Vertical integration of GP practices with acute hospitals in England and Wales: rapid evaluation

Article history

The research reported here is the product of an HSDR Rapid Service Evaluation Team, contracted to undertake real time evaluations of innovations and development in health and care services, which will generate evidence of national relevance. Other evaluations by the HSDR Rapid Service Evaluation Teams are available in the HSDR journal.

The research reported in this issue of the journal was funded by the HSDR programme or one of its preceding programmes as project number NIHR131295. The contractual start date was in October 2018. The final report began editorial review in July 2020 and was accepted for publication in October 2020. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HSDR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2022 Sidhu et al. This work was produced by Sidhu et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaption in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2022 Sidhu et al.

Summary of key points

• The evaluation reported here is of instances where organisations running NHS acute hospitals have taken on the responsibility of also running some general practices.

• This change in practice has been happening in several locations in England and Wales, with the first instances commencing in 2016, but it is not yet widespread practice.

• An acute hospital taking responsibility for running general practices is an example of ‘vertical integration’, that is, integration between organisations operating at different stages along the patient pathway.

• Vertical integration can occur in any area of the economy and can be seen as a way of attempting to reduce transaction costs and to reduce risks for the integrating entities. Vertical integration can make decision-making, monitoring and information-sharing more efficient. In health care, this might translate as better mutual understanding between general practitioners (GPs) and hospital specialists, better communication and flows of patient information between GPs and hospital specialists, and reduction of risks through better demand management.

• The NHS across England and Wales is working towards systems of care that offer better co-ordination across the primary and secondary care interface. One innovative approach to achieving stronger integration of care is via vertically integrating primary care organisations with secondary care organisations (i.e. co-ordination within a single management entity of staff, infrastructure, functions and activities that contribute to different levels of patient care).

• There has also been growing interest in extending the range of health-care services that are delivered in primary care settings. However, at the same time, in the face of growing patient demand combined with GP workforce constraints, the long-term sustainability of primary care has become an increasing focus of concern.

Background

This report presents a rapid evaluation of arrangements whereby NHS organisations operating acute hospitals have additionally taken over the running of general practices at scale in a few locations in the NHS in England and Wales (i.e. a new form of vertical integration). There is little systematic information on the rationale for, or desired impact of, such vertical integration in a UK NHS setting, or on why it is developing in some places despite not being an explicit part of current NHS policy in England or Wales. There is, correspondingly, limited understanding about how to implement vertical integration in a UK NHS setting (e.g. contractual/governance and commissioning issues) and the enablers of and barriers to its implementation. The number of examples of vertical integration is, nevertheless, growing and so it is timely to evaluate their implications.

The initial impetus for the evaluation of vertical integration was a desire to understand the logic of this form of integration in the provision of health care across primary and secondary care settings. However, as the study progressed, it became clear that a major driver of vertical integration is a response to difficulties with sustaining general practices and, therefore, local access for patients to primary health care.

In this chapter, we first describe the institutional context. We then locate the form of vertical integration that is the focus of this evaluation within the landscape of types of health-care integration, before going on to describe the current and recent past policy context relevant to such vertical integration. We finish the chapter by stating the aims of the evaluation reported here and outlining the structure of the rest of the report.

Institutional context

Within the NHS, acute hospitals do not usually run primary care services. Indeed, at the founding of the NHS in 1948, although hospital services were nationalised, GPs remained as independent practitioners. GPs were, and usually still are, contracted by the NHS to provide primary care medical services to the patients who register with them. GPs also have the responsibility of acting as ‘gatekeepers’ to other NHS services, including acute hospitals; however, GPs are not employees of the NHS. 2 This separation between primary and secondary care has been maintained through numerous NHS reforms, including the major reorganisations in 1974 and 1990. 3 Indeed, the 1990 reforms arguably deepened the divide by casting GPs in the role of fundholders and, therefore, ‘purchasers’ from secondary care providers (i.e. NHS trusts). In the past 20 years, the role of GPs as purchasers or commissioners of care for their patients from other NHS organisations, as individual practices, has reduced. However, as members of Clinical Commissioning Groups (CCGs) and primary care networks (PCNs) in England, GPs’ involvement in purchasing hospital care remains strong. At the same time, the desirability of better integration of patient care across primary and secondary care settings has become firmly established as a policy objective, as described later in Policy-drivers for integration.

Acute hospitals are providers of hospital-based emergency and/or elective specialist health care. In England, acute hospitals are run by a publicly owned organisation known as an ‘NHS foundation trust’ or an ‘NHS trust’ (hereafter referred to collectively as trusts). There are 152 acute hospital trusts in England. 4 The institutional set-up of the NHS in Wales is different from that of the NHS in England. In Wales, NHS acute hospitals are run by seven territorially defined Local Health Boards (LHBs). The services of acute hospitals are contracted for by both national and local NHS ‘commissioners’ of care. In England, the commissioning organisations are NHS England (London, UK) (nationally) and CCGs (locally). In Wales, the commissioners are NHS Wales (Cardiff, UK) (nationally) and LHBs (locally).

Until the last few years, hospital trusts in England had no role in running general practices. Although it remains most often the case that hospital trusts do not run general practices, in recent years, a small number of departures from this have occurred (i.e. a few hospital trusts in England have started to run general practices). In Wales, the LHBs directly run hospitals and community health services. LHBs also contract with independent general practices, but LHBs do not usually operate general practices directly. However, in recent years, this has started to change in a few locations in Wales. It is these recent changes in England and Wales that provide the initial stimulus for the evaluation reported here. Our focus is on those cases where organisations running acute hospitals have taken on general practices at scale (i.e. more than just one or two practices). Such vertical integration at scale is a relatively new phenomenon in the NHS – occurring over the period since 2015 – and the number of examples is currently small and scattered geographically. Nevertheless, interest in vertical integration is growing and during the study we became aware of new examples starting up.

The arrangements for general practices are essentially the same in both England and Wales; they provide primary medical care to their registered populations, which are effectively the total national population in aggregate. General practices are staffed by GPs and, increasingly, other health-care professionals (e.g. nurses and pharmacists). In England, as of September 2019, there were around 7000 general practices with 34,586 GPs (headcount), of whom there were 26,958 full-time equivalent qualified permanent GPs (i.e. excluding locums and trainees). 5 In Wales, as of September 2018, there were 421 general practices6 with 1964 GPs (headcount). 7

There are, in both England and Wales, three types of contract by which the NHS buys the services of GPs: (1) General Medical Services (GMS), (2) Personal Medical Services (PMS) and (3) Alternative Provider Medical Services (APMS) (Box 1). GMS contracts are the most common and are between GP-owned practices and NHS England (in England) or their LHB (in Wales); PMS and APMS contracts, which resulted from primary care reforms in 1997,8 are less common. However, the traditional GMS contract, where GP partners own the practice, has been declining in popularity. The UK Government’s 2019 GP Partnership Review9 stated that ‘The numbers of GP partners are falling as a proportion of the general practice workforce’ (contains public sector information licensed under the Open Government Licence v3.0) (although the report did not provide any relevant data).

In both England and Wales, there are many areas where recruitment and retention of GPs is a challenge. The number of GPs per 100,000 population has fallen in England since 2009 (from 66 then to 58 in 2018) and in Wales since 2013 (from 64 then to 63 in 2018) in a context, in both countries, of increasing demand from the population for health care, including primary care. 10

The evaluation reported here is of those instances where acute hospitals have taken on the responsibility for fulfilling GP contracts. This change in practice has been happening in several locations in England and Wales, with the first instances commencing in 2016; however, it is not yet widespread practice and so it is now timely to evaluate such arrangements.

Vertical integration

An acute hospital taking responsibility for these GP contracts and, therefore, for running general practices, is an example of ‘vertical integration’, that is, integration between organisations operating at different stages along the patient pathway. This kind of integration is distinct from ‘horizontal integration’, whereby organisations at similar stages along the patient pathway merge, such as when one acute hospital trust merges with another, or when one general practice merges with another or joins a network of general practices. 11,12 Vertical integration between acute hospitals and general practices in the NHS often entails an element of horizontal integration. When hospitals merge with more than one general practice, those practices are effectively integrated horizontally with one another, as well as vertically with the hospital. Our focus is on the vertical aspect of the integration. There are many other types of horizontal integration between general practices (Smith J and Sidhu M, University of Birmingham, 2020; Parkinson S and Harshfield A, RAND Europe, 2020; personal communication), but horizontal integration is not the focus of the evaluation reported here.

The concept and term ‘vertical integration’ has a long history in the economics literature13 and is seen as having three main aspects: (1) market power, (2) risk management and (3) transactions costs. Some associated perspectives on vertical integration have emerged in the context of competitive product markets; these perspectives are likely to be of limited relevance in an NHS context. For example, no UK NHS acute hospital has ever set up new general practices, as they might, in principle, were they in a competitive market where they were keen to channel patients to themselves rather than to rival hospitals. Nonetheless, the economic perspective does offer the following additional insights of potential relevance to the case of acute hospitals running general practices.

Vertical integration can be seen as a way of attempting to reduce transaction costs, as it can make decision-making, monitoring and information-sharing more efficient. 14 In the context of our evaluation, this might translate as better mutual understanding between GPs and hospital specialists, and better communication and flows of patient information between GPs and hospital specialists. A vertically integrated organisation can use more informal and less time-consuming procedures to resolve conflicts internally than would independent agents from separate organisations. 13

Vertical integration could also be a strategy to reduce risk. If risk grows over time, then the attractiveness of this strategy may increase correspondingly. In a competitive market, this may involve securing a segment of potential customers who might otherwise purchase from a rival, but it is also relevant in an NHS context, where the risk to be reduced may be that of demand growing faster than the capacity of the hospital and its sources of NHS funds to meet that demand. A hospital integrating with general practices from which patients are referred to it may offer the prospect of greater demand management. Less positively, vertical integration may mean that patients find themselves less able to exercise choice between alternative providers of hospital care because their GP is inclined to default to referring them to the hospital that employs them.

Laugesen and France15 reviewed the applicability of economic theories of integration to the health-care sector, with specific reference to the UK and the USA. Laugesen and France15 were particularly interested in the scope for risk management and transactions cost reduction via vertical integration. The study15 does not mention a specific example from primary care but vertical integration in the UK is such an example. Here, risk management might mean not only the risks to a hospital of insufficient or excessive demand, but also the financial risks faced by the owners of individual general practices that are facing difficulties in recruiting and retaining new partner GPs. In the US context, Laugesen and France15 highlighted Kaiser Permanente (Oakland, CA, USA) as a well-developed vertically integrated health-care organisation. Kaiser Permanente and other international examples of vertical integration in health care are briefly described in Chapter 2.

In the health-care context, vertical integration has been defined by Ramsay et al. 16 as being of six types: organisational integration (organisations merge into one or bind themselves together contractually), functional integration (non-clinical activities for multiple organisations are combined), service integration (clinical services are integrated within the same organisation), clinical integration (where patient care becomes a single integrated process both within professions and across them, e.g. by using shared guidelines), normative integration (the sharing of values in collaborating to deliver health care) and systemic integration (making rules and policies coherent across all organisational levels).

Conrad and Dowling17 have argued that vertical integration requires (1) the clinical integration of the processes involved in delivering patient care and (2) administrative (functional) integration at both inter- and intraorganisational levels. Therefore, strong administrative integration (e.g. management support, governance and strategy formation) is required to support clinical integration whereby patient care is improved over time. Subsequently, to achieve near-complete vertical integration of care across primary, community and secondary sectors, health care for patients should cut across health promotion and disease prevention all the way through to treating short-term episodic acute illness. Shortell et al. 18 identified a number of limiting factors that need to be considered prior to commencing vertical integration. These factors include (but are not limited to) the need to establish trust among clinicians and institutions, well-integrated information systems and non-clinical support services (i.e. back office functions) and consensus on practice and care delivery guidelines. Robinson and Casalino19 go further, suggesting that vertical integration needs to go beyond a physician’s desire for professional autonomy and a hospital’s desire for organisational co-ordination to show how services in primary care can be delivered through greater horizontal integration, achieving ‘integrated delivery systems’ (known as PCNs and primary care clusters in England and Wales, respectively). Theoretically, vertical integration between acute hospitals and integrated delivery systems can lead to efficient use of services across the care pathway (particularly post-acute care) and avoid duplication of administrative tasks; but, more importantly, vertical integration should lead to a single hierarchy of authority with shared goals and strategies.

Policy-drivers for integration

Despite the historical institutional separation between general practices and hospitals, integration between the multiple levels and providers of health and social care has increasingly been seen, in both England and Wales, as important for the effectiveness and efficiency of the NHS in the long term and to improving the care provided to patients. The NHS across England and Wales is working towards systems of care that offer better co-ordination across the primary and secondary care interface. One innovative approach to achieving stronger integration of care is vertically integrating primary care organisations with secondary care organisations (i.e. providing co-ordination within a single management entity of staff, infrastructure, functions and activities that contribute to different levels of patient care). 12,16,17 Such organisational integration can range from virtual integration that entails the formation of relatively flexible alliance arrangements to a fully integrated organisational model in which a single body holds contracts to deliver both secondary and primary care services. 20 In this evaluation, we are interested in the fully integrated model of vertical integration, specifically NHS acute hospitals taking over the management of general practices.

In England, the NHS Five Year Forward View21 described, among other options for stronger integration between primary and secondary care, what it termed primary and acute care systems (PACS), which would combine general practice and hospital services ‘for the first time’21 (contains public sector information licensed under the Open Government Licence v3.0),21 although not yet within single vertically integrated organisations. There were nine PACS ‘vanguards’ at a variety of locations scattered across England. 22 In 2018, a report from The King’s Fund recommended that acute hospitals should play a greater role in primary care provision:

. . . acute hospitals will need to play a fundamentally different role within local health economies. This will involve: . . . working more closely with local partners, including primary care . . . developing integrated service models that span organisational boundaries . . .

Naylor et al. 23

Evaluations of the PACS ‘vanguards’ have found evidence of considerable change activity, but have not yet yielded clear conclusions on the outcomes achieved as a result. The National Audit Office (London, UK) has found early signs that PACS and other new models of care focused on integration are having a beneficial effect on emergency admissions to hospitals,22 and there have been ‘significant amounts of innovation in terms of both front-line services and wider structures supporting system-wide collaboration’. 20 However, data and findings are limited across the vanguard sites, with questions about the reliability of outcomes data and, consequently, uncertainty about the impact of PACS on key dimensions of health service delivery. 24,25

Five years on from the NHS Five Year Forward View,21 the 2019 NHS Long Term Plan26 announced the intention that all general practices in England should combine into PCNs covering populations of 30,000–50,000. 26 Since July 2019, all but a small number of practices have become horizontally integrated in that way with other practices, although remaining separate legal entities with separate contracts. Findings from a recent BRACE (Birmingham, RAND and Cambridge Evaluation Centre) rapid evaluation (Smith J, Parkinson S, Harshfield A, Sidhu M, personal communication) of PCNs show that there have been a number of facilitators of and challenges to horizontal integration achieving sustainable primary care, addressing growing workload issues and improving the availability and co-ordination of local primary care services. The evaluation (Smith J, Parkinson S, Harshfield A, Sidhu M, personal communication) revealed a tension between the desire for local autonomy and influence over PCNs, and the top-down nature of PCN policy (i.e. the need for effective leadership). Notably, rural case study sites from the evaluation felt that PCN policy had been developed with urban practices and collaborations in mind, and did not adequately account for the experience of primary care in rural areas. Vertical integration was not specifically mentioned in the NHS Long Term Plan,26 but is, nevertheless, happening in a few locations.

