Implementation of interventions to reduce preventable hospital admissions for cardiovascular or respiratory conditions: an evidence map and realist synthesis

Inclusion criteria

Studies were included in the mapping review if they met the following criteria:

• Published in or since 2010 in the English language.

• Conducted in a relevant country (UK, USA, Canada, Australia or New Zealand). Canada, Australia and New Zealand were included as the countries with health systems most similar to that in the UK, and the USA was included because of the high volume of good-quality health research conducted there. We took a pragmatic decision to exclude as far as possible studies from other European countries because of differences in health service organisation. We recognise that this could involve excluding potentially relevant studies from some countries, but consider that the impact on the structure of the evidence map was likely to be relatively minor.

• Recruited adults with a cardiovascular or respiratory condition (not cancer).

• Evaluated or described an intervention that could reduce preventable hospital admissions or re-admissions. Based on the work of Purdy et al. ,3 the following interventions of interest were specified in advance: case management, specialist clinics, community interventions (not fitting into any other relevant category), patient education, self-management, pulmonary rehabilitation, cardiac rehabilitation and telehealth. Programmes involving combinations of these interventions were also eligible for inclusion.

• Reported admissions/re-admissions or prevented admissions as an outcome and/or reported on implementation of the intervention (e.g. barriers and facilitators, qualitative studies of staff or patient views/experiences) in the context of reducing admissions.

The main study designs of interest were experimental studies (e.g. randomised and non-randomised trials), controlled and uncontrolled observational studies, qualitative studies and systematic reviews. We attempted to exclude editorials, letters, study protocols, papers discussing study rationale and design and other papers not reporting substantive data. Given that inclusion was based on titles and abstracts, published conference abstracts were eligible for inclusion.

Literature search

Formal bibliographic searches of MEDLINE, Cumulative Index to Nursing and Allied Health Literature, Health Management Information Consortium, EMBASE, Web of Science and The Cochrane Library were conducted in September 2017 and October 2017. The search was developed from initial scoping searches and previous systematic reviews, with search terms adapted for each information source. The search comprised terms for ACSCs combined with intervention terms and terms around admissions, implementation and research dissemination. The MEDLINE search is documented in Appendix 1. The ACSCs included the following: angina, hypertension, COPD and asthma. Intervention terms were derived from The King’s Fund report Avoiding Hospital Admissions: What Does the Research Evidence Say?. 1 The search was limited to studies published from 2010, when the Purdy report1 was published, to July 2018 and research published in the English language. Focusing on this narrower time frame is further justified by the specific focus of this review on implementation; an implementation context is a continually mutable backdrop within which to evaluate the introduction of a complex intervention.

Recent initiatives and, specifically, those that have been evaluated within a UK context were prioritised. Nevertheless, the review methodology preserved the potential to engage with the wider literature through coverage of reviews that extend the time and geographical limits beyond the formal sampling frame. The UK focus was strengthened by examination of the catalogues of the Health Services Management Centre at the University of Birmingham, The King’s Fund Library and Health Management Online (NHS Scotland).

Screening of search results and coding of included records

Bibliographic records identified by the literature research were imported into EPPI-Reviewer 4 (Evidence for Policy and Practice Information and Co-ordinating Centre, University of London, London, UK) for screening and data extraction. Records were screened for inclusion or exclusion by one reviewer, with a 10% sample being screened by two reviewers to check for consistency. Screening of search results was based on information in database records (title or title and abstract) only; we did not systematically screen full texts.

Data extraction (coding) was carried out by one reviewer in EPPI-Reviewer 4 using a mixture of tick-box selection and manual data entry. Table 2 summarises the extracted data items. We did not extract study findings or authors’ conclusions because the purpose of the review was to map interventions and not to evaluate their effectiveness.

Mapping review synthesis

We produced a descriptive summary of key characteristics of the body of included studies as reported in Chapter 3. Summary tables were developed using the search, cross-tabulation and reporting functions of EPPI-Reviewer 4.

Data extraction

In contrast to the mapping review, data extraction for the realist synthesis was based on the full text of each item. Purpose-designed data extraction forms drew on appropriate frameworks as structures by which to interrogate the theories and the empirical evidence. We used an experimental methodological development, previously tried in another NIHR HSDR realist synthesis,10,11 which combined use of realist synthesis methods with elements of best-fit framework synthesis. Best-fit framework synthesis involves identification of an appropriate ‘good-enough’ framework to operate as both a vehicle for data extraction and, subsequently, an analytical lens for examination of extracted data. 12 Best-fit framework synthesis is believed to expedite the data extraction process,13 with a majority of the data being handled deductively using the framework before a subsequent inductive phase to code data not explained by the categories derived from the original framework. 14

The focus of this realist synthesis on implementation encouraged the review team to focus on frameworks or models specifically derived in an implementation, knowledge translation or evidence-based health-care context. Instead of embarking on an extensive parallel process of framework identification, which was typically the case in previous uses of best-fit framework synthesis,15 the team used the sourcebook, Models and Frameworks for Implementing Evidence-Based Practice: Linking Evidence to Action,16 and rapidly reviewed the pictorial models and accompanying textual descriptions for ‘fit’ to the purpose and context of the review. A short list of candidate models was subsequently narrowed down to the PARiHS framework. Within the PARiHS framework, successful implementation is represented as a function of the nature and type of evidence (to be examined from the mapping review), the qualities of the context in which the evidence is being introduced and the way the process is facilitated17 (to be extracted from included UK studies, both quantitative and qualitative).

Previous experience of realist synthesis within complex service delivery contexts had also revealed the value of using the TIDieR as a formal framework for identifying and describing the components of included interventions. 18 We therefore decided to use a version of this template, abbreviated in recognition of both the time constraints and the generic level at which interventions have been characterised, as a structure for describing the nine intervention types. The 12-item TIDieR checklist [brief name, why, what (materials), what (procedure), who provided, how, where, when and how much, tailoring, modifications, how well (planned), how well (actual)] was therefore abbreviated in the form of the five-item ‘TIDieR-Lite’ (by whom, what, where, to what intensity, how often). Importantly, this was to be used to summarise the generic characteristics of each intervention type across multiple intervention reports, although significant areas of variation across each generic type were prompted for identification by the framework.

Finally, data relating to programme theory were extracted using an ‘if–then–leading to’ realist logic structure, pioneered by other research teams and used in previous reviews by team members. This enabled the generation of multiple programme theory statements [ultimately five in number: programme theories (PTs) 1–5; see Programme theory development] to be used to examine data from included empirical studies and to communicate programme logic to stakeholders in the form of narrative scenarios.

In summary, data extraction comprised use of:

• an implementation framework, PARiHS, as a structure for examining how interventions are delivered

• an intervention template, TIDieR, as a format for describing intervention components

• a realist logic template, if–then–leading to, to elicit programme theory on how interventions might work.

Synthesis

Following identification of the initial programme theories, the review team extracted data into evidence tables. The resultant hypotheses operationalised synthesised statements around which we developed an explanatory narrative referenced to the underpinning evidence base. Additional searches for mid-range theories (conceptual models or frameworks relevant to one or several of the interventions covered by the review) and overarching theories (theories relevant to the phenomenon of inappropriate admissions as a whole) were conducted using Google Scholar (Google Inc., Mountain View, CA, USA). The aim of searching for these higher-level theories was to elucidate how previous researchers have understood factors underlying inappropriate admissions and their prevention at a relatively high level of abstraction. Our EPPI-Reviewer map, reference management database and accompanying data extraction spreadsheets collectively offer a comprehensive evidence base relevant to interventions to reduce unplanned hospital admissions.

Testing and refining the programme theories

Searches for programme theories relevant to avoidable admissions were conducted using Publish or Perish 6™ software. 20 This desk-based tool offers an auditable interface to searching of the Google Scholar resource as well as allowing construction of semicomplex search strategies.

The first set of Google Scholar searches combined the term ‘ambulatory sensitive’ with terms for preventable admissions and different programme theory terms. The second set of searches combined the term ‘preventable admissions’ with ‘hospitalisation or hospitalization’ and different programme theory terms. Full details of the search terms used for each individual search and the number of results retrieved are provided in Appendix 2.

Papers retrieved from the programme theory searches were reviewed by one reviewer with experience of harvesting programme theories. Prioritising conceptually and contextually rich papers, the reviewer drafted preliminary if–then–leading to statements21 [known technically as context–mechanism–outcome (CMO) configurations] for discussion with the review team. The aim of this process was not to generate an exhaustive list of possible explanations but to generate a selection of theories of change operating variously at patient/carer, health provider and health system levels. When several programme theory components seemed to be interrelated, these were combined into a more overarching explanation.

From the papers retrieved from the programme theories searches and from the papers included in the mapping review, we developed five programme theories to guide the realist review. Details of the generic programme theories are given in Box 1. Details of the programme theories expressed as if–then–leading to statements and the probable types of evidence identified by the team by which programme theories might be supported or negated are provided in Table 3. In Chapter 4, the programme theories are considered in the specific context of each of the included interventions.

The initial programme theories were tested from the theoretical literature, empirical studies and insights from the PPI group. Programme theories were examined against the nine individual intervention types and collectively as a set. A subsequent activity involved seeking to map the elements of the programme theory to potential mid-range theories that might add greater transferability to review findings. 22 Mid-range theories could be identified serendipitously, when reviewing the empirical evidence, but more typically were identified from Google Scholar searches that combined the intervention (e.g. ‘case management’) with terms relating to models or theories (i.e. ‘theor*’ or ‘model*’ or ‘framework*’ or ‘concept*’). 23

Summary

Exploration of case management reveals that the role of the specialist nurse is key to its success. The impact of the role is determined by issues relating to responsiveness (as seen in response times) and availability (as seen in the demand for 24-hour access), which themselves can be moderated by the size of the nursing caseload and the competing demands of the administrative workload. A tension may be identified between the intrinsic advantages of continuity of care, as especially evidenced in a knowledge of a patient’s symptoms and the need to provide round-the-clock coverage to patients with more severe manifestations of cardiorespiratory conditions. Case management is therefore seen to require substantive re-engineering of the health system in which it is intended to operate.

Definition

Case management is a generic term, with no single definition, described as the process of planning, co-ordinating and reviewing the care of an individual. The Case Management Society of America provides a definition on its website (www.cmsa.org; accessed 4 April 2019). This definition was operationalised by Purdy et al. 3 in their series of NIHR reviews. The literature reflects confusion between case management as an ongoing process and as an intensive time-limited intervention. 72

Case management within the NHS has been largely configured as community-based programmes, set up and funded by primary care trusts and typically (but not always) staffed by community matrons. 72 In recent years, UK initiatives have focused on multidisciplinary team (MDT)-led case management but have demonstrated little or no reduction in use of secondary care. 73 Increasingly, attention has focused on the ‘added value’ of benefits for patients and professionals. 74 Interest in case management has been revived by the new models of care initiatives, with their focus being on integrated care.

Intervention components

Rather than comprising a single intervention, case management typically describes a package of care that covers activities that vary widely between programmes; it is described as a ‘prototypical example of a complex intervention’. 75 Such variation makes case management both difficult to describe and challenging to evaluate. TIDieR-Lite components of self-management interventions are summarised in Table 6.

Several commentators72,76,77 identify the following core components as particularly important to case management programmes:

• case-finding

• assessment

• care planning

• care co-ordination (usually undertaken by a case manager within the context of a MDT), including but not limited to –

  • medication management

  • self-care support

  • advocacy and negotiation

  • psychosocial support

  • monitoring and review

  • case closure (in time-limited interventions).

Case management can include components such as self-management, patient education and disease management programmes,72 making it more challenging to distinguish this intervention from others reviewed in this report. Case management may be delivered in diverse ways that vary according to intensity (frequency and duration of the contacts), degree of embeddedness in the local care network, the background and training of case managers and the extent to which they work alone or within a team. Further variation is exhibited in whether or not the case manager is supported through reflexive group meetings with peers or supervisors, how the target population is identified and how the case management intervention is initiated.

Number and type of UK studies identified

The effectiveness review by Purdy1 drew on only one UK study of case management. 78 This randomised controlled trial (RCT) in a COPD population in West London found no difference between case management and usual care in terms of numbers of hospital admissions. Indeed, the primary impact of the intervention seemed to be a reduced need for unscheduled primary care consultations. For every one COPD patient receiving the intervention and self-management advice, there were 1.79 fewer unscheduled GP contacts. 78

The mapping review identified two further quantitative UK studies (three papers) of case management for cardiovascular or respiratory conditions published since 2010. One study examined a COPD population in a single general practice79 and the other study73,80 presented data on 20 ACSCs. The six conditions of relevance to this review were asthma, atrial fibrillation, CHD, COPD, heart failure and hypertension.

Three UK qualitative studies were identified (four papers/reports). One of these studies examined patients with heart failure and staff involved in their care81 and the other two initiatives targeted case management in ‘high-risk’ populations including conditions eligible for this review.

Details of the studies included in the analysis are presented in Appendix 4, Table 25.

Operating programme theories

Programme theory 2 proposes that IF patients feel confident and satisfied with non-secondary care health provision THEN they will not consider it necessary to access/request secondary care services LEADING TO appropriate utilisation of health resources and a reduction in unplanned admissions. Case management affects a patient’s perceived capability of staying in their own home. 82 Crisis situations can be anticipated, if not averted; this is particularly important in the context of exacerbations as, for example, with COPD.

However, because of its personalised, tailored nature and the involvement of multiple health and social care professionals, case management may be considered overly intrusive. 82 Others patients report that the case manager was perceived as an impediment to accessing their GP. 83 Furthermore, if a patient feels supported, they are less likely to feel a need to access sources of support from secondary care. One patient with respiratory problems74 reported past and probable future occasions when her needs might be best met in hospital. This patient perceived the hospital as a means for meeting not just her medical needs but also her holistic needs, making her feel safer and therefore less anxious. Other participants described their ‘confidence’ in local hospitals should circumstances arise in which they feel that some of their needs may be better met in hospital.

Case management also engages with PT3: IF primary care staff (in this case the case manager) feel confident in their own ability to refer patients appropriately and have confidence in and knowledge of services available within primary and community care THEN they will not refer patients to hospital LEADING TO an increase in use of self-management and non-secondary care health service provision and a reduction in unplanned admissions. In this context, the detailed knowledge of a patient’s condition and circumstances and the holistic perspective of their care enables the case manager to calibrate and negotiate appropriate thresholds for secondary care intervention. Previously, the Evercare evaluation found that, as advanced primary nurses’ knowledge of available services increased over time, they referred their patients to an increasing range of resources for support. 84 Gowing et al. 74 report that some patients may be content to trust the case manager’s judgement, but others may resolutely insist that hospital is the best place for them. A major issue was the lack of adequate social care support, although isolated instances of providing overnight care following hospital discharge were reported. 74

An important contextual variable is the need for adequate training to strengthen the case managers’ self-efficacy and confidence in the appropriateness of their situational assessments. 82 This confirms that, as case managers become more experienced, GPs are likely to spend less time liaising with other services and case managers are able to provide more patient care themselves. 85 One potential unintended consequence of case management is increased levels of case finding resulting in a non-reduction in hospital admissions. 85

Detailed knowledge of a patient’s condition and circumstances (embracing both clinical and social contextual factors) also mitigates operation of action according to PT4, that is the patient does not perceive a need to delay presentation to secondary care services; they feel empowered to elicit information from their case manager as and when required. The triad of case manager, patient and informal carer only perceives a need to escalate action when the personalised threshold has been exceeded. A key contextual factor here relates to case manager caseload: if a case manager holds responsibility for an excessive number of cases then they will be unable to determine appropriate personalised thresholds and will either admit a patient unnecessarily or cause/contribute to delays in seeking treatment. Delays in accessing services have been shown to lead to deterioration in patients’ health and are a probable cause of future hospital admission. Lack of available community-based services constitutes a major challenge to effective case management. 84,86

Description of putative mechanisms

Possible mechanisms for case management are summarised in Box 2. In theory, case management seeks to increase efficiency by reducing unnecessary contacts with the health system. 87 Such contacts include fragmented routine contacts, as well as emergency contacts caused by potentially preventable exacerbations. The goal of case management is ‘to better co-ordinate care, offering individually-tailored contacts and care planning’. 87

The case management model is predicated on the presence of so-called ‘super utilisers’: ‘high-risk’, high-need patients, typically with multiple health conditions, who utilise a disproportionate amount of health-care resource (with a high cost). 88 The idea behind case management is that by targeting additional and individually tailored primary care at these patients, more costly secondary care admissions (particularly emergency admissions) can be avoided. 89

However, research suggests that case management does not meet its primary aims for the patients involved, although it is associated with increased patient satisfaction. 87,90 In practice, MDT case management tends to target those identified as ‘high risk’ using a selected statistical algorithm that is validated to predict patients who are likely to have substantial future health-care use. 80 These tools generate a heterogeneous group of patients, and it may be that there are subgroups for which the direct effects of the intervention are more effective. However, as papers by Stokes et al. 73,80 reveal, it can be extremely challenging to identify so-called ‘super utilisers’ both in terms of the ‘safe’ margin of those who can be appropriately managed in primary care and, equally, in terms of those for whom the complexity of their comorbidities renders secondary care an appropriate option.

