The burden of disease worldwide is shifting to long-term conditions.1,2 Although advances have been made in effective service delivery, major challenges remain, namely projected increases in populations aged ≥ 65 years, the increases in demand associated with an ageing population and government pressure for major efficiency savings.
Current services are organised around single conditions, but many people have more than one (multimorbidity), which means that care is often fragmented and unresponsive to needs. Patient and policy consultation around care for long-term conditions has repeatedly emphasised the need for ‘integration’.3–5
What is integrated care?
There is a significant body of literature on integration, but a lack of consensus around definition; one review found 175 definitions.6 A number of different perspectives are possible on the meaning of integration, including managerial, health systems, social science and patient perspectives.7 The British Medical Association8 has highlighted that integration is a nebulous term, associated with wide-ranging definitions and processes. Analysts have distinguished between different dimensions of integration:9
- Types of integration – functional (key support and functions, i.e. human resources and financial management), organisational (contracting or strategic alliances between different organisations), professional (joint working, alliance and strategic contracting between professionals) and clinical (co-ordination of patient care services).
- Breadth of integration – this includes both vertical (bringing together organisations at different hierarchical levels) and horizontal (bringing together organisations that are on the same working level) integration.
- Degree of integration – full integration or more limited collaboration of services, working practices or organisations.
- Process of integration – this includes structural (alignments of tasks, functions and activities), cultural (values, norms and working practices) and social (the strengthening of social relationships between individuals) integration.
Models of implementing integration are also diverse. Health and social care systems are complex, with multiple providers and different levels of demand on the system, and so integration is likely to be equally variable.9 A review referred to three different models of integration:6
- System level – the focus here is on organisational change, whereby leadership plays a pivotal role in performance.
- Programme or service level – the emphasis here is to try to improve the patient outcomes by providing more co-ordinated care.
- Progressive/sequential models – integration is not a specific goal but is a means to try to improve health-care performance in general.
Some partners have adopted a person-centred definition of integrated care, focusing on the ways in which care is experienced by patients.5 This definition is supported by a number of ‘I statements’, which set out what integrated care should feel like to those in receipt of it. It is suggested that delivering care in this way will fulfil a number of goals (e.g. people feeling more confident to manage their conditions, improved sharing of decisions and relationships, and better sharing of information with the patient and among different services), which will in turn lead to improved outcomes (such as fewer admissions and, crucially, lower costs). However, it has been suggested that a person-centred model of integration might be better achieved through policy innovations such as personal health budgets and direct payments (allowing individuals to join up services in ways that make sense to them), rather than organisational and professional integration.10 Such a conception has been supported by recent qualitative work within integrated care pilots (ICPs).11
What is the review evidence for the benefits for ‘integrated care’?
There have been a variety of reviews and syntheses around the effectiveness and cost-effectiveness of integrated care. A ‘stocktake’ of the integrated care literature in 2009 found discussions to be dominated by potential benefits, with a lack of clarity over definitions and standardised outcomes.6 Although the scope of the literature has improved since that time, there is still a lack of clarity over the main findings in this area.
A metareview12 (or ‘review of reviews’) that included 27 separate reviews explored integrated care for adults across a range of long-term conditions. The authors coded 10 key principles of integration and reported a range of positive outcomes across the reviews, including in relation to hospital admissions (in heart failure and diabetes mellitus), adherence to guidelines [diabetes mellitus, chronic obstructive pulmonary disease (COPD) and asthma] and quality of life (diabetes mellitus). Reductions in costs were far less frequently reported. Another review13 of integrated delivery systems found 25 reports, with the majority showing an increase in quality of care associated with integration. Again, there was more limited evidence showing reductions in utilisation associated with integration. A recent ‘review of reviews’14 on integrated care for chronic diseases synthesised data from 50 reviews, which included a wide range of interventions (e.g. case management, variants of the chronic care model,15 multidisciplinary teams and self-management). As with other reviews, there was evidence of positive impacts in many outcomes, but results were not consistent, and the authors again highlighted the gap between the importance of the concept of integration in health policy and the strength of the evidence concerning its benefits.
