Facilitators and barriers of implementing the chronic care model in primary care: a systematic review

Background The Chronic Care Model (CCM) is a framework developed to redesign care delivery for individuals living with chronic diseases in primary care. The CCM and its various components have been widely adopted and evaluated, however, little is known about different primary care experiences with its implementation, and the factors that influence its successful uptake. The purpose of this review is to synthesize findings of studies that implemented the CCM in primary care, in order to identify facilitators and barriers encountered during implementation. Methods This study identified English-language, peer-reviewed research articles, describing the CCM in primary care settings. Searches were performed in three data bases: Web of Knowledge, Pubmed and Scopus. Article abstracts and titles were read based on whether they met the following inclusion criteria: 1) studies published after 2003 that described or evaluated the implementation of the CCM; 2) the care setting was primary care; 3) the target population of the study was adults over the age of 18 with chronic conditions. Studies were categorized by reference, study design and methods, participants and setting, study objective, CCM components used, and description of the intervention. The next stage of data abstraction involved qualitative analysis of cited barriers and facilitators using the Consolidating Framework for Research Implementation. Results This review identified barriers and facilitators of implementation across various primary care settings in 22 studies. The major emerging themes were those related to the inner setting of the organization, the process of implementation and characteristics of the individual healthcare providers. These included: organizational culture, its structural characteristics, networks and communication, implementation climate and readiness, presence of supportive leadership, and provider attitudes and beliefs. Conclusions These findings highlight the importance of assessing organizational capacity and needs prior to and during the implementation of the CCM, as well as gaining a better understanding of health care providers’ and organizational perspective.


Background
The prevalence of chronic diseases is globally on the rise, with cardiovascular diseases, respiratory disease, diabetes, cancer, and other chronic illnesses being major contributors to disability [1,2]. In Canada, two out of five people have at least one chronic disease. Chronic disease is a major driver of health care expenditure, reaching approximately $68 billion in Canada in 2010 [3]. The current health care system is oriented towards episodic and acute care, making it unprepared to address the multi-faceted and complex needs of those with chronic diseases [4,5]. Given the need for continuity, comprehensiveness and coordination, primary care has been suggested as potentially playing a central role in effective management and integration of care [6]. However, literature on current practice suggests that patients often receive inadequate care, with limited physician involvement in disease management, and little coordination and communication among care providers [7].
In response to these challenges and the call for redesigning care delivery for chronic diseases, Wagner and colleagues developed the Chronic Care Model (CCM) [8,9]. The CCM was developed to bridge the gap and translate knowledge between evidence-based chronic disease care and actual care practices. The framework which is centered in primary care, 'conceptualizes care as prepared practice teams in productive interactions with informed, activated patients' [10]. It posits six interrelated elements that are key to high quality chronic disease care: self-management support, redesigning delivery systems, decision support that is system wide, clinical information technology, linkages to community resources, and health care system organization [10,11]. The components seek organizational change at the systems' level, promoting and delivering care that is evidence-based through using clinical tools such as guidelines, utilizing information systems that improve patient data sharing across the organization and between providers, engaging and empowering patients in their care, and mobilizing community resources to meet patient needs [11]. The CCM and its various components have been widely adopted and evaluated, with results showing that it improves patient care and clinical outcomes, and reduces care utilization and costs [12][13][14][15][16].
Despite the extensive evaluation of quality improvement (QI) initiatives, and research on CCM-based interventions, particularly across the United States, little is known about different primary care experiences with its implementation, and the factors that influence its successful uptake [10,14,17,18]. The model provides no clear blueprint on how each component can be implemented in practice, and there is considerable heterogeneity in the types of interventions implemented in primary care in support of the CCM [10]. Previous reviews synthesizing evidence on the CCM have focused on associated care changes, clinical outcomes and costeffectiveness [10,14,19]. Although a recent systematic review by Holm and Severinsson identified barriers and facilitators of successful CCM implementation in primary care, it was specific to depression management in the US [20]. An understanding of the barriers and facilitators of implementing the CCM, in different care settings is important for several reasons. A barrier in this context is defined as any factor that hinders or impedes care change processes of CCM implementation. First, there are numerous contextual factors that enable organizational change and successful translation of evidence into practice [21,22]. Some of the factors previously identified include: evidence fit and relevance to the organizational context, staff relationships and collaboration, availability of resources, strong and committed leadership, and a culture supportive of change [22][23][24]. Second, given the complex and multifaceted nature of the model, primary care organizations can face difficulties with its implementation [12]. This is particularly the case given that there are no guidelines available on how to effectively operationalize CCM elements across different settings [25]. We therefore aimed to identify and review evidence on the challenges and barriers encountered while implementing the CCM in primary care.