Systematic review evidence identifies continuity of care as an important influence on admissions for long-term conditions. 91 Commentators seek to distinguish continuity of relationship (a continuous caring relationship with clinicians) from continuity of management (all aspects of integration, co-ordination and sharing of information). Both mechanisms can be considered important in the context of preventable admissions. A patient must feel that they can trust the judgement of the HCP and that the HCP has a sufficient understanding of their unique personal circumstances. Practically, continuity of management is important in the context of 24-hour care and delivery of services across organisational and professional boundaries.

Case management is centred on the premise that targeted, proactive, community-based care is more cost-effective than downstream acute care. Delivering such care requires that an intervention is integrated across care providers to avoid overlap and to ensure that each care provider knows and realises what each other care provider does. Wagner et al. ’s92 chronic care model (CCM) has been proposed as a framework to restructure the health system towards integrated, proactive, consistent and continuous care, and, thus, anticipate acute exacerbations or lessen their consequences. 93 The CCM draws on six interacting elements: links with the community, the health system, self-management support, tailored delivery system design, help for decision support and adequate clinical information systems. 94 Case management for people with complex care needs offers a potential strategy for delivering this type of integrated care.

Time-limited case management targets those with the greatest risk of emergency admission. A stepped approach means that people at lower risk of admission can be targeted with disease management programmes or supported in self-management. Case management shares the patient orientation of self-management75 (see Self-management). Indeed, patients in the study by Gowing et al. 74 reported being able to take a more active role in their own planning and care, thereby promoting independence. However, independence appears context dependent: a respiratory patient in the same study described being given a rescue pack for COPD and struggling with her own judgement about when to use it. 74

Previous evaluations report that policy-makers assumed that case management would stimulate ‘service redesign beyond the introduction of case management itself’. 85 However, evidence of wider local re-engineering of primary and secondary care for older people has proved challenging to establish.

Conceptually, case management can be understood in the context of integrated care. The six dimensions of services integration suggested by the National Collaboration for Integrated Care and Support (2013)95 are:

1. consideration of patient and family needs

2. communication with the patient and between HCPs

3. access to information

4. involvement in decision-making

5. care planning

6. transitions between various HCPs.

Roland et al. 96 advance possible explanations as to why using case management to improve care integration is not guaranteed to improve outcomes. The first is a potentially faulty underlying programme theory (i.e. because supply-induced demand increases appropriate admissions and does not simply decrease inappropriate admissions). Alternatively, the implementation of case management interventions may be wanting – an explanation on which Goodwin97 and Ling et al. 98 draw to explain suboptimal achievement of effects.

Efforts to strengthen coping capability are closely linked to the self-efficacy theories propounded by Bandura99 (see Self-management).

Contextual factors

In common with other complex interventions, case management studies generally lack contextual detail. As a complex intervention, case management includes various components interacting in a non-linear way to produce outcomes that are highly dependent on context and variables across settings. Attention should focus on analysing not only if and how case management works for frequent users of health-care services, but also in what contexts it works.

Supporting evidence

This is a brief summary of evidence from systematic reviews, concentrating on hospital (re)-admission as the outcome of interest. Key results from UK studies included in this analysis are also presented.

Case management is an area that is rich in systematic reviews and evidence syntheses. Small numbers of systematic reviews briefly addressed enabling factors of successful case management interventions in the discussion sections of their papers. In a review of the effectiveness of case management among frequent emergency department users, Kumar and Klein108 noted that ‘frequency of follow-up, availability of psychosocial services, assistance with financial issues and active engagement of the case manager and the patient were important characteristics of CM [case management] interventions’.

Huntley et al. 109 conducted a systematic review to identify observational studies conducted at a practice level that describe factors and interventions that have an impact on levels of utilisation of unscheduled secondary care. Their review was limited by the challenges of trying to review across different health systems in different contexts. They found a benefit from seeing the same HCP, thus informing debates about continuity of care. Proximity to health-care provision was another major factor. However, they found it difficult to determine factors affecting quality of care.

Huntley et al. 109 subsequently conducted a systematic review and meta-analysis of case management interventions for heart failure. They included case management within a hospital context and also studies targeting nursing homes and long-term care settings. None of their included UK studies therefore met our tighter inclusion criteria. They identified four studies of community-initiated case management versus usual care (two RCTs and two non-RCTs), with only the two non--RCTs showing a reduction in admissions. 109

Stokes et al. 87 conducted a meta-analytic review of 36 studies. Meta-analyses showed no significant differences in total cost, mortality or utilisation of primary or secondary care. They found small yet significant effects for case management in terms of self-reported health status and patient satisfaction. Interestingly, their secondary analyses indicated that ‘the effectiveness of case management may be increased when delivered by a MDT, when a social worker was involved, and when delivered in a setting rated as low in initial “strength” of primary care’. 87

In an effectiveness review that included 10 studies, none from the UK, Joo and Liu110 reported that three studies documented statistically significant reductions in hospital re-admissions. Two of these statistically significant studies fell within our 2010–18 time period. Melton et al. 111 report that participants with multiple chronic illnesses in a nurse-led case management intervention group demonstrated lower 30-day and 60-day hospital re-admission rates than participants in the control group (p < 0.05 vs. p = 0.01, respectively). Chow and Wong112 similarly report that a nurse-led case management focused on older adults with chronic illnesses in China demonstrated a significant reduction in hospital re-admission rates in the intervention group compared with the control group (p = 0.018). 112 The characteristics of these interventions indicate potential overlap in this review with the intervention labelled ‘specialist clinics’, with such clinics typically being nurse led.

Joo and Liu113 conducted a qualitative evidence synthesis, using thematic synthesis, to review 10 qualitative studies published between 2007 and 2016 for attitudes of individuals with chronic illnesses and their caregivers towards case management. Access to health-care resources, health status supports and emotional aid were identified as facilitators of case management. Low information about case management and time constraints were identified as barriers. A complementary synthesis by Joo and Huber114 looked at barriers perceived by case managers when implementing case management. This thematic synthesis of 10 qualitative studies (2007–16) identified five barriers to implementation: unclear scope of practice, diverse and complex case management activities, insufficient training, poor collaboration with other HCPs and client relationship challenges.

In summary, there is little recent evidence to suggest that case management is effective across multiple diverse contexts. The UK research is limited in both quantity and rigour and suggests that such interventions are very context specific, with the surrounding environment playing a major role in the effect of the case management intervention. The intervention is well liked by patients and, generally, by staff, although concerns persist about caseload and the ability to deliver care in a timely and accessible manner. Economic evaluations are relatively rare and it would seem that the performance of case management depends very much on how wide the evaluation framework is cast in terms of both costs and benefits.

Potential enhancements of intervention/implementation

Commentators on case management consistently point out that, although the case manager role is pivotal, ultimately, success is determined by the support available to back up the initial management. This depends not only on available primary care resources but also on further social care involvement. 74 A further consideration is the availability of a case manager. Optimally, 24-hour access to case management74 can combat the uncertainties that arise when the patient is unable to contact necessary advice and assistance. Below this very intensive level of provision, uncertainties about whether or not help is available and what to do when it is unavailable serve to subvert the very mechanisms by which case management achieves its success. Complementing availability concerns are anxieties about responsiveness. Issues relating to the overall caseload of the case manager, the level of dependency of patients within that caseload and the amount of time shared with other duties, for example administrative responsibilities, are critical to a timely response. So, specifically in connection with community matrons in a case manager role, Brown et al. 83 reported concern at the ability of community matrons to be able to respond within the patient’s perceived time scale.

Although opportunities to co-create care plans with members of the primary care team are generally welcomed by patients and family members,74 this requires sufficient time to prepare for care planning and to discuss those plans. It is therefore essential to make this planning process feel meaningful to the patient. Some patients felt that by articulating their views and voicing concerns on behalf of other patients they could advocate for those who felt unable to take an active role.

Case management is a complex activity that operates over multiple care settings. Some commentators suggest that greater integration of information technology (IT) systems, including shared electronic medical records and access to resource directories and clinical guidelines, could lead to improved patient outcomes (e.g. Lynch et al. 115). However, this may be an artefact of earlier evaluations and may now be offset, to a substantive degree, by technological developments in electronic health records, care plans and remote servers accessed through user authentication and secured, encrypted transmissions.

Previous evaluations emphasise the need for care pathways to be interpreted flexibly because patients do not fit into standard care pathway approaches. At a practical level, this meant that strict eligibility criteria and inefficient administrative procedures often led to delays in service delivery. 84 Another reason for system failure was that nurses did not always have adequate access to alternatives to admission.

In a novel attempt to bring external theory to bear on the case management phenomenon, Swanson and Weissert116 explore the principal agent framework and street-level bureaucratic theory. 116 They conclude that incentives, as well as ‘green tape’, clear rules, guidelines and algorithms relating to resource allocation among patients, would have an impact on the greater effectiveness of case management.

Evaluations have found little evidence of the systematic redesign of care aspired to by many case management programmes. The Evercare evaluation observed that ‘poor integration between primary and secondary care, and out of hours services were not focused on keeping patients out of hospital’. 84 It concluded that more radical system redesign is needed to achieve a greater impact on admissions. Such radical system redesign has not generally been achieved, notwithstanding the potential offered by new models of care initiatives. 11,117

Summary

Patient education is seen as a key component to several interventions designed to reduce preventable admissions and, given demands on GP time, the role of the nurse is seen as critical to the delivery of such information. Three particular factors can be seen to mediate the success of patient education-based interventions (1) patient education is delivered in anticipation of exacerbations or other health incidents and so patients may not see the direct relevance of the information at the time provided or may forget it completely, (2) patient education needs to be situated within the relative experience of individual patients (e.g. ‘if your exacerbation is worse than the last one then . . .’) rather than in some impersonal absolute for all patients and (3) for many patients, self-efficacy, rather than education retention, is the issue. Many patients experience anxiety about their own ability to manage a situation and to make a correct decision. Unless this anxiety is alleviated, they will always seek a second opinion for reassurance.

Definition

A minimal definition of patient education is the teaching or training of patients concerning their own health needs. In their report, Purdy et al. 3 distinguish traditional patient education, which offers information and technical skills, from self-management education, which teaches problem-solving skills. From this perspective, self-management education complements traditional patient education in supporting patients to live the best-possible quality of life with their chronic condition. 118 Nevertheless, such a distinction is not implemented consistently throughout the literature, making the co-existence of patient education and self-management as separate intervention categories in this report particularly problematic. Self-management of multiple chronic illnesses, educating patients to monitor their own health and being able to recognise illness severity feature in current health-care policy. The boundary between providing patient information and delivering patient education may also be difficult to perceive.

Controversy in the literature surrounds whether or not adult education is an appropriate model for patient education, particularly given that patients’ health needs are contingent to particular time periods or stages in a disease or treatment pathway. The above definition conveys formality, contrasting with the ‘teachable moment’, the idea that a HCP may harness any opportunity during their therapeutic encounter to impart some education that meets a patient’s health needs. 119 Key to this process is recognition of what health needs are; a patient may not know what they need to know while a HCP may not be able to time their education intervention to a point when a patient is most attentive, receptive or responsive to a particular message.

Kongstvedt120 defines patient education as a process that involves imparting information to patients that will alter their health behaviours or improve their health. This concept of behavioural change, very visible in the literature of the ‘teachable moment‘, is fundamental to contemporary views of patient education.

The National Institute for Health and Care Excellence (NICE) guidelines for heart failure121 recommend that people with chronic heart failure should be offered:

personalised information, education, support and opportunities for discussion throughout their care to help them understand their condition . . .

© NICE 2010. Management of Chronic Heart Failure in Adults in Primary and Secondary Care (Clinical Guidance 108). 121 Available from www.nice.org.uk/guidance/cg108. All rights reserved. Subject to Notice of rights. NICE guidance is prepared for the National Health Service in England. All NICE guidance is subject to regular review and may be updated or withdrawn. NICE accepts no responsibility for the use of its content in this product/publication

The reference to ‘personalised information’ is particularly significant in this context; patients need to be able to situate the information that they have received within their own context of symptoms and disease progression.

Key challenges for patient education as an intervention relate to the fact that patient education is often an adjunct to other interventions that might be used to manage preventable admissions (e.g. cardiac rehabilitation, self-management and telehealth) and so it may be challenging to separate patient education as a discrete intervention/intervention component. The Cochrane review on patient education in the management of CHD122 does attempt to separate the educational component from other aspects of cardiac rehabilitation. Furthermore, patient education may constitute an ongoing iterative process that takes place over frequent patient–clinician interactions and therefore may be difficult to quantify.

Intervention components

The key components of patient education interventions as reflected by UK studies identified for the mapping review are summarised in Table 7. As mentioned previously, a key challenge relates to how to operationalise the concept of patient education. Within a research context, the Cochrane review definition122 requires the presence of all four of the following elements: (1) instructional activities organised in a systematic way, (2) an inpatient, outpatient or community setting, (3) structured knowledge transfer about CHD for secondary prevention and (4) face to face (group or one to one) or interactive delivery.

Number and type of UK studies identified

The mapping review identified three UK quantitative studies of patient education for cardiovascular or respiratory conditions. One study examined patients admitted with acute asthma,123 a further study explored COPD patients with breathlessness124 and the remaining study examined multiple conditions including the cardiac and respiratory conditions,125 which are the focus of this report.

Five UK qualitative papers were identified. Four of these papers reported findings from the HoldFAST study,126–129 and some of those identified were not from the mapping review but from supplementary follow-up of citations and references. The HoldFAST study was a multicentre study co-ordinated across the University of Bristol, the University of Oxford and Keele University and funded by NIHR. It aimed to explore the experiences of patients with heart failure from ‘multiple standpoints using ethnographic methods, combining observations, interviews and documentary data sources’. 128 It therefore offers a rich source of data on the experience of heart failure patients and particularly their experience of interventions that were commonly encountered in their ongoing care (e.g. patient education and specialist nurses). The remaining study130 examined the experience of patients who have encountered an exacerbation of COPD, particularly in relation to fear and anxiety.

Details of the studies included in the analysis are presented in Appendix 4, Table 26.

Operating programme theories

Programme theory 1 is ‘People with chronic conditions are frequently admitted to hospital when hospital is not the optimal destination for them. They may have symptoms that could be self-managed or anxieties that could be addressed by patient education or information.’. However, patients not only lack information about heart failure and self-care but, most importantly within this context, they also lack knowledge on when to seek help. 128 Patients may often, but not always, comprehend the information that they have been provided, particularly if it is given in a standardised format. However, what they may typically lack is situational knowledge (i.e. how to relate the information they have previously been given to the signs and symptoms that they are experiencing). Patients may feel confident about self-management of their medication in general, and yet not feel confident about their actions in an ‘emergency’ situation.

