An evidence synthesis of the international knowledge base for new care models to inform and mobilise knowledge for multispecialty community providers (MCPs)

The aim of this synthesis is to provide decision makers in health and social care with a practical evidence base relating to the multispecialty community provider (MCP) model of care. The synthesis, based on realist principles, will develop a “blueprint” which articulates how and why MCP models work, to inform design of future iterations of the MCP model. We believe this synthesis will support decision makers in a range of settings, by:

Specifically, the objectives of the synthesis are to:

A full list of programme theories, interventions and mechanisms will be extracted from the logic models developed by the Vanguards [11]. In accordance with our identified methodology, we will identify an a priori “best fit” framework which is both relevant to the context of integrated health and care services and meaningful to the intended audience for our final synthesis. We will use the a priori framework to code underlying programme theories of the current MCP Vanguards and mechanisms identified. These will be prioritised with the support of the Project Advisory Group. The logic models developed by the Vanguards typically follow the model recommended in the Magenta Book [12] which lends itself to the CIMO (context, intervention, mechanisms, outcomes) [13] framework with which we intend to plan the search strategy.

A systematic search will be conducted to identify research and practice-derived evidence between January 2000 and December 2016. This will achieve a balance that captures the historical legacy of MCP models but focuses on contemporary evidence. Sources will include bibliographic databases including MEDLINE, PreMEDLINE, CINAHL, Embase, HMIC and Cochrane Library; and grey literature sources such as the King’s Fund and Nuffield Trust.

Candidate search terms will be identified by analysing documentation, including logic models, from the current MCP demonstrator sites. These will be reviewed by the project team, with particular support from the Vanguards Relationship Lead; we will also seek advice from the Advisory Group setup to support the synthesis. The Advisory Group represents key stakeholder groups, with membership including senior leaders from three MCP Vanguards, a community NHS Trust, a Clinical Commissioning Group and NHS England; members of the public; and the University of Sheffield (the objectives of the Group are included as an Additional file 1).

Given the difficulties in searching for evidence in social science fields [14], particularly with regard to concepts related to integrated care (i.e. inconsistent definitions, changing terminology), we will check the reference lists of included evidence, to identify additional evidence. As evidence is likely to be distributed across a wide variety of professional and managerial evidence sources, we do not plan to use formal hand searching of a select list of resources. Instead, we will systematically follow up citation networks using both the Google Scholar “cited by” function and formal Web of Science citation searching.

The scope of the search will include international literature within a developed country context to ensure inclusion of relevant literature of appropriate mechanisms—defined as [15]: “the interaction between what the programme provides and the reasoning of its intended target population”. For example, a model based on mutuality, perceiving service users/patients as partners involved in decision making might involve a range of mechanisms relating to relationships with local communities. However, we acknowledge that findings from other health systems do not always transfer well to NHS settings [16] and may yield indicative rather than definitive findings.

We do not propose to include non-English language studies for the following reasons:

Based on early plans of existing MCPs, we have identified a core set of potential search terms which capture the common features and activities of MCP models of care (these are included as an Additional file 2).

We have described our methods as per Preferred Reporting Items for Systematic Review and Meta-Analysis for protocol (PRISMA-P) recommendations [17], and this checklist is included as an Additional file 3.

We aim to identify relevant evidence on the most significant (prioritised with our Advisory Group) contexts, mechanisms and outcomes relevant to the MCP demonstrator sites. As it is likely that there will be, as yet, little evidence explicitly on UK MCP models of care, we propose to explore the literature which could be described as the “intellectual heritage” of the MCP model (e.g. [18, 19]).

In terms of defining evidence, decision makers in large-scale change programmes need to draw on multiple types of evidence. We will adopt Williams and Glasby’s [20] definition which describes “evidence” from an evidence-based management rather than evidence-based medicine perspective, comprising empirical evidence from research; practice-based and experiential evidence from service delivery; and theoretical evidence. Our search strategy therefore will include research studies (trials and reviews), service evaluations and case studies, in addition to thought-leading papers.

We will focus on a core of highly relevant literature, using this to identify secondary terms and other relevant papers, in an iterative cycle. This approach, which is well accepted within a realist synthesis context [21], is methodologically stronger than an optimally sensitive search strategy which runs the risk of identifying such a large volume of marginally relevant and irrelevant papers, that time for analysis will be compromised by time spent sifting candidate references.

The population targeted by the evidence synthesis will include patients, carers and communities receiving care via interventions which feature in the MCP model of care. We will also focus on staff across health and social care involved in the design, delivery and evaluation of interventions delivered within the MCP model of care.

The synthesis will focus on the 14 MCP Vanguards within the NHS England’s Vanguards programme. These MCP Vanguards aim to develop extended primary care services, “offering some care in fundamentally different ways, making fuller use of digital technologies, new skills and roles, and offering greater convenience for patients”, essentially becoming “the focus of a far wider range of care needed by their registered patients” [1]. Primary care is a priority development area for local health economies, driven by the shift towards co-commissioning, integrated care and patient empowerment.

The MCP models of care aim to provide wrap-around and coordinated services for patients, which, whilst following some standard principles, will adapt to fit with the local context. The evidence base on system integration is variable, characterised by a lack of consensus on what constitutes integrated care [22‐24]. It is generally accepted that no single model or approach to integrated care can be applied universally [25, 26] and this is evident in the models developed in the 14 MCP demonstrator sites. A rapid desk-based analysis of the aims and objectives of the MCP demonstrators suggests that these models of care incorporate a number of different interventions at different levels (micro, meso and macro) within the system, including extensivist primary care services, multidisciplinary case management and social prescribing.