Methods
We conducted a systematic literature review to synthesize findings of studies that implemented the CCM in primary care, in order to identify facilitators and barriers encountered during implementation. Barriers and facilitators were interpreted using the Consolidated Framework for Implementation Research (CFIR) [26]. As this research did not involve human subjects, we did not seek ethics clearance for the project.

Data sources
This study identified English-language, peer-reviewed research articles, describing the CCM in primary care settings. Searches were performed in three data bases: Web of Knowledge, PubMed and Scopus. These databases include Medline, EMBASE and the National Library of Medicine. The PubMed and Scopus search strategy used the following MeSH terms to describe 'primary care': primary health care, general practice and family practice. Since there were no MeSH terms for Chronic Care Model, the term was put under quotation marks during the search. In order to ensure a comprehensive search that included all studies that implemented the CCM, MeSH terms for 'implementation' were not used in the search. This strategy was also used to avoid excluding studies that might not have identified the term in their titles and abstracts. Search terms and concepts were combined using the Boolean and Proximity operator 'OR' , while concepts were combined using ' AND' and 'Near' (Table 1).
A second strategy adapted from Coleman and colleagues involved searching articles from Web of Knowledge Science Citation Index, which cited the five foundational CCM articles by Wagner and colleagues and Bodenheimer and colleagues [8][9][10]14,27,28].
In addition, hand searching of the reference lists in all articles that met the inclusion criteria outlined below was performed to identify any missed relevant articles. Search terms used in both search strategies are described in Table 1.

Study selection
Citations were downloaded and screened in Refworks, an online citation manager tool. Article abstracts and titles were read based on the exclusion and inclusion criteria detailed below. If the reviewer could not determine whether to exclude an article based on its abstracts and title, then it was retrieved for full text reading. Figure 1 displays the process involved in study selection.
Exclusion criteria: 1) Articles published before 2003 and in languages other than English; this year was chosen as the search cut-off to follow the publication date of the last CCM foundational paper by Bodenheimer and colleagues [10], thus reflecting studies that implemented a more mature conceptualization of the model 2) Articles that solely described the CCM conceptually, i.e., did not report on an actual implementation of the model, commentaries and opinion pieces, study protocols, reviews including: systematic and narrative reviews, and meta-analyses 3) The target population of the study was not adults aged 18+ with chronic conditions Figure 1 Exclusion and inclusion criteria for article selection.

4)
Articles arguing or providing the rationale for implementation of CCM in primary care, but which were not based on empirical studies.
Inclusion criteria: 1) Articles describing or evaluating the implementation of the CCM. Implementation had to refer to efforts which used change strategies to promote use of evidence-based practices or programs [29] 2) Implementation of the CCM had to occur in primary care, which is defined as integrated and accessible healthcare, delivered in the context of family and community [30]. 3) Articles identifying barriers and/or facilitators of CCM implementation.

Data abstraction
The methods used for the study selection and data abstraction in this systematic review are aligned with those in the PRISMA statement. The PRISMA statement provides an evidence-based checklist intended to improve the standards of reporting in systematic reviews [31]. Given that the focus was on implementation, rather than study outcomes, not all aspects of the PRISMA statement were adopted. Data abstraction involved two stages. First, articles were categorized by reference, study design and methods, participants and setting, study objective, CCM components used, and description of the intervention. The next stage of data abstraction involved qualitative analysis using the Consolidating Framework for Research Implementation (CFIR), which has five domains: intervention characteristics, outer setting, inner setting, characteristics of the individuals involved, and the process of implementation [26]. It provides a conceptual framework which can be used to understand factors that influence successful implementation in health care, and is based on theories identified by Greenhalgh and colleagues' widely cited systematic review [26,32]. The CFIR was selected because it includes multiple constructs and theories from peer reviewed studies on evidence-based knowledge dissemination and translation, organizational change and implementation, and uptake of research. It has also been suggested as a framework that can be used to guide the implementation of CCM components in interventions: therefore, it was deemed most appropriate for our study [26]. Table 2 provides summarized descriptions of the CFIR domains.
Using qualitative content analysis, implementation barriers and facilitators in 22 articles were mapped on to the CFIR framework. When articles described barriers or facilitators of CCM implementation, they were regarded as "attributive statements", which were coded under the appropriate constructs and domains. These statements were often found in the discussion and results section of the articles. If the statement was beyond the domains and constructs of the CFIR, then it was still documented. Our approach was modeled after the data abstraction method used in a systematic review by Mair and colleagues [33]. The data abstraction and coding was performed by one reviewer. Interpretative and inductive reasoning were used to map out the attributive statements to the framework.