Access to written or verbal advice from a HCP, delivered remotely via technology and without presenting for face-to-face consultation or admission, may offer one route by which to address patient symptoms and concerns. However, there is increasing recognition that such information is effective only if the patient is receptive and the information is received at the time of need. If such information is not available, or if the patient perceives that it is not readily accessible, they may resort to easier channels of access, such as a telephone call to their GP. There is substantial evidence to indicate that human sources of information are considered more accessible than other information sources in most contexts. Increased facility to look up information, for example on the internet, may change perceptions about accessibility, particularly out of hours, but this raises associated questions about whether or not such information is authoritative and credible and concerns that such information is not typically tailored to the context-specific needs of the individual patient. Glogowska et al. 127 documents how education messages were not always received and acted on by patients, which could lead to unplanned admissions.

Patients may find it particularly challenging to assess whether or not they have reached a legitimate trigger point when first experiencing symptoms (e.g. coughing in relation to COPD131), when symptoms are of a severity not previously encountered or when symptoms include a new and unfamiliar feature. Patients may therefore delay seeking help [‘I’ll just wait to see if the symptoms get better (or worse) before seeking further help’] or they may respond to the novelty or unfamiliarity of the symptoms (‘I’ve not had one of these before, it must be bad’). Being unable to normalise their symptoms against their own experience, patients may need to normalise what they are encountering against the experience of others, either as encountered through self-help groups or secondhand via their nurse or doctor.

People with chronic conditions could lack knowledge about alternative health provision and therefore draw disproportionately on well-signposted channels, such as their GP or the emergency department (PT2). Similar concerns relate to information required for directional purposes, as opposed to information required to manage a clinical problem. Directory-type information may not be easily accessed or navigated and may not be maintained and kept up to date. If the credibility of such a source is damaged, albeit in a minor way, a patient may find themselves pursuing channels that are more likely to yield a successful response.

When patients have a good relationship with their GP, they may perceive that they are more accessible at an appointment or a home visit. 132 Other patients may feel guilty if they have to call the GP. As a consequence, they may leave the decision to the agency of another: a carer or family member. 132 Having missed the initial window for timely non-hospital intervention, they may find themselves at a point at which they are too late to avoid hospital admission.

If patients do not have written or verbal information on when to telephone for an ambulance, they may be less likely to use emergency services appropriately. One possible approach is to include such information in a written self-management plan. 132 Both primary and secondary care physicians in the HoldFAST study acknowledged that they were unsure about what services were available or that there were few services on which they could call. 129

People with chronic conditions may also use health services inappropriately, delaying their presentation to a GP or hospital (PT4) because of perceptions of the service either anticipated or based on the past experience of either themselves or others. Some patients may also feel a reluctance to seek medical advice if they perceive that their previous actions may have contributed to their condition (as in the case of smoking and COPD131).

Description of putative mechanisms

Possible mechanisms for patient education are summarised in Box 3. As noted by Fry et al. ,126 Bury133 describes the effects of chronic illness as ‘biographical disruption’ to everyday life. Disruption extends beyond the patients themselves to have an impact on their families and wider social network. In response, patients and their families may look for information, support and the most effective strategies to manage their symptoms, hoping to minimise potential future disruption.

Symptom unpredictability reportedly leaves patients feeling helpless and completely dependent on those around them. This, in turn, can lead to a lack of control of the illness and an increased burden on families and the health-care system (Morris et al. 134). Implications for carers are that responsibility for care may shift from the patient to their spouse or immediate family. Commentators suggest that these factors may precipitate a patient to adopt the ‘sick role’, relying on their family for support with both their illness and previous responsibilities. 135

Contextual factors

The HoldFAST study found that, notwithstanding almost universal recognition of the importance of providing education to patients to help them manage their condition among secondary care physicians (specialists) and GPs, both secondary care physicians and GPs found that time pressures during consultations restricted their contribution to patient education.

Participants acknowledged that not all patients would take up the education offered. Patients who find the information difficult to assimilate would find self-management difficult. These patients were more likely to be those for whom English was not their first language, those too ill to benefit from education or in denial about their condition, people with learning disabilities and those experiencing cognitive decline or living with addictions. The specialist heart failure nurses felt that it was necessary to adapt the education they offered patients to the individual’s ability to receive it. They tried to identify issues of importance to patients as a way of personalising the information. The nurses stressed that even when patients were able to understand information, it would still be necessary to repeat those messages regularly. 126

Supporting evidence

Numerous studies have identified a lack of patient understanding of heart failure including a lack of patient knowledge of medications and self-care. A Cochrane review of patient education in the management of CHD122 concludes that, overall, the evidence is of only very low to moderate quality. It finds that patient education, as part of a cardiac rehabilitation programme, ‘does not contribute to fewer deaths, further heart attacks, heart by-pass or angioplasty, or admission to hospital for heart-related problems’. 122 It does acknowledge that there is evidence of fewer other heart-related events and improvements in health-related quality of life with education-based interventions. It was not possible to determine an effect on individual mortality.

The current evidence base seems largely interpreted within a ‘first do no harm’ frame. In the absence of sufficient information at present, ‘to fully understand the benefits or harms of patient education for people with heart disease’, the Cochrane review122 endorses current guidelines that ‘people with heart disease should receive comprehensive rehabilitation that includes education’. Further research is needed to evaluate how education for people with heart disease can be delivered in an effective and cost-effective way.

Potential enhancements of intervention/implementation

Provision of specialist heart failure nurses, alongside GP and specialist clinician input, was recognised as an essential mechanism for ongoing patient education. Time spent in outpatient clinics is brief and highly pressurised and HCPs have to select judiciously the two or three items of information that they want the patient to recall when they get home. 127 Specialist heart failure nurses, either in clinic or in the community, were seen as having the optimal opportunity to exploit opportunities for ongoing education.

Glogowska et al. 127 report that education represented a considerable portion of the specialist heart failure nurse role. The ongoing relationship between the specialist nurse and the patient, in particular during routine home visits, where patients are more relaxed, in control and have more time to assimilate information, offered an opportunity for delivery of patient education. This was considered particularly important given that the patient can take in only a small percentage of what the HCP shares on any one occasion. 127

Further exploration is required on the degree of tailoring and personalisation required in order for patient education to be effective. Some commentators have suggested the value of the ‘information prescription’, analogous to a medication prescription, which combines a formality of exchange with personalisation to patient perspectives. 35 In connection with tailoring, a review by the Joanna Briggs Institute34 suggests that future studies might consider interventions inclusive of more diverse ethnic populations of varying literacy and socioeconomic levels.

Summary

Self-management follows from and complements patient education. Written action plans, including information on when to seek medical help, are a key component of many self-management programmes. The key mechanism for self-management is thought to involve increasing a patient’s self-efficacy and confidence to be actively involved in managing their condition. There is strong evidence from systematic reviews of randomised trials supporting the effectiveness of supported self-management for reducing hospital admissions in respiratory conditions, such as asthma and COPD, and cardiovascular conditions, such as chronic heart failure. UK studies did not find a decrease in admissions with self-management relative to usual care, although one study suggested that self-management for COPD is likely to be cost-effective.

Self-management support is most frequently delivered by nurses, but GPs and other physicians are also involved. Some studies stress the importance of the whole team in embedding self-management support at the level of the general practice.

Definition

Purdy et al. 3 define self-management in terms of education in skills needed to manage the disease, behaviour change and emotional support for patients. Self-management education complements traditional patient education by teaching problem-solving skills alongside information and technical skills. Written action plans are a key component of many self-management programmes, especially for respiratory conditions. The plans support patients in managing both chronic conditions and acute exacerbations of their condition, including when to seek medical help.

An early review paper136 identifies three key self-management tasks (medical management, role management and emotional management) and six self-management skills (problem-solving, decision-making, resource use, formation of a patient–provider partnership, action planning and self-tailoring). Role management involves adapting behaviour and/or activities as required to manage the health condition and self-tailoring involves applying self-management knowledge and skills appropriately to a patient’s individual context.

Intervention components

The key components of self-management interventions are summarised in Table 8. This table reflects UK practice, as described in the studies selected from the mapping review (see the following section).

Number and type of UK studies identified

The mapping review identified 25 studies of self-management for cardiovascular or respiratory conditions. The largest group of studies dealt with COPD (12 studies), followed by asthma (six studies) and hypertension (five studies). One study covered both asthma and COPD. 142 The studies selected for inclusion in this analysis focused on asthma or COPD (or both) because these were the conditions with most data on the role of self-management for reducing avoidable admissions in the UK. Two studies on heart failure were excluded: one dealt largely with the development of an intervention146 and the other was used earlier in the review process to develop initial programme theories. 128 Details of the studies included in the analysis are presented in Appendix 4, Table 27.

Operating programme theories

By definition, PT1 is applicable to self-management education and support. However, the included UK studies provided mixed evidence relating to this theory. Studies generally failed to provide clear evidence of a reduction in hospital admissions associated with self-management interventions for patients with COPD or a mixture of long-term conditions. 138,139,144 There was evidence of patients feeling satisfied with non-secondary care provision of self-management support (PT2), but this referred to a service with enhanced support. 140 Implementation of asthma action plans was hindered by lack of confidence by GPs and other clinicians in the usefulness of the intervention (PT3). 141 PT4, referring to the possibility of patients delaying seeking treatment leading to greater overall use of health services, may also operate in some circumstances but we did not find evidence of this in our sample of studies.

Several studies related to PT5, which stresses the influence of the wider health-care system and the availability or otherwise of support and incentives. These included factors operating at the level of general practices, local NHS commissioners and national policies, such as the Quality and Outcomes Framework (QOF), that encourage practices to prioritise mandatory targets at the expense of other work. 137,138,144,147

Description of putative mechanisms

The key mechanism for self-management is thought to involve increasing a patient’s self-efficacy and confidence to be actively involved in managing their condition. Lorig and Holman136 state that increasing self-efficacy can be achieved through four key mechanisms: performance mastery, (role) modelling, reinterpretation of symptoms and social persuasion. Some relevant mechanisms highlighted in this analysis of UK studies are summarised in Box 4.

Contextual factors

Supporting evidence

This is a brief summary of evidence from systematic reviews, concentrating on hospital (re-)admission as the outcome of interest. Key results from UK studies included in this analysis are also presented.

For asthma, a Cochrane review published in 2003148 found that optimal self-management for adults with asthma almost halved the risk of hospitalisation [risk ratio 0.58, 95% confidence interval (CI) 0.43 to 0.77]. Optimal self-management was defined as provision of a written action plan for self-management of exacerbations together with self-monitoring and regular medical review. The finding was based on 36 randomised trials involving 6090 participants. Supported self-management is strongly recommended in national and international guidelines, including those of the British Thoracic Society/Scottish Intercollegiate Guidelines Network. 149 These guidelines cite 261 randomised trials included in 22 systematic reviews in support of the recommendation for self-management to be provided.

In a 2014 Cochrane review,63 self-management for patients with COPD was associated with a lower probability (compared with usual care) of respiratory-related hospitalisations [odds ratio (OR) 0.57, 95% CI 0.43 to 0.75, nine studies, 1749 participants, moderate-quality evidence] and all-cause hospitalisations (OR 0.60, 95% CI 0.40 to 0.89, six studies, 1365 participants, moderate-quality evidence). The authors calculated that, over 1 year of follow-up, 8 (95% CI 5 to 14) participants with a high baseline risk of respiratory-related hospital admission needed to be treated to prevent one participant with at least one hospital admission. 63 The corresponding figure for those at low baseline risk was 20 (95% CI 15 to 35) participants.

Although heart failure was not covered in this section of the report, systematic reviews and meta-analyses suggest a benefit of self-management interventions to reduce hospital admissions in this condition also. However, interpretation of findings is complicated by heterogeneity arising from issues such as varying trial designs, intervention components, follow-up periods and methods of outcome assessment. 150

In summary, there is strong evidence from systematic reviews of randomised trials supporting the effectiveness of supported self-management for reducing hospital admissions in respiratory conditions such as asthma and COPD and cardiovascular conditions such as chronic heart failure. This evidence has resulted in strong recommendations in clinical guidelines in favour of implementing self-management as widely as possible. However, guidelines also recognise the challenges of implementing a complex intervention with limited resources and of involving patients as active partners in managing their condition.

The UK research did not entirely support the findings of the systematic reviews. None of the three studies that evaluated the outcome found a decrease in hospital admissions with self-management relative to control groups. 138,139,144 The influence of these findings on UK practice is unclear and it should be noted that we did not attempt to identify all relevant UK studies or even a representative sample of them. The only economic evaluation included in our analysis suggested that self-management for COPD is likely to be cost-effective in the UK, with an estimated cost of £280.00 per QALY gained compared with usual care. 145

Potential enhancements of intervention/implementation

Suggestions for improving the implementation of self-management came from three qualitative studies. The suggestions originated from HCPs participating in the studies or from study authors rather than from patients or the public.

Suggestions from the study by Morrow et al. 137 included emphasising the evidence for benefit and improving teamwork (including team-based education) and organisational strategies (including remote consulting), which would need to fit within existing practice routines. Technology was thought to offer some potential solutions but these must be integrated with existing practice IT systems. In Ogunbayo et al. ’s147 study, normalising self-management into routine practice was identified as a theme at the practitioner level. Participants perceived that this involved use of various self-management planning tools as well as changing practitioner or practice culture.

Finally, Ring et al. 141 discussed the potential role of a multifaceted intervention to facilitate implementation of personalised action plans by tackling identified barriers (e.g. using standard templates, electronic prompts to review plans and encouraging patients to bring plans to review appointments).

Summary

Telehealthcare appears to offer inherent advantages in terms of feedback mechanisms, condition monitoring and communication between health providers and patients. However, some systemic issues remain: either the patient has to make the decision as to when emergency admission is appropriate, based on feedback data that need to be calibrated according to probable severity, or the HCP remains in control of that decision. In the latter case, it is critical that the HCP has all the information that is required to facilitate the decision. For example, they may wish to corroborate the telemonitoring data with observation of physical signs. Provision of telemonitoring data does not address the essential issues of patient self-efficacy or practitioner efficacy, namely the confidence to make a decision about the appropriateness of admission within a risk-averse culture or a safety-first psychological state.

Definition

‘Telemedicine’ is frequently used as an umbrella term. However, this term carries unwelcome connotations by placing the doctor at the centre of the interaction. Terms such as telehealthcare, telehealth and telecare are to be preferred. Telehealth involves the remote exchange of data between a patient and HCPs as part of a patient’s diagnosis and health-care management. Examples include monitoring of blood pressure or of blood glucose for diabetes. Telehealth may facilitate an improved patient understanding of their health conditions, providing them with tools for self-monitoring, encouraging better self-management of health problems and alerting additional professional support if devices signal a problem.

A Cochrane review of services for patients with asthma151 combines both the ‘health’ suffix and the ‘care’ suffix to define ‘telehealthcare’ as ‘the provision of personalized healthcare at a distance’. 151 Telehealthcare embodies three components: (1) information obtained from the patient, (2) electronic transfer of data over a distance and (3) personalised feedback from a HCP.

‘At a distance’ requires use of a ‘tool of distance communication that works without the simultaneous physical presence of the participants in the interaction’. Such a technology might be ‘telephone, e-mail, the internet or any other networked or mobile device’. 151 The emphasis is on the nature of the interaction; the novelty or sophistication of the technology is irrelevant to an understanding of how the interaction is successful, although unfamiliarity and complexity of the technology may add further barriers to implementation. Feedback from the HCP to the patient could be synchronous or asynchronous (i.e. by store-and-forward technology, in which a patient’s data are kept in an electronic repository and forwarded to a HCP on request). Advice should be tailored to the consulting patient. Many telehealthcare interventions are designed to help patients manage their condition and, therefore, several of the underpinning mechanisms may relate to the facilitation of self-management.

There are no recommendations for telehealthcare for the management of chronic heart failure in the NICE, European Society of Cardiology or American College of Cardiology/American Heart Association guidelines. 152 Despite the lack of UK guidance, the government has vigorously advanced the telehealth agenda through the 3 Million Lives initiative. This policy initiative aspired to enhance the lives of the population by encouraging and supporting (but not directly funding) the use of telecare and telehealth applications as a vehicle for providing more person-centred and integrated care.

Intervention components

The key components of telehealthcare interventions are summarised in Table 10. This table reflects UK practice, as described in the studies selected from the mapping review (see the following section).

