Team interventions in acute hospital contexts: protocol for the evaluation of an initial programme theory using realist methods

Introduction: Literature on multi-disciplinary healthcare team interventions to improve quality and safety of care in acute hospital contexts tends to focus on evaluating the success of the intervention by assessing patient outcomes. In contrast, there is little focus on the team who delivered the intervention, how the team worked to deliver the intervention or the context in which it was delivered. In practice, there is therefore a poor understanding of why some interventions work and are sustained and why others fail. There is little emphasis in the literature on how the team delivering the intervention might impact success or failure. Given that team is the vehicle through which these interventions are introduced, it is important to understand interventions from their perspectives. This research seeks to deepen understanding of enablers and barriers for effective team interventions. Using two case studies, we will evaluate previously developed initial programme theories to understand, what worked for whom, in what conditions, why, to what extent and how? Methods and analysis: A realist evaluation approach will be employed to test the previously formed set of initial programme theories. Two multi-disciplinary acute hospital team interventions in two different geographical and organisational contexts will be identified. In case study 1, a theory based approach to interviewing will be used. In case study 2, interview transcripts obtained using a semi- structured approach for primary research purposes will undergo secondary analysis. This will enable a more sensitive look at patterns and variations in patterns of multi-disciplinary team interventions. Researchers will first iteratively interrogate each respective dataset to identify the characteristics or resources present within the specific context that influenced how the team intervention worked to produce particular outcomes. Data will then be synthesised across contexts in order to produce middle range theories and thereby more generalisable insights.


Introduction
In acute hospital contexts, the primary purpose of most multi-disciplinary teams (MDTs) is to provide quality patient care in a co-ordinated way as patients require interventions from across several clinical areas and functions. Choi and Pak 1 describe MDTs as teams "where disciplines operate within their own boundaries" (p.225). Hospital staff may, however, identify with several multi-disciplinary teams including those that provide other functions; for example, governance and leadership, service or quality improvement, innovation or re-design and project or change management functions.
Given staff turnover in hospitals, and the dynamic system that requires teams to form and re-form, MDTs tend to be construed as multi-dimensional constructs with fluctuating team structures and processes depending on team composition, team purpose and other clinical teams and functions with whom they interact 2 . As a result, teams have many inter-dependencies and operate within uncertain conditions. Many are formed on the need for professional role representation and membership may be fluid and ad hoc 3-6 . By their nature, MDTs are therefore complex and operate within complex dynamic open systems 7 that reflect matrix management structures and hierarchical professional structures.
West and Lyubovnilova 8 caution that in some instances, it is difficult to define multi-disciplinary healthcare teams , i.e., they do not meet traditional definitions of what constitutes a 'team' and refer to their operation as "pseudo-like groups" (p. 8). These teams are further complicated by their composition of multiple healthcare professionals each having their own identity, culture, educational background and objectives. Consequently, professional boundaries, status and power differences will affect interprofessional collaboration [8]. These factors may help to explain why healthcare literature often lacks specificity in descriptors of multi-disciplinary teams involved in interventions to improve quality and safety of care 8,9 .
Teamwork failures are increasingly cited as significantly impacting on patient safety with concomitant costs to patients, hospitals and consequently the wider economy 10-12 .
There has been a sizeable growth in the area of implementation science, and the planning and implementation of team interventions to support the delivery of high quality and safe patient care 13,14 . Over the past decade in particular, team interventions have attracted increased research attention in comparison to the previous decade with an emphasis on interventions to improve quality and safety in areas including acute care 15 ; emergency departments 16 maternity units 17 , intensive care units 18 and trauma units 19 .
Whilst numerous studies address interventions by teams to improve quality and safety of care in hospitals 20-24 , there is still a dearth of high quality evidence on interventions to improve team effectiveness 13 .
Emphasis to date has largely been on whether effective teams yield positive results for patients or whether team interventions work or not to produce specific outcomes. In contrast, little attention is given in the literature to the team delivering the intervention and the context in which it is being implemented e.g. detail of team composition, team dynamics, team communication or organisational supports. As MDTs are the vehicle for improvement in quality and safety, understanding how and why team members engage with innovation and improvement interventions is important for both their implementation 25 and adaptation to open systems or 'real world' contexts 26 . It is recognised that there is a dynamic interplay between team intervention and context 27 . It is therefore necessary to understand contextual details of team interventions in order to understand the mechanisms that influence the outcomes of interventions. The dearth of such research, however, means there is little understanding of how and why the team itself impacts on the delivery of these interventions and their success or failure.
For the purpose of this research, team interventions have been defined as: An intervention where a team of two or more disciplines is trying to improve how the team delivers patient care-for example: quality improvement, service improvement or change initiatives; process re-design or team training events 28 .
As illustrated in their systematic review, Buljac-Samardzic et al. categorise team interventions into four primary categories: training tools, organisational re-design and programmes or a combination of these three 14 . These interventions tend to be multi-layered and complex as teams involved in their introduction are affected by cultural, leadership, financial and other organisational factors making them highly variable and context dependent 29 . Exploring team interventions and their effectiveness without appropriate consideration of context therefore seems meaningless. Attendance to the interplay between contextual factors and aspects of the intervention could illuminate why and how an intervention may be more impactful in one setting compared to another and should constitute valuable learning for intervention designers and for researchers.
Each hospital context has a uniqueness and a specific workplace culture and therefore its own specific requirements to support change and improvement 30 . Identification of patterns in these unique settings that can subsequently be extrapolated to general principles should help to guide implementation of multi-disciplinary team interventions in hospitals. Understanding the conditions under which teams tend to enact certain types of co-ordination mechanisms is critical to creating the conditions for effective performance and delivery of successful outcomes 31 .
This paper is the third in a series of papers which explore enablers and barriers to team interventions. Previous papers focused on the development of initial programme theories (IPTs) through a systematic search of the literature using realist synthesis 32 and interviews with key informants 28 . These IPTs describe the conditions in which multi-disciplinary team interventions appear to work best and why team interventions work best in these conditions The next phase of the research will elaborate on previous findings by testing these previously developed IPTs in two diverse acute hospital contexts. Findings will therefore be novel. This paper sets out the protocol for this phase of the research.

