The Chronic Care Model

The Chronic Care Model developed by Wagner provides a framework for describing the elements needed in a system that aspires to provide high-quality care for people with chronic diseases (Wagner, Reference Wagner1998). The Model was based on a review of available literature about promising strategies for chronic illness management, much of which derived from experience in European health care systems. Key elements in the Model are the community, the health system, self-management support, delivery system design, decision support and clinical information systems. Later refinements of the Model by Wagner and colleagues included an emphasis on patient safety, care coordination and case management. The Model has been described as “a functional blueprint or template, as well as a set of organising principles, for basic changes to support care that is evidence-based, population-based and patient-centred. It defines the broad areas that must be considered (e.g. information systems, self-management support), but not a specific set of interventions; rather, it is a framework in which improvement strategies can be tailored to local conditions” (Glasgow et al., Reference Glasgow, Orleans, Wagner, Curry and Solberg2001: 581) (Figure 1).

Figure 1

The Chronic Care Model

The Chronic Care Model has been used in a range of settings to support the reorientation from acute care to chronic care (Singh and Ham, Reference Singh and Ham2006). The WHO’s Innovative Care for Chronic Conditions Framework is an example. Other approaches include the Expanded Chronic Care Model developed in British Columbia, which gives greater attention to population health than the original Model, and the NHS and Social Care Model used in the English NHS. Australia and New Zealand have also drawn inspiration from the Chronic Care Model, both in the development of national frameworks and in strategies developed by individual states and regions (National Health Priority Action Council, 2006; National Health Committee, 2007).

There have been a number of evaluations of the Chronic Care Model. A systematic review of 41 studies found that multifaceted professional interventions and organisational interventions that facilitate structured and regular review of patients with diabetes were effective in improving the process of care (Renders et al., Reference Renders, Valk, Griffin, Wagner, van Eijk and Assendelft2001). The addition of patient education and the enhancement of the role of nurses led to improvements in the process and outcomes of care. Another systematic review reported that in most studies interventions based on the elements of the Model improved at least one process or outcome measure for diabetic patients (Bodenheimer et al., Reference Bodenheimer, Wagner and Grumbach2002). The same review found evidence of reduced costs or lower use of services for patients with congestive heart failure, asthma and diabetes.

A meta-analysis of a number of studies found that no single element in the Model was essential to improving outcomes but that changes to delivery system design significantly improved processes and outcomes, as did self-management support (Tsai et al., Reference Tsai, Morton, Mangione and Keeler2005). Other work has also emphasised that it is the cumulative effect of the different elements of the Model that is likely to have the greatest impact, rather than individual elements (Bodenheimer et al., Reference Bodenheimer, Wagner and Grumbach2002). A series of studies by researchers at RAND reported better processes and outcomes of care from use of the Model in relation to adults and children with conditions such as asthma, diabetes, heart failure and depression (http://www.rand.org/health/projects/icice/findings.html). RAND also found that teams participating in a collaborative programme were able to make improvements in care in line with the Model and most of these improvements were sustained after a year (Pearson et al., Reference Pearson, Wu, Schaefer, Bonomi, Shortell, Mendell, Marsteller, Louis, Rosen and Keeler2005).

Evidence on chronic care programmes

These findings echo the results of other reviews of the evidence on chronic care programmes. A project undertaken by the Australian Primary Health Care Research Institute used the Chronic Care Model as a framework for undertaking a systematic review of the evidence of interventions for chronic disease management in primary health care and found benefits of self-management support, particularly for diabetes and hypertension (Zwar et al., Reference Zwar, Harris, Griffiths, Roland, Dennis, Powell Davies and Hasan2006). The review also found that delivery system design, through the use of multidisciplinary teams, improved disease measures and adherence to guidelines. The combination of self-management support and delivery system design was noted to be particularly effective, for example, nurses acting as case managers for people with diabetes. Decision support and clinical information systems were effective in that evidence-based guidelines and educational meetings for professionals improved professional adherence to guidelines and some patient outcomes.