In 2015, the Welsh government published a plan for primary care services in Wales over the following 3 years. The plan set out a specific aim for the 64 ‘primary care clusters’ (i.e. collaborations between general practices) that had recently been set up in Wales, each covering a population of 30,000–50,000, to ‘play a significant role in planning the transfer of services and resources out of hospitals and into their local communities for the benefit of their local populations’ (© National Assembly for Wales Commission Copyright 2017. Inquiry into Primary Care: Clusters). 27 Reviewing progress with primary care clusters, the National Assembly for Wales Health, Social Care and Sport Committee reinforced the desirability of that form of integration. 27 The subsequent Strategic Programme for Primary Care in 2019 built further on these initiatives through an enhanced community-based model of primary care. 28 This model arose, in part, from the national Pacesetter Programme, in which LHBs received funding to support innovations within primary care. Research by University of Birmingham (Birmingham, UK) found that, although the Pacesetter Programme had enabled some positive developments, overall, it was hampered by a lack of clarity about its objectives, insufficient evaluation capacity and underdeveloped opportunities to share learning. 28 Similarly to the situation in England, vertical integration between acute hospitals and general practices is not currently being specifically advocated by NHS policy-makers in Wales, but it is happening spontaneously in a small number of places.

Alongside the push for better integrated care, there has been growing interest in extending the range of health-care services that are delivered in primary care settings. 21 However, at the same time, the long-term sustainability of primary care has become an increasing focus of concern in the face of growing patient demand combined with GP workforce constraints. 29,30

With this background of pressure to better integrate secondary and primary care services and to increase the range of services available in primary care settings, combined with concerns about the sustainability of general practices, there are a number of potential pragmatic reasons that might drive vertical integration between acute hospitals and general practices. Vertical integration brings the opportunity to redesign services; address problems with governance, funding, differing objectives and drivers; reduce transactions costs; enhance the ability to involve both primary care and secondary care clinicians in the design of effective and efficient clinical pathways; and, for patients, improve continuity of care and foster more seamless transitions between teams and services. Organisational vertical integration between hospitals and general practices might help to overcome co-ordination challenges that result from the discrete financial ‘silos’ in which primary care and secondary care otherwise find themselves (e.g. where investment in one sector yields cost savings in the other). Vertical integration, which entails practice staff becoming salaried employees of an acute hospital NHS trust (England) or LHB (Wales), may particularly suit younger GPs, many of whom are showing reluctance to buy into the traditional ‘partner model’ of general practice and are showing a preference to be sessional or salaried. 16,31

There are also potential downsides to the implementation of acute hospital and general practice vertical integration, including fears among some GPs of reduced GP autonomy and, therefore, damage to what they see as the entrepreneurialism and innovation of general practice;16,32 mistrust from local primary care stakeholders;33 and doubts over how much inappropriate health service utilisation can be reduced through vertical integration, especially compared with other models of integration. 34

Summary of key points

• The study team took a narrative-based descriptive thematic approach to synthesising findings drawn from a rapid scoping review guided by existing theory on vertical integration.

• The reasons most commonly cited in the literature for integrating acute hospitals and primary care services (i.e. vertical integration) were concerned with expectations of providing better-quality, more effective health care, with better patient experiences of care being delivered at lower, or at least no higher, costs to the health-care system as a whole.

• Vertical integration can be undertaken by recourse to different types of contractual and governance arrangements, but evidence as to which are more effective, and in which circumstances, is not yet available.

• Core characteristics and enablers of vertical integration include establishing continuity of care; the management of finance, human resources (HR), information technology (IT) and planning being closely co-ordinated across the units being integrated; and building relationships with doctors who actively participate in management and are economically linked to their organisation.

• Barriers to the successful implementation of vertical integration include professional silos, unlinked and differing information systems, difficulties sharing patient records between organisations and organisations’ different approaches to case management.

• There is a lack of robust evidence on the impact and effectiveness of vertical integration in health care, particularly with respect to patient outcomes.

The first stage of the evaluation comprised a pragmatic rapid scoping review of the literature, supported by interviews with key informants, discussion with the members of the BRACE Steering Group (which includes experienced health services evaluators, commissioners and health-care practitioners) and a project design workshop with policy-makers and external researchers. This was a key stage in the evaluation to understand the existing evidence, develop research questions and identify potential case study sites. We obtained an overview of the evidence already reported on vertical integration of acute hospitals with general practices and identified knowledge gaps to be filled and propositions to be tested later through comparative case studies, following principles outlined by Munn et al. 35 In this chapter, we present our findings from the scoping of the literature and key learning from interviews with informants, and the research questions that subsequently emerged.

To scope the evaluation, we reviewed published literature, including grey literature (i.e. literature that is not formally published in sources such as books or journal articles such as NHS trust reports and/or working papers from policy groups or committees). We used a selective but systematic approach to searching and provided a descriptive summary of our findings.

The study team took a narrative-based descriptive thematic approach to synthesising findings drawn from the rapid scoping review, guided by existing theory on vertical integration. 17 We thematically categorised literature according to differing models of vertical integration to aid the identification of core characteristics that act as enablers of successful implementation. We then sought to draw connections between core characteristics and desired outcomes for vertical integration within a UK NHS context. A more detailed description of our literature review method is provided in Appendix 1, along with our search terms (see Appendix 1, Box 2) and a PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow chart with regard to screening results (see Appendix 1, Figure 8).

Discussions with key experts

In parallel with the scoping review, members of the study team completed telephone interviews (using a structured topic guide informed by the literature; see Report Supplementary Material 1) and face-to-face meetings with academics, policy analysts and NHS staff involved in the implementation of vertical integration across different sites in the UK (total n = 13). The study team sought (1) their initial insights and perspectives on why vertical integration was introduced and (2) their views on which research questions should be prioritised in a rapid evaluation. The team also sought to better understand the current climate of primary and acute care working more closely together and what the possible advantages and disadvantages of vertical integration would be for acute trusts and general practices. Telephone interviews with NHS staff helped the research team to identify potential case study sites across the UK that would be eligible to partake in a rapid evaluation.

Following the review of the literature and the scoping interviews, the research team held a stakeholder project design workshop to consider the scope of an evaluation of vertical integration between acute hospitals and general practices, and to refine the research questions. 36 The workshop took place in Birmingham in March 2019. In addition to research team members, participants included an independent researcher expert in the field of integration in the NHS and employees of the Department of Health and Social Care (England) and NHS England.

Findings from the literature

Several examples of vertical integration between organisations involved at various stages of the care pathway were identified in the UK, some of which are examples of an NHS acute hospital taking responsibility for the delivery of GP services.

Findings from discussions with key experts

Policy experts and academics expressed their concerns about the rationale given for vertical integration, as it appeared to be driven more by structural/organisational integration than better co-ordination of care delivery. Policy experts and academics perceived a paucity of evidence for vertical integration between acute hospitals and general practices, noting that some of the impetus was from international examples, which may not be transferable to UK settings, and they were keen to hear more about primary care-related outcomes and patient views, as distinct from the hospital perspective. In addition, NHS staff were concerned about the impact that vertical integration might have on general practices working in the same area but outside the vertically integrated arrangement and, therefore, having to compete with acute hospital-managed practices, which would have a greater pool of resources (e.g. finance, clinical staff and administrative capacity) to rely on. All stakeholders raised a number of questions with regard to the impact of vertical integration on the wider health economy, specifically how the delivery of community care services might change as a direct result of integration.

During scoping interviews, NHS staff and policy experts recognised benefits of vertical integration, especially for younger GPs, as younger GPs could be given greater opportunities to acquire specialist training across the primary and secondary interface and would be more likely to have a set/defined workload and greater back office support (e.g. legal, indemnity, procurement), creating more time for clinical contribution. This relates strongly to the current shift in primary care working among GPs, where almost half of GPs in the NHS in England are employed on a salaried or sessional basis (as opposed to having equity ownership of the practice) and a majority are women (many of whom work part-time and/or wish to have portfolio careers, as do some of their male colleagues). 52 However, a number of academics were cautious that there would be a marked difference of leadership styles between salaried GPs and GPs as partners. The closer integration of primary and secondary care might also lead to increased data-sharing aligned with improved management in primary care of high-risk patients who are more likely to access emergency secondary care. Policy experts also expressed a desire to understand how the ‘success’ of vertical integration would be measured. Would success be measured by general practice sustainability, by understanding the ‘right’ professional mix of multidisciplinary teams (MDTs) in primary care or by improved financial and outcomes-based performance by the acute trust? As a result, one academic interviewee felt that it would be appropriate for the study team to develop individual programme theories for each vertically integrated organisation so as to better understand the rationale for, and consequences of, vertical integration in a variety of approaches.

The project design workshop was an opportunity to consider, as a whole, the material gathered from the scoping interviews and literature review, and to identify and prioritise, in discussion with policy and academic stakeholders in the area, which questions to pursue for evaluation of vertical integration between acute hospitals and general practices. These discussions led the team to identify the following six research questions for an initial evaluation:

1. What are the drivers of and rationale for acute hospitals taking over the management and governance of general practices? What does this type of vertical integration aim to achieve?

2. What models/arrangements exist for acute hospital organisations to manage general practices (including different contractual/legal/organisational arrangements across primary, secondary and community health services)?

3. What is the experience of implementing this model of vertical integration, including barriers to and enablers of vertical integration, as well as the lessons learnt?

4. In what ways, if any, has this model of vertical integration influenced the extent and type of health service provision delivered in primary care?

5. What are the views of the primary and secondary care workforces about working together in this way across the care interface?

6. In what ways, if any, has this model of vertical integration had impact so far? What are the expected longer-term impacts? How is progress being measured?

Addressing these questions informs the development of a theory of change for vertical integration between acute hospitals and general practices, describing its desired outcomes and the mechanisms by which these are expected to be achieved.

A conclusion of the discussion at the workshop was that a rapid evaluation could not address all of the main knowledge gaps and that it would be better to adopt a phased approach. This phased approach would comprise an initial evaluation (which is the subject of the present paper) followed by a second phase evaluation somewhat later. The second phase would start from the theory of change and then more fully assess the impact of vertical integration, in particular on patients and on the efficiency and effectiveness of the health-care system.

The method adopted for the initial rapid evaluation is described in the next chapter.

Summary of key points

• This rapid evaluation had two distinct aims.

  • Aim 1: to understand the early impact of vertical integration, including its objectives, how it is being implemented, if and how vertical integration can underpin and drive the redesigning of care pathways, if and how services offered in primary care settings change as a result, and the impact on the general practice and hospital workforces.

  • Aim 2: to develop a theory of change for vertical integration, including identifying what outcomes this model of vertical integration is expected to achieve in the short, medium and long terms, and under what circumstances.

• We completed a qualitative cross-comparative case study evaluation that comprised three work packages (WPs).

  • WP1: rapid review of the literature, telephone scoping interviews and a stakeholder workshop.

  • WP2: comparative case studies of three vertical integration sites with interviews with key staff involved in the conceptual design, implementation and analysis of this model of vertical integration; analysis of key documentation (both internal and publicly shared and that related to patient experience); and non-participant observation of strategic meetings.

  • WP3: development of a theory of change for each case study site and an overall theory of change for vertical integration.

• We undertook a purposive sampling to select three case study sites, based on identifying appropriate sites where vertical integration was happening at scale.

• A content analysis approach to documentary reviews and observations was undertaken. Data analysis for interviews was informed by a framework method for the analysis of qualitative data in multidisciplinary health research. Our analysis was guided by theoretical literature on vertical integration between acute trusts and general practices.

General approach

This rapid evaluation had two distinct aims.

Comparative case studies of three vertical integration sites

In January 2019, through grey literature searching (of Pulse, the Health Service Journal and GPonline), during the scoping of the evaluation, the research team identified five sites (Table 3) across England and Wales where vertical integration between an acute hospital and general practices was already being delivered at scale (i.e. four practices or more integrated with the local trust/LHB). Members of the study team sent e-mail correspondence to vertical integration organisational strategic and/or clinical leads (or their equivalent) on behalf of the principal investigator (JS).

The selection of case study sites was purposive, with the aim of seeking variation across three sites in terms of (1) their rationale for implementing vertical integration and intentions for growth (i.e. management of more general practices operating under this model in future), (2) their geographical location and population served (i.e. a mix of rural and urban locations), (3) their legal and governance working frameworks and (4) the amount of time that has elapsed since vertical integration was introduced. One site declined to participate as it was already taking part in another research study, whereas another site was addressing pertinent local challenges to health-care delivery. Three sites were selected and approached to seek their agreement to participate in the study through face-to-face meetings at each respective site. All three sites agreed by sending formal confirmation of participation letters/e-mails.

Modes of analysis/interpretation

Summary of key points

• This chapter provides a brief overview and context for each of the three case study sites, along with their core characteristics.

• Urbanville comprises an acute hospital trust that is one of the largest acute and community providers of health services in a socioeconomically disadvantaged, ethnically diverse urban area of England. Integration with primary care practices was initiated in 2016, when three general practices approached the acute trust. As of March 2020, the acute trust had integrated 10 general practices, treating approximately one-third of the local primary care population.

• The acute trust–general practice vertical integration model in the Greenvale case study serves a dispersed rural population, a relatively high proportion of whom are elderly compared with the national average in England. The vertical integration model takes the form of the acute hospital trust 100% owning a limited company, which by late 2019 was running GP services for a group of 13 practices covering a total registered population of approximately 80,000.

• Seaview comprises a LHB in Wales that provides a full range of primary, community, mental health and acute hospital services for a population of around 700,000 people across a largely rural and coastal area. As of March 2020, the LHB had received resignations from 22 GMS contract-holding practices, which were, consequently, being directly managed by the LHB.

In this chapter, a brief overview is given of the background and context to each of the three case study sites, along with their core characteristics. Particular attention has been paid to the context into which vertical integration has been introduced. The way in which these sites were selected for study is set out in Chapter 3. Taken together, the three case studies offer a mix of very urban and very rural geographies, and a variety of organisational arrangements to vertical integration. In that way, we are able to examine a range of vertical integration experience, providing relevant learning for much of the NHS in England and Wales and, indeed, in the rest of the UK and, to some extent, internationally.

The site-specific information on each location has been derived from local, public domain documentation, including published reports and minutes, plus supporting papers from trust board meetings. Interviewees corroborated much of the information during the conversations we had with them.

Urbanville

The Urbanville acute hospital trust is one of the largest acute and community providers of health services in a socioeconomically disadvantaged and ethnically diverse urban area of England. There is a significant gap between healthy life expectancy and life expectancy, which is reflected in a high demand on health and social care services in the area. This is coupled with a large ageing population living with a range of long-term health conditions, and often with multiple conditions. The trust delivers care across three acute sites and more than 20 community settings working with third-sector organisations and the local authority. The acute trust is the largest employer in the local area, with nearly 10,000 staff.

Integration with primary care practices was initiated in 2016, when three general practices approached the acute trust. Vertical integration offered the opportunity to redesign patient services and clinical pathways across primary and secondary care in a better co-ordinated way to improve (appropriate) access and positively impact on patient outcomes. As of March 2020, the acute trust had integrated 10 general practices, treating approximately one-third of the local primary care population.

As part of the integration, the acute trust has established a governance team, comprising both primary care clinicians and secondary care managerial staff, to oversee the merger. A diagram of Urbanville’s vertical integration organisational structure is provided in Figure 1.

FIGURE 1.

Urbanville vertical integration organogram.

Greenvale

The acute trust–general practice vertical integration model in the Greenvale case study serves a dispersed rural population, a relatively high proportion of whom are elderly compared with the national average in England. There is a small acute hospital in a market town. There is also rural deprivation and a limited transport infrastructure, with substantial journey times between villages and the market town. Some general practices increasingly struggle to recruit and retain GPs and have to rely heavily on the use of locums.

In 2015, prior to the official formation of the acute trust–general practice vertical integration model, the local health economy became a vanguard for a PACS new model of integrated primary and secondary care. The main focus of the vanguard was on:

• complex care teams focusing on patients with complex and multiple conditions who are heavy users of health-care resources (analysis showed that 50% of resources in the local health economy were being used for just 4% of patients)

• enhanced primary care to offer greater local practice-based support for people with long-term conditions [enhancements included a wider range of health-care professionals (e.g. musculoskeletal practitioners, mental health workers, pharmacists), use of health coaches to support patients and free up GP time, and providing GPs with easier access to advice from secondary care specialists].