Although comprehensive, these very broad reviews necessarily include a very wide range of patients and interventions and, therefore, can lack precision. Other reviews in the literature have had a more restricted scope in terms of interventions, populations and outcomes, providing greater specificity over outcomes. A review of integration at the primary–secondary care interface found 10 studies that demonstrated some benefits in terms of process of care (care delivery, disease control), but these generally did not extend to clinical outcomes and were achieved at some increase in costs.16 A review of co-ordinated and integrated care for the frail elderly found nine studies, with a slim majority reporting improved outcomes and reduced health-care utilisation, but with few data on the effects on caregivers.17
Case management is a popular method of integration, and a review18 of case management for older people found that the majority of trials showed no reduction in admission rates compared with usual care. A review19 of case management for at-risk patients in primary care reported 36 studies, but the only consistent benefits were in terms of patient satisfaction, with no demonstrable benefit in utilisation, costs or mortality. In contrast, a review20 of hospital-initiated case management for heart failure reported reductions in readmissions and length of stay, although those benefits did not translate to reduced costs. Interventions initiated from the community were less prevalent and showed less evidence of benefit.
A number of previous reviews have suggested that the economic benefits of integrated care are less consistently demonstrated than impacts on the process or quality of care. A review21 restricted to the economic impacts of integrated care identified 19 studies of relevance. As well as identifying a lack of clarity about definitions, the evidence was mixed, with some positive findings; generally, however, the evidence was characterised as ‘weak’. A review22 focused on integrating funding for health and social care found 38 studies. Health outcomes were frequently assessed, but evidence of benefits were limited and only a minority of studies found reduced secondary care costs.
The evidence is clearly mixed.23 There is a significant amount of evidence, such that a number of studies have used a metareview method, which is an efficient instrument for summarising large numbers of data, but not a very precise method of quantifying gain or assessing patterns in the data, or identifying fruitful approaches to integration. Interpreting the reviews is a challenge because of the complexity of the concept of integrated care and the different scope of the reviews, which is clearly demonstrated in the different numbers of studies included in each review. However, the overall impression is that benefits are inconsistent and most regularly associated with process measures (e.g. quality of care). When impacts on admissions are reported, they are most likely to be related to certain conditions (such as heart failure), rather than demonstrated across broader groups of long-term conditions. Reductions in utilisation may not translate to reduced costs, which may reflect the fact that integrated care is associated with its own costs; benefits of reduced utilisation in one part of the system may be lost when other costs are taken into account.24 Cost savings may require radical changes such as closing hospital beds,25 which may be unpopular and difficult to implement.23
Recent empirical evaluations in the UK
The previous section has outlined reviews of the effects of integrated care and highlighted the inconsistency in the evidence. The reviews have been international in scope. Although that brings major benefits in terms of the size and scope of the evidence, it does lead to additional complications in interpretation. Integrated care may have different meanings in different health systems, and the comparator conditions may also vary widely. It is generally accepted that context is an important moderator of the effects of complex interventions,26–28 and the context in which integrated care is introduced may also be very different.25 This section has a focus on empirical evaluations conducted in the UK.
The Evercare evaluation29 explored the case management of older people at high risk of emergency hospital admission. Although not a formal integrated care intervention, it shares a number of features in terms of the eligible population and the nature of the intervention. Evaluation showed no effects on admissions or other outcomes, although the service was popular with patients and carers.30
The Partnerships for Older People Projects involved a wide range of community- and hospital-facing services, with a significant focus on prevention. Evaluation using data from the British Household Panel Survey suggested some improvements in quality of life, although the comparator was not particularly strong in methodological terms. Although overall analyses31 suggested that the investment led to savings, more detailed analyses32 of a subgroup of services found no evidence of reductions in hospital admissions, and even suggested some increases.