A rapid review of the evidence about interventions to improve chronic care undertaken to inform work being done in the English NHS confirmed the benefits of self-management support and multidisciplinary team working while also highlighting evidence to support broad chronic care models such as those developed by Wagner and colleagues. In addition, the review found evidence to support integrated community and hospital care, greater reliance on primary care and the use of nurse-led strategies. On the other hand, there was less evidence to support case management and the use of care pathways to improve chronic care, and insufficient evidence to draw conclusions about the use of the voluntary sector. The review emphasised the uncertainty associated with the interpretation of these findings, particularly relating to the components of multifaceted interventions that contributed to improved outcomes (Singh, Reference Singh2005a).

Despite this caveat, the review underlined the conclusions of other studies in highlighting the actions that can be taken to meet the needs of people with chronic diseases. These actions include giving greater attention to supporting people with chronic diseases to care for themselves, increasing the emphasis on the provision of high-quality primary care by multidisciplinary teams, and achieving effective integration of primary care and specialist care. Moreover, as the WHO has argued in its assessment of the global challenge of chronic disease, there is an urgent need to tackle modifiable risk factors (World Health Organisation, 2005). This requires a renewed emphasis on evidence-based preventative interventions to reduce tobacco use, increase the use of healthy food and promote physical activity. These interventions are as relevant in high income countries as in the rest of the world, particularly given adverse trends in risk factors such as obesity and the likely consequences for the health of the population in future if these trends are not reversed (Olshansky et al., Reference Olshansky, Passaro, Hershow, Layden, Carnes, Brody, Hayflick, Butler, Allison and Ludwig2005).

The ten characteristics of the high-performing chronic care system

The first and arguably most important characteristic of the high-performing chronic care system is ‘ensuring universal coverage’, for without universal coverage it is difficult to act consistently on the other characteristics. Recent confirmation of the critical importance of universal coverage can be found in a study of the health of Medicare recipients which reported dramatic improvements in health of previously uninsured adults who gained health coverage, especially those with cardiovascular disease and diabetes (McWilliams et al., Reference McWilliams, Meara, Zaslavsky and Ayanian2007). This provided evidence, if it were needed, that lack of universal coverage contributes to poorer health outcomes. The importance of this characteristic has been recognised and acted on in almost all developed countries, USA being the most significant exception.

The second characteristic is the provision of ‘care that is free at the point of use’, or at least care that is provided at a cost that does not act as a major deterrent to sick patients seeking medical help. The RAND study of the Health Insurance Experiment, which highlighted the deterrent effects of co-payments on service use, especially among people in the poorest health, underscores the importance of this characteristic (Newhouse, Reference Newhouse1993). The Commonwealth Fund’s surveys have found wide differences between countries in this regard, with USA again appearing as an outlier in the countries studied. The cost of care is also reported as a barrier to use by between one-quarter and almost two-fifths of respondents in Canada, Australia, Germany and New Zealand. In recognition of the adverse effects of co-payments, even those levied at a modest level, policy makers in some systems have sought to counter these effects by increasing public subsidies of care. New Zealand illustrates this trend with the Labour Government elected in 1999 providing extra funding for primary care with the aim of reducing out-of-pocket payments, particularly for children and those on low incomes.

The third characteristic is that ‘the delivery system should focus on the prevention of ill health’ and not just the treatment of sickness. Despite progress in a number of systems in implementing population health measures such as controls over tobacco consumption, and in targeting medical support at individuals deemed to be at high risk of chronic disease, policy makers have found it difficult to invest in prevention to the same degree that they have invested in treatment services. A partial exception to this that holds pointers for the future is the pay for performance contract for family physicians in the UK introduced in 2004 that links the payment of physicians to the achievement of targets for the provision of high-quality preventative chronic care (Doran et al., Reference Doran, Fullwood, Gravelle, Reeves, Kontopantelis, Hiroeh and Roland2006) (Box 1).