As a result of the vanguard programme, strong relationships were developed between primary care and secondary care in the area.

The vertical integration model began in 2016, after three general practices experiencing difficulties approached the acute trust. The vertical integration model takes the form of the acute hospital trust 100% owning a limited company, which by late 2019 was running GP services for a group of 13 practices (together holding nine GP contracts, following mergers between some practices), covering a total registered population of approximately 80,000. The NHS trust-owned primary care company reports directly to the hospital board. A diagram of Greenvale’s vertical integration organisational structure is provided in Figure 2.

FIGURE 2.

Greenvale vertical integration organogram.

Seaview

Seaview comprises a LHB in Wales that provides a full range of primary, community, mental health and acute hospital services for a largely rural and coastal area. The area serves a large elderly population, many of whom live with long-term health conditions, as well as a transient population of holidaymakers during the summer months.

There are three main hospitals, along with a network of community hospitals, health centres, clinics, mental health units and community team bases. The LHB also co-ordinates the work of more than 100 general practices and NHS services provided by dentists, opticians and pharmacists. The LHB was, at the time of the study, in the second half of 2019, under ‘special measures’ by the CQC.

Since 2015, the LHB has been directly managing general practices as a result of GP resignations. There was no strategic intention for the LHB to directly manage general practices. As of March 2020, the LHB had received resignations from 22 GMS contract-holding practices, which were consequently being directly managed by the LHB. Some of these 22 practices have been merged or their patients have been dispersed to other LHB-managed practices.

The LHB has created a governance team of staff from the LHB and managed practices, and also GPs from non-managed practices in the area. In addition, regional teams have been developed (east, central and west) to oversee general practices. Managed practices continue to work with regional primary care clusters comprising both managed and non-managed practices. A diagram of Seaview’s vertical integration organisational structure is provided in Figure 3.

FIGURE 3.

Seaview vertical integration organogram.

Summary of key findings

• The rationale for vertical integration for Urbanville was to better manage demand in the local health economy through (1) improving patient access to primary care services, (2) increasing quality of care and (3) increasing opportunities for innovation.

• GP contracts have been directly taken on by the acute hospital trust.

• The vertical integration strategy was built on shared values between primary care and acute trust colleagues about how best to deliver care to patients and the ability to adapt pathways so that more patients could be treated in the community rather than in acute hospital settings.

• Urbanville was able to develop a primary care-specific division within the acute trust to support the restructuring of back office functions to serve the vertical integration organisation. Despite this, many in primary care felt that back office tasks now took longer to complete than they did before integration.

• Primary care leads have taken up prominent positions within the acute trust to oversee and govern the process of integration, alongside secondary care managerial colleagues.

• Initial investment from the acute trust prioritised the recruitment of salaried GPs and locums, alongside upskilling the primary care workforce.

The study team carried out 18 interviews with stakeholders and a single observation of a strategic planning event hosted by the acute hospital with primary and secondary care practitioners in attendance. Further observations were scheduled in March 2020; however, because of the COVID-19 pandemic, the research team was advised not to attend meetings in clinical settings. The acute hospital shared documents that were discussed in the meetings.

Theme 1: understanding the need for and purpose of acute hospital integration with primary care

For a number of years, the traditional partner model of primary care was considered unsustainable in the face of such challenges as a limited primary care workforce, fewer GPs and the increasing number of patients living with complex multimorbidities. Furthermore, greater integration of primary care with secondary and community health provision was understood to be one remedy to alleviate pressures elsewhere in the health economy. For example, poor access to primary care often meant patients accessing A&E departments in secondary care. Therefore, it was imperative to break this cycle:

I mean, the rationale that was given to me, and I think it was the rationale that was out there just in the landscape, was . . . firstly, primary care isn’t sustainable as it is . . . we were already talking about an ICA [Integrated Care Alliance] anyway.

A14, secondary care, organisational management

. . . so, basically they wanted to integrate primary and secondary care and try to break that divide between purchaser/provider, that was their main thing.

A03, primary care, clinical

. . . and I came to the conclusion the only way I can manage demand in secondary care is from primary care.

A04, secondary care, senior management

A key premise for Urbanville was that vertical integration would allow practitioners in primary and secondary care to create a shared vision of how best to design health services and manage back office functions (e.g. payroll, HR, estates and legal planning) to improve population health and manage resources effectively. Therefore, a key driver of vertical integration was how to make the most of resources through collaborative working and the development of a formal legal arrangement between general practices and the local acute trust, which would reduce primary care workload:

. . . there was that shared vision of ‘How is this going to work? What is it going to look like?’ We got that shared vision that primary care was . . . sinking. We were doing our best, but we just could not get on top of the amount of work we were being asked to do.

A09, secondary care, organisational management

Many participants, both in primary care and in the acute trust, considered that, by moving GPs from being partners to salaried positions as employees of the trust, the maximisation of personal/practice income was no longer an incentive to improve delivery of services to the local population:

. . . so the whole idea that you merge with the hospital and you become salaried – so the money, you take the money out of the equation.

A09, secondary care, organisational management

The GPs like the idea of having a regular salary because they’re all employees of the trust, they liked the idea of having a regular salary.

A02, secondary care, organisational management

However, the nature and introduction of vertical integration did give rise to concerns as to whether or not a shared vision could be reached between general practices and the acute trust. It is not uncommon for there to be a cultural divide and mutual mistrust and lack of understanding between those working in primary care and those working in secondary care. For this reason, many GPs whom we interviewed felt that they had much to lose by entering into a legal arrangement that might lead to a loss of autonomy:

I think there’s always been a concern about the big monster, which is the trust, so it’s about being engulfed into an organisation and losing to some degree your autonomy. So the one advantage of being a partner and being a GP is it’s your business; you’re autonomous.

A15, primary care, clinical

It was a new project; we were three practices, and of course, secondary care don’t know anything about primary care. We find them an anomaly; they find us an anomaly.

A02, secondary care, organisational management

Although former GP partners shared their genuine concerns about the sustainability and future of primary care in the area, they also noted that vertical integration is one way of addressing this and that there were few, if any, alternative options to consider. In addition, GPs may have recognised that there was limited gain for the acute trust too:

If you can’t get any partners and your estates are falling apart and you can’t do anything about it and you’ve got no way of expanding, it might be the right thing for you. It really is down to individual practices.

A12, primary care, clinical

Other GPs in the patch, obviously, became very nervous. Where a big organisation starts to throw their weight around primary care, you know, it’s kind of worrying for their business model. But our counter view to that is, ‘Actually, we’re trying to provide better care. We know there’s no money in primary care, so we’re not doing it for the money’.

A05, secondary care, senior management

Theme 2: developing an integration model and implementation strategy

The implementation of the vertical integration model was instigated by three general practices collectively approaching the acute hospital to discuss how best to deliver primary care services and improve patient management in the local area. Although these practices were, relatively speaking, financially stable, they were encountering other challenges:

. . . the practices that had joined seemed to be those that have had difficulty recruiting staff and, therefore, were struggling generally; they’ve had problems with their estates maybe and felt that the hospital could help them with that, and that seems to have been the main driver for people moving. The practices that are quite stable, are happy running their own businesses, seem to be making money, having no staff problems, don’t seem to be that interested.

A12, primary care, clinical

Very early stages of the integration were characterised by due diligence meetings, whereby a team from the acute trust would visit an interested practice to understand more about its financial and legal viability:

It’s looking at their income and expenditure, it’s looking at salaries, anything to do with finance, so it would be looking at, you know, all of their bills, you know, things like leases and things like that, just to get an overall picture of financial viability and, actually, almost mapping that against, for example, another similar practice of similar size to see what that looks like and what the financial envelope would look like.

A14, secondary care, organisational management

The due diligence process enabled both parties to understand, and develop an offer of, how best to integrate, and establish whether or not becoming salaried employees would be financially viable for the general practice partners. According to senior managers at the acute trust, integration was not restricted to financially better-performing practices, but was extended to those that had the ethos and infrastructure in place to support redesigned primary care. The acute trust recognised (and accepted) that there would be significant financial impact when integrating practices. Furthermore, the process of integration often highlighted other concerns too:

. . . why has that person been off sick for that amount of time? Is there a problem here . . . are they ill, is there another thing going on? . . . I think there’s also something about understanding clinically their current way of working with the way of working that we aspire to, and how far away are they? What is the shift or change we need to make and what’s that going to mean for us in terms of resource, whether it’s time or bodies or money? Because that’s what we actually find. When they come to us, there’s a huge amount of work and could we have known that, could we have planned for that work better? Or, actually, some of it you just don’t find out until you’ve got them.

A11, secondary care, organisational management

Depending on the nature and complications associated with a practice, due diligence and integration took between 3 and 6 months.

The expansion and sustainability of vertical integration has faced challenges with respect to NHS England and NHS Improvement’s (NHSE&I’s) introduction of PCNs. All but one of the practices aligned with the acute trust to form a single PCN to discuss service delivery and quality improvement, whereby financial resources that would otherwise go directly to general practices instead go to the acute trust. The PCN model, more contextually, has strengthened the position of non-integrated practices in the local area that felt uneasy about the potential influence the acute hospital would have by having greater control in primary care through their managed practices model. Conversely, GPs working within the acute hospital felt that PCNs were supporting the work of the vertical integration model. For example, the PCN model may support the introduction of a single contract (rather than GPs individually subcontracting their contracts) that covers all practices, and it may offer better employment prospects for new types of staff joining general practice:

. . . they’re looking at the PCN holding the GMS [contract] for all of our practices, so they’re looking at one GMS contract for the PCN, which is all the vertical integration practices.

A17, primary care, organisational management

The recruitment of staff has been an interesting one . . . one new person working in a PCN in a new type of role is really vulnerable to getting them to stay. You’ve got to be a particular type of person who can work in a new role, change work and be on your own within a PCN, and I think that’s one of the failings of it. I think the – and this is what we argue – as [a] PCN, we can recruit six of those types . . . they have the care and support of each other.

A11, secondary care, organisational management

To support the implementation and integration of general practices, the acute trust restructured staff and resources to specifically manage back office functions (March 2018). This will be explored further in the next theme.

Theme 3: making the change – from General Medical Services to subcontracted providers of primary care

Throughout this theme, we provide further details of how practices transitioned from being GMS contract-holding providers of primary care to providers within a subcontracting model with the acute trust.

Theme 4: impact on patient management – changes to primary care and secondary care delivery

The nature of clinical integration between primary and secondary care at Urbanville was driven by the acute trust, with the aim of treating a larger number of patients in the community and reducing utilisation of secondary care services. However, it had taken considerable time to set up back office support functions, with the consequence that attention had only recently been directed to service redesign. This centred on primary care services. The following example highlights possible savings that could be made by moving community services into primary care:

So phlebotomy at the moment is delivered through the community contract, community services, which is actually very expensive. If you start to train up your primary care nurses and, actually, you’re delivering through extended access or extended hours or putting on specific specialised clinics, actually, you start to pool your resource. Those nurses then that are in community that were doing phlebotomy . . . you’re freeing them up then.

A14, secondary care, organisational management

In addition, the acute hospital being integrated with general practices allowed the development and application of innovative methods for managing highly complex patients who are likely to be frequent users of secondary emergency care. The vertically integrated organisation was able to develop its own database to identify and track high-risk patients across primary and secondary care, sharing real-time information with practitioners across the care interface:

It’s like a dashboard, so it tells us what patients have been admitted overnight, what procedures they’ve had done, and we get a copy of that each day and the GPs look at it and think, ‘Oh, Mrs so-and-so was in there last night; I might give her a call and see if she’s OK’. So it links the care up better.

A02, secondary care, organisational management

Although new roles were being introduced in primary care, the acute trust was still struggling to rectify GP shortages. Many of the GP partners in the practices that were integrated had either moved to being salaried or had left primary care prior to or shortly after integration. This led to locums being used as a short-term measure across managed practices until suitable salaried GPs could be recruited. For some in primary care, the use of locums was a source of tension:

. . . salaried partners, they worked as hard as us and contributed to the teamwork in exactly the same fashion, which is different than locums when you have to use them.

A07, primary and secondary care, clinical

Many of the non-clinical staff working across practices felt that both locums and salaried GPs had a different approach to managing and delivering general practice services compared with GP partners. In general, non-clinical staff felt that former partners of their practice (which was now part of the vertical integration organisation) had been much more involved in, and focused on, the everyday challenges of delivering primary care, whereas locum and salaried GPs focused only on delivering care to patients. This was seen as leading to a lack of cohesiveness among primary care teams:

I mean, none of the GPs here have been a partner before but they behave like partners, if you like. So, like, with holiday, we all say: ‘Unwritten rule: no one has time off at Christmas, because it’s not fair for one or the other’. They do what they like and they book time off, which upsets everybody else, which upsets the girls, because I’m saying, ‘You can’t have time off’. They’re, like, ‘Well, they’re salaried like I’m salaried, so why do they get it and not me?’.

A17, primary care, organisational management

However, salaried GPs could focus more on patient care, which might be a good thing for both the GPs and the patients:

I think from a clinician’s perspective, especially the GP’s, they’re actually able to concentrate on their clinical work because all of that admin side of things, that has been taken away; so they’re actually able to be GPs and see patients and provide that clinical care.

A14, secondary care, organisational management

One fundamental and very significant opportunity that vertical integration created was for training and career development, for both clinical and non-clinical staff in primary care. Some staff from primary care had already moved into positions at the acute trust (e.g. to manage and provide quality assurance for the running of managed practices):

When you get to the top of the band, there might be an opportunity to change, you know, to a higher band. You know, it’s just a bit more organised, I guess, and fair. But, yeah, there’s loads of opportunities within the hospital.

A17, primary care, organisational management

Primary care itself, there’s no career prospects; basically, it’s a dead-end job. You start off as a receptionist, and if you want to be a practice manager, it’s dead man’s shoes. But within this [vertical integration], you’ve got the opportunity to move into the trust, perhaps on the same grading, and develop your career, and you’ve also got more education and training. A couple of the staff have been supported through NVQs [National Vocational Qualifications]. There’s two people in my team who have come from primary care; they’re doing their degrees, and the trust is financing that for them. Things they would never have been able to do before.

A02, secondary care, organisational management

Therefore, vertical integration opened a number of opportunities for those working in primary care, especially those currently in non-clinical positions. Notably, perhaps as a result of our sampling frame, there were very few references within the participants’ accounts of staff from secondary care finding opportunities to work with, or in, primary care. However, we heard, anecdotally, a view that some younger medical students at the acute trust were becoming more interested in working in primary care as a result of the vertical integration model. The following quotation is from a GP working as part of the vertical integration model, relating being approached by a medical student:

‘I came to your talk, I was there, but now I think I want to come and talk to you about what it’s like to be a GP and what other roles could I have’. So it’s almost putting those seed thoughts in place, isn’t it, about what this is going to look like for the future? So I think it’s playing to the strengths of being part of a much bigger organisation.

A09, secondary care, organisational management

Theme 5: measuring success and identifying any areas of unintended impact

Given the stage that the acute trust had reached with vertical integration at the time of our evaluation, much of the described impact focused on non-clinical outcomes. Financially, the acute trust was carrying a significant burden, which was a result of having to use locums to cover former GP partners who had left and difficulties in recruiting salaried GPs. In addition, the trust encountered high staff costs, with the majority of primary care staff being ‘TUPE-ed’ [i.e. Transfer of Undertakings Protection of Employment (TUPE)] over onto Agenda for Change (AfC) NHS contracts, resulting in increased salary levels for some and increased National Insurance contributions:

So employer’s National Insurance is more expensive than self-employed. That’s the first obstacle. Then you’re taking a lot of risk. So if a partner leaves or becomes ill or a salaried GP leaves or becomes ill, the risk is then on us to fill that with locums, agency, which can be very expensive. To recruit we may need to pay more than has been paid in the past because of the increasing shortage of GPs.