Context -realist evaluation
Having explored use of realist evaluation in studies relating to complex interventions in healthcare 7,33,34 , realist evaluation 35 Table 1 below for an explanation of realist terminology used in this protocol paper. Figure 1 below depicts an overview of the realist evaluation framework adopted for this research.
As outlined in Figure 1, this research involves three phases. Phase 1 and phase 2 have already been completed and are reported in detail elsewhere 28,32 . These are summarised below in order to provide background context for phase 3.

Phase 1 -systematic search of the literature
The first phase of the realist evaluation 32 involved a systematic search of the literature using realist synthesis. Consistent with Table 1. Realist terminology.

CMOC Definition
Context Those features of the situation into which programmes are introduced that affect the operation of programme mechanisms 36 .

Mechanism
A combination of resources offered and the participants reasoning in response

Outcome
The intended and un-intended consequences of the intervention.

Context-Mechanism-Outcome-Configuration (CMOC) -Patterns and variations in patterns
Demi-regularity Semi-predictable pattern of occurrences within the data

Initial Programme Theory
The programme architect's articulation of how the intervention is expected to lead to its effects and in which conditions it should do so Middle Range Theory "Theories that have a common thread running through them traceable to more abstract analytic frameworks" 35 . p. 123 the realist evaluation approach, this was driven by the primary researcher's (UC) own knowledge and experience of team interventions in an acute hospital context. The primary researcher's assumptions led to rough programme theories which formed the basis for the search strategy for the review. Relevant literature on team interventions in acute hospital contexts was explored via systematic search processes to determine what worked for whom in what conditions, why to what extent and how. Using realist synthesis, five plausible hypotheses were identified and presented in the form of context, mechanism, and outcome configurations (CMOCs) as per Table 2 below 32 .