Nurses working in primary care carry a major part of the workload involved in achieving the quality targets contained in the contract and in thereby enabling family physicians to increase their incomes and profits. This includes maintaining registers of patients with specified conditions and implementing review and recall systems to ensure these patients are checked at regular intervals and have their care plans updated as appropriate. Beyond this specific policy initiative, Glasgow et al. (Reference Glasgow, Orleans, Wagner, Curry and Solberg2001) have shown how the Chronic Care Model can act as a template for improving prevention.

The fourth characteristic is that ‘priority is given to patients to self manage their conditions with support from carers and families’. The importance of self-management support has been captured well by David Sobel in his argument that people with chronic diseases are the most important primary care providers in a context in which chronic diseases represent the major burden of disease. As Sobel shows, self-management support is critical because most care is and always has been self care, and a small reduction in the propensity of people to self care will result in a significant and potentially insupportable increase in demand for organised care services (Sobel, Reference Sobel1995). The pioneering work, in this area, of Kate Lorig et al. (Reference Lorig, Seleznick, Lubeck, Ung, Chastain and Holman1989) at Stanford University has been extended and adapted in other health care systems through a range of generic and disease-specific training programmes designed to enable people with chronic disease to take control of their conditions. An example is the Expert Patient Programme in England, which had delivered training in self-management support to an estimated 30,000 people by the end of 2007. There is evidence to show the benefits of this programme in terms of greater self-efficacy compared with patients not receiving this kind of support (Kennedy et al., Reference Kennedy2007).

A number of reviews have summarised the evidence on the benefits of self-management support (Singh, Reference Singh2005a; Zwar et al., Reference Zwar, Harris, Griffiths, Roland, Dennis, Powell Davies and Hasan2006). One of the most comprehensive studies focused on randomised trials of self-management interventions relating to type 2 diabetes, asthma and arthritis (Newman et al., Reference Newman, Steed and Mulligan2004). This study underlined the variety of interventions that have been used, both between and within illnesses, and the range of objectives they have addressed. It also highlighted variations in the outcomes that were achieved from different interventions. One of the conclusions of this study was that it was important to establish the extent to which generic chronic disease self-management programmes such as those developed by Lorig and Sobel were sufficient compared with disease-specific programmes. This conclusion is important in underlining the uncertainties that exist in relation to the specific interventions needed to turn some elements of the Chronic Care Model into practice.

The fifth characteristic is that ‘priority is given to primary health care’. This characteristic derives from evidence on the contribution that primary care makes to overall system performance (Starfield et al., Reference Starfield, Shi and Mackino2005), and to the everyday reality that most chronic disease management takes place in primary care in most health care systems (and commensurately that most primary care entails the management of chronic disease, especially, but not exclusively among older people). To be more specific, there is evidence of the value of multidisciplinary team work in primary care with much of the responsibility for the management of chronic diseases resting with nurses working as part of the team (Grumbach and Bodenheimer, Reference Grumbach and Bodenheimer2004; Singh, Reference Singh2005b). The role that nurses play alongside physicians in the implementation of the new contract in the UK (see above) illustrates this point well.

Primary health care is important in many systems in providing a medical home for patients. Evidence from the Commonwealth Fund’s 2007 international health policy survey found that in all seven countries in the survey, between 45% and 61% of respondents said they had a medical home, defined as a regular source of care with easy contact by phone, a physician with a knowledge of the patient’s medical history and a physician able to coordinate care. Respondents who reported having a medical home were more positive about their experiences in areas like primary care access and communication with physicians than those who reported not having a medical home (Schoen et al., Reference Schoen, Osborn, Doty, Bishop, Peugh and Murukutla2007).

The sixth characteristic is that ‘population management is emphasised’ through the use of tools to stratify people with chronic diseases according to their risk and offering support commensurate with this risk.

The risk pyramid used within Kaiser Permanente and adapted in other systems, for example, the NHS and Social Care Model developed in England, illustrates the practical application of a population management approach. The pyramid organises the chronic disease population into three categories: people requiring usual care and support to self manage; people needing regular contact with a multidisciplinary team to ensure effective management of their disease; and people requiring more intensive support, perhaps from a specialist case manager, often when they are coping with the complications of co-morbidities, and are at risk of their condition deteriorating and needing to make use of specialist hospital care. Disease registers play an important part in population management by enabling primary health care teams to recall patients for regular checks and to review their treatment and care plan (Figure 2).