A13, secondary care, organisational management

Meanwhile, the stakeholders interviewed from the acute trust felt that it had taken time for the trust’s finance managers to understand how best to maximise primary care income and ensure primary care activity was recorded accurately owing to the rather different ways in which acute hospitals and GPs are funded:

I think there’s something around, actually, in terms of how primary care funding also works. It’s been a real learning curve for the finance department within [the acute trust], because from an acute perspective, this is how much you’re getting. That’s a block contract, so you’ve only got that much to spend. There’s a little bit of variation here or there but, actually, that’s the envelope that you have to work with, and that’s the constant money that you know that you’re going to get. Whereas in primary care, you’ve got your GMS contract money, which more or less stays the same but, actually, if you then don’t do any in-house services, direct enhanced services, public health initiatives . . .

A14, secondary care, organisational management

In an attempt to maximise income for its primary care activity, the acute trust established a back office function that was focused on ensuring complete, accurately coded recording of activity. Some of the vertical integration general practices had been continuing to record some of the details of their activity as ‘free text’ within patient records, rather than fully coding that activity. As a result, there was a significant amount of lost revenue, for example through failing to ensure that eligible patients were invited for consultations:

. . . patients with mental health or patients that were reaching the end of the age eligibility so that they could call them in and not miss that point of care. By inviting those patients in, [the practice] would get a set amount for an invitation. So our reporting structure highlights if a patient has actually had a health check but hasn’t had an invitation coded.

A16, secondary care, clinical

For the acute trust, the main outcome that vertical integration intended to achieve was a reduction in secondary care emergency admissions, while, at the same, reducing the number of ‘did not attends’ and increasing the number of appointments available in primary care. Comments on early data collected from the trust had indicated a reduction in referrals to A&E following vertical integration, as well as fewer patients being readmitted to hospital within 30 days of discharge. However, the trust was unwilling to share these data with the research team until their analysis was complete and ready for wider dissemination.

Nevertheless, there have been some notable successes with regard to other outcomes, specifically recruitment. As a larger organisation with greater availability of resources, the trust has been able to fill GP vacancies more quickly; however, this came at a significant financial cost:

So there was one practice that lost quite a lot just because they could never fill the vacancy, because there were three GPs required. They could never recruit anybody. Cost us a fortune. I think we may have recruited now.

A13, secondary care, organisational management

Notably, since the commencement of vertical integration, only one practice had left the collaboration (mutually agreed by both parties). There is engagement across practices. Back office functions have been established and contribute to a more cohesive and collective approach to improved service provision across the area, and this has led to lower staff turnover and to some staff reporting of higher levels of job satisfaction. For those in primary care, the greatest success of vertical integration is that the acute trust ultimately provides additional financial resources:

I think success otherwise for me is how engaged are the practices when we get them all together to talk about what does the future look like and things like that. You know that is big; we moved away from some basic level conversation to, actually, we’re now talking about doing something quite different, and we’ve gone up a level . . .

A11, secondary care, organisational management

So, personally, primary care has probably been underfunded. If we had to put in a certain amount of money into a practice, lose a bit of money on the practice, but it makes it a better practice that supports people that are frail, that diverts patients from A&E, all those good things, if we spend a bit of money on that, that’s great. So that probably is value for money.

A13, secondary care, organisational management

Theory of change

A theory of change illustration for this case study site is presented in Figure 4. The central rationale for integration between the acute trust and primary care practices was to better manage demand in the local health economy through (1) improving patient access to primary care services, (2) increasing quality of care and (3) increasing opportunities for innovation. Collectively, this would, it was hoped, lower the whole-system cost of treating patients. The acute trust was feeling the impact of poorly resourced general practices on secondary care, with many patients accessing emergency care at the trust because GPs were experiencing increasing workloads with fewer full-time clinical staff.

FIGURE 4.

Urbanville vertical integration theory of change. PPG, patient participation group; STP, Sustainability and Transformation Partnership.

Integration required stakeholders from primary and secondary care to establish a strategy, built on shared values, on how best to deliver care to patients and have the ability to adapt pathways so that more patients could be treated in the community rather than in acute hospital settings. This was achieved, in part, by sharing real-time data from secondary care with primary care practitioners to manage high-risk patients (i.e. those most likely to require emergency secondary care). This strategy was further strengthened with ties built by the trust with the local CCG and Sustainability and Transformation Partnership, by establishing a vertical integration-specific PCN and by working with NHSE&I, which advised the trust on how this model of integration could legally occur. Therefore, significant time was spent establishing a clear line of communication between primary and secondary care.

Initial investment from the acute trust prioritised the recruitment of salaried GP and locums, as well as the introduction of non-GP clinical staff who would be better suited to address the needs of patients living with long-term complex health conditions. This coincided with skilling up the primary care workforce. Positively, integration created a number of opportunities for clinical and non-clinical staff to further their careers in a larger organisation that could afford to fund time away from their main roles. Such opportunities were welcomed by staff, alongside the transition to AfC contracts. In addition, there was substantial investment in managing estates-related costs where buildings were owned and not leased.

Urbanville prioritised, early in the implementation of its vertically integrated model, functional integration and developed a primary care-specific division within the acute trust to support the restructuring of back office functions, with a mixture of new and existing staff to support financial management, procurement and estates. This shared ethos was largely achieved by certain primary care leads taking up prominent positions within the acute trust to oversee and govern the process of integration alongside secondary care managerial colleagues. This was largely received as a positive step to having balanced representation and ensuring that the interests of primary care remained at the foreground of vertical integration decision-making, with effective communication with the acute trust board. However, despite large-scale restructuring, many in primary care felt that back office tasks now took longer to complete when compared with pre-integration. This was despite the acute trust appointing a number of senior primary care clinical staff to oversee the management and governance of integration.

As a result, the acute trust achieved a number of short-, medium- and long-term outcomes with regard to implementation (see Figure 4), but some areas still required further exploration to comprehend their effectiveness. The trust continues to bridge the cultural divide between primary and secondary care, establishing shared values among its managed practices to make the most out the resources available in the local health economy. Yet, in the long term, it remains to be seen whether or not the high cost of locums can be sustained and whether or not linking patient data across the interface leads to a reduction in secondary care demand.

Summary of key findings

• The main rationale for vertical integration at Greenvale was to enable general practices to continue to operate in the face of recruitment and financial difficulties, and in the context of rising demand for care.

• Vertical integration was implemented by setting up a limited company to run the GP services, with the company 100% owned by the acute hospital trust.

• The financial backing of the acute hospital trust, centralised back office functions within the primary care company and the creation of a central management team for the primary care company have contributed to the viability of primary care at Greenvale.

• Vertical integration benefited from active and supportive leadership, both at the acute trust and in the local primary care community. Relationships had developed and strengthened during a prior PACS vanguard in the area.

• Recruitment and retention of GPs, now all salaried, has been supported by the vertical integration arrangement, and no practices have had to close.

• Primary care practices have more MDTs than previously, share staff across practices and have access to some specialist services via outreach from the acute hospital.

The study team completed 17 interviews with stakeholders and two non-participant observations of meetings. The two meetings observed were of the management team of the primary care company owned by the acute trust and of the practice managers from the practices within the company.

Theme 1: understanding the need for and purpose of acute hospital integration with primary care

It was a clear theme across all of the interviews at Greenvale that the principal driver of vertical integration was the risk of general practices closing and the scope for vertical integration to prevent that happening. Therefore, the main purpose of vertical integration at Greenvale was to sustain primary care in the area; subsequently, it would be possible to continue to develop better integrated care, not least to carry on the changes commenced by the PACS vanguard, with its focus on better managing care for people with complex needs and multiple long-term conditions.

The acute trust-owned company that was to operate a number of general practices in its catchment area was set up in April 2016. According to our interviewees, it was clear by that time, and has become clearer in the years since, that the traditional model of general practice was unattractive to some GPs:

GPs going in just don’t want that responsibility any more. They don’t want to run a business and have a mortgage and manage people. They just don’t want it.

B08, primary care, organisational management

Not only do potential GP recruits not want to take on the responsibilities and personal financial risks of co-owning a small business (i.e. a general practice co-owned by partner GPs, which is the traditional model of general practices in England and Wales), but some existing GP partners would like to relinquish them. This was particularly the case for practices where, owing to a recent departure, only one GP partner remained or where the remaining partner(s) were approaching retirement age and becoming increasingly concerned about how they might sell on their practice. The consequence was that some practices in the area were on the verge of handing back their contract to the CCG:

. . . it was quite apparent there was several practices that would fold, because – well, as you’ve heard a million times before – so elderly GPs having to retire, no one to take over the practice, or no desire to move it forward. Expensive locums to fill people, long-term vacancies.

B14, primary and secondary care, organisational management

. . . certainly our local experience, it’s been largely a response to crisis, although there have been some very good practices where the partners have simply said, you know, ‘We trained to be doctors, we don’t want to be partners, we don’t want to own the business, we don’t want to employ all these staff, we’d just like someone to come and do that, and we’ll stay on and serve our patients’ and that’s fine, but I think the majority of practices that both [name of another vertical integration organisation] and [name of case study B organisation] have taken on have been, you know, in a bit of a pickle, really.

B15, primary care commissioning, organisational management

The formation of the acute trust–general practice vertical integration, in April 2016, and its subsequent enlargement were in direct response to this pressure. In 2015, the acute trust was approached by a group of three general practices with a proposal to integrate. Links between practices in the area and the acute hospital were already good, in part because of the work around the PACS vanguard model. In addition, the acute trust’s chief executive had been active in integrating health and social care in his previous post in another part of the country and was known to be supportive of integration:

So, basically, they applied for the vanguard, got all this research done about new models of care, and, at that point, as they started to work with primary care and were looking at the benefits of sustaining primary care, some practices came forward to the hospital . . .

B09, primary care, organisational management

. . . because of the vanguard work, the relationships between primary care and the acute trust were very strong, and I think that is a very, quite an important point.

B01, primary care, organisational management

[Name of individual] was very much around sustaining primary care when he was chief executive in the hospital. And he was approached by the [name of general practice] surgery, and then also [name of another general practice], who owned [names of two health centres], about formally taking that step further to support primary care within [area]. And that’s when they integrated as part of [acute hospital] . . .

B11, primary care, organisational management

Alongside the strong pre-existing relationships between primary care and secondary care in the area, one interviewee said that the vertical integration model was formed partly because of the appetite of the acute trust to take on the challenge:

The management at the time, and this continues today, have a probably higher appetite for risk-taking and transforming at pace than perhaps elsewhere . . . So there was an appetite amongst the leadership to take this on, which perhaps doesn’t exist everywhere.

B14, primary and secondary care, organisational management

By sustaining practices that were in particular difficulty, the vertical integration arrangement not only helped the practices that were integrated with the acute hospital, but also avoided the imposition of a burden on the general practices that remained (i.e. practices that would otherwise have had to take on patients from the failed practices):

Yeah, the whole premise was around sustaining primary care. And, obviously, not just about sustaining those practices, but, obviously, taking the pressure off the other practices in [name of county], because if you disperse lists, that then has pressure on the other practices because they’ve then got to take a cohort of patients . . .

B11, primary care, organisational management

Vertical integration of general practices with the acute hospital trust was, from the outset, clearly seen to be beneficial to the trust, as well as to the GPs concerned:

. . . there was a clear and sort of symbiotic relationship, I guess, between primary care and the acute trust. So if primary care starts to get into difficulty, then that has significant impacts across the health and care system. So if practices start to hand back contracts and start to dissolve, then that has a big impact on the hospital, because those patients end up going somewhere else. And for a lot of them on their, sort of, urgent, on-the-day demand, that would come here if there wasn’t access to primary care. So that was the first reason.

B07, secondary care, organisational management

They’d looked at lots of other options but were at the point of handing back their lists to the CCG, and if they did that, we know within the hospital the impact would have been most greatly felt within the hospital. And if those patients aren’t being cared for appropriately in primary care, where will they go? They’ll rock up to A&E.

B08, primary care, organisational management

Therefore, in addition to safeguarding the survival of some general practices that were in difficulties, another major purpose of vertical integration at Greenvale was to manage demand for acute care in the local health economy:

. . . where practices are understandably focusing on their cost base, actually, what they haven’t traditionally done is focused on their impact on the wider health and care system. And so . . . if you can start to standardise and you can start to make some resource decisions that are different, then you can start to affect the way that cost is attributed across the healthcare system in quite a dramatic way. That was kind of part of the logic.

B07, secondary care, organisational management

. . . the acute trust wanted strong primary care, and they wanted to make sure that the people that came into the hospital were the right people that came into hospital, and when they came in, that they were there for the minimum amount of time, and we got them home, living independently in the community, supported by their GPs and the community staff.

B01, primary care, organisational management

This joint rationale of sustaining primary care and thereby enabling better responding to, and management of, patient demand has remained unchanged so far.

Theme 2: developing a model of integration and an implementation strategy

Various models of integration were considered for the local health economy, particularly before and during the PACS vanguard, which commenced in 2015. When the acute trust was approached by three general practices in difficulty and wanting to integrate, various specific models for achieving that were considered, drawing on the research that was undertaken during the vanguard programme. The decision was made to establish an organisation to run the general practices that was at arm’s length from the acute trust to lessen the likelihood that it would be perceived as a secondary care takeover of primary care:

. . . it was very strongly felt by all those that primary care needed to be an arm’s-length body from the hospital, so that it had, it was still very clear that it was a primary care organisation.

B08, primary care, organisational management

So the advantage of having a limited company, which was arm’s length from the hospital, is it meant that the focus would be kept on primary care by a team of people that would be running that organisation at arm’s length, as opposed to the practices being swallowed up by the hospital and then the focus becomes all about secondary care, not primary care.

B09, primary care, organisational management

The arm’s-length relationship with the acute hospital was reported by some in primary care to have achieved that aim:

If you’re on the ground . . . you don’t even particularly think about [the acute trust] owning you.

B05, primary care, organisational management

We heard no reports from interviewees about the general public being concerned about who was running their local GP services, and it seems unlikely that clinicians at the hospital were concerned about whether or not any particular patient was a patient of a hospital-owned practice. Indeed, these seemed to be non-issues:

So I think as soon as you start discussing that relationship between the hospital, [the vertical integration company] and the practice – I’m not being uncharitable – I think it goes over people’s heads or, quite frankly, they’re not interested. It’s the ‘So what? Can I get an appointment tomorrow?’ that I’m interested in.

B16, primary care, patient participation group

I suspect a lot of the secondary care teams don’t know who are in [the vertical integration company] and who aren’t when it comes to the practices.

B13, primary care, clinical

The arm’s-length limited company that managed the general practices had its own board, which was chaired by a non-executive director from the acute trust’s board, but had more primary care than secondary care representatives. The company board reported directly to the acute trust board. Voting power on the company board was effectively split 50/50 between primary care and the acute trust. The organisation had a central management team and staff, as well as the staff in each of the constituent practices, all of whom focus on the running of primary care, while drawing on the resources of the acute trust where appropriate:

. . . if you have a subsidiary of the hospital which is at arm’s length, has its own management team, it can focus solely on the day-to-day running of the primary care, which has its different constraints, different contracts. But being under the same hood, it can borrow resources.

B14, primary and secondary care, organisational management

Another feature of the implementation of the vertical integration model at Greenvale was the concern to ensure that the GP contracts taken over by the primary care company owned by the acute trust could be reversed, that is, the GPs of the practice can take back the contract (i.e. leave the company) if they wish to at some future date, including, for example, if the vertical integration company decided it no longer wished to run the practice or wanted to sell it to a non-NHS-owned company:

As with any new initiative, there were concerns that it would not work out. Practices have retained their contracts to alleviate this concern, i.e. reversibility.

B04, primary care, professional representation

. . . what was put in was that if at any point [the vertical integration company] decided to stop doing what it does and potentially sell up to a private provider, within the contract, the partners that were there before have first opportunity to take it back.

B08, primary care, organisational management

However, concerns were raised when the acute trust–general practice vertical integration model was designed and established. Some primary care stakeholders in the region worried that secondary care was taking over primary care, leaving the traditional partnership model under threat:

I think there’s still a lot of negative sentiment and a belief that [the vertical integration company] was set up as a trojan horse to kind of take over general practice as part of the overall kind of conspiracy against the partnership model.