This study will review evidence relating to the stated outcomes of MCP models, as articulated in the MCP logic models. These relate to the quadruple aim: patient experience, the health of the population, healthcare costs and staff experience. The synthesis will include qualitative data and quantitative measures.

Retrieved papers will be managed using Endnote and will be screened by the Chief Investigator and Vanguards Relationship Lead, for relevance according to the mechanisms identified and prioritised in the a priori framework.

We anticipate including all evidence deemed to have satisfied the relevance criteria; several researchers (e.g. [27]) contend that exclusion of qualitative research on the basis of quality risks missing important insights. Studies will be appraised using standard tools [28, 29].

Data will be extracted using a standard form to capture key characteristics necessary to understand the context of the evidence; relevant concepts identified from the a priori framework and which have emerged from the evidence; and important findings on mechanisms and outcomes. Thematic analysis across the selected evidence base will identify confirming and conflicting theories which will form the basis of a draft conceptual model. NVivo will be used to manage data analysis.

To illustrate how the “best fit” framework analysis will work in practice, we conducted a rapid desk-based analysis of the MCP Vanguard applications which suggests multiple interventions and mechanisms are being developed at micro, meso and macro levels, e.g. extensivist primary care, multidisciplinary community teams, social prescribing (interventions), community assets and social capital (mechanisms). A number of potential themes also feature in guidance [11] from NHS England: design, evaluation, integrated commissioning, patient and community empowerment, technology, workforce, leadership and engagement.

Potential “umbrella” mechanisms and themes will be used as a framework for data extraction against which the literature on new models of care will be extracted and analysed. Within each of these umbrella mechanisms, we will derive if-then causal statements, based on the literature, that seek to unpick what exactly it is about these mechanisms that is likely to result in improved outcomes or process variables. For example we will be able to identify the putative active ingredients for extensivist services, accountability, reduced fragmentation etcetera that likely result in the intended (or indeed unintended) outcomes. We will seek preliminary verification of these if-then statements with a convenience sample of staff working within existing Vanguard sites. This convenience sample will be recruited from the current Vanguards; we will aim to represent a minimum of five different sites.

We will seek to verify the draft conceptual model with key stakeholders from within the MCP Vanguards and NHS England. This will be achieved via the project’s Advisory Group and a dedicated focus group with key stakeholders. This is an important phase of the project, ensuring that the model is fit for purpose and meaningful to decision makers and practitioners.

As the MCP model comprises a broad range of activities and mechanisms, the potential evidence base is vast and diffuse. We anticipate that the main types of evidence sources identified will be of four main types:

Only category 1 material is suitable for formal risk of bias assessment. Quality appraisal will be performed on a case-by-case basis using appropriate tools from the Cochrane Effective Practice and Organisation Group (EPOC) to examine bias in this category of studies (Suggested risk of bias criteria for EPOC reviews http://​epoc.​cochrane.​org/​sites/​epoc.​cochrane.​org/​files/​uploads/​14%20​Suggested%20​risk%20​of%20​bias%20​criteria%20​for%20​EPOC%20​reviews%20​2015%20​09%20​02.​pdf). However, we acknowledge that these tools are designed simply to explore potential bias and, unlike Cochrane risk of bias tools for randomised and non-randomised studies, have not had their validity demonstrated. The remaining three categories of evidence are anticipated to make either a theoretical or contextual contribution, and therefore it is not meaningful to undertake a formal quality assessment of such items. For these conceptual and contextual types of evidence, we will go beyond the conventional focus on the methodological quality of studies to employ a more reflexive approach to locate the contribution of each item within the overall body of evidence and to assess its specific contribution, according to accepted principles of realist synthesis. Where evidence is considered to be of marginal rigour, we will iteratively and purposively seek additional confirmatory or disconfirming evidence. Thus we will combine the rigour of the systematic review process for category 1 materials with the acknowledged trade-off between relevance and rigour for the remaining categories of material [21], whilst being sure to clearly delineate between materials selected for rigour, for context and for conceptual contribution.

This synthesis will employ best fit framework synthesis, as a rapid tool by which to facilitate the data extraction and analysis process, combined with realist synthesis principles to maximise the value of the interpretative process resulting in practicable and feasible recommendations for practice. Dixon-Woods [30] suggests that framework synthesis is “especially suitable in addressing urgent policy questions where the need for a more fully developed synthesis is balanced by the need for a quick answer”. The best fit framework synthesis methodology was developed by Carroll et al. [31] as a pragmatic variation on framework synthesis. Best fit framework synthesis introduces the deductive step of developing an a priori framework, thus “harnessing the recognised strengths of both framework and thematic synthesis” [32]. Best fit framework synthesis will be combined with a synthesis following realist principles which are particularly suited to exploring what works, when, for whom and in what circumstances. The realist principles are derived from work by Pawson and Tilley [33], who recognised the need for methods suited to the inherent complexity within change programmes and their evaluation. The realist synthesis approach [34] has been consolidated and extended in more recent studies such as Rycroft-Malone et al. [35]. The realist approach acknowledges that interventions do not necessarily transfer easily from one setting to another and offers deeper insights into the contextual factors involved in change. Following synthesis, we will offer practitioner relevant dissemination activities to mobilise knowledge and support decision makers [36].