Results
Twenty two studies were included in this review. Study descriptions and methodological procedures were summarized in terms of design, measurements, sample size and context, as shown in Table 3. In Table 4 statements reflecting implementation barriers and facilitators from each article were analyzed and coded to their respective domains and constructs under the CFIR framework.

Networks and communication
Strong networks and increased communication between health care providers and organizations were fostered by collaboration across disciplines and specializations during care change processes [39,40,44,50,51]. Communication was reportedly supported by regular group meetings with supervisors and managers to discuss implementation issues, computerized information sharing and clinical assessment tools [41,45,52].

Culture
An organizational culture that promotes multidisciplinary, or patient centered care, was identified as important during implementation [45,51,52]. Support from clinical providers and the recognition of their importance in care change efforts was found to increase uptake of the CCM in primary care [35,37,39].

Implementation climate
Studies found that implementation climate was influenced by commitment and recognition for the need for change from the organization [40,45]. Willingness to advance and manage change was evident through incentivizing provider buy-in using financial reimbursement and work credit for project involvement [37,42,51].

Structural characteristics
Operationalization of CCM components was facilitated by health care providers, particularly specialists and non-physician staff such as nurse practitioners, who had to expand their responsibilities and scope of practice [45,53]. This sometimes required changing organizational policies and development of care teams to meet implementation needs [40,44].

Engaging
Strong, committed and engaging leadership in the form of supportive administration and supervisors, with clear goals, was cited as a facilitator [40,45,50]. This included the appointment of an intervention champion to promote uptake of the model within the organizations [19,37,51]. Leadership roles were not limited to physicians, other health care providers such as nurse practitioners were found to play a major role in implementation [40].

Knowledge and beliefs about the intervention
Provider knowledge about CCM interventions was promoted through observing the execution process by other staff and education about project goals [42,50,51]. Strategies used to foster beliefs of the CCM effectiveness in care providers, particularly physicians, included demonstration of its benefits to their practice and sharing reports of patient improvements [37,51].

Barriers Executing
Many studies identified barriers related to executing intervention processes. Implementing the multiple components of CCM into practice created additional responsibilities for staff who were limited by time constraints [19,40,48,50]. Pearson & colleagues found that operationalizing the model elements at a high level of intensity, within a short time frame to be challenging [46]. Sustainability of the intervention was found to be difficult in some studies; in some instances, staff buy-in, an important aspect of implementation, was not enough to ensure program longevity [48].

Structural characteristics
Characteristics of the healthcare organization such as its size, whether it adopted a team-based approach and its flexibility in reorganizing care, were seen to influence the success of CCM adoption [40,45,48,52]. Institutional factors such as staff turnover and loss meant increased burden of responsibilities on existing providers [19,44] 10). Leadership turnover, particularly that of a medical director, was cited as a barrier towards implementing care change processes [38].

Readiness for implementation
Organizational readiness for the CCM was found to be impacted by the lack of interest and commitment from leadership and unavailability of resources for implementation [40,45]. Lack of resources that influenced readiness included low funding, lack of provider reimbursement strategies and low staff numbers [34,43,45,50].

Engaging
Many studies found that execution of the intervention processes was challenging without support and accountability from senior leadership [19,20,44]. Without the presence of an intervention champion, endorsement of the CCM initiative was found to be limited in healthcare providers [19].