B15, primary care commissioning, organisational management

There was also a concern that if the acute hospital trust was looking to take over primary care in the area, then this would create a single monopoly provider of primary care to the local population:

I mean, obviously, patients still need to have choice where they go, so in terms of GP practices, you wouldn’t want [the acute trust] . . . to take all of the practices in [the area].

B11, primary care, organisational management

There were still concerns that, even if a single monopoly provider of primary care did not materialise, general practices in the area that were not part of the integrated organisation would suffer. As a result, the central management team of the trust-owned primary care company sought to reassure other general practices in the area that this would not be the case. Furthermore, of the three GP positions on the vertical integration company board, one position was specifically for a local GP who was not part of the vertical integration arrangement.

By taking over general practices in difficulty, the acute trust was thereby taking over the financial losses those practices were incurring. Nevertheless, senior trust managers considered this to be ultimately beneficial to the trust, and to the local health economy overall, because it ensured that the continuance of primary care would support better demand management and more efficiency in care provision to the local population. Analysis undertaken by the trust indicated that the cost to the wider system (i.e. through increased secondary care utilisation as a result of failing primary care) would be much greater without vertical integration:

. . . there’s a question whether you ever can get [the general practices] financially stable, but the cost to the system would be so much greater [without integration], and that’s where the argument, you know, goes back and forth.

B08, primary care, organisational management

With respect to the implementation of the vertical integration model at Greenvale, each practice was still responsible for running itself, while drawing on the support of the central team within the primary care company. This central team provided back office support functions, such as HR, payroll and accounts, to all integrated practices:

. . . practices still very much run themselves, with support from a central team; it is different to how it first started out, as the central management team is much smaller now.

B04, primary care, professional representation

. . . we are a central team, and we brought in, like, our [the acute trust’s] HR and our finance and corporate governance and, like, management policy procedures, which is now managed centrally.

B11, primary care, organisational management

Within each general practice, there was an operational manager who was responsible for running the practice on a day-to-day basis, supported by a clinical lead and, often, a nurse. The clinical leads of each general practice were supported by a MDT of other health-care professionals:

. . . we’ve got a mixture of not just GPs, but we’ve got advance nurse practitioners, nurse practitioners, practice nurses, highly skilled HCAs [health-care assistants] and apprentice HCAs. We’ve got MSK [musculoskeletal] practitioners, and we’re piloting mental health workers. And we have health coaches, which is, obviously, which came from the vanguard programme. So we really do work as a multidisciplinary team.

B11, primary care, organisational management

Practices within the vertical integration organisation also collaborated with one another, as they would in a horizontally integrated network. Staff worked across different sites where necessary (e.g. if there was a shortage of staff in one practice):

. . . one of the things that can happen is to phone another [vertical integration company] practice and say, ‘Have you got a bit of slack on the system today? Have you got a GP that’s got an admin[istration] morning that might sacrifice that to do a couple of hours for us?’ I mean, I use ‘GP’ as an example. It can be any of the clinical staff, where the pressure is on in one place. Now I think that it’s beginning to operate in an informal way, and without treating all the clinicians as a pool, because I don’t think that’s helpful, but to have a protocol or, if you like, contingency plans, where the sharing of resources at times of pressure becomes the norm rather than something that kicks in at the last minute.

B16, primary care, patient participation group

After the first three general practices approached the acute trust, which led to the formation of the vertical integration model, several others joined, so that by late 2019 the primary care company had 13 general practices (i.e. 13 sites), which were covered by nine contracts with NHS England. The acute trust and the primary care company it owns had not actively sought to recruit additional practices. All of the new additions had approached the trust/company. A due diligence process had been developed around the decision as to whether or not additional general practices were suitable for integration. There were three core perspectives considered in the current due diligence process: (1) quality regulation, (2) finances and (3) workforce.

Interviewees acknowledged that, initially, the due diligence process was not stringent enough, leading to unexpected issues arising after some practices had been integrated. However, the process had since been tightened up:

. . . one of the initial risks was around due diligence and our due diligence process – the risk being that our due diligence process wasn’t robust enough to pick up some of the issues. And so then you, kind of, you think you’re doing one thing and you’re actually, once you get in there, you find there’s a whole heap of other things that you should have found out about that before.

B07, secondary care, organisational management

Since the NHS Long Term Plan’s21 formalisation of PCNs and their near-100% coverage across England since July 2019, the vertical integration model has operated within a new landscape of horizontal networks of general practices. At Greenvale, approximately half of the vertical integration practices together formed a single PCN that was made up solely of those vertical integration practices. The other vertical integration practices were spread thinly across other PCNs that also incorporated non-integrated practices. These latter vertical integration practices had been allowed the freedom to support their PCN in whatever way necessary, with one interviewee even suggesting practices should look towards their PCN, rather than the vertical integration company, to provide some collaborative services:

. . . we’ve given our practices that aren’t in the main PCN, if you like, permission to just go and support their PCN to do whatever it needs to do and not worry about the fact that they’re a [vertical integration company] practice.

B03, primary care, clinical and managerial

It makes much more sense for our practices to be looking at their PCN as the way to then provide services collaboratively, rather than look at [the vertical integration company].

B03, primary care, clinical and managerial

However, the view was expressed that an arrangement such as the vertical integration company model would continue to be important because, unlike with PCNs, the acute trust that owns the vertical integration company is able, in effect, to underwrite general practices in financial difficulties:

. . . when you get to a point where you’re trying to support practices that get into big financial difficulties, the PCN isn’t the risk-holding entity. So the question then is: who holds the financial liability for that practice? So I think the PCN will change the nature of the relationship with [the vertical integration company], and it will almost put a step in place before you get to an entity like [the vertical integration company], that scale provider of local practices trying to support a practice that is in difficulty and kind of keep it stable. But there’s likely to still be a requirement for an at-scale provider like [the vertical integration company], who can then come in.

B07, secondary care, organisational management

Theme 3: making the change – from General Medical Services to subcontracted providers of primary care

Theme 4: impact on patient management – changes to primary care and secondary care delivery

In spring 2020, at the time of writing, the vertical integration company had been in operation for 4 years. We have described how the main purpose of the vertically integrated model at Greenvale had been to sustain primary care in the face of severe staffing and financial difficulties being experienced by some practices in the area. Changing the way that health care was provided to patients had evidently not been the number one preoccupation. Changes were happening, but it was unclear to what extent these were due to the vertical integration arrangement rather than a result of the PACS vanguard and county-wide initiatives that were also affecting non-vertical integration general practices in the area, or if these were purely in line with more general trends in the rest of the NHS in England and Wales.

For example, in primary care, the health-care teams at the vertical integration company’s general practices had become more multidisciplinary, but this was a widespread trend:

I think we’ve really focused on the multidisciplinary team, so how we probably differ from most traditional practices is we’ve got a much smaller ratio of doctors to patients . . . So we’ve got more of multi-disciplinary team members and less GPs . . .

B08, primary care, organisational management

. . . what we’ve done in [vertical integration company] is to, firstly, create a much flatter multidisciplinary team structure, where the GP effectively sits in the middle of it rather than the top of it and acts partly as a consultant to the other members of the team to support them, but also focuses on complex patients, with the idea that most of the other work that GPs historically provided is done by other members of the team.

B03, primary care, clinical and managerial

By contrast, two interviewees who were not part of the vertical integration organisation stated that, as far as they knew, there had been:

. . . no changes to health service provision due to [the vertical integration company].

B04, primary care, professional representation

No, I think . . . it feels like a – I won’t say a missed opportunity – it feels like an opportunity we haven’t yet grasped.

B15, primary care commissioning, organisational management

We heard from several interviewees about health coaches working at general practices and a complex care team focused on people with complex and/or multiple conditions. However, these coaches and teams had been introduced before vertical integration, as part of the PACS vanguard.

We also heard about initiatives to bring some specialist care out of the acute hospital and into general practices, which were being introduced with the vertical integration practices:

We have got some links with the hospital, so we have got a couple of the specialist nurses coming and helping us with doing chronic disease clinics, for example, in our practices and supporting our practice nurses who run these clinics to provide slightly more complex care for who are the most difficult patients to manage.

B03, primary care, clinical and managerial

. . . a good example of that is we’ve got a respiratory specialist nurse, OK? So she’s based at the hospital, she was doing her job completely at the hospital, right? So she now comes out into [vertical integration company] practices to see the complex patients . . .

B06, primary and secondary care, clinical and managerial

We also heard from one interviewee a view that primary care staff in the vertical integration model had been given more scope than before to develop special interests/skills for the benefit of patients:

I think there’s been more creativity and imagination about special provisions . . . when I say special provisions, let’s think in terms of diabetic clinics, provision for the elderly, special needs and so on. That a great deal of attention has been given to what needs need special, particular attention and can we provide clinics for that . . . And this may be . . . that some of the existing staff are given special responsibilities, are trained, which is attractive to them. If they have – I’ll give the example of diabetes – they’re particularly interested in diabetic care, one of those practitioners will be trained or focused on that.

B16, primary care, patient participation group

None of the interviewees we spoke to felt that the vertical integration model had, as yet, had any impact on service delivery in secondary care.

Theme 5: measuring success and identifying unintended impacts

We asked interviewees how the success of the vertical integration model, or the opposite, was being, or could be, measured. We also discussed unexpected or unintended impacts. Measures of success related to the maintenance of patient access to local general practices, the net financial impact on the local health economy and recruitment and retention of primary care staff. The last of these three categories (i.e. recruitment and retention of primary care staff) was where unintended and undesired consequences were noted.

The vertical integration model had, so far, been successful in sustaining primary care in the short term, apparently saving at-risk general practices from closing and recruiting new staff there. No practices taken on by the vertical integration company had closed, despite several of them having been in severe staffing and financial difficulty when they were taken on by the acute trust-owned primary care company. Keeping local practices open is essential to maintaining patient access to primary care in rural areas with poor transport links, that is, areas such as Greenvale:

. . . if [the vertical integration company] didn’t exist now and I looked at our practices, I think three of them definitely would not be in the NHS.

B03, primary care, clinical and managerial

I know that morale in our team is good. People feel listened to. We’ve had, the practice I’ve had, hasn’t been [vertical integration company] for a huge amount of time, but going from being a practice that was really struggling to getting very positive patient feedback now is very nice. And to have people happy where they’re working, feel that they can actually make a difference, can develop things themselves, I think that’s a sign of success.

B17, primary care, clinical

Taken together, the vertical integration general practices are not yet breaking even, with a total deficit (expenditure in excess of income) of nearly £1M expected for financial year 2019/20, despite savings on the cost of back office functions. This reflects, in part, continuing high costs of employing locum GPs.

Offsetting that, there appears to be preliminary evidence (from Greenvale summary reports identified from documentary analysis) that the vertical integration model is supporting a reduction in growth of attendance and admittance to secondary care among patients of integrated practices. The creation of the vertical integration model took place on the basis that, although expenditure on general practice is much less than on acute hospital care, how general practice works can have a large impact on the finances of acute hospitals:

If you just measure primary care on their contract value, which actually in comparison to how much it would cost to run an acute trust is miniscule, you know, is a fraction of the cost, but if you look at their impact and influence on spend in the system, i.e. how many referrals they make, how many they signpost, how many they recommend someone to go to residential nursing care, how many they refer, you know, into other services, to how much they try, you know, how much testing they do, their costs are something like seven times what their contract value is.

B01, primary care, organisational management

I know that things like admission rates and readmission rates and all the classic kind of markers of control had improved because of what [the vertical integration company] were doing.

B02, primary care, professional representation

Senior managers at the acute trust argued that the vertical integration model had saved them money as a result of lower secondary care utilisation, achieved by stopping patients inappropriately entering secondary care (often as emergencies), as a result of better primary care provision. This also saves the local health economy overall, as the primary care provided is less costly than the secondary care it obviates:

[The financial impact on the acute trust] has been positive in the sense that we’ve avoided all of those emergency costs.

B15, primary care commissioning, organisational management

I firmly believe that if [the vertical integration company] wasn’t here, the cost to the system and the cost to the commissioner would be far greater, because a number of these practices would have handed back contracts and they would have been subject to transition funding and dispersal costs and all the rest of it . . . The counterfactual case is that we think the [vertical integration company] has saved the system between £6M and £8M each year. So there’s a £1M hit sitting on the [acute trust’s] books because that’s kind of where it sits, but the counterfactual deficit to the system could have been up to £6–8M, so £5–7M more than that.

B07, secondary care, organisational management

There were indications that the vertical integration model was having a net positive effect on recruitment and retention of GPs to the area, as the vertical integration arrangement enables more attractive career options for GPs and other staff, which aids recruitment and retention:

We can offer that the career is working partly in the hospital, partly in primary care, which a lot of GPs seem to want. We can offer portfolio careers, where you’re not just working in one practice. You can be moving around the practices, which, again, some people that might appeal to . . .

B10, primary care, organisational management

Although recruitment of new GPs by the vertical integration company had proved quite successful, GPs remained the main staffing problem and it had proven impossible to recruit GPs to work full time:

GPs are the key issue, definitely. We are able to recruit. We do, unlike a lot of partnerships, the traditional practices, we have been able to recruit. We’ve recruited quite a lot of GPs over time . . .

B10, primary care, organisational management

We’ve recruited something approaching 40 GPs in the last two and a half years, but not one single one of them is full time.

B03, primary care, clinical and managerial

The recruitment of non-GP staff seems to have been little affected either positively or negatively by the vertical integration model:

In terms of other clinicians, so nurse practitioners, practice nurses, again, it really varies in our practices – overall there’s probably been a balanced exodus and recruitment – but we have some practices where at the moment we’re really struggling to provide chronic disease management, for example, because we don’t have enough practice nurses. But there are others that are completely fully staffed.

B03, primary care, clinical and managerial

So, yeah, I don’t think on the non-clinical, I don’t think that’s any different than any other practice. People do come and go; that’s the way of life.

B05, primary care, organisational management

One negative unintended effect of the vertical integration company taking over some practices, which was remarked on by several interviewees, was that the remaining GP(s) at practices joining the vertical integration company had, on occasion, taken the opportunity to then retire or move to become locums soon afterwards. Likewise, GPs had been more inclined to take sick leave (for which they get paid when they are salaried) once the practice ‘business’ had been taken off their hands. Consequently, it is possible that, in the short term, the need to recruit new GPs had been increased in a few practices:

When we took on that practice, we knew they were in trouble because they’d tried to recruit for 4 years and they hadn’t been able to. There were three partners there and a salaried GP, and, at the time, we had assurances from everyone, including the CCG, that those partners would stay if [the vertical integration company] took them on and helped stabilise the practice. Within . . . a week of taking them on, the salaried GP retired, and then one of the partners resigned, and then subsequently the other two partners both went off sick.

B09, primary care, organisational management

I think integration is a big shock, and we do seem to lose GPs sometimes at integration. I think they’re not used to not being their own boss maybe. And usually I think they have plans to leave that we don’t necessarily know about.

B05, primary care, organisational management

Theory of change

The research team’s analysis of the findings from Greenvale is summarised in a theory of change map, which is presented in Figure 5. The predominant rationale for vertical integration between the acute hospital and several general practices in its catchment area was to enable those practices to continue in the face of recruitment and financial difficulties, where the traditional partnership model was no longer sustainable as a result. The preceding PACS vanguard in the area, which was focused on improving care for people with multiple conditions or other complex health needs, had reinforced and strengthened relationships, thereby providing fertile ground for the vertical integration idea, which also benefited from active and supportive leadership, both at the acute trust and in the primary care community locally.

FIGURE 5.

Greenvale vertical integration theory of change. CEO, chief executive officer.