Number and type of UK studies identified

The mapping review identified one UK quantitative study of telehealthcare, from the Whole System Demonstrator Project,71 for COPD, diabetes and heart failure. The quantitative TELESCOT RCTs153–155 that accompanied several UK qualitative studies predated our cut-off date, being published in 2009.

The mapping review identified a total of seven UK qualitative studies: three examined COPD, one examined hypertension and the remainder explored multiple conditions. Three UK qualitative studies had been conducted alongside the TELESCOT trials. Further qualitative studies examined COPD patients receiving a community respiratory service156 and patients with hypertension,157,158 both in the West Midlands. Ure et al. 159 conducted a qualitative study examining a COPD telemonitoring service in NHS Lothian and a critical commentary160 examined transferable lessons from this experience. Finally, Williams et al. 161 examined patients’ expectations and experiences of a mHealth application for COPD in Oxfordshire.

Details of the studies included in the analysis are presented in Appendix 4, Table 28.

Operating programme theories

Crundall-Goode and Goode152 identify four types of barriers to successful implementation of telehealthcare applications for chronic health conditions: patient related, health and social care organisation related, technology related and evidence/economic related. The first and second of these barriers figure prominently among our candidate programme theories, with technology related being specific to telehealthcare interventions and evidence/economic related being picked up across all interventions in our PARiHS framework.

Specifically, telehealthcare, in the form of telemonitoring, operates within PT1. PT1 states that if patients are equipped with knowledge and information for self-management, they will access health services only when required, leading to appropriate resource use and a reduction in unplanned admissions. Telemonitoring addresses the fear and anxiety that a patient experiences if they are not sure whether or not to call the health services. They have the reassurance that their condition is being monitored, thereby accelerating the feedback loop between patient and clinician. However, telemonitoring is limited, given its focus only on specific signs, whereas the patient’s symptoms may cross a wide range of circumstances, either linked to or independent of what they are being monitored for. In this case, the telemonitor may be offering an alternative or contradictory picture that increases, rather than reduces, the uncertainty.

Alternatively, people with chronic conditions may perceive that presentation to an emergency department holds relative advantage (e.g. quality, ease of access, response) over GP-based or other primary or community care services (PT2). People with chronic conditions may pressure GPs to admit them to hospital. All of the above explanations assume that people behave in a ‘rational’ manner when seeking information. However, information need has an important affective dimension. As has already been demonstrated, much inappropriate utilisation by patients is attributable to anxiety and uncertainty. Provision of information does not always allay this; its effectiveness depends on whether it is trusted or whether it needs further verification or validation. Under such circumstances, a patient may consult an information source but have an outstanding need for a ‘second opinion’, particularly if the information they have to hand is not situated in their specific circumstances. The perceived authority of the local hospital may be seen to eclipse the less satisfactory response from local primary care provision.

Alternatively, GPs or other HCPs may lack confidence in their own diagnoses or may lack confidence in or knowledge of alternative sources of health-care provision and so may refer people with chronic conditions or admit them directly to hospital. GPs or other HCPs may feel under pressure to admit people with chronic conditions directly to hospital (PT3). In the context of patient education, and not the educational needs of the HCPs themselves, health providers may find it challenging to identify and locate information for patients, which is, again, situated within the patient’s specific needs. Alternatively, they may encounter contradictions between what the available information recommends and their own clinical judgement.

People with chronic conditions may also use health services inappropriately, delaying their presentation to a GP or hospital (PT4) because of perceptions of the service either anticipated or based on their own or others’ past experiences. Such mental imprints may exert a stronger and more powerful influence over their behaviour than the provision of patient education.

Salisbury et al. 162 identify the existence of the CCM as a dominant way of conceptualising chronic disease management. However, they highlight that this model was not designed with telehealthcare in mind and that it does not, therefore, readily accommodate considerations relating to this type of intervention.

Description of putative mechanisms

Box 5 provides a summary of putative mechanisms. We believe that telehealthcare is unique among the interventions examined in this report, given that a realist synthesis has already been conducted to identify the core mechanisms underpinning telehealthcare interventions. Vassilev et al. 45 examined three chronic health conditions, including COPD and heart failure (which are eligible for this review), and identified three concepts that suggest how telehealthcare engages with patients, carers and HCPs:

1. relationships – if and how a telehealth intervention enables or limits the possibility for relationships with professionals and/or peers

2. fit – how well telehealthcare interventions can be integrated into everyday life and health-care routines and the extent to which they are easy to use; compatible with patients’ existing environment, skills and capacity; and do not significantly disrupt patients’ lives and routines

3. visibility – visualisation of symptoms and feedback has the capacity to improve knowledge, motivation and a sense of empowerment, to engage network members and to reinforce positive behaviour change, prompts for action and surveillance.

These three concepts (communication, fit and visibility) clearly have a critical role in acceptance of the intervention. Our own programme theories have a different emphasis and particularly confirm aspects of visibility, namely the closure of a feedback loop so that patients and professionals can respond to current symptoms and their severity and make a judgement on whether emergency admission is indeed appropriate or whether appropriate alternatives exist. The fact that this assessment can be conducted in real time provides the additional benefit that neither the patient nor the professional has to worry about the deterioration or, indeed, improvement of the symptoms in the interval between the ‘trigger’ symptoms (e.g. an exacerbation) and contact with a HCP.

Contextual factors

Supporting evidence

This is a brief summary of evidence from systematic reviews, concentrating on hospital (re-)admission as the outcome of interest. Key results from UK studies included in this analysis are also presented.

For COPD, Polisena et al. 170 showed lower rates of emergency admissions for patients receiving home monitoring plus telephone support. Their systematic review tentatively suggested that home telehealth could reduce the rate of hospitalisation for patients with COPD, but was limited by the quality of the studies included. 170

An important review outside our inclusion period, by Paré et al. ,171 reviewed 65 empirical studies across four conditions and distinguished the effects of telehealth between different conditions. They suggested that effects on reduced visits to emergency departments, hospital admissions and average length of hospital stay were more consistent in pulmonary and cardiac disorders than in diabetes and hypertension. 171

A Cochrane review of studies of heart failure reported that telemonitoring and structured telephone support reduced admissions for heart failure. 172 However, findings were based on generalising a large number of studies with a mean sample size of 330. A study of 1653 patients with heart failure found no significant effect on hospital use or mortality. 173

Data supporting the efficacy of telemonitoring in reducing hospital admissions is mixed. Clinical trials of telemonitoring reported within the inclusion period have not demonstrated the positive impact on hospital admissions found in a systematic review based on smaller studies. 174–176

Potential enhancements of intervention/implementation

Peirce et al. 177 draw on stakeholder interviews to conclude that telemonitoring systems have not overcome the drawbacks of conventional care because ‘they existed as an off-the shelf “bolt-on” and could not be connected with many of the different care services that these patients engage with’. They assert that a future early-detection telemonitoring system should be designed with high levels of user input and, therefore, be characterised by high usability and integration with multiple services. To this they add technical requirements such as more versatility of monitoring capabilities and proven automated methods of pattern detection. Crundall-Goode and Goode152 highlight how there is little guidance outside a trial context on the optimal duration of telemonitoring arrangements, citing evidence that mortality benefits may be lost past a certain duration of intervention. They therefore suggest the enhancement of a care plan that proposes when to step down telehealth and have devised their own example of this.

Summary

Cardiac rehabilitation is offered to people with cardiovascular disease, typically those who have survived a heart attack but also patients with angina or heart failure. Cardiac rehabilitation programmes typically involve tailored exercise together with education and advice to help patients reduce risk factors and recover their confidence. Cochrane reviews of cardiac rehabilitation concentrate on explaining the benefits of the intervention through the physiological effects of exercise on the cardiovascular system. However, given that most cardiac rehabilitation programmes involve interventions other than structured exercise, other mechanisms are likely to be involved in most cases.

Cardiac rehabilitation should be delivered by a MDT led by a clinical co-ordinator (a senior clinician responsible for co-ordinating, managing and evaluating the service). There is a strong international evidence base for cardiac rehabilitation but interpretation of the evidence for the UK is complicated by multiple factors, including a trial casting doubt on its effectiveness as delivered in UK practice. At a national level, the National Audit of Cardiac Rehabilitation (NACR) has provided leadership for cardiac rehabilitation programmes in most countries of the UK. The publication of regular reports and data has been accompanied by a campaign of support for cardiac rehabilitation backed by the British Heart Foundation.

Definition

Cardiac rehabilitation is an intervention offered to people with cardiovascular disease, typically those who have survived a heart attack but also patients with angina or heart failure. Cardiac rehabilitation programmes typically involve tailored exercise together with education and advice to help patients reduce risk factors and recover their confidence. More detailed definitions are available from the NACR (www.cardiacrehabilitation.org.uk/patient-information.htm; accessed 9 March 2018) and the British Association for Cardiovascular Prevention and Rehabilitation (BACPR). 178

Intervention components

The six core components of cardiac rehabilitation as defined by the BACPR are health behaviour change and education; lifestyle risk factor management; psychosocial health; medical risk management; long-term strategies; and audit and evaluation. Table 12 summarises the key features of the intervention.

The BACPR standards for cardiac rehabilitation specify that the intervention should be delivered by a MDT led by a clinical co-ordinator (a senior clinician responsible for co-ordinating, managing and evaluating the service). 178 The standards acknowledge that the composition of the MDT may vary but state that, overall, the team must have appropriate knowledge, skills and competencies in line with the standards and be able to deliver the core components. 178 The disciplines considered key to the delivery of cardiac rehabilitation include dietitian, exercise specialist, nurse specialist, occupational therapist, pharmacist, physician with special interest in prevention and rehabilitation, physiotherapist and practitioner psychologist. The BACPR standards also stress the importance of active engagement with a patient’s wider care team, for example the GP, other general practice staff and social worker. 178

The content of cardiac rehabilitation programmes is designed to deliver the six core components mentioned earlier in this section while meeting individual goals and taking account of patient preference and choice. The BACPR recommend that patients should be offered a choice in terms of venue (e.g. home, community setting or hospital) and timing of sessions. Cardiac rehabilitation may be delivered in a variety of ways (centre based, home based, manual based, web based, etc.). However the programme is delivered, patients are encouraged to engage in structured exercise at least two to three times per week, with documentation of regular review, goal-setting and exercise progression. The BACPR standards specify that the patient and MDT should work together for at least 8 weeks. 178 Detailed standards for the non-exercise components of programmes are also provided.

Number and type of UK studies identified

The mapping review included just two UK studies coded for cardiac rehabilitation: a qualitative study of web-based cardiac rehabilitation in primary care179 and a pilot study of early rehabilitation for patients hospitalised for heart failure. 180 Two systematic reviews of cardiac rehabilitation provided evidence for the effectiveness of cardiac rehabilitation and identified further UK studies181,182 (see Appendix 4, Table 29), although many of these were outside the date range of interest for this review. A key UK study identified from the reviews was the Rehabilitation After Myocardial Infarction Trial (RAMIT), which concluded that ‘the value of cardiac rehabilitation as practised in the UK is open to question’. 183 The publication of RAMIT was accompanied by editorials and commentaries that, although not empirical research, provided important insights into the implementation of cardiac rehabilitation programmes in the UK. 184–186 Another randomised trial questioned the effectiveness and cost-effectiveness of exercise-based rehabilitation for elderly patients with heart failure. 187 Finally, we included the latest report from the UK NACR, which provides detailed statistical information about the delivery of cardiac rehabilitation in different regions of the UK (excluding Scotland). 188 A recent publication using NACR data identified factors influencing patient engagement with cardiac rehabilitation programmes, including service-level factors. 189

Operating programme theories

Cardiac rehabilitation is primarily concerned with preventing further events and re-admissions for people hospitalised for CHD or heart failure. As the intervention includes behaviour change and education, the overarching PT1 appears relevant. This states that IF patients are equipped with knowledge and information for self-management, they will access health services only when required, LEADING TO appropriate resource use and a reduction in unplanned admissions. The main evidence for this comes from RAMIT, in which the majority of cardiac rehabilitation patients rated 7 out of 11 elements (information on heart disease, risk factors, medication, advice on diet, exercise training, relaxation training and stress management training) as being very or fairly helpful on a four-point scale. 183 However, this finding is difficult to interpret as the study did not find evidence of a reduction in admissions in the cardiac rehabilitation group.

If patients are undergoing cardiac rehabilitation as part of a co-ordinated programme of care at home or in the community, the overarching PT2 may operate. Briefly, IF patients are satisfied with provision outside secondary care, THEN they will not feel it necessary to access secondary care services, LEADING TO appropriate use of services and reduction in unplanned admissions. Evidence on provision outside secondary care was lacking in the included studies, making the role of PT2 difficult to evaluate.

Programme theory 3, relating to clinicians’ confidence in diagnosis and referral, is of limited relevance because cardiac rehabilitation is mainly used to reduce re-admission in patients who have already been diagnosed and hospitalised. A similar statement may be made about PT4 (adverse effects of patient delays in accessing treatment), but there is evidence from the included studies of patients failing to take up cardiac rehabilitation or dropping out before completing the programme, which would prevent them from obtaining the benefits claimed for cardiac rehabilitation. 188,189

Description of putative mechanisms

Cochrane reviews of cardiac rehabilitation concentrate on explaining the benefits of the intervention through the physiological effects of exercise on the cardiovascular system. For example, Anderson et al. 181 list six direct benefits of exercise training on the heart and coronary vasculature. They also note possible indirect effects of exercise on risk factors for atherosclerotic disease, such as blood lipids, smoking and blood pressure. 181 These indirect effects include changes in behaviour as well as physiological changes.

The Cochrane review of exercise-based rehabilitation for heart failure also stresses the physiological benefits of exercise training while acknowledging that the precise mechanism(s) leading to benefits in health outcomes is unclear. 182

Given that most cardiac rehabilitation programmes involve interventions other than structured exercise, other mechanisms are likely to be involved in most cases (Box 6). Core components of cardiac rehabilitation aim to support people to change their behaviour and hence reduce their modifiable risk factors for hospital admission or re-admission and other adverse outcomes. This includes increasing participants’ knowledge but also increasing their confidence so that they continue to exercise, abstain from smoking, etc., after the end of the formal cardiac rehabilitation programme. As most cardiac rehabilitation programmes are delivered in groups, support from other group members as well as programme staff is likely to be important. 190

Contextual factors

Supporting evidence

This is a brief summary of the evidence for the effectiveness of cardiac rehabilitation in terms of reduction of hospital (re-)admissions, based on systematic reviews and guidelines.

Cardiac rehabilitation is recommended in the NICE guidelines on secondary prevention after myocardial infarction192 and management of chronic heart failure,121 and in other national and international guidelines. 193 The Department of Health and Social Care published guidance on commissioning cardiac rehabilitation in 2010. 194

There are Cochrane systematic reviews of cardiac rehabilitation (described as ‘exercise-based rehabilitation’) for CHD181 and heart failure,182 as well as numerous other systematic reviews and meta-analyses. In the review of cardiac rehabilitation for CHD, cardiac rehabilitation reduced overall risk of hospital admissions by 18% compared with no-exercise controls [15 trials; risk ratio 0.82, 95% CI 0.70 to 0.96). 181 A larger effect was seen in the Cochrane review of cardiac rehabilitation for patients with heart failure. 182 Compared with a control, exercise training reduced the rate of overall (15 trials, 1328 participants; risk ratio 0.75, 95% CI 0.62 to 0.92, fixed-effect analysis) and heart failure-specific hospitalisation (12 trials, 1036 participants; risk ratio 0.61, 95% CI 0.46 to 0.80, fixed-effect analysis).

Numerous factors complicate the interpretation of this evidence for the UK NHS setting. As noted previously, RAMIT questioned the effectiveness of cardiac rehabilitation programmes as delivered to patients with CHD in NHS hospitals. This prompted a considerable amount of discussion. Some authors questioned the relevance of the Cochrane reviews on the grounds that they include older studies that do not reflect the improved treatment offered to patients receiving ‘usual care’ in modern practice. However, advocates of cardiac rehabilitation point out that there is no strong evidence of a decline in the effect of cardiac rehabilitation on mortality over time. 185 The issue remains unresolved and appears to be a potential barrier to full implementation of cardiac rehabilitation in the NHS.