Phase 2 -critical incident interviews (Use of key informants to refine plausible hypotheses)
Phase 2 of the research involved building of the IPTs by seeking the views of key informants (KIs) (hospital workers directly involved in the design or delivery of team interventions) on the plausible hypotheses which had been developed in Phase 1. Flanagan's Critical Incident Technique was adapted to seek the views of 17 KIs who were asked to recall both a positive and negative experience of a team intervention with carefully selected probes used to seek their views on the 5 plausible hypotheses. Adhering to RAMESES guidelines 37 for realist evaluation, data were analysed using a retroductive approach 38 from a total of 31 incidents. The plausible hypotheses were refined iteratively via a series of consultation sessions between the primary researcher and research team with a methodology expert panel i .
This phase of the research resulted in the production of seven IPTs outlined below in Table 3.
The next phase of this research (Phase 3) will involve testing of theses IPTs.

Ranking exercise
The seven IPTs developed via phase 1 and phase 2 of the research were first presented to a content expert advisory panel for discussion and refinement. Please refer to Table 4 below for further information on composition and expertise of this content expert advisory panel.
Following a brief presentation and discussion of the 7 IPTs with the content expert advisory panel, a ranking exercise was undertaken to enable them to prioritise five of the seven IPTs for testing.

Testing IPTs via case studies
In order to develop an in-depth understanding of how and why contexts interacting with mechanisms produce the intended and/or unintended outcomes in team interventions, it was agreed with both advisory panels that these five IPTs should be tested in two different acute hospital contexts using two different team interventions. This will result in further refinement of the initial programme theories in order to progress towards development of middle-range theories (MRTs) that are more widely generalisable.
As per Pawson, workplace interventions are 'active' rather than "passive programmes" continuously responding to contextual factors and emerging processes 39 . Exploring two different team interventions in two different contexts will allow for a broader array of factors, thus providing more rigorous testing of the IPTs. Conditions in one case may enable some mechanisms and consequently trigger intended outcomes whilst contextual conditions have potential to impact these mechanisms differently in the second case resulting in different outcomes. In essence, this will determine how the theories "hold up" within and across both contexts and will yield information that indicates why teams under certain conditions work (generative mechanism) and the conditions that are needed for a particular mechanism to work (specification of contexts) 35 .
It will be important to first understand how the team members respond to the respective intervention in each of the two case contexts in terms of their reasoning and the subsequent behavioural change that occurs. 27,28 . The team intervention and the conditions within which the intervention was implemented will have determined the outcomes of the intervention 40 . Team members within the same context may have different understanding of contextual conditions in that context, for example the impetus for change or detail of the specific intervention process. Their individual interaction and reasoning with these different contextual conditions therefore needs to be understood in terms of "generative causality" i.e. from their perspective how and why outcomes came about.
The underlying social and psychological drivers that drive both intended and unintended intervention outcomes for team i an interdisciplinary group of researchers and academics with a specific interest in, and experience in applying, realist methods  members in each of the two different contexts will therefore first be unpacked in the form of CMOCs. Patterns of regularity will be extrapolated from across team member interviews within each case and will be evaluated to discern whether they support, refute or require further refinement of the initial programme theories.
The refined programme theories (again in the form of CMOCs) from each of the individual cases will then be synthesised in terms of their usefulness and efficacy across both cases. The CMOCs will thus iterate backwards and forwards in this process of refinement towards development of a middle range theory (MRT): "Theories that have a common thread running through them traceable to more abstract analytic frameworks" 35 . p. 123 and "are close enough to observed data to be incorporated in propositions that permit empirical testing" 35 . p. 22 Moving CMOCs from initial programme theories to refined programme theories and subsequently to middle range theories in this iterative process will allow for the development of a generic set of principles that will be broadly transferable to other acute hospital contexts. The MRTs will provide valuable information to support design, facilitation and implementation of team interventions in acute hospital contexts.