Implementing the local vertical integration solution – which entails a limited company running the GP services, with the company 100% owned by the acute hospital trust – required a substantial up-front investment of time and effort by senior management at the trust and leading GPs. The trust’s then chief executive and one of the leading GPs were singled out in several of the interviews as having provided strong and active leadership. Creating the novel legal and governance structure preferred at Greenvale required detailed discussions with NHSE&I and the Department of Health and Social Care, as well as expert legal and VAT-related advice. The provision of centralised back office functions, drawing on those of the acute trust, to achieve both functional and systemic integration, along with the creation of a central management team for the vertical integration company’s constituent general practices, appears to have been successful and may have contributed to the viability of primary care financially, but it is still operating at a loss.

Recruitment and retention of GPs, now all salaried, had been supported by the vertical integration arrangement and no practices had to close. This is important for the maintenance of patient access to primary care in this rural area with poor transport links. Recruitment and retention of non-medical staff at primary care practices had been maintained.

The success in sustaining primary care and stabilising recruitment had enabled the continuation of the PACS model’s innovations for people with multiple morbidities and the gradual introduction of other changes in patient care. Primary care practices had more MDTs than previously and they had access to some specialist services via specially trained practice staff and some limited outreach from the acute hospital.

Finally, the acute trust and its board were strongly of the view, based on their own analyses (not in the public domain), that, despite the financial deficit for both their own trust and the vertical integration primary care company (with the latter contributing to the former), the trust’s financial position was considerably stronger than it would have been without implementation of vertical integration. Comparing vertical integration practices with patient flows from non-vertical integration practices in the county, the acute trust sees apparent reductions in the volumes of both emergency and elective work they are called on to deliver to the vertically integrated practices, which imply net savings to the local health economy to the tune of a few millions of pounds per year relative to the position that would be expected without vertical integration.

Summary of key findings

• The rationale for implementing vertical integration at Seaview was to address the potential problem of a large number of general practices closing by integrating them within the LHB, as there were no other viable alternatives.

• There was a shift in strategy from a desire to return practices to GMS contract-holding providers, to holding onto practices and applying a MDT model of working.

• The LHB was able to set up a governance team specifically to address the management of integrated practices, comprising primary care and secondary care stakeholders in the region.

• Attracting GPs to a rural/coastal area remains difficult, especially in remoter locations, and locums remain costly and scarce.

• In the vertical integration arrangement, the LHB was able to create attractive portfolio-based career opportunities for salaried GPs, and practice managers were given opportunities for greater responsibility.

• Little is known so far about the impact on demand for secondary care.

The study team carried out 17 interviews with stakeholders and a single observation of a governance meeting hosted by the LHB, which had primary and secondary care practitioners in attendance. Further observations had been scheduled in February and March 2020; however, because of the COVID-19 pandemic, the LHB advised the research team not to attend meetings in its clinical settings to prevent the unnecessary risk of spreading the virus.

Theme 1: understanding the need for and purpose of acute hospital integration with primary care

Integrating with, and managing, primary care practices was a predicament that was placed on the LHB, with a considerable number of general practices handing back their contracts over a sustained period of time. Initially, the LHB was unprepared to manage a large number of practices more centrally. As a result, there was no initial model for integrating the practices with the health board, rather the strategy developed as the number of managed practices grew. The key goal for the LHB was to return practices to independent contractors:

In the last four years, we’ve received – I can get the number for you – so we’ve got 16 managed practices, we’ve probably received around 20 – over 20 resignations from contract. So for some of them I’ve managed in the central area, two of them that didn’t become managed were very small, single-handed practices, that had been ticking over for years – the GP has just retired, and they’re so small you wouldn’t do anything else but disperse them [their patients].

C02, secondary care, organisational management

There were a number of reasons why the LHB was being asked to manage practices. The main reason was the inability of some independent providers to recruit and retain GPs. Often, these practices were led by GPs who were nearing retirement and who were unable to find suitable successors to take over their practice, whereas, for a significant number of other GPs, the pressures of running a practice, increasing workloads and tensions arising with their partners were all contributory factors:

. . . there’s sometimes things going on behind the scenes that we don’t know about, in terms of partnership fall-outs and things like that – but I think the main reason is burn-out. It’s: ‘Don’t want to be responsible for the whole business anymore, had enough, but I’m quite happy to continue as a GP, seeing my patients, but I don’t want to run a, you know, work 12-hour days and run a practice’. Can’t recruit GPs; the age profile of GPs in [Seaview] is going up. You know, they’re getting older and older. And we’re not – I don’t think we’re replacing as quickly as they’re retiring . . .

C02, secondary care, organisational management

However, according to accounts from LHB employees, GPs and primary care staff, not all the practices had been performing poorly:

. . . it was a very good practice and certainly not a practice you would have expected to have resigned.

C06, secondary care, organisational management

Notably, as the number of general practices being managed by the LHB increased, in parallel with being unable to find suitable independent primary care providers to take over practices, the LHB changed its rationale and its understanding of how to make best use of the opportunity of having direct involvement in the delivery of primary care services:

. . . there was always the feeling this is the way the health board wants to go, so there was an awful lot of discussion about that: ‘Is this now the health board strategy? How hard have they tried to find GMS providers?’ And I think certainly early on, it was ‘No this is about us; don’t want anymore, please, no more!’.

C03, primary and secondary care, organisational management

Yet, given this new opportunity, there were still a number of challenges that persisted, primarily how to achieve standardisation across managed practices over a very dispersed and varied geographical landscape, where local knowledge of service provision was required to treat a complex patient population. Therefore, the immediate challenge was to address the nature of horizontal integration, before attempting to make gains through vertical integration between the LHB and general practices:

. . . if we have them all in one area, in one geographical area, and it’s much easier to make that horizontal integration than it is separately, so it’s working out, I think, going forward, how we do that horizontal integration. The vertical integration, we’ve just scratched the surface of, and that would be my next bit, I suppose, primary, community, but we keep saying, ‘Actually, the vertical integration is where we’ve got the greatest gains potential, because of clinical pathways being changed’.

C03, primary and secondary care, organisational management

Theme 2: developing a model of integration and an implementation strategy

The strategy and processes of taking over a practice at Seaview became more refined over time. The first step for the LHB was to speak with a practice to understand more about the specific challenges that it was facing at the time and to try to encourage the partners not to return their contract to the LHB:

I mean, with any practice, when they first resign, if we didn’t know they were going to resign, in the situation we’re in at the moment with this latest one, the first thing we would do is meet with the GPs, first of all to try and see if there’s any way that possibly they might not wish to change their minds, to try and understand why they’ve done it, and for us to get a better feel of what the problems are and also to understand whether they themselves, as GPs, are interested in continuing in any shape or form.

C06, secondary care, organisational management

In most cases, when a practice returned its GMS contract to the LHB there was a three-pronged approach, among which integration was considered the third and final option. (However, in a small number of cases, a decision to directly manage a practice following conversations with primary care clusters was taken immediately.) First, the LHB would attempt to readvertise the practice to see whether or not there were GPs in the local cluster, the UK or internationally who would be interested in working in the area. Up to autumn 2019, this option had yielded only one former GP from the local area, who agreed to acquire and manage three practices in a staggered sequence. Second, if there were no interested parties, then the LHB would decide whether or not the practice patient list could be dispersed to neighbouring practices. Third, the LHB would decide whether to integrate the practice as it is or to merge the practice with another geographically local practice that had already integrated with the LHB:

. . . the process was fairly established, so for single-handers, you obviously only have 3 months to sort it out, for group practices 6 months. So the single-handers – it had to be moved quite quickly. So it went out to the cluster GPs to see what their feeling was about what should happen next, and the options on the table of course at any time you get a resignation are: dispersing, taking over, re-advertising or possibly combining two lists, so if there was a willing nearby practice.

C03, primary and secondary care, organisational management

A fundamental action soon after integration was to invest in the workforce to address the shortage of GPs. This led to the development of greater and more widespread use of models of MDTs working across the LHB area. As a result, in one area within the LHB, a model was developed to encourage interdisciplinary working, supporting patients living with complex multimorbid conditions and reducing medicine prescribing. The model was brought about following the quick resignation and return of GMS contracts of five general practices in a small geographical area covering a population of approximately 20,000 patients. Clinical teams were established within this model to treat groups of 5000–6000 patients, using a range of primary care staff:

So within one of the normal teams, there’s – depending on the hours that they work – sort of perhaps one and a half full-time equivalent GPs. We have nurse practitioners, possibly some ANPs’ [advanced nurse practitioners’] involvement, there’s an occupational therapist linked to that, a pharmacist and then we also have a team co-ordinator, who tries to manage the team’s work and direct it according to where’s best to be dealt with for the patient.

C09, primary care, clinical

There have been early attempts, using this model, to integrate into primary care settings some specialised services that are usually delivered in secondary care settings. However, there were certain caveats:

. . . secondary care diabetic care and doing it much more holistically in the community, it’s a piece of work, but it is not . . . it’s very difficult to realise the potential of doing that in other subjects without a lead who’s previously worked in secondary care, coming out to primary care and going ‘Woo!’ So the separate management structures, I think, between the hospital specialties and the centralisation of them limits it . . .

C03, primary and secondary care, organisational management

A number of primary and secondary care colleagues were critical of this MDT primary care model. First, many interviewees felt that the nature of the additional funding the first area received could not be replicated widely across the LHB. Second, recruitment of staff is difficult in rural and/or coastal areas farther away from the English border. Third, many interviewees questioned the effectiveness of the MDT model:

I mean, the idea when [subsite within Seaview] was set up was that it was going to offer a different type of service, there’s mixed messages about how that’s worked, you know, the idea, I know everybody talks about it, but there are some, sort of . . . sometimes you do get a bit of feedback that, actually, there are problems with it.

C06, secondary care, organisational management

Last, although practices have been integrated with the LHB, they continue to maintain a reasonably integral role in local primary care, working via local clusters. Managed practices’ involvement in cluster working has been a cause of tension with non-managed practices, which feel disadvantaged because managed practices continue to have access to cluster funding and also receive additional resources from the LHB:

This is going to sound crude, but because the GP clusters have – because they are still GP clusters – but they are going to have to change – and some of them are much more mature than others – but they have funds, and of course you’ll have the GMS practices who, some of them, tend to think, ‘Well, why are the managed practices getting cluster funds?’.

C06, secondary care, organisational management

Theme 3: making the change – from General Medical Services to subcontracted providers of primary care

Theme 4: impact on patient management – changes to primary and secondary care delivery

A significant challenge faced by the LHB was how to best treat an ageing population living with complex multimorbidities who were dispersed across a large geographically rural/coastal area, along with more transient tourists who access health services during seasonal peak times throughout the year. In response to this challenge, there was a strong focus on developing services that support patient self-management, with the hope that this would bring about a subsequent reduction in health service demand:

We work with people with chronic long-term conditions, so looking at self-management. So things like COPD [chronic obstructive pulmonary disease], pain – pain is a big one – and looking at trying to reduce the impact of those conditions on people’s lives. So, essentially, increase their function in the community and reduce demand on health services . . .

C04, primary care, clinical

The mechanism, as stated in Theme 2: developing a model of integration and an implementation strategy, was the introduction of distinct clinical and non-clinical allied health professionals, using a model of MDT working within managed practices. The majority of primary care staff welcomed the introduction of additional resource to manage patient demand. However, this introduction of new staff contradicted patients’ perceptions of how care should be delivered, with many patients still continually requesting to be seen by GPs:

I think surgeries are trying to get away from, in some parts, of always seeing the same patients, because you’re not, actually – at one time you used to be registered with a particular GP – and some patients still think you are, so they want to see that GP. So we’re obviously trying to move away from that and get patients to see other clinicians, so signpost them. They don’t always have to see a doctor, but some people have got it in their head they’ve got a problem, they need to see a doctor. So it’s educating, trying to educate the patients as well.

C08, primary care, organisational management

And so we have had, probably the biggest difference would be, having the occupational therapists that have particularly linked with the community and a lot of things like social prescribing have raised the profile of that, I think, within the practice.

C09, primary care clinical

However, although patients might be seen more readily, the new system does not necessarily preclude the patient being seen by their preferred choice of health professional:

So, I think, for me, availability of appointments is always what everyone talks about in practice, but actually getting the person to the right practitioner is something that perhaps patients didn’t appreciate right at the beginning, whereas now they probably do. So I think over time probably patient opinions have changed. And there’s a lot of noise initially; if you get a lot of noise in the system that ‘That’s a managed practice, so you won’t be able to get an appointment’, it ends up becoming the truth, even though, actually, you’re getting an appointment with the appropriate person.

C03, primary and secondary care, organisational management

Overall, there remained a dichotomous relationship between primary and secondary care, whereby there remained some lack of communication and a lack of data/IT integration across the interface. There were instances where secondary care specialists were becoming more involved in delivering primary care-based services; however, such instances remained rare:

Just small little pockets of, so, for example, one orthopaedic consultant is coming to work at one of the managed practices alongside. But, again, really difficult to get the processes, just things, like, because it’s a GP practice, you can’t get notes delivered to it and they work on clinical notes system, you know. The barriers sometimes between historically very different organisations are difficult.

C03, primary and secondary care, organisational management

I think we need more connection with the secondary care physicians as well. I think, you know – and I recognise it’s really difficult because there are vast differences and secondary care’s under a lot of pressure as well, so I do recognise that – but that would be helpful that we’re actually working together.

C09, primary care, clinical

With regard to staffing, the main concern for the LHB was how to attract health professionals to the area and get them to remain for a sustained period of time. The greatest challenge was the employment of locums, which was creating competition across the three regional areas. Therefore, the LHB tried to develop attractive packages to entice GPs, while, at the same time, co-ordinating across its whole area so that practices were not competing against one another within the LHB area for the limited pool of potential recruits:

One of the main things we sorted out early, one of the initial things we sorted out as a group, was terms of conditions of locums – because we’re all fishing in the same pond for the locums – and making sure they don’t play us off on each other, and we’re all paying the same rate and setting expectations . . .

C02, secondary care, organisational management

I described the training offer, support with training, special-interest development. Obviously, they get the terms and conditions that are better. We try and put support in, so they are there doing what they want to do, . . . the running of the practice isn’t their problem anymore, if that’s what they want. We offer, yeah, all sorts of different things, anything, really, just to incentivise them!

C02, secondary care, organisational management

Non-clinical staff faced challenges working with locum GPs compared with permanently employed GPs. Some staff commented on locums who were working to rule, placing greater pressure on the remaining staff:

So I feel bad for some of them because some of them, they get away with not – they’re like, ‘Oh, I’m a locum; I won’t do that kind of thing’. And then there’s other ones that are long-term locums and I forget that they’re long-term locums but they’re not a salaried GP, because they act almost like a salaried GP, and then they’re, like, ‘Oh, we’ll be around now for a few months’, and it’s, like, [sighs]

C07, primary care, clinical

And just the admin[istration], like looking at letters and things, relying on salaried GPs when you’ve only got, like, one salaried GP and they only work 2 or 3 days a week, it’s just . . . you just rely on the regular locums, and when they all go to go off in the holidays and stuff, sometimes it’s just not safe because you don’t have the consistency and . . . they can just be, like, ‘Yeah, I’m not coming in for the next few weeks; good luck’ kind of thing, and we don’t have anyone. And then you feel bad for the . . . because if there’s just one locum that’s on and they end up leaving because they’re, like, ‘Well, I’m a locum, like, this is more salaried GP job’ kind of thing.

C07, primary care, clinical

The greatest tension for existing staff was the transition to AfC contracts and the recruitment of new employees in primary care being paid more, as a consequence of that, than those who have been working for longer as part of independent practices:

Well, it’s a mess, basically. There’s an understandable want from the staff in the practices when they’re transferring to us [the LHB] that, obviously, they transfer on their own terms and conditions, and then the problem always comes in when we recruit new staff who are on Agenda for Change and then you’ve got one member of staff who’s been there 20 years on this much money and then you bring in a new kid of about 16 and they’re on more money and you think, ‘Well, I can understand why you don’t like that, but it’s because of AfC’.

C06, secondary care, organisational management

. . . we’ve got some existent staff whose morale is very low. For me, I would love them all on Agenda for Change, but it’s a higher decision than me, and we are working towards getting them all over there. So, for example, we have one of our practices has three members of staff. They’re the eldest, the longest-serving members of staff, on their own terms and conditions. And we have a junior who’s just started on an AfC getting better paid, better terms and conditions, and they’ve never known the role before; they’re new to the role.