Potential enhancements of intervention/implementation

Factors favouring successful implementation were identified as follows:

• Giving patients a firm date for initial assessment. 189

• Web-based programmes may promote wider engagement with cardiac rehabilitation. 179,191

• Early (within 4 weeks of hospital discharge) rehabilitation for patients with heart failure. 180

Summary

The British Lung Foundation’s definition of pulmonary rehabilitation is an exercise and education programme for people with long-term lung conditions. 195 UK programmes are diverse in setting; length, type and intensity of exercise; and provision of education. A key mechanism for the effect of pulmonary rehabilitation appears to be increased self-efficacy for both exercise and self-management. There is strong evidence for the effectiveness of pulmonary rehabilitation (including reducing re-admissions) from systematic reviews of RCTs of pulmonary rehabilitation for COPD. 70,186,187,192,193,196 However, further research is needed to investigate the components of COPD that are most important, how long the programme should be and the frequency of the training.

Definition

The British Lung Foundation’s definition of pulmonary rehabilitation is ‘a programme of exercise and education for people with a long-term lung condition’. 195 Pulmonary rehabilitation consists of a tailored physical exercise programme combined with general health advice and education on managing the condition and associated symptoms. 195

The British Thoracic Society Guideline on pulmonary rehabilitation for adults (2013) has recommendations for the physical exercise training and education components of pulmonary rehabilitation. 197

Intervention components

The updated Cochrane review highlights how pulmonary rehabilitation programmes (PRPs) offer diverse components. 70 They can be based in inpatient or outpatient settings, or at home. The length of the programmes can also differ, ranging from 6 weeks to 6 months, and training intensity can range from daily to twice weekly. The exercise training component can include endurance or strength training or both, with specific exercise training selected to best meet the needs of each individual patient. In addition, the patient education component of the programme can extend from basic advice through to more extensive self-management programmes. All of these differences in the components can influence their effectiveness and contribute to the complexity of implementation. Key components of pulmonary rehabilitation are detailed in Table 13. This reflects UK practice as described in the studies selected from the mapping review.

Number and type of UK studies

The mapping review identified 17 papers of 15 UK studies for pulmonary rehabilitation for COPD; one of the studies was for exacerbations of chronic respiratory disease but the majority of the patients had a primary diagnosis of COPD. 203 In addition, three systematic reviews of pulmonary rehabilitation were identified. 69,196,208 These reviews demonstrated the effectiveness of pulmonary rehabilitation and helped with the identification of further UK studies, although many were outside the date range of interest for this review. Five qualitative studies and 11 quantitative studies are included in the realist synthesis (see Appendix 4, Tables 30a and 30b). One of the quantitative studies is reported in three papers. 198–200

Operating programme theories

Pulmonary rehabilitation programmes aim to improve the length of time that patients with COPD can exercise for without feeling out of breath and to improve their symptoms, self-confidence and emotional well-being. According to the Cochrane review by McCarthy et al. ,196 pulmonary rehabilitation is likely to achieve its effect through several mechanisms. Collectively, pulmonary rehabilitation seeks ‘to reduce COPD symptoms, re-establish and improve functional ability, enhance participation in everyday life, promote autonomy and improve [quality of life]’. Pulmonary rehabilitation focuses on systemic aspects of the disease common among patients with COPD. The three principal components are (1) an exercise component, (2) an educational component and (3) a behaviour change element.

The exercise component of pulmonary rehabilitation targets physical improvement by increasing inspiratory volume and reducing dynamic hyperinflation. Exercise increases muscle function, delaying fatigue and resulting in increased exercise tolerance. 196 Simultaneously, the educational component of pulmonary rehabilitation focuses on collaborative self-management and behaviour change. Patients are provided with information and knowledge regarding COPD, building skills such as goal-setting, problem-solving and decision-making, and developing action plans that allow them to better recognise and manage their disease. The behaviour change element focuses on modifying nutritional intake and smoking patterns; adhering to medication and regular exercise; and utilising effective breathing techniques and energy-saving strategies. These last two elements emphasise the similarities of pulmonary rehabilitation to other interventions, such as patient education and self management.

As PRPs include patient education, PT1 resonates as one possible theory of change. When patients are equipped with knowledge/information for self-management of COPD, including when and where to seek help as appropriate, they will hopefully access the hospital/health services as required. This, in turn, could lead to appropriate utilisation of health resources and a reduction in unplanned admissions. Participants in one programme198–200 described how their relationship with HCPs changed to a more collaborative relationship and that they felt that they now knew what to ask HCPs for and how to self-manage their condition, which was empowering for them.

One study worked with GPs and practice nurses to develop strategies that would influence and consequently increase their referrals to pulmonary rehabilitation. 209 This relates to PT5 when GPs and other HCPs are influenced by the wider context of the health-care system and the availability or otherwise of support and incentives may influence their adoption of pulmonary rehabilitation, which could help to avoid unnecessary referrals and admissions to hospital.

Programme theory 4 refers to the possibility of patients delaying seeking treatment, which then leads to greater overall use of health services; this could happen with patients with COPD, and one of the UK studies investigating the way in which professionals introduce the possibility of a pulmonary rehabilitation-affected uptake discussed this. 202 The study found that COPD patients are often self-conscious about their condition; this can be related to shame or stigmatisation and they can believe that the COPD is their own fault. This can be associated with a reluctance to ask for help from HCPs, which could lead them to not taking up pulmonary rehabilitation or not completing it, which could then mean that their health deteriorates and they then need to access other services. Their reluctance to ask for help could therefore lead to them making greater use of the health service or to unavoidable admissions.

Description of putative mechanisms

Pulmonary rehabilitation consists of two components: exercise and patient education (Box 7). The patient education component can be basic information or advice about COPD or medication or a more extensive self-management programme. One mechanism in self-management involves enabling participants to successfully understand the need for change, to appreciate the benefits of change and to live with the change long term. This mechanism took place in the UK study in which occupational therapists delivered a participant-empowered approach to pulmonary rehabilitation that built on traditional models of self-management. 201 The programme was highly beneficial and patients improved their physical fitness, well-being and confidence to manage their condition.

Self-management programmes also work on increasing the patients’ self-efficacy and confidence in managing their own condition. 210,211 Through completing a self-management programme and the exercise component, patients could also increase their confidence in their ability to exercise, meaning that they will then exercise more. Enhanced understanding of the benefits of regular activity as part of disease management prompted increased participation. 212

Pulmonary rehabilitation programmes are provided to patients in groups and exercising in a group can have important benefits for the patients:

The key premise of social identity theory is that group membership (e.g. exercise group) to which a person belongs can provide an individual with a sense of who they are in terms of a defined group identity (i.e. ‘we’ and ‘us’ rather than ‘I’ and ‘me’), that is, the way person feels and thinks about self is derived from their social groupings.

Reproduced with permission from Levy et al. 213 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/

The operation of social identity theory can be demonstrated in UK studies on pulmonary rehabilitation. A qualitative study considering the barriers and facilitators for patients continuing to exercise following the programme found that continued peer and professional support was important to participants. 212 Participants on another programme199,200 described how working in the group helped them by making them motivated to continue with the programme and helped them to focus on what they wanted to achieve. One qualitative study found that professional and peer support were important aspects of the programme. Participants want to exercise in a peer group and saw this as an opportunity for social interaction. 212

Contextual factors

Supporting evidence

Three systematic reviews of pulmonary rehabilitation were identified. 69,70,196 Two are updated Cochrane systematic reviews, published in 2015196 and 2016. 70

The 2015 updated review reviewed RCTs that found that pulmonary rehabilitation improves the health-related quality of life of people suffering from COPD and that it should be part of the management and treatment of patients with COPD. 196 Further research in this area could beneficially consider which components of pulmonary rehabilitation are the most important, the ideal length of the programme, the intensity of training required and how long the programme benefits actually last.

The second updated Cochrane review70 and the other review69 considered the effect of pulmonary rehabilitation following acute exacerbations of COPD. Puhan et al. 70 found that pulmonary rehabilitation improves quality of life and exercise capacity and is a safe intervention for patients with COPD following an exacerbation. However, the reasons for diverse effects on hospital re-admissions and mortality shown in the studies are not fully clear and future studies should research whether or not the extent of the rehabilitation programme and the organisation of such programmes in specific health-care systems (e.g. in the rehabilitation setting vs. embedded in the continuum of care from hospital to home to outpatient care) influences the effects of rehabilitation after COPD exacerbations. The third review69 reviewed the results from 10 RCTs, which suggested that pulmonary rehabilitation reduces subsequent admissions, but pooled results from the three cohort studies did not. This could possibly be explained by the heterogeneous nature of individuals included in observational research and the varying standard of PRPs.

Potential enhancements of intervention/implementation

A study investigating the impact of deprivation on pulmonary rehabilitation completion found that patients living in more deprived areas are less likely to complete a PRP. 215 However, people in deprived areas who completed a PRP had similar clinical outcomes to patients in more affluent areas. Therefore, interventions targeted at enhancing referral, uptake and completion of pulmonary rehabilitation among patients living in deprived areas could reduce morbidity and health-care costs in this hard-to-reach population. Another study found that introducing strategies that are relatively easy to implement lead to the potential for more patients to access the health and quality-of-life benefits that pulmonary rehabilitation offers. 209

A large well-conducted RCT compared the efficacy of pulmonary rehabilitation delivered in a community setting with pulmonary rehabilitation delivered in a more traditional hospital-based setting. 204 Both settings produced significant improvements in terms of exercise capacity and quality of life acutely and after long-term follow-up. The costs of the two programmes were similar. The choice of model will therefore depend on local factors of convenience, existing availability of resources and incremental costs. Another RCT found that a primary care-based structured education PRP is feasible and may increase local accessibility to people with moderate and severe COPD. 198–200

Occupational therapists can deliver a participant-empowered approach to pulmonary rehabilitation that builds on traditional models of self-management. Critically, this approach enables participants to successfully understand the need for change, appreciate the benefits of change and live with the change long term. 201

Introducing the group opt-in session improved the graduation rates at The North Bristol Lung Centre pulmonary rehabilitation course and reduced wasted assessments. 205

Professionals should facilitate an open discussion, which offers time, compassion and understanding, with patients as a means of facilitating pulmonary rehabilitation uptake. 202 For patients who refused referral to pulmonary rehabilitation, had not completed a course or had yet to be referred, the way the service was introduced was an important determinant of willingness to participate. 216 Recognition of the role that uncertainty plays in patients with COPD is the first step towards developing interventions focused on reducing this uncertainty, thereby reducing the burden of the disease for the individual patient and facilitating pulmonary rehabilitation attendance. 217

Smoking status, availability of social support and markers of disease severity were predictors of attendance and adherence to pulmonary rehabilitation. 218

Summary

Specialist clinics provide advanced diagnostic or treatment services for specific conditions, often in primary care or community settings, with care delivered by specialist nurses or MDTs. Possible mechanisms for effectiveness of specialist clinics include improved self-efficacy and open communication with HCPs, especially specialist nurses. Limited UK research on specialist clinics was identified. Systematic review evidence supported their effectiveness for heart failure but evidence for hypertension was less clear. Potential enhancements to improve effectiveness include clinics being run by nurses already known to the patient and a high intensity of contact immediately after discharge, decreasing over time.

Definition

Specialist clinics provide advanced diagnostic or treatment services for specific conditions. Such clinics exist in both primary and secondary care settings. They may use nurses to lead clinics or MDTs to help manage long-term conditions. 3

Intervention components

Specialist clinics are provided by a range of HCPs trained in the specialist care of the specific condition or patient type. The clinics are provided in the community and can include staff from primary and secondary care. The clinics are of varying intensity and comprise different components of care. More details of the key components of specialist clinics are provided in Table 14. This reflects UK practice, as described in the studies selected from the mapping review.

Number and type of UK studies identified

Two UK research studies on specialist clinics published from 2010 onwards were identified from the mapping review. 219,220 One of the clinics was for patients with hypertension219 and one was for high-risk respiratory patients with COPD or asthma. 220 The research studies were both quantitative: one was a RCT219 and the other was an uncontrolled observational study. 220 A summary table of the included studies is provided (see Appendix 4, Table 31).

Operating programme theories

Specialist clinics included education about the chronic condition and any medication that relates to PT1. If the specialist clinics can educate patients about their condition, this can lead to self-management, which could then lead to a reduction in unplanned admissions. The RCT for hypertension219 found a reduction in systolic blood pressure; this reduction potentially means that the condition is being appropriately managed and thus should mean that these patients will not have unplanned admissions. The observational study found a reduction in acute respiratory exacerbations, unscheduled visits to GPs and hospital admissions for acute respiratory exacerbations. 220

Linking to the GP programme theory, PT5, potentially several aspects of the UK RCT219 had a beneficial effect on patient blood pressure. One aspect was a letter about the clinic and explaining that they had been invited as their last recorded blood pressure reading was higher than recommended levels. Although they might not attend the clinic, they might more regularly take their tablets and see their own GP for regular blood pressure checks. Patients who did attend might start taking their medication more regularly as they learnt more about the risks of cardiovascular disease and because their blood pressure was being monitored on a regular basis. Motivational interviews conducted by the nurse could also have helped to reinforce regular taking of medication and lifestyle changes. In addition, in this study, the patients with blood pressure above the target were being regularly followed up, which might not be the norm in primary care. This links to the idea of clinical inertia, and if nurse-led specialist clinics can overcome clinical inertia, then they could be effective. 221

In addition, for PT5, if clinicians and other health service staff believe that specialist clinics provide appropriate support, then they will refer patients to them if they are established or will implement the setting up of specialist clinics where appropriate, which, through the changes to practice and their professional role, will lead to appropriate utilisation of health resources and, potentially, a reduction in unplanned admissions.

The two included studies offered specialist clinics in the primary care setting, which links to PT2. The provision of a specialist clinic in the primary care setting will hopefully mean that the patients feel confident and satisfied with the care offered to them and will continue to use these services and not request secondary care services. Positive written feedback from patients indicated that the patients found the joint clinics provided by primary and secondary care professionals for high-risk respiratory patients acceptable and that they were satisfied with health services provided. 220 This observational study found a reduction in hospital admissions, indicating that the clinic was potentially effective. Satisfaction with the clinic was also demonstrated by the high attendance levels, and some of the patients attending had previously missed their routine annual COPD check-up.

Programme theory 3 could potentially be relevant in terms of GPs and primary care staff. If GPs and primary care staff have confidence in and knowledge of specialist clinics available within primary care, then they will be more likely to refer patients to the specialist clinics instead of secondary health-care provision. Diffusion of innovation is relevant here: if roles (i.e. specialist nurse clinics) are not supported and embedded into the existing system, then patients will not be referred to them. A case study of consultant nurse posts discussed how they developed in an ad hoc manner and that new posts need to be supported by HCPs and embedded into the system to succeed. 222 This diffusion of innovation is based on an earlier systematic review by Greenhalgh et al. 223

Description of putative mechanisms

Possible mechanisms of action are summarised in Box 8.

Contextual factors

Both of the included studies discussed how patient preference could work as either a barrier to or a facilitator of the operation of specialist clinics. Feedback from patients who attended the joint clinics for high-risk respiratory patients220 was positive. All patients described attending the clinic as a positive experience in terms of patient experience of the clinic and the interventions made. The reduction in hospital admissions indicated in this study suggests that the clinic was potentially effective. Satisfaction with the clinic was also demonstrated by the high attendance levels, and some of the patients attending had previously missed their routine annual COPD check-up.

The RCT researched patients from two inner-city practices who were invited to a specialist nurse-led hypertension clinic or received usual care. 219 Eighty-two patients were randomised to the specialist nurse-led hypertension clinic; of these, 27 (33%) chose not to attend. The study authors thought that the attendance would potentially have been higher if the clinic was run by a practice nurse with whom patients were familiar.

Supporting evidence

Evidence for specialist clinics was found in the series of systematic reviews by Purdy et al. 3 and one conference abstract of a review on specialist clinics for hypertension. 224

Purdy et al. 3 found that specialist clinics for patients with heart failure can decrease the risk of unplanned admission, being most effective when patients had frequent appointments close to the time of discharge, which then reduced during follow-up.