An early pilot scheme in England33 involved the establishment of 16 ICPs. It should be noted that although these were all introduced into a single health-care system, the pilots did vary, being based on local circumstances in which the care included in the ‘integration’ project was dependent on the local context. Overall, the evaluation found that there was an increase in emergency admissions in the pilot areas and there was mixed evidence about whether or not the ICPs were able to reduce costs. Among the 16 ICPs, case management was perceived to be the best option for reducing secondary care costs (a net reduction in combined inpatient and outpatient costs were reported). Such comparisons lack the rigour of randomisation. The findings were also difficult to interpret as the key outcomes that the services were trying to change (emergency admissions) showed increases in activity, whereas reductions occurred in untargeted elective services. Assessments of patients were also conducted as part of the evaluation of the ICPs. Patients reported that they were more likely to be told that they had a care plan, to feel clear about follow-up arrangements and to know whom to contact, and were less likely to report problems with medication. All of these are relevant outcomes of an integrated care initiative. However, somewhat surprisingly, they also reported being less likely to see the health professional of their choice, being less involved in decisions about their care and being less likely to report that their preferences had been taken into account. Again, all these are relevant outcomes for a person-centred integrated care service. The fact that patients reported reductions in some patient experience measures and improvements in others highlights the difficulties of improving outcomes in this area.33
The North West London Integrated Care Pilot was a large-scale programme that had an initial focus on people with diabetes mellitus and patients aged > 75 years. The intervention involved information technology to support case finding and multidisciplinary groups (MDGs) to deliver care planning. Although implementation was generally successful (albeit somewhat delayed) and there were some impacts on process of care (including rates of care planning), a matched controls analysis of effects showed no impact on emergency admissions, although the analysis was judged to be preliminary.34,35
There was also a call for ‘ambitious and visionary’ local areas to become integrated care pioneers, with 14 sites starting in one wave in 2013 and another 11 sites starting in a second wave in 2015.36 Pioneers were tasked with the conventional outcomes of integration initiatives (improved patient experience, outcomes and financial efficiency), with expert support and some very limited additional funding. Early results from the pioneers (largely on the basis of interviews and self-reports from stakeholders) found a common focus on a particular cohort (older, multimorbid or frail patients) and a wide range of potential interventions (including interventions focused on those in need, as well as longer-term prevention work).
Early evaluation has identified a number of barriers to and facilitators of progress, leading to slow progress and a reining in of ambitions concerning any rapid demonstration of improved outcomes. Patient experience was judged to be the area in which initial gains were most likely to be made. The authors of the report into the pioneers highlighted the ‘integration paradox’, whereby financial and other service pressures both increase the pressure for integration (to manage those pressures) and act as a barrier to its effective implementation.37
In some ways, the evidence from the UK studies is less positive than the international literature. Although some positive impacts have been observed, these have been matched by some negative findings (including increases in admissions and reductions in some aspects of patient experience). It is not clear why this should be. The UK has a fairly strong primary care system with which patients are generally highly satisfied.38 It is possible that changes that lead to disruption in existing arrangements can cause difficulties for patients, even if the intention is to improve integration.
Summary
Integration remains a cornerstone of current health policy, but evidence concerning the benefits of integration, optimal methods of achieving it and the factors that influence success is still limited. The identification of models of integration in the UK that are feasible, sustainable and cost-effective remains a priority.
In that context, the National Institute for Health Research (NIHR) Health Services and Delivery Research (HSDR) programme advertised a call for ‘ambitious research studies assessing the cost-effectiveness of new and innovative models of care or clinical pathways for people with long term conditions. The aim is to generate high-impact research which will provide commissioners and providers with useful evidence when re-designing services’.39
The Salford Integrated Care Programme (SICP) was judged to be ‘a new and innovative model of care . . . for people with long term conditions’.39 The aims of the SICP were to improve integration of care to provide better health and social care outcomes, improved experience for services users and carers, and reduced health and social care costs.
The broad aims of Comprehensive Longitudinal Assessment of Salford Integrated Care (CLASSIC) were to assess the clinical effectiveness and cost-effectiveness of the SICP, with the following research questions.
Implementation
- How do key stakeholders (commissioners, strategic partners) view the SICP, what do they expect from it and how is it aligned with their objectives and incentives?
- What is the process of implementation of two key aspects of the SICP [the MDGs and the integrated contact centre (ICC)]?
Outcomes
- What is the impact of the MDGs on the outcomes and costs of people with long-term conditions?
- What is the impact of health coaching in the ICC on the outcomes and costs of people with long-term conditions?