Figure 2

The Kaiser Permanente Risk Pyramid

A wide range of tools have been developed to support population management, designed to analyse patient data to enable care to be provided pro-actively. These tools are important in ensuring that case management is targeted at patients most at risk and also those most likely to benefit from intensive support. There is a cautionary tale here from the experience in England of adapting the Evercare approach to case management, originally developed in USA to support at risk nursing home residents, to the population served by the NHS. Specialist nurses were employed in the NHS to provide care to people who had experienced two or more unplanned admissions in the previous year. The evaluation of this scheme reported no significant effect on hospital admissions, although the people receiving care valued the support they were given (Boaden et al., Reference Boaden, Dusheiko, Gravelle, Parker, Pickard, Roland, Sargent and Sheaff2006). Part of the explanation of these findings is that there is regression to the mean in the use of health services by the chronically ill population, with periods of high use being followed by periods of average use even in the absence of interventions. The lesson from this example is that the method of identifying the recipients of case management needs to use indicators other than (or in addition to) hospital admissions. In England this is being done through the use of the patients at risk of rehospitalisation (PARR) tool (Billings et al., Reference Billings, Dixon, Mijanovich and Wennberg2006).

The seventh characteristic is that ‘care should be integrated to enable primary health care teams to access specialist advice and support when needed’. The importance of this characteristic relates to the dynamic and fluid nature of chronic diseases and the tendency for people to require different types of support and intervention at different times depending on the progression of their condition. It also relates to the particular needs of people with two or more conditions who may find themselves seeking a range of advice from different specialists while also having a continuing relationship with a primary health care team. In these circumstances, there is a risk of care not being coordinated and of complications arising, for example, through drug interactions resulting from polypharmacy. Evidence summarised above from the Commonwealth Fund’s surveys underlines the reality of this risk, and the importance of overcoming professional, organisational and financial barriers to the integration of care. Integrated delivery systems like Kaiser Permanente, the Veterans Health Administration and Group Health Cooperative appear to be better placed than other systems in addressing this challenge (Box 2). Partly in recognition of this, New Zealand has reorganised its health care system to create integrated district health boards with responsibility for the full range of health care for the populations they serve.

The eighth characteristic, closely linked to the last point, is ‘the need to exploit the potential benefits of information technology in improving chronic care’. Not least, information technology underpins effective population management in enabling primary health care teams to develop disease registers and to stratify the population according to risk. In some applications, it also supports patient self-management through easing communication between patients and health care professionals and facilitating shared decision-making.

Two other aspects of information technology appear to have a particularly promising part to play at this stage in their application to health care. First, there is the role of the electronic patient care record in supporting communication between health care professionals and in reducing the possibility of errors. This includes the use of clinical decision support systems to provide prompts and reminders to professionals in relation to adherence to guidelines on tests and treatment reviews. Data captured on these support systems can in turn be used to compare the performance of providers in undertaking tests and reviews as a tool for continuous quality improvement. Integrated systems like Kaiser Permanente are already utilising their investment in information technology to prepare ‘missed opportunity’ reports that highlight ways in which providers can do more to systematically implement planned care for patients with chronic diseases.