The other review224 considered the involvement of nurses in the management of hypertension, particularly in the UK. The systematic review found evidence of improved outcomes with nurse prescribers from non-UK health-care settings. However, the review was able to identify only one trial of adequate size recruiting patients with hypertension in UK primary care, and found that, although it is currently practice for hypertension management to move from GPs to nurses in the UK, there is actually no good-quality trial evidence from UK primary health care that supports the widespread employment of nurses in hypertension management.

Potential enhancements of intervention/implementation

The RCT authors220 suggested that a specialist clinic run by a practice nurse known to patients might increase uptake.

Purdy et al. 3 found that specialist clinics were most effective when there was a high intensity of appointments immediately after discharge from hospital that then steadily reduced.

Summary

Community interventions are interventions that take place in community settings but do not primarily involve case management, attending a specialist clinic, education or exercise/rehabilitation. Examples were identified from the UK for heart failure and COPD, including a ‘natural experiment’ comparing different COPD service models commissioned in demographically similar areas of Wales. A key component is providing support to people living in the community, often through home visits. Possible mechanisms of action depend on formation of a trusting relationship between the patient and the HCP (usually a nurse) delivering the service.

Overall, the systematic review evidence identified by our mapping review is not strongly supportive of the effectiveness of community interventions for reducing hospital admissions. However, this may reflect the inherent clinical heterogeneity of the interventions in this group. Study authors have suggested a diverse range of possible enhancements to interventions, without any clear themes emerging.

Definition

Community interventions are defined by exclusion as those that take place in community settings but do not primarily involve case management, attendance at a specialist clinic, education or exercise/rehabilitation. 3

Intervention components

Key components are summarised in Table 15 based on the TIDieR-Lite checklist. This is based on our sample of included UK studies.

Number and type of UK studies identified

The mapping review identified eight UK studies covering community interventions. Two were qualitative studies of specialist heart failure nurse services in the community. 127,226 Both studies presented HCPs’ perspectives only.

Three other studies dealt with COPD. A qualitative study of barriers to and facilitators of physical activity for patients in primary care227 appeared relevant to designing community interventions that might be less demanding for patients than full pulmonary rehabilitation. An evaluation of a telephone alert service provided by the Met Office228 differs from other community intervention studies in that the intervention was not provided by the NHS (although it was commissioned by multiple primary care trusts) and did not involve home visits. The service reduced hospital admissions when admissions with a comorbid diagnosis of COPD were included. 228 It was discontinued in 2013 and evaluated retrospectively. Flood-Page225 used a ‘natural experiment’ to compare different models of community care implemented in neighbouring areas with similar demographic characteristics. The authors suggested that service models with close links between primary and secondary care may be more effective in reducing hospital admissions than those involving primary care alone.

Appendix 4, Table 32, summarises the included studies. A news article,229 a narrative literature review230 and a conference abstract231 were excluded on the grounds of being too short to provide relevant data for this analysis.

Purdy et al. ’s3 review of community interventions included only one UK study. This was a randomised trial of home-based support for older patients with heart disease following hospital discharge. 232 The study fell outside the time period for inclusion in our review. A citation search of this study did not identify any more recent relevant community intervention studies.

Operating programme theories

The relevant programme theories are as follows:

• PT1 – IF patients are equipped with knowledge/information for self-management, THEN they will access hospital/health services as required, LEADING TO appropriate utilisation of health resources and a reduction in unplanned admissions. Increased patient knowledge could result from working with specialist nurses or other community-based HCPs, but this was not reported in our included studies, which concentrated on the perspectives of HCPs.

• PT2 – IF patients feel satisfied with non-secondary care provision, THEN they will not consider it necessary to access secondary care services, LEADING TO appropriate utilisation of health resources and a reduction in unplanned admissions. The included studies from the UK do not directly support this theory because they do not include data on patient satisfaction and related outcomes.

• PT3 – IF GPs/primary care staff feel confident in their own ability to diagnose and/or refer patients appropriately and have confidence in and knowledge of services available within primary and community care, THEN they will not refer patients to hospital, LEADING TO an increase in use of self-management and non-secondary care services and a reduction in unplanned admissions. This theory appears relevant to some of the studies included in Supporting evidence. Flood-Page225 found that services with close links between primary and secondary care appeared to be most effective for reducing COPD admissions. 225 Similarly, MacKenzie et al. 226 highlighted the improving relationships between GPs and specialist nurses over time following implementation of a specialist nurse service and the role of the nurses in providing co-ordination and continuity of care for patients with heart failure.

• PT5 – IF clinicians and other health service staff perceive that the wider health system provides appropriate support and incentives, THEN they will feel confident in implementing (and evaluating) interventions that involve changes to practice and professional roles, LEADING TO appropriate utilisation of health resources and a reduction in unplanned admissions. Again, support for this programme theory comes from studies of HCPs’ perspectives on the implementation of new services, generally based on specialist nurses working in the community as part of a MDT covering COPD or heart failure. 127,226

Description of putative mechanisms

Box 9 summarises key mechanisms to explain how the intervention is believed to work.

Contextual factors

Supporting evidence

Evidence on the effectiveness of community interventions (including hospital at home) comes from Cochrane and other systematic reviews. Cochrane reviews have assessed hospital at home interventions for admission avoidance233 and early discharge,234 and specifically for acute exacerbations of COPD. 235 Although the interventions were delivered in the patient’s home, staff delivering the service could be community based, hospital based or a mixture of the two. Of the three reviews, only the one covering acute exacerbations of COPD found a significant effect on reducing hospital re-admissions compared with hospital inpatient treatment. 235 It should be noted that Purdy et al. 3 treat hospital at home separately from community interventions.

Health Quality Ontario33 systematically reviewed studies of early follow-up (within 7 or 30 days) after hospital discharge for heart failure or COPD in reducing re-admissions. Follow-up could be conducted by telephone or by home visits, so the review scope was relatively broad. Early follow-up was associated with reduced all-cause re-admissions after adjusting for confounders, but the strength of evidence was classed as low to very low on the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system. A Cochrane review by Wong et al. 236 examined home care by outreach nurses for people with COPD. This review also failed to find a significant effect of the intervention on hospital admission or mortality. Significant heterogeneity was present for hospital admissions, possibly related to the presence of one outlying study with a large decrease in admissions whereas other studies showed an increase.

Other reviews identified by the mapping review focused on community-based programmes for hypertension but with a focus on Canada237 and on transitional care interventions, which are dealt with in the next section.

Overall, the systematic review evidence identified by our mapping review is not strongly supportive of the effectiveness of community interventions for reducing hospital admissions. Most of the evidence relates to heart failure or COPD. The definition of community interventions covers a mixed group of interventions and the limitations of the evidence base may reflect the inherent clinical heterogeneity of this group.

Potential enhancements of intervention/implementation

Study authors suggested a diverse range of possible enhancements without any clear themes emerging. Flood-Page et al. 225 emphasised the importance of links between primary and secondary care, whereas MacKenzie et al. 226 pointed to the need to consider urban versus rural settings, as GPs see their roles differently in the two settings. Participants also mentioned the need for effective planning and preparation before implementing a new community service. Glogowska et al. 127 stressed the importance of both the nurse’s ongoing relationship with the patient and their role in co-ordinating and ensuring continuity of care.

Summary

It is difficult to provide a formal definition of this group of miscellaneous interventions. Most examples of multiple interventions involve co-operation between different elements of the health-care system/workforce. Other interventions are generally simpler interventions not covered by any of the other groups of interventions in the review. Five UK studies were included, of which three involved multiple interventions (two for COPD and one for heart failure) and two were classified as ‘other interventions’. Two studies included explicit statements about how the intervention is believed to work, both relating to PT1 and PT2. The disparate nature of the interventions in this group makes it difficult to identify relevant systematic reviews and other sources of synthesised evidence. Contextual factors around implementation were relatively well reported for this group of studies, with differences in facilitation between interventions implemented across a whole area and those implemented at a smaller scale (e.g. a single hospital or practice).

Definition

It is difficult to provide a formal definition of this group of miscellaneous interventions. Most examples of multiple interventions involve co-operation between different elements of the health-care system/workforce, in some cases affecting all stages of the patient pathway from diagnosis to the end of life. 238 ‘Other’ interventions are generally simpler interventions not covered by any of the other groups of interventions in the review.

Intervention components

The key components of each included UK study are summarised in Table 16 based on the TIDieR-Lite checklist.

Number and type of UK studies identified

Five studies were included, of which three involved multiple interventions238–240 and two were classified as ‘other interventions’. 241,242 Two of the multiple intervention studies were for COPD and one was for heart failure. All three studies used an observational design and only one had a control group (see Appendix 4, Table 33).

The intervention reported by Roberts et al. 238 was an integrated COPD service covering the inner-city area of Salford. Ghosh et al. 240 evaluated a complex intervention involving telehealth, health coaching and a specialist nurse service, also in an urban setting (Leicester). The heart failure transitional care service evaluated by Williams et al. 239 involved discharge planning, patient education and follow-up. Patients received regular visits during their hospital stay and follow-up in nurse-led clinics or home visits. This study had a historical control group.

The two other interventions involved a consultant-led community-based service for COPD patients with frequent admissions and education, spirometry training and case-finding for primary care241 and a hospital in-reach service to optimise management and support early discharge to community services. 242

Operating programme theories

The overarching programme theories appear to operate to different extents in different interventions. PT1 (patient knowledge for self-management) was part of the interventions reported by Williams et al. ,239 Cope et al. 242 and Ghosh et al. 240 The studies by Cope et al. 242 and Ghosh et al. 240 reported increases in patient confidence to self-manage their condition, although based on a small sample, whereas the study by Williams et al. 239 assessed only patient satisfaction. As this was a short-term study of a transitional care service, it appears to have limited relevance to PT2 (patient satisfaction with non-secondary care provision). Other studies did not report data relevant to PT2.

Programme theory 3 (GP/primary care staff confidence) appears relevant to those studies in which enhanced education and training for HCPs was part of the intervention. 238,241 However, outcomes related to increased HCP confidence were not reported in these studies. PT4 (patient delay in seeking treatment) appears relevant to the work of Wilkinson et al. ,241 who identified a group of patients with frequent admissions who had become disengaged from primary care and were more likely to visit the emergency department. PT5 (wider health system support and incentives) appears to operate in those studies in which new services were implemented across a broad area with input from a range of relevant stakeholders. This is exemplified by the studies of Ghosh et al. ,240 Roberts et al. 238 and Wilkinson et al. 241

Description of putative mechanisms

Two studies included explicit statements about how the intervention is believed to work. Williams et al. 239 invoke the health belief model, arguing that support in the transition from hospital to community would increase patients’ self-efficacy and develop their confidence to make decisions about their health. Cope et al. 242 state that early assessment followed by supported discharge is thought to work through patients being more aware of the community services available to them and therefore having the confidence to access such services again in the future. Patients have a better understanding and ability to self-manage their respiratory condition. These mechanisms relate to PT1 and PT2. Other relevant factors identified from the included studies are summarised in Box 10.

Contextual factors

Supporting evidence

The disparate nature of the interventions in this group makes it difficult to identify relevant systematic reviews and other sources of synthesised evidence. For example, one of the included studies evaluated a transitional care intervention for heart failure patients. 239 Although there are several published systematic reviews of transitional care interventions,243–245 their scope is generally wider than that of Williams et al. ,239 making their relevance to implementation of this specific intervention uncertain.

Potential enhancements of intervention/implementation

In relation to community services for COPD, suggested enhancements were improved access to smoking cessation services, especially in deprived areas,238 and support for practice nurses in primary care. 241

For transitional care and early supported discharge, authors suggested support for 7-day working239,242 and agreed diagnostic criteria and treatment guidelines for heart failure between primary and secondary care. 239

Evidence

The clinical conditions selected as the focus for this report, namely cardiac and respiratory conditions, were purposively prioritised to represent areas with strong evidence of effectiveness in terms of the reduction of avoidable admissions. However, studies that demonstrated a positive effect had not necessarily demonstrated an effect in a UK context. Indeed, only a very small proportion of the studies identified by Purdy et al. 3 and adopted by us as a supposed robust evidence base had been conducted in a UK setting. Furthermore, the types of evidence available to explore implementation constitute a far less robust source of evidence than that for interventions.

Context

A typical limitation of the health services and delivery evidence base is its dependence on studies from the USA. This complicates interpretation of intervention studies, because the US-managed care system is so markedly different from the UK NHS, and this challenge is accentuated when examining implementation evidence. We have had to restrict implementation evidence used, at least to a large part, to insights from the UK. However, such a pragmatic decision should not be allowed to mask the fact that, even within the NHS, contexts may be markedly different in terms of culture, organisation and resources, so pilots or initiatives evaluated in one area may not readily transfer even to adjacent NHS constituencies.

Facilitation

The PARiHS framework emphasises the need for appropriate facilitation to increase the likelihood of a successful outcome. 17 The type of facilitation and the role and skills of the facilitator are determined by the specific needs of the organisation. Facilitators work with individuals and teams to enhance the implementation process. In the included studies, the most common mode of facilitation was training of staff. This can be challenging to achieve in an NHS environment alongside ongoing delivery of health services. Training was largely procedural, with little evidence of attempts to change the organisational culture in which preventable admissions take place. This deficiency may be addressed in the new models of care initiative,247 which targets widespread organisational change with preventable admissions as a shared priority.

Much of the evidence on facilitation related to the initiation of the intervention in the context of an experimental evaluation, with the expectation that additional resources for facilitation might be available in the trial setting, rather than in subsequent sustainability and spread. Despite extensive attempts to follow up citations from trials and other experimental studies, we found few examples of the implementation of interventions as trial results were rolled out more widely. Those accounts that do exist tend to focus on packages of care that have progressed a relatively long way down the diffusions of innovations curve, such as case management and cardiac rehabilitation. This probably reflects the focus on publishing generalisable results from empirical studies in the research literature rather than on examining implementation issues related to sustainability and spread.

Programme theory 1

Programme theory 1: people with chronic conditions are frequently admitted to hospital when hospital is not the optimal destination for them. People with chronic conditions may have symptoms that could be self-managed or anxieties that could be addressed by patient education or information.

We found evidence to support PT1. Challenges for people with chronic conditions related less to being ill-equipped for self-management and more to difficulties in situating their symptoms within the information with which they have been provided. People with chronic conditions are typically able to interpret their symptoms relative to only their own past experience or the observed experience of others, not against some absolute ‘trigger point’.

There was some evidence to suggest that being equipped with action plans or personalised self-management plans can help people with chronic conditions to assess whether or not external intervention is required. A similar challenge with patient education related to the need to locate available information within the context of what the patient already knows and what they perceive they need to know. There was some evidence to suggest that patients are not typically well equipped to process information on a just-in-case basis, particularly as this may result in a form of information overload. Again, the facility to access contextually sensitive information at the time it is required, whether this be available as a personalised action plan or as telephone advice, rather than being expected to absorb it within a short period of time and then recall it when needed, was positively indicated. The likelihood that a patient’s personal experience serves as a benchmark for their decisions about the need for future intervention indicates the potential value of using a trajectory-based approach to the identification of trigger points (e.g. ‘if you either experience these symptoms that you have previously experienced again or start to experience these specific new symptoms then you should . . .’).

Programme theory 2

Programme theory 2: people with chronic conditions lack knowledge about alternative health provision and therefore draw disproportionately on well-signposted channels, such as their GP or the emergency department.

Alternatively, patients perceive that presentation to an emergency department holds relative advantage (e.g. quality, ease of access and response) over GP-based or other primary or community care services (PT2). People with chronic conditions pressure HCPs to admit them to hospital.