C05, secondary care, organisational management

Theme 5: measuring success and identifying unintended impacts

The revenue generated from primary care practices was considered an important marker of success. However, of greatest importance was the amount the LHB was currently spending to bring practices into a sustainable model. Therefore, a fundamental marker for measuring success was to control the expenditure being made to manage and redesign health service delivery across managed general practices:

. . . we’ve got a, sort of, a set budget based on what we would generate as a GMS practice. They’re all overspent . . . But they’re overspent for a plethora of reasons. One of the obvious ones is reliance on locums, so – for [subsite within Seaview], we’re not too bad now, so that overspend has come right down. So their overspend has come down. Reliance on locums.

C02, secondary care, organisational management

. . . there’s always that because we have restrictions within the health board, so we can’t always spend the money as freely as GMS practices. For instance, if we wanted a new phone system, GMS can just go out and purchase whatever. We have to actually go through the procurement process . . . , so we have a lot more restrictions.

C05, secondary care, organisational management

Monitoring financial expenditure was important to show potential interested parties that the general practices that were being managed by the LHB were both financially stable and a viable business opportunity if someone was willing to take them over:

So you’d want to see the accounts, effectively, wouldn’t you, of a new business? So I think although, you know, I’ve heard people say, ‘Well, what’s the point in doing the enhanced services, because the health board are paying themselves?’ but I think it’s also important to have that service still running because then at least there is – and for the health board as well – to see how much that the practice are worth and that it is a viable option having a practice within the health board.

C08, primary care, organisational management

One of the opportunities the LHB envisaged when taking over general practices was the ability to train staff to undertake a larger number of responsibilities and also to identify candidates who were already skilled enough to progress their careers across the LHB:

I think in terms of development for staff, that’s clinical and administrative, I think there’s a lot more opportunities. I think within an independent practice you’re very restricted, because, for example, if you’re a practice nurse, unless the practice, depending on its size and, sort of, financially, whether they’d be able to encourage them to work towards an advanced nurse practitioner role, or if you were a nurse practitioner, where do you go after that?

C08, primary care, organisational management

There were some notable drawbacks. One drawback was a lack of leadership at practice level, seen as a result of the shift within the GP workforce from partnered GPs to part-time salaried GPs or locums. In addition, there was a perceived sense of competitiveness and tension across staff in certain areas within the LHB, who found themselves, unintentionally, competing with one another for the same small pool of resources:

And one of the huge differences is the onus on the GPs working within the practice to take on clinical leadership, such as they would do were they partners, and not just turn up and see the patients, as they would do under a salaried GP contract. So I think the big – for me, one of the huge – differences is that establishing clinical leadership within the practices. There are huge differences of course: as a partner, you’re running the business, so there’s no timesheets or whatever, so you expand the work – no, you expand your time to fit the work – whereas one of the big differences, I think, in directly managed practices is people are on salaried contracts, so they’ve got no responsibility to stay when the work is not done. And so the shift of leadership towards management, I think, is one of the biggest challenges that if you’d had more time you could have planned.

C03, primary and secondary care, organisational management

I think it’s a very ‘us and them’ kind of scenario . . . Central on one table, East on another table, West on another table.

C07, primary care, clinical

Yet there were some widespread successes from the integration. Although many of our stakeholders recognised that there was still plenty to do to improve the nature of functional integration with the LHB, many felt that noteworthy changes had been made to health service delivery and, having the opportunity to deliver beyond the GMS contract (known locally as ‘GMS plus’), better communication across the care interface had been established (mostly between primary and secondary care, rather than primary and community care), ratings with the CQC had improved and patient satisfaction had improved (i.e. fewer complaints had been received):

I think communication has improved. Referrals are quicker. The access to information is there. Unfortunately, IT systems aren’t as cohesive as they could be.

C01, secondary care, organisational management

And from a CQC perspective, we know that all of our practices have got higher ratings since they’ve been with us.

C01, secondary care, organisational management

. . . we also try to do a lot of engagement with patients about satisfaction and things, so we’ve got some of that information.

C02, secondary care, organisational management

More recently, the LHB has been committed to understanding more about the changes that the integration of practices with the LHB had achieved. This played a role in allowing access to the research team to speak with stakeholders:

. . . we know from [subsite within Seaview], although I haven’t looked at it for the last 12 months, is medicines, pharmacy spend has come right down, prescribing spend. And in terms of the cluster, they are within the lowest users of the ED [emergency department], that practice population.

C02, secondary care, organisational management

Theory of change

A theory of change map for this case study site is presented in Figure 6. The rationale for Seaview was to address the immediate crisis of a large number of general practices closing and, consequently, integrate them within the LHB because there were no other viable alternatives. The reasons for closure were an ageing profile of existing GPs, being unable to attract new GPs within a traditional partner model and increasing GP workload.

FIGURE 6.

Seaview vertical integration theory of change. BMA, British Medical Association; LMC, local medical committee.

Different from our other case study sites, the desired impact of integrating and managing general practices at Seaview was to eventually be able to return general practices to GMS contract-holding providers. However, owing to the ever-increasing number of general practices returning their contracts to the LHB, the strategy for managed practices has now shifted. Therefore, although general practices remained integrated with the LHB, the rationale was to improve patient care through the regional application of a MDT model. There was little or no reference within interviews and observations of achieving clinical integration across primary and secondary care in the immediate term.

This came with a number of caveats. There remained significant variation in the management of integrated practices across the LHB’s area. A ‘one size fits all’ approach to MDT care was considered difficult by both primary and secondary care stakeholders. Nevertheless, many staff recognised that having integrated practices made it easier to make changes to clinical pathways for patients, especially between primary and community care (more so than between primary and hospital care because of geographical constraints faced by patients attending distant acute hospitals). In particular, stakeholders found it difficult to comprehend how a MDT model should be implemented if the rationale was based on ultimately returning practices to GMS contract-holding providers. Therefore, a clear long-term strategy for the sustainability of primary care was desired.

There was a substantial cost associated with integrating practices, especially for a LHB in special measures, as was the case at Seaview. Attracting GPs to a rural/coastal area remained difficult, especially in remote locations, and locums remained an expensive and scarce resource. There were also significant estates-related costs, with a number of managed practices being moved to new premises that were fit for purpose and could accommodate a MDT model.

Nevertheless, integration had a number of advantages. The LHB was able to create attractive portfolio-based career opportunities for salaried GPs and practice managers were given opportunities for greater responsibility. Nearly all staff would eventually receive improved contractual terms on AfC, which brought perceived greater employment security. In addition, the LHB was able to set up a governance team specifically to address the management of integrated practices, with primary and secondary stakeholders, both within and outside the model.

As a result, Seaview had achieved some sustainability with regard to the number of general practice closures, but little is known so far about the effect on secondary care demand or the financial cost of integration.

Summary of key themes

• The main rationale for the introduction of a vertically integrated model, evident in all three case studies, was to sustain primary care by supporting local general practices that are in financial and recruitment difficulties and, at the same time, manage demand effectively and efficiently across primary, secondary and community care.

• Our overall theory of change shows that the early implementation of vertical integration is focused more on achieving functional than clinical integration.

• Closer integration was attributed to previous good relationships between primary and secondary care locally, and to historical planning and preparation for integrated working across the local health economy.

• Our initial evaluation of three case studies, two in England and one in Wales, implies that vertical integration may, indeed, have a role to play as a route to better integration of patient care, at least in some geographical areas. Although there have been innovations in patient care, these do not appear to be specifically attributed to the vertically integrated model.

• The inevitable corollary of salaried doctors replacing independent GP contractors is a degree of loss of autonomy for GPs.

• Making the shift to vertical integration imposes a significant short-term challenge on acute trust/LHB staff (who are used to operating in large organisations focused on secondary care) to learn about and understand the practicalities and culture of running primary care.

• Improving care pathways, and the efficiency of the local health economy, for patients who are high users of emergency secondary care and/or living with complex or multiple morbidities was a particular focus at two of our three case study sites. We will explore impact and effectiveness in greater depth in our phase 2 evaluation.

Chapters 5–7 demonstrate overlapping, but not completely congruent, findings from the three case study sites. Our study design deliberately selected three sites that, although similar in the core feature that general practices were effectively being run by local acute hospital services, had clear differences in the ways that they were set up. At Urbanville, in England, the acute trust ran the integrated general practices as a division of the trust, and employed the GPs and other primary care staff of those practices just as it employed the medical and other staff at its hospitals. At Greenvale, in England, the acute trust set up a wholly owned subsidiary company to run the integrated GP services and employ the primary care staff. At Seaview, in Wales, the LHB, which had previously been focused mainly on managing secondary care, took over the direct management of GP services in practices that otherwise would have handed back their contracts, alongside the management of hospital services.

A strong theme at all three sites was that the rationale for vertical integration was to enable the continued existence of general practices in difficulty and thereby support a local health-care system in which patient demand would be better managed and patient pathways more appropriate than if general practices were allowed to fail. At all three sites, the establishment of vertical integration created opportunities for better integration of patient care across secondary and primary care settings and for rationalisation of common functions, such as back office functions. However, the extent to which these opportunities were being exploited differed across sites.

In the following section, we present and discuss an overall theory of change for vertical integration between acute hospitals and general practices, before going on to consider similarities and differences across the case studies, referring to the same themes as in the individual case study chapters. The chapter concludes with reflections on the strengths and weaknesses of the evaluation, and highlights issues that are candidates to be the focus of a second evaluation of vertical integration.

An overall theory of change for vertical integration

Overview

With reference to Weiss’59 four main purposes when conducting evaluations on programme theories (as detailed in Chapter 3), we have spoken to a range of stakeholders across sites to make their assumptions of vertical integration explicit. In developing our overall theory of change and accompanying illustration, following Vogel’s60 criteria, we have provided a detailed description of the wider context of vertical integration in the light of wider primary care policy in both England and Wales, and provide the long-term outcomes that vertical integration aims to achieve, detailing as many assumptions of how these outcomes are perceived to be achieved. An overall theory of change map is presented in Figure 7, drawing out the common features of the individual theories of change for the three case study sites. Although no single theory of change can do full justice to the specific aspects of all cases of vertical integration, there are some core aspects, and this overall theory of change map provides a framework to understand how, at a high level, the implications of vertical integration between primary and acute care in the UK can vary considerably across sites with regard to local feasibility (e.g. setting, context and resources). In addition, our theory of change illustrations remain in development and reflect that different sites are at different stages (i.e. from conception to achieving clinical integration).

FIGURE 7.

Overall vertical integration theory of change.

As the theory of change map illustrates (see Figure 7), the main rationale for the introduction of a vertically integrated model, a rationale that was evident across all three case study sites, was to sustain primary care by supporting local general practices that are in financial and recruitment difficulties and, at the same time, manage demand effectively and efficiently across primary, secondary and community care. This focus differs, to some extent, from international examples (e.g. Kaiser Permanente and the Alzira model)42,45,64 where rationale and implementation were focused principally on closer integration of primary and secondary care to improve quality of care and increase efficiency, with no explicit reference to sustaining primary care in the face of growing recruitment/retention and workload pressures.

In our case studies, the introduction of vertical integration across sites led to increased interdependencies among individual general practices and the acute trust or LHB with regard to controlling costs, responding to patient needs and developing legitimacy throughout the local health economy. Similar to their international predecessors, UK examples of vertical integration are challenged with initial costs and barriers, and with limited co-ordination of care across settings. However, despite these challenges of implementation, the desire to integrate both clinically and functionally across primary and secondary care has persisted.

The theory of change shows that the early implementation of vertical integration is primarily centred on achieving functional integration as opposed to clinical integration. 65 As Lasker et al. 66 further emphasise, key to achieving vertical integration is integration of information systems and financial arrangements. Therefore, functional integration can be synergistic with clinical integration. At all three case study sites, stakeholders sought to achieve greater ‘closeness of working’ to achieve functional integration,67 whereby organisational arrangements and designated staff to oversee the introduction of vertical integration led to more cohesive working, both among general practices and between them and the acute trust/LHB.

Synthesis of case study findings

As part of the analysis, we identified five major themes, as well as four subthemes, that provided a framework for evaluation of all three case studies. In the following section, we reflect on the cross-case study implications within each theme.

Feedback workshops with case study sites

In July 2020, the study team approached all three case study sites to feed back results from the rapid evaluation and learn how the vertical integration arrangements fared during the COVID-19 pandemic. Workshops were delivered with two sites (Greenvale and Seaview) in August and September 2020, sharing and discussing overall and site-specific findings and theory of change illustrations. The study team made a number of attempts to arrange a workshop with Urbanville, but was unable to arrange and deliver a brief session with key stakeholders prior to submitting this report.

At Greenvale, the workshop was attended by six colleagues, including medical, HR, operations directors, business and finance managers and a GP, all of whom were from within the primary care company. Attendees, in general, agreed with the findings presented, but stressed the development of forthcoming changes to patient care and the collaborative innovation happening across the limited company and general practice. However, the attendees recognised that it was difficult to distinguish innovation as a direct result of vertical integration rather than stemming from working as part of PCN. In addition, attendees felt that the governance and managerial framework that had been established was vital to sharing services across primary care and would create opportunities to do more with secondary care in the future.

In relation to the COVID-19 pandemic, general practices within the vertical integration arrangement felt that the were able to have a ‘greater voice with commissioners’ than might otherwise have been the case and, therefore, were able to help shape the local response to the pandemic better as part of the vertical integration limited company. Furthermore, the GP in attendance felt that the centralised governance team was crucial with regard to working collaboratively with neighbouring PCNs (as opposed to working more closely with the local acute hospital). GPs employed by the limited company were happy to move across practices to provide greater clinical oversight and care during early stages of the pandemic. A novel finding from the workshop was that the limited company had experienced a significant surge in the number of GPs recruited on permanent contracts since 1 April 2020. In total, 18 GPs, the majority of whom were working as locums within managed practices, had been recruited on, predominantly, permanent but part-time contracts. Anecdotal evidence suggests that GPs may have been seeking greater job security as the pandemic continues.

At Seaview, the workshop was attended by three heads of managed services for east, west and central areas, and two assistant directors of primary care and community services. Between the time data collection was completed (February 2020) and the time of the workshop (September 2020), a number of new practices were being managed, whereas other practices had returned to independent providers of primary care. As a result, as of September 2020, five practices in the eastern area, four in the central area and five in the western area were being managed. In addition, there had been greater integration with community services, such as district nursing and medicines management, but less so with mental health services, for all general practices under the remit of the LHB, but still no integration with acute hospitals.

Attendees welcomed the presentation and were surprised to learn that their experiences of managed practices had such strong similarities with English sites. The attendees were generally in consensus with the study team with regard to ‘overall’ findings. One attendee expressed challenges with working on the basis of GMS contracts and the need to stay within global sum payments, which can vary widely between practices, even on a per registered patient basis, and, therefore, suggested a need to move away from working towards this contractual model. Another attendee felt that the LHB was still able to deliver services beyond the contract model, but that provision of additional services comes at a greater cost for the LHB. Notably, the two attendees felt that the process of managing practices was a challenging learning experience, whereby ‘rolling up your sleeves and running primary care is quite different from commissioning it’.

During the COVID-19 pandemic, the scale and spread of technology in primary care had been very rapid and encouraged by Welsh government, with far greater use of telephone triage. Although appointments using such technology can be much longer because of patient unfamiliarity with using technologically grounded services, in managed practices, the LHB was able to redeploy staff to areas of most need to create ‘upstream prevention’ to a greater extent than was possible in non-managed practices. However, there remains a patient backlog to work through across managed practices (e.g. annual reviews for those living with long-term conditions), which will be challenging with respect to treating COVID-19, as well as managing seasonal flu vaccinations and cases.