Case study selection
For the most rigorous testing, together with the content and methodology expert advisory panels, the primary researcher and research team agreed to the choice of two different team interventions from two different hospitals, operating in two different health systems, one in Ireland and one in the USA. Table 5 below includes descriptors of the two cases.
Through the analysis and testing of the IPTs in two diverse cases geographic and healthcare contexts, it will be possible to develop a deeper understanding of the contextual enablers and barriers for team interventions at the team level, as well as exploring whether enablers and barriers differ according to the respective national healthcare contexts.
Case study 1 (Irish context) Intervention descriptor and primary goal. This team intervention was designed to change the process for daily general internal medicine (GIM) takeover of care from the daily post-call round in an academic teaching hospital context in Ireland. Prior to this team intervention, the practice was that all medical patients were automatically assigned to the care of the "GIM on-call team" for that night and remained under their care with consults requested from other specialties or requests made to take over care if deemed appropriate. The primary goal of the intervention was to ensure patients were cared for by the most appropriate medical specialty for their needs within 24 hours of admission (where possible) and to ensure that there was a more even distribution of workloads across specialties on a daily basis. This intervention was introduced because of a very large caseload for the respective medical specialty on the day post-call and delays in terms of takeover of care and/or inpatient consults from other specialties. These in-efficiencies were resulting in delays with clinical decision making and discharge planning and consequently resulted in protracted lengths of stay for medical patients. Larger caseloads also had potential to impact quality and safety of patient care.

Sample and recruitment. Members of this "GIM project team"
will be invited to opt-in and to register their consent to participate in the study by the primary researcher (UC) via e-mail correspondence two weeks in advance of scheduled interviews. As the primary researcher was involved in delivering the GIM project, interviews will be conducted by another member of this research team who is an experienced qualitative researcher (EMcA).

Data collection.
Interviews using an interview guide informed by the IPT will be used to collect data from the individual participants in case study 1. During the first part of these oneto-one interviews with the GIM project team members, information will be gathered about the team intervention, the

Descriptor
Experts on teamwork subject matter and quality and safety in healthcare.
Senior Healthcare Managers with operational expertise in acute hospital contexts.
Individuals who are currently conducting research in the Irish healthcare context or are renowned for their experiential knowledge in the subject.
Service users with knowledge of acute hospital contexts from a user's perspective Intervention descriptor To change the process for daily general internal medicine (GIM) takeover of care from the daily post-call round.
To strengthen inter-professional collaborative practice and facilitate practice transformation through development and implementation of structured inter-professional bedside rounds (SIBR).

Primary goal
To ensure care of patients admitted from the Emergency Dept. via the un-scheduled care medical pathway is taken over by the most appropriate medical specialty within 24 hours of admission where possible and that there is a more equitable daily distribution of workloads across all medical specialties to ensure safer and better quality of patient care.
To improve relational co-ordination (team communication and relationships) because of high Registered Nurses (RN) turnover; low patient satisfaction and high re-admission rates for patients.

Intervention driver Internal-Division of Medicine and hospital management
Internal and external-An academic practice partnership between the School of Nursing and the AHF care team with funding from the Health Services Resources Administration (HRSA).

Leadership Support for the intervention Leadership support included active participation and attendance of the Chief Executive Officer (CEO), Chief Operations Officer (COO) and Executive Clinical Director (ECD) at all meetings and workshops related to the intervention.
Leadership support included attendance during project initiation and close out and at a celebratory workshop.
Team Structure and Composition The intervention involved formation of a "GIM project team" Following an invitation from facilitators to the GIM group for representation from each medical specialty, medical consultants self -selected to participate in the project team. The CEO, COO and ECD were considered core team members and the team was facilitated by internal facilitators with expertise in lean methodology and organisational change.
The intervention involved formation of a "change team" comprised of inter-professional front-line care team members and grant team members.
This purposefully selected change management team comprising multiple disciplines (medical nursing and allied health professionals) from across the advanced heart failure (AHF) faculty and was facilitated by an external research team.