We did not find substantive evidence to suggest that patients exert direct pressure on the decision for emergency admission primarily on the basis of relative advantage. On the contrary, we found that people with chronic conditions often resisted admission to an emergency department. Structural constraints, such as distance from home, distance from family, the alien (i.e. non-homely) hospital environment and the complexity and duration of the transport and admission process, may combine to make emergency admission an unattractive option. It therefore seems that concerns associated with anxiety and risk may constitute a more important driver. Hospitals are seen as a safe place that can offer security and reassurance. However, the presence of perceived, implicit or indirect pressure cannot be underplayed. HCPs may be influenced by personal factors in their relationship with a specific patient (‘soft’ factors) or by fear of litigation, and, ultimately, this could have an impact on the appropriateness of hospital referrals.

Programme theory 3

Programme theory 3: HCPs lack confidence in their own diagnoses or may lack confidence in or knowledge of alternative sources of health-care provision and so may refer people with chronic conditions or admit them directly to hospital. HCPs feel under pressure to admit people with chronic conditions directly to hospital (PT3).

Simmonds et al. 128 use the expression ‘decision flashpoints’ and highlight how clinical misdiagnosis of symptoms (typically breathlessness in the case of comorbid heart failure) can send the patient along a different disease pathway until they require emergency admission. Simmonds et al. 128 emphasise that inappropriate resolution of ‘decision flashpoints’ is perhaps more likely to happen when there is a lack of relational continuity with a GP.

Simmonds et al. 128 further report that ‘some GPs lacked knowledge of heart failure and specialist services, resulting in “mismanaged patients” ’. 128 Conversely, they report that GPs could also be ‘possessive of their patients’,128 reducing the likelihood of referral to specialist heart failure nurses.

Programme theory 4

Programme theory 4: people with chronic conditions use health services inappropriately, delaying their presentation to HCPs or hospital because of perceptions of the service either anticipated or based on the past experience of either themselves or others.

Programme theory 5

A weakness of the above programme theories is the extent to which they imply the availability of choice and rational decision-making in determining the decision to admit. This is addressed, at least to a certain extent, by the inclusion of PT5, namely that GPs and other HCPs are influenced by the wider context of the health-care system, and the availability or otherwise of support and incentives may influence their adoption of interventions and pathways designed to avoid preventable referrals and admissions to hospital (PT5). An exploration of organisational aspects of the management of heart failure concludes that:

Different service providers and professional groups had unintentionally co-created structures, systems and professional hierarchies that militated against the provision of seamless care.

Simmonds et al. 128

Such fragmentation and compartmentalisation not only presents barriers within particular episodes of care (e.g. interactions between GPs and specialist heart failure nurses or between GPs and medical specialists), but also undermines more whole-system-based interventions, such as patient education, which requires coherence and consistency across organisational boundaries and telemedicine, which logistically requires IT systems and procedures to be joined up. 166

Simmonds et al. 128 offer evidence that carers adopt an unofficial ‘ambassadorial’ role to help the patient navigate organisational fragmentation and unravel problems caused by poor communication and co-ordination. Carers grew in confidence and proficiency, subsequently being able to become more proactive in challenging the organisation and provision of care on behalf of the person they cared for. However, this added undesirable extra stress and responsibility for carers who often had their own health problems with which to contend.

Within a publicly funded health service such as the NHS, structural constraints and limited availability of alternatives may influence the ultimate decision as to what care is provided. Few resources can be accessed urgently (i.e. within 1 day), and those urgent-access resources that do exist have limited capacity. 249 Clinicians and patients may factor in anticipated system delays into their decision-making processes and develop compensatory strategies, for example if a ‘tipping point’ is likely to occur within a particular period. For some patients, admission to an acute hospital becomes the only option available if other services cannot be accessed sufficiently rapidly.

Resource-related constraints, such as the limited capacity of health and social services, and suboptimal communication between primary and secondary care clinicians may equally affect the clinical pathway. Initiatives that offer better coverage of clinical care, yet at the associated expense of a loss of continuity of care, may increase the likelihood of preventable admissions. For example, changes to the provision of out-of-hours services have led to concerns that deputising locums (with less familiarity with the patient’s medical or social history) might refer patients to another service or hospital more often than established counterparts. 250 Conversely, specialist nursing provision, seen by many as a potential way of reducing inappropriate admissions, may constitute an early and easy target for budgetary cuts when resources are limited. 249

Case management

The search located four theoretical/conceptual papers, which were published between 2001 and 2018 (Table 18). The majority of the frameworks and models focused on the health provider, although the earliest theories were targeted at an individual patient level and the most recent focused on a complex health system model. A total of seven models were identified in the four included papers.

A recent realist synthesis82 focusing on the case management of a different population, namely the frail elderly, nevertheless reveals two potentially useful frameworks/models. According to the conceptual framework of integrated care,253 the patient’s perception of their own capability, informed by the perception of informal caregivers, can lead them to accept the primary care services offered as an alternative to hospital admission and to believe that they are able to manage their condition at home. 82 This framework links closely to the self-efficacy beliefs of Bandura,99 which figure prominently in connection with other intervention types, such as self management and patient education. Indeed, these efficacy expectations are central to PT1 when a patient does not feel equipped with sufficient situational knowledge to manage their own decision on hospital admission. The converse of this is that self-efficacy, when coping with one’s chronic conditions, will have an impact on the extent to which a patient, and their caregiver, feels sufficiently equipped to continue to manage from home.

For the second theory, from the realist synthesis of case management of the frail elderly,82 the review team draws on Wagner et al. ’s222 CCM to identify perceived social support as a mediator for helping patients to stay at home. Several of our own programme theory components draw on perceived levels of support, available in primary and community care, as an important factor in the decision of whether or not to admit to hospital. These perceptions can be held by the patient, the informal caregiver or the health provider. Key for this intervention are the skills of the case manager, who should be facilitated by access to adequate training and sufficient time to fulfil the case management function. A related consideration is the extent to which patients associate their routine care with a local primary/community care community or the extent to which the acute hospital is viewed as the hub for their care. This latter consideration can be managed by engaging with the expectations of peers within the community and may be modified by the ongoing interactions with the case manager.

Patient education

Patient education is extremely well theorised and our review team felt that there would be little benefit to be gained by engaging with the totality of this theorisation. Instead, the search for theories, conceptual frameworks and models in connection with patient education focused on reviews or compendia covering multiple frameworks. We located three reviews, published between 2008 and 2017 (Table 19). All of the frameworks and models focused on the patient (e.g. by seeking to understand and/or explain patient behaviour in relation to the operation of patient education). We were surprised not to identify more transactional-based models regarding the mutual reciprocity of information exchange between the patient and the health provider.

Self-management

The search located six papers, published between 1998 and 2016 (Table 20). The majority of the frameworks and models focused on the patient (e.g. by seeking to understand and/or explain patient behaviour in relation to the intervention).

The earliest paper to be published (in 1998)260 covered a wide range of topics under the broad heading of ‘social cognitive theory’, suggesting that theoretical treatment of self-management has a long history. The frameworks used in five further studies were diverse (see Table 20). These studies involved patients with chronic health conditions, diabetes or asthma. Heisler et al. 261 described a conceptual model that focused on how provider behaviour can influence self-management via its influence on self-efficacy. Williams et al. 258 used a self-determination theory based on the concepts of autonomous or controlled motivation and perceived competence or incompetence in a study of people with type 2 diabetes. The theory was useful for relating changes in self-management behaviour to changes in glycaemic control. The individual and family self-management theory described by Ryan and Sawin257 emphasises the role of dyads within a family and the family as a whole, as well as individuals, in supporting self-management.

Finally, in a recent empirical study, Hoskins et al. 143 evaluated a self-management intervention for asthma based on multiple theoretical models. The study suggested that the theory-based intervention was promising but the need for additional time in appointments was a potential barrier to implementation.

Telehealthcare

Telehealthcare, particularly telemonitoring, is a well-theorised area requiring some selectivity when mapping interventions to underpinning mid-range theories. The following list is, therefore, an indicative summary that seeks to extend across different foci and disciplines. A recent review of numerous theoretical frameworks extends this coverage further. 262

The search located five papers, published between 2010 and 2016 (Table 21). The majority of the frameworks and models focused either on the patient (e.g. by seeking to understand and/or explain patient behaviour in relation to the intervention) or on health providers whose attitudes to adoption and acceptance of technology are critical to the intervention success.

The earliest paper to be published was published in 2010, meaning that more long-standing theories of telehealthcare (particularly telemonitoring) have not been picked up by this report. Most of these frameworks are used in providing a theoretical basis for intervention development, with a recent review demonstrating a higher prevalence of theory-informed interventions than in many comparable areas of health care.

Cardiac rehabilitation

The search located eight papers, published between 1994 and 2010 (Table 22). The majority of the frameworks and models focused on the patient (e.g. by seeking to understand and/or explain patient behaviour in relation to the intervention).

The earliest paper to be published (in 1994)273 used the Bandura self-efficacy theory99 as a basis for predicting patient behaviour and planning interventions, including methods of increasing self-efficacy and, hence, optimising outcomes. The frameworks used in three further studies, although named differently, were similar in structure, consisting of predisposing, enabling and reinforcing/need factors. 272,275,276 These studies involved diverse populations, including white and African American women and rural US populations.

Blanchard et al. 271 found that the theory of planned behaviour was a useful framework for understanding adherence to cardiac rehabilitation, with perceived behavioural control making the largest contribution to adherence. Lau-Walker274 combined illness representation and self-efficacy theories to produce an interactive care model as a framework for research into intervention design for rehabilitation.

Other conceptual models focused on the wider health system. Kitson et al. 278 used cardiac rehabilitation as an exemplar to test an early version of the PARiHS framework for research implementation. Finally, Goud et al. 269 used a framework developed by Cabana et al. 270 to study HCPs’ adherence to guidelines related to cardiac rehabilitation. The framework includes both internal and external barriers to guideline implementation.

Pulmonary rehabilitation

We identified two mid-range theory papers for pulmonary rehabilitation,212,213 both of which used social identity theory (Table 23).

Specialist clinics

We found no relevant mid-range theory papers for specialist clinics.

Community interventions

Only two relevant papers were located for this group of interventions (Table 24). This may reflect a paucity of literature or the challenges of searching in this area in a limited time frame, or both.

Falk Rafael279 used Watson’s theory of human caring280 to investigate community nursing and the importance of the nurse–patient relationship. The normalisation process model described by May281 can be applied to any complex intervention in health-care settings. The paper includes the example of nurse-led home telecare for COPD, which is described in terms of the theory’s constructs as moderate for interaction, high for relational integration and skill-set workability, and moderate for contextual integration between primary and secondary care.

Multiple/other interventions

Transitional care is an example of a multiple-component intervention not covered by any of the other categories. Wee and Vrijhoef282 described a conceptual framework for evaluating the development, implementation and performance of transitional care interventions (see Table 24). The model seeks to explain why and how the interventions should work to minimise re-admissions for patients admitted to hospital who need to be transferred between different settings or levels of care during their admission.

Clinical inertia and admission as ‘the default position’

Engaging with the included studies at the level of mechanisms, rather than simply at an intervention level, revealed the importance of the concept of ‘clinical inertia’ in relation to the phenomenon of emergency admissions, especially re-admissions. Clinical inertia is defined as the failure to establish appropriate targets and escalate treatment to achieve treatment goals. This concept appears to have gained initial traction in the context of diabetes,283 which, along with the conditions covered by this review, figures among the most prominent of ACSCs. It has subsequently been promoted in evidence-based health care as an explanation for lack of clinical uptake of clinical decision tools rooted in a lack of confidence in their content and concerns about potential disruption to workflows. 284 Could a similar lack of confidence, this time in primary care provision, and concerns about disruptive change offer one possible explanation for clinicians’ apparent ‘reluctance’ to explore alternatives to emergency admission? Elwyn et al. 284 subsequently linked clinician inertia to the related phenomenon of ‘organisational inertia’, which seems to be at play in this particular arena.

It is interesting to observe that other comparable forms of inertia, attributable to other key protagonists in the decision to admit, are evoked in other included studies. In their Cochrane review, Glynn et al. 285 seek to explain why frequent contact with HCPs does not guarantee better blood pressure control by drawing on the observations of some commentators to suggest that this may be attributable to ineffective management and inadequate practice organisation (i.e. ‘clinical inertia’). When monitoring is shifted from HCPs to patients themselves, Jones et al. 157 remark on a patient phenomenon comparable to clinicians’ clinical inertia, with patients being reluctant to increase medication further in the face of borderline readings, a form of participant inertia.

Generically, the phenomenon of ‘clinical inertia’ is addressed by encouraging patients and physicians to communicate more effectively and respond appropriately to changes in patient circumstances. It may appear paradoxical to be considering clinical inertia in the context of inappropriate escalation of treatments, namely in referring a patient from primary to secondary care. However, failure to establish appropriate targets and to respond appropriately to changes in patient circumstances are equally at the heart of what we observe for inappropriate emergency admissions. We have identified a further variant of clinical inertia of particular relevance in the context of emergency admissions, namely as inertia by remaining within admission as the ‘default position’. As a health improvement document286 states, the culture of an organisation can often be to admit patients as a default position, which can result in admission rates being higher than they need to be.

Consequently, County Durham and Darlington NHS Foundation Trust has implemented measures to minimise admission rates to accident and emergency (A&E), with particular focus on the interaction between the A&E department and the acute medical unit (AMU). These measures include:

• ensuring that all admissions are discussed with a senior decision-maker prior to admission, including out-of-hours admissions

• developing processes to allow direct referrals to the AMU rather than requiring admittance to A&E as a preliminary, unnecessary step

• ensuring a regular, ongoing two-way dialogue between A&E and AMU staff to ensure that referral/admittance decisions are reviewed and challenged

• AMU consultants regularly coming into the A&E department to aid assessment and ‘pull’ patients from A&E when appropriate.

Early access to senior decision-makers and appropriate selection processes287 feature prominently in many of the intervention types identified in this report.

Supplementary ways of addressing clinical inertia are identified in the literature288 (Box 11) and may serve as a basis for developing new or improved interventions to address this ‘default’ variant.

According to Phillips et al. ,288 clinical inertia is due to at least three problems: overestimation of care provided; use of ‘soft’ reasons to avoid intensification of therapy; and lack of education, training and practice organisation aimed at achieving therapeutic goals. For that variant of clinical inertia whereby HCPs adopt emergency admission as a default position, there is a related pathology, but one that may relate to the converse of the original ‘clinical inertia’. For example, the primary health-care professional may underestimate the care that is currently being provided (or that might be available) in primary care, or overestimate the additional care that may be provided in secondary care. In addition, the HCP may use ‘soft’ reasons to justify and initiate intensified intervention in an emergency department. Finally, they may require recalibration, through education and training, with regard to realistic expectations from intervention. Phillips et al. 288 locate the overcoming of clinical inertia in the context of the benefits of treating to therapeutic targets and the need to structure routine practice to facilitate effective management of disorders for which resolution of patient symptoms is not a sufficient indicator of optimal care. Such systems require sophisticated and reliable mechanisms of feedback and performance evaluation.

Although the emphasis on education of GPs, supported by evidence-based guidelines, clearly offers a potential to overcome a lack of knowledge of the condition and the available services to which GPs might refer, it is worthy to note that a qualitative synthesis of GPs’ barriers to managing heart failure in primary care289 reports ongoing difficulties with information overload. It appears that, just as patients do not experience an absence of information, so much as it not being tailored and personalised to their own specific requirements, GPs do not require yet more educational strategies around greater numbers of evidence-based guidelines so much as the facility to have guidelines tailored to their own local context.

In the context of emergency department-based interventions, the concept of ‘appropriate deliberate clinical inertia’ has been framed in recognition of the need to resist inappropriate or unnecessary interventions. 290 It appears that this concept may be appropriated in primary care to recognise situations in people with chronic conditions that do not result, for example, in any change in the dose or nature of medications. Clear agreement between patients and HCPs that emergency admission will not resolve or in any way modify their symptoms or treatment can act as a sound basis for ‘shared decision-making to improve patient care with the use of clinical judgement’. 290

Candidacy and ‘deservedness’

Another important concept relates to ‘candidacy’, which has previously been explored in relation to access to and utilisation of health care in general. The concept suggests that an individual’s identification of his or her ‘candidacy’ for health services is structurally, culturally, organisationally and professionally constructed. 291 An unintended consequence of several interventions identified in this report is the ‘legitimisation’ of use of services, leading to a possible increase, rather than decrease, in the number of inappropriate admissions. Exactly how this works may vary; initial use of emergency services, either appropriately or inappropriately, may be seen as opening up a pathway of ‘deservedness’ for future entitlement. The initial episode may be seen as legitimised as a ‘one-off’ or it may open the way for ‘similar’ utilisation in the future. More constructively, appropriate use, as confirmed or ratified by HCPs during admission or subsequently during follow-up, may help a patient to ‘benchmark’ their personalised symptoms and set thresholds for external involvement. The complication here is that patients with chronic conditions typically have multiple symptoms and so the ‘benchmark’ directly relates only to those specific symptoms that precipitated admission. This may result in empirical ‘trial and error’ around emergency admissions for distinct yet potentially related symptoms.