Second, information technology is beginning to demonstrate its value in enabling people to be supported at home through telecare and telehealth applications. The Veterans Health Administration (VA) has been at the forefront of these developments and has pioneered the use of technology in this way through the Care Coordination/Home Telehealth (CCHT) programme. CCHT was piloted between 2000 and 2003 and targeted at a specific population of patients at risk of long-term institutional care by virtue of their resource use and disease-specific factors related to one or more of the following conditions: diabetes, chronic heart failure, hypertension, depression and chronic obstructive airways disease. In the approach adopted by the VA, a care coordinator (usually a nurse or social worker) uses telehealth systems to monitor vital sign and disease management data entered into technology placed in the patient’s home. Through regular monitoring of these data, the care coordinator can provide services to a panel of approximately 125 patients. The promising early results from the pilot programme led the VA to commence implementation of CCHT on a national basis in 2003 (Darkins, Reference Darkins2007). Partly influenced by the experience of the VA, the NHS in England has recently instituted a demonstration programme covering one million people to test the impact of telecare and telehealth in the management of chronic diseases (http://www.dh.gov.uk/en/Healthcare/Longtermconditions/DH_4140328).

The ninth characteristic is to ensure that ‘care is effectively coordinated’. Coordination is particularly important in the care of people with multiple conditions who are at much greater risk of hospital admission than people with single diseases (Wolff et al., Reference Wolff, Starfield and Anderson2002). The role of primary care physicians in providing coordination has been emphasised in a number of studies (Starfield et al., Reference Starfield, Lemke, Bernhardt, Foldes, Forrest and Weiner2003). Also important are patient activation through coaching interventions, and the role of key staff such as specialist nurses in supporting patients at level three in the Kaiser pyramid and during care transitions, for example, from hospital to home. Trials of coaching and transitional care interventions have demonstrated benefits including reduced hospitalisation rates. In his review of the various models being used to improve care coordination, Bodenheimer places particular emphasis on the role of primary care, arguing that “the most efficient structure for coordinating care is a system with a strong primary care foundation in which the primary care practice, in partnership with its patients, consciously assumes the responsibility for coordinating care throughout the health care system” (Bodenheimer, Reference Bodenheimer2008: 1069).

In the UK and other European countries, efforts to strengthen care coordination encompass social care as well as health care. These efforts reflect the need of people with chronic conditions to access a range of support, not just medical interventions. Various approaches have been adopted including the development of intermediate care in the UK, the use of individual budgets and direct payment systems to enable people to purchase their own care, and the establishment of agencies that integrate health and social care. These agencies are broadly analogous to organisations like On-Lok in San Francisco that have been recognised as providing innovative care for their populations. A European example of care coordination is the Esther project in Jonkoping County Council in Sweden which focuses on how to improve the quality of chronic care for a fictitious older woman (Esther) through more effective team working, and improved communication between providers involved in Esther’s care (http://www.ihi.org/IHI/Topics/Flow/PatientFlow/ImprovementStories/ImprovingPatientFlowTheEstherProjectinSweden.htm).

The tenth characteristic, alluded to in the review of the evidence above, is to ‘link these nine characteristics into a coherent whole as part of a strategic approach to change’. This is important in view of evidence that it is the cumulative effect of different elements that explains the degree of impact of the Chronic Care Model rather than individual components. By extension, the argument of this paper is that chronic care policy needs to address several issues at the same time, linking reforms to health financing (universal coverage and limited or zero co-payments), an increased focus on prevention, self management and primary health care, a commitment to achieve closer integration of care and more effective care coordination, and the greater use of tools such as population management and information technology. Action on some of these characteristics and not others is likely to limit and slow the implementation of models of care that are required to meet the needs of the chronic care population.

In identifying these ten characteristics, it is worth adding that the evidence on disease management, while much discussed in the literature, is equivocal. As a review undertaken by the Congressional Budget Office in 2004 showed, studies of disease management programmes for congestive heart failure, coronary artery disease and diabetes found that there was insufficient evidence that they could reduce the overall cost of health services, although there was evidence that they improved adherence to practice care guidelines and led to better control of the disease (Congressional Budget Office, 2004). Advocates of the Chronic Care Model such as Wagner argue that disease management programmes that are not part of a coherent approach to chronic care may accentuate the problems of care coordination and are therefore likely to be less effective than programmes that focus on improving primary care provision through multidisciplinary teams. As the Institute of Medicine put it, “One of the main concerns associated with disease management programs is the potential for fragmenting care, especially if the patient’s primary care physician is not involved in the program” (Institute of Medicine, 2001: 100).