Methodology
The team intervention was underpinned by lean six sigma Define Measure Analyse Improve Control (DMAIC) methodology. An intensive data collection phase was followed by a workshop to co-design a new way of working and was subsequently followed by a series of monthly meetings and workshops interspersed with smaller stakeholder engagement sessions. A new process was trialled and iterated over three Plan Do Study Act cycles. This was followed by a six month control phase once the intervention was embedded.
Following a grant application process and formation of the change team, composition of the team, how the team operated and team processes. Subsequently, theory-driven interviews using an adapted form of the Teacher-Learner style interview technique 34 will be used to test the initial programme theories that have been informed by the extant literature and the data collected from KIs in Phase 1 and 2 of the research. Interviewees will be invited to comment on theories that are introduced by the interviewer thereby allowing them to confirm or refute them 34 and in this way, the IPT will be refined.
Please refer to extended data for a detailed outline of the interview format 41 .
Data will be collected at a location and time suitable for participants. All interviews will be audio-recorded. The qualitative data accumulated from these interviews will provide insight into how and why the multi-disciplinary team intervention was enabled or inhibited and give insight into the experiences of those affected by the intervention, as well as the intended and unintended consequences of the intervention 35 .
Case study 2 (US context) Intervention descriptor and primary goal. This team intervention was designed to strengthen inter-professional collaborative practice and facilitate practice transformation through development and implementation of structured interprofessional bedside rounds (SIBR) at a large medical centre in the Pacific Northwest, USA [42][43][44] . The primary goal of the intervention was to improve healthcare team, healthcare system, and patient outcomes for hospitalised patients with heart failure with particular emphasis on relational co-ordination (team communication and relationships) because of high staff turnover, low patient satisfaction and high re-admission rates for patients.

Sample and recruitment.
This case was identified by the authors as meeting research criteria and constituting a suitable team intervention in a contrasting context that will enable further testing of the initial programme theories. Once the appropriate members of the US research team were identified, an overview of the research including: the research question, the methodology and the IPT was given via a power point presentation by the primary researcher (UC). The goal of the secondary analysis proposed was explained in detail i.e. to examine the stability of the IPTs in terms of what enabled and/or inhibited the multi-disciplinary healthcare team intervention using the US case which differed in terms of health care context, team composition and intervention detail. Following a comprehensive discussion, it was agreed that the data from interview narratives (n= 16) conducted 31 with the change team upon completion of the intervention would be suitable for this purpose.
Data transfer. Data from the 16 interview narratives for case study 2 will be transferred as per a data sharing agreement. Confidential information will be protected through encryption.
Please refer to extended data 41 for a detailed outline of the interview format that was used in the primary study and which has been reported elsewhere 43 .
Organisational contexts case studies 1 and 2 Prior to commencing the research study, the researcher will develop an understanding of the broader hospital context in each case at the time of the intervention being completed. This will be done both reflexively by reviewing relevant documentation for example, relevant publications, and minutes of meetings or e-mail correspondence relating to the intervention and more pragmatically by developing field notes from meetings with appropriately identified staff. Details of the drivers of the intervention and how they aligned with the overall hospital's strategic plan, quality and safety agenda and key performance indicators will be sought.
Relevant data sets relating to the intervention, for example quantitative data relating to intervention impacts will be reviewed as required. Depending on how the evaluation evolves and requirement for deeper insights relating to the generative causation, further meetings may be scheduled to clarify specific pieces of information.
Data analysis case studies 1 and 2 Data analysis and synthesis will be informed by Gilmore et al.'s guidelines for data analysis and synthesis within realist evaluation (Phases 3-5) which are outlined in Table 6 below 45 . It is expected that all data will be extracted and analysed by June 2021.
Data preparation. Data from the audio files will be transcribed (CS1) and uploaded (CS1 and CS2) to NVivo 12 software 46 . Transcripts will be read and initial observations and annotations made.
CMOC extraction and elicitation. CMOCs will be used as the units of analysis. As per realist evaluation, best practice guidelines 37 , using deductive reasoning and inductive reasoning -CMOCs will be extracted and/or new CMOCs will be elicited from the interview narratives and coded to corresponding NVivo nodes that reflect the 5 IPTs or newly created nodes for additional CMOCs elicited. Table 6. Data analysis and synthesis.