Continuity and personalisation of care

Continuity of care is revealed as an important influence on admissions for long-term conditions. 91 The distinction between continuity of relationship (a continuous caring relationship with clinicians) and continuity of management (all aspects of integration, co-ordination and sharing of information) was helpful when completing this analysis, recognising that both mechanisms are revealed as important in the context of preventable admissions. Systematic reviews have found that relational continuity was associated with lower rates of emergency department attendance and hospital admissions. 3,109 Other likely benefits include improved outcomes for patients, including in medication adherence. 128,292

However, Simmonds et al. 128 also illustrate how relational continuity may be unhelpful, for example when a GP is reluctant to refer a patient to a specialist nurse either because they wanted to hold on to management of the patient themselves or because they had little knowledge of the available services. This results in an unhelpful ‘default’ position and can lead to yet another form of ‘clinical inertia’.

Sick dependency role and reliance on primary health-care professionals

Interventions introduced to prevent emergency admissions can have a disruptive effect on the sickness role of the affected patients. For example, the Whole System Demonstrator project165 reported that patients believed that dependence on technology would reaffirm their sickness role and lack of independence. It would also be a visible reminder of their own condition, with the potential to cause increased anxiety.

HCP-mediated intervention roles, such as case manager, also hold a disruptive potential. HCPs were concerned that such roles would increase patient dependence rather than facilitate self-management. In particular, it is challenging to manage the balance between offering improved access to support and reassurance when needed and creating a dependence on the HCP that subverts self-management. This links back to the earlier consideration of candidacy: provision of a specialist service and a ‘passport’ to entitlement of follow-up services when necessary and appropriate may be seen as legitimising contact with secondary care and lead to increased, rather than reduced, utilisation.

In other cases, the technology was seen, typically by carers, as a way of seeking to replace care already provided through face-to-face interaction with an inferior care package delivered by technology. The carer saw the visible dependence of the patient on the personalised health care as appropriate recognition of their illness state.

Finally, the addition of specialist intervention alongside existing provision can lead to perceptions that levels of health care are excessive or superfluous. 165 For example, when a patient is already in regular contact with HCPs, perhaps for comorbid conditions, there may appear to be little need for additional telemonitoring. Such views may reflect actual redundancy and overlap, due to poor integration of specialist services with usual care, or may simply reflect the patient’s perception (i.e. where similar-looking interventions have clearly distinct clinical functions), resulting in poor compliance. Experimental interventions, from pilot studies through to full trials, are more likely to result in such perceived duplication than established services but, nevertheless, the integration of specialist and general care, or between different forms of specialist care for comorbid conditions, is an issue of which those implementing specialist interventions should be aware.

The role of integrated care mid-range theories

The fundamental move towards transformation of current models of primary care by means of implementing proactive integrated care hints at the future potential of integrated care approaches to analysing and resolving the systemic issues revealed by the phenomenon of preventable admissions. Attempts to understand the effects of integrated primary care approaches and underlying mechanisms in relation to other populations, such as frail community-living older adults, may translate to the condition-specific contexts of COPD and heart failure that are the focus of this report. For example, a theory-based evaluation called ‘Finding and Follow-up of Frail older persons’293 offers a theoretical model that incorporates interrelated elements of integrated primary care approaches: (1) proactive case-finding, (2) case management, (3) medication review, (4) self-management support and (5) working in multidisciplinary care teams. Furthermore, this model recognises the importance of cognitive and behavioural components of both HCPs and patients, targeted by so many of the individual mid-range theories identified for each intervention type. Seemingly intractable problems may well be amenable to analytical lenses developed in alternative arenas.

Mapping review

The mapping review of studies of interventions to reduce preventable admissions published since 2010 identified 569 publications, predominantly related to heart failure or COPD. The interventions identified by Purdy et al. 3 as having the best evidence of effectiveness (or no effect) were well represented in the map. The largest group of studies originated from the USA. The included studies from the UK showed a similar distribution of studies by intervention and population to that of the map as a whole, but there was evidence of some country-specific features, such as the prominence of studies of telehealth. The findings of the mapping review were helpful in informing the sampling frame for the subsequent realist synthesis.

The mapping review and subsequent analysis of UK studies for each intervention (see Chapter 5) revealed that, in many cases, a strong international evidence base for effectiveness of an intervention existed alongside limited evidence specific to the UK context. Cardiac rehabilitation (see Chapter 4, Cardiac rehabilitation) perhaps illustrated this phenomenon most clearly. We often found limited information to help understand how particular interventions have been implemented in the UK NHS and which approaches to implementation work best in NHS contexts. Information on the role of leadership and approaches to facilitating the implementation of interventions was often lacking.

Implementation framework

Within the PARiHS framework, successful implementation is represented as a function of the nature and type of evidence (to be examined from the mapping review), the qualities of the context in which the evidence is being introduced and the way the process is facilitated17 (to be extracted from included UK studies, both quantitative and qualitative). We found that interventions with strong evidence of effectiveness overall had not necessarily demonstrated effectiveness in UK settings; that the largest proportion of the evidence came from the USA, where the context for delivery of health care is very different from that of the UK; and that facilitation of the implementation of interventions was often not reported or inadequately reported in UK studies, which generally focused mainly on effectiveness or qualitative evidence of the patient and HCP experience of service delivery.

Descriptive framework

The TIDieR-Lite framework provided a useful descriptive framework for recording key elements of the interventions and their delivery. Many of the included interventions were highly diverse in the way they were delivered, the main exceptions being cardiac and pulmonary rehabilitation. There was also considerable overlap between interventions in terms of their key components. The role of specialist nurses in providing continuity of care and links between primary and secondary care was highlighted in multiple studies.

Programme theories

We identified and tested five programme theories (using the sequence CMO) to explain how the interventions might work (see Box 1). The programme theories, expressed as scenarios, were refined and endorsed by our PPI group. We found evidence to support PT1, which suggests that hospital admissions could be reduced by optimal self-management. Considering PT2, we did not find substantive evidence to suggest that patients may seek hospital admission primarily on the basis of relative advantage. It seems that concerns associated with anxiety and risk may constitute a more important driver, hospitals being seen as safe places that can offer security and reassurance. However, the presence of perceived, implicit or indirect pressure cannot be ruled out. PT3 relates to clinicians’ confidence in their own diagnoses and ability to refer appropriately to services that might avoid admission. In the context of cardiovascular and respiratory disease, this is relevant to patients with symptoms, such as breathlessness, that could result from various underlying causes.

Direct evidence for PT4 (admissions resulting from patient delay in seeking treatment) was limited in our sample of studies. One study128 noted that a lack of timely and accurate diagnosis contributed to exacerbations of heart failure. Finally, PT5 (influence of the broader health system context) addressed the limitations on rational decision-making around hospital admissions. This was reflected in our studies. For example, heart failure care delivered across multiple services, confusion about eligibility for specialist care and relational/managerial discontinuity of care increased the likelihood of suboptimal management and unplanned admissions. 128

Overall, we believe that the programme theories considered in this realist synthesis are valuable for understanding why unplanned and avoidable admissions occur and the facilitators of and barriers to reducing them.

Mid-range theories

We found numerous examples (both descriptive and empirical studies) of mid-range theories relevant to the interventions under review. The largest group focused on the patient, for example factors influencing adherence to recommended interventions, but theories related to HCPs’ behaviour and the overall health system were also located. Some theories were cited in relation to several interventions (e.g. Bandura’s self-efficacy theory). However, it is unclear whether the prevalence of efficacy/coping theories reflects their greater utility in this context or simply their higher profile. Nevertheless, self-efficacy was revealed as a key component to PT1, not only determining the level of comfort that a patient had with coping with their own condition but also, more importantly, how well equipped the patient feels they are to be able to manage their own exacerbations without resorting inappropriately to primary care health provision or to admission to an acute hospital.

Overarching theories

The overarching theories discussed in Chapter 5 may be considered as more exploratory than the programme and mid-range theories. In general, these theories may help to understand the underlying mechanisms at the level of the patient and HCP in the presence and absence of interventions designed to avoid admissions.

The double evidence mismatch

In theory, seeking UK-focused implementation evidence to accompany interventions that had previously been demonstrated as effective in preventing inappropriate hospital admissions for cardiorespiratory conditions appears both a coherent and an easily manageable literature review task. In practice, however, we encountered a logic problem that we have labelled the ‘double evidence mismatch’. First, the 2010 review by Purdy1 demonstrated the average effectiveness of the candidate interventions based on a comprehensive sample of international studies, within which UK studies constituted a small and largely insignificant part. The quest for implementation evidence might therefore largely focus on UK studies that had a small effect, negative effect or a statistically insignificant effect. Second, the fact that our review team focused on studies published since 2010 means that interventions being implemented during this period may not resemble the interventions being trialled during evaluation of the original intervention. Indeed, this situation was compounded by the profound shortage of implementation evidence for UK initiatives between 2010 and 2018. Qualitative evidence constituted a large part of our recent evidence base and may not necessarily be intervention focused, nor linked to relevant trials. This type of evidence ‘mismatch’ is considered rare; we have previously encountered such a mismatch only when trying to map international RCTs from a Cochrane review against a qualitative research evidence base restricted to the UK and similar health systems. We therefore make the methodological recommendation that the scope of intervention and implementation evidence, or indeed quantitative and qualitative evidence, seeks to be as coterminous as possible so as not to artificially constrain the evidence base. However, we also acknowledge that, by its nature, implementation evidence may need to be more current than the original effectiveness studies.

Value of the PARiHS framework

We were unable to exploit the full value of the PARiHS framework, not through any limitations of the framework itself but because of constraints in the reporting of the individual interventions. Limitations encountered tended to fall into three categories:

1. Published reports did not articulate implementation issues in terms of evidence, context and facilitation or did not cover the range of aspects of these dimensions included in PARiHS (e.g. ‘evidence’ is not limited to research evidence).

2. Published reports covered one or more of these issues but reports were not structured in a way in which such data were straightforward to extract.

3. Published reports focused on implementation at a level that excluded some of the important PARiHS concepts. For example, only a study specifically on organisational aspects included data on organisational culture, whereas leadership issues were almost entirely absent.

This confirms observations previously made in relation to frameworks (i.e. that frameworks derived for use in primary research studies may not represent a good match for the level of granularity sustained by their use in synthesis activities, with syntheses typically covering issues at a broader level of detail). The version of PARiHS that we chose was not the most recent but we believe that limitations in the reporting of implementation in the included studies mean that little would have been gained by using a more refined version of the framework.

We had reviewed a range of candidate frameworks before selecting the PARiHS framework as our eventual choice. Analysis of included literature during the subsequent realist process revealed use of alternative frameworks,98 specifically when exploring implementation issues, notably May et al. s’294 normalisation process model295 and a framework of barriers to and facilitators of quality improvement projects from a systematic review led by Kaplan et al. 296 It is presently unclear whether or not these frameworks would share similar limitations with regard to the granularity of synthesised data or, indeed, if selection of these alternative evaluation ‘lenses’ would have yielded more insightful observations.

Value of the TIDieR-Lite template

In contrast, the TIDieR-Lite framework proved particularly useful both as a means of exploring variation within intervention groups and when seeking to delineate between interventions. It became apparent that several interventions labelled in their own right were also included within other interventions; for example, patient education is a key component in self-management. Similarly, self-management is a key function of much telehealthcare. Specialist clinics may hold a primary function in providing patient education and supporting self-management. Use of the framework therefore revealed the difficulties in attributing an effect to particular interventions and in isolating which components are most essential in delivering an effect. The limitations of the classification of interventions further confirmed the value of moving from an intervention-based approach towards the realist synthesis based on mechanisms. Phenomena such as clinical inertia and the default position, the effects, intended and unintended, of candidacy and the pivotal role of continuity and personalisation of care transcend the seven intervention types and therefore offer more transferable messages.

Search strategy

1. ambulatory care sensitive condition$.ab,ti. (399)

2. “Ambulatory care sensitive conditions”.kw. (37)

3. ACSC.ab,ti. (210)

4. primary care sensitive condition$.ab,ti. (39)

5. PCSC.ab,ti. (63)

6. exp Angina Pectoris/ (44,306)

7. angina.ab,ti. (51,535)

8. heart failure/or heart failure, diastolic/or heart failure, systolic/ (108,637)

9. ((heart or cardiac or myocardial) and fail$).ab,ti. (207,554)

10. Hypertension/ (227,011)

11. hypertension.ab,ti. (347,343)

12. Asthma/ (118,479)

13. asthma$.ab,ti. (143,461)

14. Pulmonary Disease, Chronic Obstructive/ (32,875)

15. (copd or “chronic obstructive pulmonary disease$”).ab,ti. (52,948)

16. Lung Diseases, Obstructive/ (18,640)

17. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 (903,975)

18. Case Management/ (9789)

19. case management.ab,ti. (9810)

20. “specialist clinic$”.ab,ti. (1226)

21. “Drug Utilization Review”/ (3572)

22. ((medication or drug) adj3 review$).ab,ti. (7588)

23. self care/or self administration/ (41,630)

24. self-care.ab,ti. (14,579)

25. self-management.ab,ti. (14,098)

26. pulmonary rehabilitation.ab,ti. (2824)

27. Influenza Vaccines/ (20,563)

28. influenza vaccin$.ab,ti. (15,110)

29. flu vaccin$.ab,ti. (1062)

30. Practice Guidelines as Topic/ (104,685)

31. care pathway$.ab,ti. (3000)

32. care guideline$.ab,ti. (1663)

33. Community Health Services/ (30,426)

34. community intervention$.tw. (1653)

35. House Calls/ (3093)

36. home visit$.ab,ti. (7640)

37. Patient Education as Topic/ (81,677)

38. patient education.ab,ti. (15,166)

39. cardiac rehabilitation/or exercise therapy/ (36,370)

40. exercise based cardiac rehabilitation.ab,ti. (169)

41. Telemedicine/ (17,189)

42. telemedicine.ab,ti. (8638)

43. or/18-42 (381,151)

44. 17 and 43 (31,399)

45. Patient Admission/ (22,420)

46. (prevent$ or avoid$ or reduc$ or unplanned or unnecessary or unscheduled).ab,ti. (4,125,882)

47. 45 and 46 (4775)

48. ((prevent$ or avoid$ or reduc$ or unplanned or unnecessary or unscheduled) adj3 admission$).ab,ti. (4908)

49. hospitalization/or patient readmission/ (104,518)

50. 46 and 49 (25,222)

51. 47 or 48 or 50 (32,244)

52. (implementation or implementing).ab,ti. (230,025)

53. (dissemination or disseminating).ab,ti. (52,450)

54. (research adj2 integration).ab,ti. (589)

55. (transfer$ adj2 knowledge).ab,ti. (2249)

56. barrier$.ab,ti. (237,473)

57. facilitator$.ab,ti. (19,882)

58. sustainability.ab,ti. (15,412)

59. ((change or changing) adj (behavio?r or practice)).ab,ti. (2509)

60. (research adj2 utili?ation).ab,ti. (1214)

61. “research into practice”.ab,ti. (783)

62. “knowledge to action”.ab,ti. (559)

63. 52 or 53 or 54 or 55 or 56 or 57 or 58 or 59 or 60 or 61 or 62 (529,378)

64. 51 or 63 (559,087)

65. 44 and 64 (3868)

66. limit 65 to (english language and yr = “2010 -Current”) (1943)