Reflections on the robustness of the results

The study team completed a qualitative mixed-method comparative evaluation, following established methodology and guided by previous evidence of implementation, 69–71 while iteratively engaging with published literature. The robustness of our findings has been shown in a number of ways throughout the evaluation. The use of triangulation (i.e. collecting data through interviews, non-participant observation and documentary analysis)72 enabled the team to develop a comprehensive understanding of the rationale and the process of implementation demonstrated by our theory of change maps. We completed a large number (n = 52) of interviews with a range of stakeholders across primary and acute care settings (however, we had fewer interviews with clinical staff working in acute settings and those working within the community). Given the extent of our data collection (despite recruitment of interviewees guided by local gatekeepers and fewer completed interviews of secondary care professionals at one site) and the scope of the evaluation, we are confident that data saturation was achieved. 73

Member validation occurred throughout data analysis among study team members,74 senior members of the BRACE executive team and external colleagues with expertise in theory development and in undertaking evaluations. In addition, theories of change were developed with significant input from evaluation experts (however, these have not been confirmed/checked by case study sites) and consolidated with numerous analysis workshops.

The team encountered a number of challenges that resulted in a longer than anticipated period to complete the evaluation (i.e. 18 months between the identification of sites to the submission of the report). Notably, the team encountered difficulties in obtaining documents for analysis due to data-sharing governance issues between an acute trust and the University of Birmingham, which led to the team relying on publicly available information. Therefore, the team was often subject to meeting the criteria for studies categorised as research when dealing with NHS trusts and university governance teams, with standards and processes for service evaluation remaining unclear. As a result, processes for ethics, data-sharing and appropriate governance remains difficult to complete for rapid evaluations within 12 months in NHS settings. With respect to project management, a small three-person team with a clear hierarchy, roles and responsibilities (i.e. principal investigator, project manager and researcher) consistent throughout the project was a significant positive, alongside weekly communication to facilitate rapid working.

Finally, there was significant cross-learning with another BRACE project75 that is specifically exploring the early implementation of PCN working across England. This cross-learning supported nuanced interpretation of how vertical integration was being implemented in tandem with wholesale primary care policy changes involving horizontal integration.

Limitations

There were a number of limitations when completing our rapid evaluation. The team found it challenging, when deciding the scope of this qualitative rapid evaluation and the nature of exploration, to determine what could (1) be completed in a relatively short timescale and (2) be responsive to NHS health care policy. Following the project design workshop, research questions were identified and, alongside discussions with senior members of the BRACE centre, it was deemed appropriate to split our research questions into two separate evaluations (i.e. implementation and impact). In comparison with the wider BRACE programme, this rapid evaluation has informed early learning on best practice with regard to engagement with stakeholders in relation to design, patient and public involvement (PPI) and data collection, as well as resources required. Across all three sites, fewer than intended non-participant observations were completed because of meetings being rescheduled/cancelled at short notice as a result of the onset of the COVID-19 pandemic. Access to such data would have further strengthened the robustness of our findings, as well as potentially giving the study team increased access to a wider number of stakeholders to approach for interview, rather than being reliant solely on our gatekeepers.

The success of our interviewing approach was largely dependent on the relationships members of the study team established with gatekeepers over a relatively short period of time. In addition, the success of our recruitment approach also depended on the strength of the relationship between the gatekeeper and the interviewee. This may explain why the study team had limited success in recruiting NHS colleagues working outside vertical integration, which would have counteracted potential bias in our sample of interviewees and further strengthened our critique. In addition, the study team recognises that the balance of interviewees from primary and secondary care varied between sites. Despite concerted efforts by researchers, poor recruitment of secondary care practitioners could be explained by their limited contribution to the vertical integration model thus far, as well as gatekeepers potentially highlighting participants who they implicitly felt were more involved with the implementation of vertical integration. This is despite the study team completing a stakeholder mapping exercise with gatekeepers to identify a range of stakeholders. Notwithstanding contacting a range of stakeholders at each site, a small number of stakeholders did not respond to initial e-mail invitations. On reflection, the team does not believe that this imbalance across primary and secondary participants influenced the overall data set, but rather focused our interpretation on the most pertinent purpose for vertical integration (i.e. creating stability and sustainability in primary care). However, exploration of the impact on secondary care will be a key area in any possible impact evaluation (phase 2) and we remain cautious over the conclusions drawn with regard to this. Yet, given the nature of the relationships established with gatekeepers and senior colleagues across all three sites, all three sites remain interested in taking part in a further evaluation on the effectiveness of vertical integration.

Another methodological limitation was the varied pace of data collection across case study sites. The study team experienced a ‘stop-start’ approach to early data collection, which made it difficult to share learning simultaneously; specifically, the study team experienced considerable delay in collecting data from one site. We encountered significant difficulty with regard to signing a data-sharing agreement between one case study site and the University of Birmingham (study sponsor) because of differing interpretations of how best to share relevant documentation that may contain sensitive and/or personally identifiable information. Last, owing to the onset of the COVID-19 pandemic during project write-up, the team was unable to undertake a site-specific and an overall face-to-face workshop with policy experts.

We recognise that there are a number of advantages and drawbacks of using theory of change methodology and developing accompanying diagrammatic illustrations as part of this evaluation. The approach provides a rigorous systematic method of determining how, why and if a programme works. Given that acute hospitals managing general practices is a recent phenomenon and with limited literature, there was a clear need to determine the rationale, inputs, processes and desired outcome as to what vertical integration is expected to achieve in a UK-wide setting. Furthermore, the study team has outlined all contextual factors drawn from documentary analysis and interviews, although there may be other political, social and economic factors at play and having an impact on vertical integration at each site. In addition, theory of change is able to bridge process (phase 1) and effectiveness (phase 2) evaluations.

We also recognise that our understanding of theories of change is based on post hoc rationalisation by stakeholders who we were able to interview at the time of data collection, and may not be wholly representative of vertical integration as it happened. However, we were able to obtain feedback from two sites with regard to our illustrations and narrative.

Reflections on the experience of conducting a rapid evaluation

We have summarised our experiences of conducting a National Institute for Health and Care Research-funded rapid evaluation, as defined by following three overarching principles (see also Chapter 2).

Summary of key points

• Vertical integration is enabling locally accessible primary care to continue to be provided for patients in areas in which general practices have faced particular difficulties.

• Vertical integration appears to provide an opportunity to further develop and integrate care for general practice patients.

• Vertical integration developed further where there were good pre-existing relationships between primary and secondary care, and where key individuals were active in providing leadership, energy and focus for the integration.

• The vertical integration sites have acted as pioneers, and we have heard that they are being approached by others in the NHS who wish to learn about such models of care organisation.

• The net impact of vertical integration on health system costs is argued by local stakeholders to be beneficial.

• We propose a follow-up evaluation, quantitative as well as qualitative, returning to one or more of our original case study sites and potentially including other vertical integration sites in the UK to address the following questions.

  • Is vertical integration improving recruitment and retention of primary care staff, including, but not limited to, GPs?

  • To what extent has there been service redesign as a result of the vertical integration arrangement, as distinct from horizontal integration via PCNs?

  • What are the views of patients in relation to their experience of accessing primary care services?

  • Are these views different among patients with multiple morbidities?

  • What impact is vertical integration having on secondary care service utilisation (e.g. A&E admissions, readmissions, length of stay in bed-days) and patient access (e.g. GP and practice nurse appointments) to primary care?

  • Is this impact different for patients with multiple morbidities?

  • Are there economies of scale in provision of back office functions?

Aims

This rapid evaluation had two distinct aims:

1. To understand the early impact of vertical integration, namely its objectives, how it is being implemented, if and how vertical integration can underpin and drive the redesigning of care pathways, if and how services offered in primary care settings change as a result, and the impact on the general practice and hospital workforces.

2. To develop a theory of change for vertical integration, identifying what outcomes this model of vertical integration is expected to achieve in the short, medium and long terms, and under what circumstances.

In addressing these aims we have attempted to answer six research questions suggested to us by our initial scoping work. We return to these questions, and how far we have found answers to them, in the following paragraphs and, in addition, we identify where further research would be desirable.

Overall assessment and implications for decision-makers

In this rapid evaluation we have been able to develop a theory of change for each of the case study sites. In addition, there has proven to be sufficient commonality between sites for it to be possible to derive an initial overall theory of change for vertical integration (see Figure 7). We intend to test and develop these theories of change in a follow-on phase 2 study of vertical integration.

Although vertical integration is being implemented in a number of locations in England and Wales, some urban, some rural, it is not evident that it is a generally desired way forward for health-care provision. Even where it has been introduced, many general practices choose to remain outside the vertical integration arrangement. The spread of (horizontally integrated) PCNs and primary care clusters may, if successful, enable sustainable primary care without vertical integration as well.

The implication is that, based on the rapid evaluation reported here, vertical integration is a valuable option to consider when general practices look likely to fail. However, it is not an option that should be imposed from the top down. Many GPs evidently do not wish to join such arrangements, as they have not taken that option even when available to them locally. By contrast, there are clearly some GPs who prefer to give up the autonomy and responsibility of owning and running a business (as necessitated by the partnership model) to retire or take up other opportunities. Some GPs and other practice staff are clearly content to be salaried employees. Primary and secondary care workforces are divided by decades of separate working and bringing them together is easier to achieve when strong and positive relationships already exist or can be built up between primary and secondary care leaders. Even then, we have not seen much evidence of greater clinical integration in providing care to patients.

A variety of specific forms of legal and governance arrangements for vertical integration is evident. All three varieties we have studied are viable and all entailed a good deal of detailed work to design and agree the specifics of arrangements, including staff terms and conditions, to enable previously independent practices to be integrated with NHS trusts and LHBs.

It remains to be seen whether the major changes that have occurred in primary care practice as a result of the COVID-19 pandemic, with patients communicating with primary care professionals by telephone and online, rather than face to face, affect the case for or against vertical integration. If GP workloads and lifestyles were to become more manageable as a result, the pressure from failing practices would abate; however, the need and opportunities for better integration of primary care with secondary care would likely remain.

Whatever the future, vertical integration can claim some successes so far in having maintained patient access to primary care in some locations at a higher level than would have otherwise happened. Improvements to patient care, including better integration of that care, would not be possible in the absence of such foundations. Therefore, although vertical integration has not been highlighted in current NHS policy in England and Wales, it appears to have an important role to play in some locations in enabling the continuation and development of locally accessible patient care.

Recommendations for future research

Our rapid qualitative evaluation based on three case studies has found much of interest, but important questions remain. Vertical integration shows promise, particularly in areas where GP retention and recruitment is particularly problematic. It may well remain a feature of the NHS in England and Wales, and may spread, given expectations that pressures on the GP workforce are set to increase over the coming years. It is unclear how vertical integration might develop in a primary care world where PCNs (England) and primary care clusters (Wales) are general, but our a priori expectation is that vertical integration will continue. After all, vertical integration has developed in an NHS where alliances, federations, networks, clusters, etc., of general practices have been very common, if not quite universal. Therefore, vertical integration has come about simultaneously with horizontal integration, rather than in advance of it.

There remain questions about vertical integration and its impacts and we intend to address some of them in a follow-up phase 2 quantitative and qualitative evaluation. This evaluation would investigate how far the current local instances of vertical integration are proving to be sustained and would test and develop the case-specific and overall theories of change created in the initial evaluation.

In a future mixed-methods follow-up evaluation, we intend to test and develop further the theories of change set out in this report. In particular, we intend to examine outcomes and impacts, including patient and carer experience, associated with hospitals managing GP services, and to explore the impact of vertical integration on local health-care systems in total. Within that, we will investigate how well vertical integration is adapted to the health-care needs of people with multiple long-term morbidities, for whom local access and well-integrated care are particularly important and in respect of whom the impact of vertical integration on the health-care system might be expected to be most prominent.

We propose to return, if possible, to one or more of the original case study sites and potentially to include other vertical integration sites in the UK to address some, or all, of the following evaluation questions:

• Is vertical integration improving recruitment and retention of primary care staff, including, but not limited to, GPs?

• To what extent has there been service redesign as a result of the vertical integration arrangement, as distinct from being a result of horizontal integration via PCNs?

• What are the views of patients in relation to their experience of accessing primary care services?

• Are these views different among patients with multiple morbidities?

• What impact is vertical integration having on secondary care service utilisation (e.g. A&E attendances, emergency admissions and readmissions, length of stay in bed-days) and patient access (e.g. GP and practice nurse appointments) to primary care?

• Is this impact different for patients with multiple morbidities?

• Are there economies of scale in provision of back office functions?

Where possible, quantitative analysis based on administrative (primary care and secondary care) data will be included, alongside further qualitative research.

Any future evaluation of vertical integration would necessarily take place in the context of the great changes to primary care practice that have come about as a result of the COVID-19 pandemic. Therefore, a second evaluation could provide an early indication of whether or not those changes have affected the desirability or otherwise of vertical integration in different types of local health economy.

A longer-term assessment of the impact of vertical integration in a world where telephone and digital consultations in primary care have become more common than face-to-face consultations could represent an important body of work for the wider research community in future. Other knowledge gaps that could also be targeted in a future research agenda for the wider research community include the following:

• the role of changing organisational culture, both as an enabler and a consequence of vertical integration

• the impact of vertical integration on innovation in health services

• the impacts of other types of vertical integration than that between acute hospitals and general practices.

Many interesting questions remain about the scope for integration across primary and community care, and between community and hospital care, whether for physical or mental health services, respectively.

The potential future research agenda around vertical integration is large, but we hope that the initial rapid evaluation presented in this report represents an informative first step.

Patient and public involvement

The proposal for the evaluation was reviewed by a PPI member of the BRACE Health and Care Panel. The Plain English summary of the report has been reviewed by another PPI adviser. Project findings will be shared and discussed at a meeting of the full BRACE Health and Care Panel, which includes eight patient and public members, and we will seek the advice of those members concerning how best to communicate the evaluation’s findings to patient and public audiences.

The phase 2 vertical integration evaluation, which was planned to follow up on the initial evaluation that is the subject of this report, will include assessment of the impacts of vertical integration on patients’ experience of, and access to, secondary and primary health care. Therefore, we intend to involve PPI members from the BRACE Health and Care Panel in the design of the phase 2 project at the protocol stage and, later on, in the design of project materials as the evaluation progresses.

Contributions of authors

Manbinder Sidhu (https://orcid.org/0000-0001-5663-107X) (Research Fellow, University of Birmingham) was the project manager and a researcher for the study. He contributed to the conception and design of the study, the theoretical framework, data collection at two study sites, overall data analysis and interpretation, and he is the lead author of the final report. He contributed to integrating the findings of the study and to all chapters, and he made critical revisions to the report for important intellectual content and approved the final manuscript.

Jack Pollard (https://orcid.org/0000-0002-2403-2635) (Researcher in Health Economics, University of Oxford) was a research analyst (employed by RAND Europe) for the study. He contributed to the conception and design of the study, the theoretical framework, data collection at one study site, overall data analysis and interpretation, and he is a co-author of the final report. He contributed to integrating the findings of the study and to Chapters 2 and 6.

Jon Sussex (https://orcid.org/0000-0001-8970-4546) (Chief Economist at RAND Europe and Co-Director of the Cambridge Centre for Health Services Research, a collaboration between RAND Europe and University of Cambridge) was the principal investigator for the study and led the study. He contributed to the conception and design of the study, the theoretical framework, data collection at one study site, overall data analysis and interpretation, and he is corresponding author of the final report. He contributed to integrating the findings of the study and to all chapters, and he made critical revisions to the report for important intellectual content and approved the final manuscript.

Publications

Sidhu MS, Pollard J, Sussex J. Acute Hospitals Managing General Practice Services in the UK: Why, How and What are the Consequences for Primary Care Delivery? Oral presentation at the Health Services Research Virtual Conference, 1 July 2020.

Sidhu M, Pollard J, Sussex J. Vertical integration of primary care practices with acute hospitals in England and Wales: why, how and so what? Findings from a qualitative, rapid evaluation. BMJ Open 2022;12:e053222.

Data-sharing statement

Owing to the consent process for data collection at case study sites within this evaluation, there are no data that can be shared. Please contact the corresponding author with any queries.

Disclaimers

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