Data analysis and synthesis within realist evaluation 45
Phase 3 Step 1 Data preparation Step 2 CMOC* extraction and elicitation Phase 4 Step 1 Using CMOCs* to refine IPTs Step 2 Collating evidence and refinement verification Phase 5 Step 1 Synthesis across studies for MRTs Using CMOCs to refine IPTs. Using deductive reasoning 45,47,48 , the IPTs will be tested to determine whether the perspectives and account of interviewees support or refute the IPT. In addition, via a process of inductive reasoning 45,47,49 , new information may result in refinement of the existing theories and the development of further new theories if a series of observations are made and new patterns of regularity emerge in terms of generative causation of outcomes and un-intended outcomes.
Collating evidence and refinement verification. A retroductive approach i.e. a process of moving backwards and forwards between the data within each case searching for clarification of support, refute or refinement will be used to determine how the CMOCs align with the original IPTs. All decisions and thought processes will be logged in linked memos for the purposes of transparency.
In order to ensure rigour and robustness of the process, a random sample of four narratives from each case study will be double coded by another member of the research team and co-author (ADB).

Synthesis across studies for MRTs.
Following data analysis within cases, data analysis will then move to synthesis and refinement of theories across cases in order to reach middle range theories. This will be informed by the results of data analysis within each respective case study and will incorporate a search for demi-regularities (semi predictable patterns occurring in the data) across the two case studies.
As the evaluation progresses, the methodology expert advisory panel may be contacted with regard to data analysis in order to assist and challenge decision making and in so doing, to optimise quality of research design and methodological rigour. The evaluation therefore will not progress in a linear fashion.
Further engagement with the content expert advisory panel may also be warranted as well as refinement based on focussed reviews of relevant literature. This iterative process of seeking advice at each stage from the expert advisory panel is a recommendation from the RAMESES guidelines for realist evaluation 37 .

Ethics and dissemination
Favourable ethical opinion has been received from University College Dublin Ethics Committee (HREC-LS-16-116397) for this research without requirement for further ethical review (LS-E-19-109) for testing in external contexts. Written permission was secured from the organisation involved in the first case study and recruitment of participants and other data collection did not begin until this was in place. Human subject's approval from the US-based institution was not needed given that the initial study was deemed exempt from the Human Subjects Review Board and only de-identified data were to be transferred. A data sharing agreement between the authors and the research team from the US academic institution was subsequently drawn up, agreed and signed by both parties.
De-identified transcripts were not shared until this was in place.
In accordance with University College Dublin's policy on data protection and storage, any paper versions of notes will be anonymised and will be stored securely and only accessible to the members of the research team.
Results will be disseminated via peer-review journals, national and international conferences and presentations to relevant stakeholders and interest groups for example: quality and safety governance groups, clinical audit and effectiveness committees and HSE Quality and Information Division (Ireland) and US research fora as deemed appropriate by the US research team. The findings will also be published in peer review journals.

Study status
All participants have been contacted and consent has been obtained for case study 1. Ethical approval has been obtained. Data for case study 2 has been anonymised and prepared for transfer for the purposes of secondary analysis. Ethical approval has been obtained to support this secondary analysis.

