An introduction. ESRC Research Methods Programme Working Paper Series, University of Manchester. Pawson, R., Greenhalgh, T., Harvey, G., & Walshe, K. (2005). Realist review – A new method of systematic review designed for complex policy interventions. Journal of Health Services Research & Policy, 10, 21–34.

J Interprof Care, 2014; 28(6): 501–506

Pawson, R., & Tilley, N. (1997). Realistic evaluation. London: Sage Publications. Pethybridge, J. (2004). How team working influences discharge planning from hospital: A study of four multi-disciplinary teams in an acute hospital in England. Journal of Interprofessional Care, 18, 29–41. Piquette, D., Reeves, S., & Leblanc, V.R. (2009). Interprofessional intensive care unit team interactions and medical crises: A qualitative study. Journal of Interprofessional Care, 23, 273–285. Reeves, S., & Lewin, S. (2004). Interprofessional collaboration in the hospital: Strategies and meanings. Journal of Health Services Research and Policy, 9, 218–225. Reeves, S., Lewin, S., Espin, S., & Zwarenstein, M. (2010). Interprofessional teamwork for health and social care. Chichester: Wiley-Blackwell. Rice Simpson, K., James, D. C., & Knox, G. E. (2006). Nurse-physician communication during labor and birth: Implications for patient safety. Journal of Obstetric, Gynecologic, and Neonatal Nursing, 35, 547–556. Salhani, D., & Coulter, I. (2009). The politics of interprofessional working and the struggle for professional autonomy in nursing. Social Science & Medicine, 68, 1221–1228. Sargeant, J., Loney, E., & Murphy, G. (2008). Effective interprofessional teams: ‘‘Contact is not enough’’ to build a team. Journal of Continuing Education in the Health Professions, 28, 228–234. Shaw, L., Walker, R., & Hogue, A. (2008). The art and science of teamwork: enacting a transdisciplinary approach in work rehabilitation. Work (Reading, MA), 30, 297–306. Silen-Lipponen, M., Tossavainen, K., Turunen, H., Smith, A., & Burdett, K. (2004). Teamwork in operating room nursing as experienced by Finnish, British and American nurses. Diversity in Health & Social Care, 1, 127–137. Sims, S., Hewitt, G., & Harris, R. (in press a). Evidence of collaboration, pooling of resources, learning and role blurring in interprofessional healthcare teams: A realist synthesis. Journal of Interprofessional Care. Sims, S., Hewitt, G., & Harris, R. (in press b). Evidence of a shared sense of purpose, critical reflection, innovation and leadership in interprofessional healthcare teams: A realist synthesis. Journal of Interprofessional Care. Wilson, V.J., McCormack, B.G., & Ives, G. (2005). Understanding the workplace culture of a special care nursery. Journal of Advanced Nursing, 50, 27–38. Wittenberg-Lyles, E.M., & Parker-Oliver, D. (2007). The power of interdisciplinary collaboration in hospice. Progress in Palliative Care, 15, 6–12.

Appendix: Data extraction form Reviewer: EndNote library number: Reference: Objective of study: Description of paper (include location, setting, field of health, participants, brief methods): Mechanisms discussed: Shared sense of purpose Pooling of resources Collaboration and coordination Efficient, open & equitable communication Shared responsibility and influence Role blurring Tactical communication

Support and value Critically reviewing performance & decisions Generating and implementing new ideas Individual learning Leadership Team behavioural norms

Summary of main findings related to original mechanisms (including page nos. for reference); highlight which mechanisms findings relate to, any important context factors and whether outcomes relate to staff, team, patients or organization: Summary of main findings related to alternative mechanisms (including page nos. for reference); as above: Study quality issues: Any other comments (including reasons for exclusion, if applicable): Other references identified from paper:

Copyright of Journal of Interprofessional Care is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.