Discussion
Understanding the contextual conditions under which team interventions are undertaken and how these contextual conditions interact with team members' reasoning as individuals and as a collective will be helpful in order to understand how and why implementation of some interventions might fail or flourish. Realist evaluation is a complex research design and allows deep exploration and insights to be developed which consider the influence of contextual factors when exploring the enablers and barriers to multi-disciplinary team interventions in acute hospital contexts.
This work is novel as unlike other research which focuses on whether interventions work or not, it will explore how and why interventions work, what specific contextual and team factors enable team members' as individuals and as a collective to work effectively to produce successful outcomes. Given the importance of teamwork to delivering healthcare, a better understanding of these factors will be valuable for education, training and development of hospital teams.
Realist evaluation is theory driven and is in keeping with an interpretative process. It seeks to deepen understanding of 'what works, for whom, in what conditions, why, to what extent and how, as opposed to more traditional empirical studies, which look for more definitive answers of whether an intervention works or not 7 . Examining different contexts by using realist evaluation allows for a more rounded comprehensive approach and takes into account a broader range of perspectives.
Realist evaluation is being employed for this research because it is innovative and insightful and will allow deconstruction of the causal web of conditions underlying team interventions whilst grounding it in the 'messy reality' of healthcare. A realist evaluation yields information that indicates how the intervention works and the conditions that are needed for a particular mechanism to work and, thus, it is likely to be more useful than other types of evaluation in making recommendations for the design of team interventions.
Due to the variation in the contexts in which multi-disciplinary healthcare teams operate and the teamwork mechanisms enacted in those contexts, there may be many different outcomes from team interventions. Realism is not obsessed with the target of a single pass or fail outcome. Instead, use of this methodology to test the IPTs will enable a more sensitive look at patterns and variations in patterns of multi-disciplinary team interventions, for example: This research will therefore have a practical application for educators, managers, policy makers and decision makers in terms of providing recommendations on how to enable team effectiveness when delivering team interventions and thereby improve quality and safety in delivery of care for patients. This will help to ensure its relevance and application in ultimately improving team performance and enhancing patient safety cultures.
Outputs of this work will be applied directly to the implementation of interventions to improve team working in acute hospitals, will inform local work in healthcare transformation as well as influencing work on development of multi-disciplinary team interventions in the national and international context.
As the IPTs were informed by hospital workers directly involved in the design or delivery of team interventions and will be tested using case studies from two different hospital systems, this ensures that the middle range theory reached will be grounded in the reality of everyday experiences of hospital staff. Use of data from interviews in the two case studies will enable a comprehensive assessment of the team intervention from several perspectives. By understanding the contextual factors and the mechanisms through which outcomes are mediated, realist evaluators conclude that findings and recommendations are therefore more relevant 37 .
This phase of testing the IPTs will help to further topic development in terms of understanding how and why multi-disciplinary team interventions in acute hospital contexts are impacted by various contextual conditions in terms of generating specific outcomes. The engagement of hospital staff and expert advisory methodology and content panels consisting of senior academics, patient representatives and senior hospital managers as well as the researchers will help to ensure robustness, relevance and rigour of the research.
There is a diverse group of institutional partners involved in this research and it is intended in addition to peer reviewed publications that each research partner will utilise their existing networks and partnerships to discuss and disseminate findings. The influence and impact of this study will thereby extend beyond a single context.

Patient and public involvement
Two patient advocates were involved in the ranking of initial programme theories for testing as part of the content expert advisory panel.

Reporting guidelines
RAMESES II reporting standards for realist evaluations 50 will be adhered to for reporting purposes of this study.

Data availability
Underlying data No underlying data are associated with this article. This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Bronagh Blackwood
The Wellcome-Wolfson Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK This research is highly relevant considering the continuing interest in healthcare research in the evaluation of complex interventions: particularly those that involve a behaviour change from an established practice to adoption of a new approach. In such cases, the gold standard randomised controlled trial alone provides insufficient evidence to explain why a team-based intervention does or does not work. Thus this approach, that places the emphasis on evaluating the context into which a team is based and the mechanisms of their interactions, is highly important.
The protocol is well written, clear and robust in its presentation of the plan to conduct a realist evaluation of two team-based interventions. The authors present a clear overview of the literature and a logical justification for the need to approach the evaluation of team-based interventions using this method. The chosen case studies represent two different interventions in different contexts and in two countries with very different healthcare organisations and resources. This is a strength in that evaluation of the initial programme theories is being tested more broadly and the findings may be more relevant for developing a mid-range theory. Had the authors been able to sample an additional case study from a lower income country, it could have been more interesting and broadly applicable. Perhaps something to be taken forward in future research.
Is the rationale for, and objectives of, the study clearly described? Yes

Are sufficient details of the methods provided to allow replication by others? Yes
Are the datasets clearly presented in a useable and accessible format? Yes