Knowledge and beliefs
Provider buy-in was greatly influenced by knowledge and beliefs about the intervention, particularly if they had misconceptions, were unconvinced of its effectiveness or Table 2 Description of CFIR domains and constructs [26] Domain Definition

Intervention characteristics
The characteristics of the intervention being implemented include whether: the intervention is perceived to be developed external or internal to the organization, there is evidence supporting its effectiveness, and its implementation will be advantageous to its alternatives. Other characteristics include how the intervention is presented, its adaptability, complexity and whether it can be tested on a smaller scale.

Outer setting
The external context of the organization includes: patient needs and the ability to meet them, networks with other organizations, pressure to implement the intervention and external policies and incentives to adopt the intervention.

Inner setting
Features of the organization including its structural characteristics (such as size, age of the organization and division of labour), networks and communication (such as connections and information sharing between individuals, units and services), cultural norms and values, implementation climate, organizational capacity and readiness for change.

Characteristics of individuals
Staff knowledge and belief about the intervention, their ability to execute their respective aspects of the implementation, and their individual stage of change. Other characteristics include individual identification with the organization and other personal attributes.

Process
Active change process, the purpose of which is to promote uptake of the intervention by the organization. This is influenced by the level of planning prior to implementation, and engaging organization stakeholders through appointing implementation leaders and champions of the intervention. This includes the ability to execute the implementation of the intervention as planned and to continuously reflect on and evaluate the quality of implementation and intervention as it progresses.    "Limited guidance on prepared practice team development" [40] C. Relative advantage "Patient screened by staff before seeing physician " [39], "Structured assessment in patient education" [39] D. Adapability "Integrating Guided Care nurse in work flow" [36], "Processes integrated in to existing clinical operations" [43], "CCM adaption within context of daily practice" [48], "Program tailored to region needs" [50], "Adapting communication system to local context" [52], "Integrated project to routine care" [52] E. Trialability F. Complexity "Intervention was too complex, targeted different components resulting in many priorities" [50] G. Design quality & packaging "Nurse training for components of intervention" [35], "Curriculum should be specific to CCM intervention" [36], "Different intervention model options were offered" [19], "Structured learning sessions and support by health collaborative" [44], "Guideline development" [50] "Intervention was too disease specific and did not address chronic care principles" [45] H. Cost "Low-cost program relied on community health workers, mentors and non-clinical staff" [43], "Financially viable" [48], "Sufficient funding" [37] 2. Outer setting

A. Knowledge & beliefs about intervention
"Increase awareness and education about program to providers" [41], "Observation of program processes by providers" [42], "Patient registry received interest in providers" [44], "Clinical assessment tool accepted and endorsed" [45], "Information campaign to increase awareness and knowledge" [50], "Education about project goals & process" [51], "Demonstration of project benefit to physicians" [51], "Staff morale and burnout reduction associated with reports of improved care outcomes" [37] "Needed more information on structured assessment" [39], "Unconvinced of usefulness of structured assessment for diagnoses" [39], "Lack of program information from providers that were not full time" [41], "Physician buy-in and adoption of intervention was not uniform" [47], "Fear of losing patient control to education program" [50], "Time needed for provider trust in program" [50], "Clinicians sensitive to workload and time commitment" [45] B. Self-efficacy "Fatigue and apathy from pace of change" [40], "Decreased staff participation in intervention results in low morale" [37] C. Individual identification with organization D. Personal attributes

Process
A. Planning "Realistic expectations for measureable results" [40], "Consultation with focus groups for change process priorities" [50], "Physician involvement in planning" [51], "Utilized patient and physician experience in project development" [51], "Goals of QI as drivers of planning " [52] "Lack of details on care change goals & outcomes" [40], "Too many priorities and uncoordinated change processes" [40],"Need for stronger program goals delineation" [41], "Lack of clear program aim at the start of campaign" [50] lacked information [39,47,50]. Acceptance of the interventions by clinicians required time, and was also affected by the workload associated with implementing and executing the intervention components [45,50].

Discussion
This review identified multiple barriers and facilitators of implementing the CCM across various primary care settings. The major emerging themes were those related to the inner setting of the organization, the process of implementation and characteristics of the individual healthcare providers. These included: culture of the organization, its structural characteristics, networks and communication, implementation climate and readiness, supportive leadership, and provider attitudes and beliefs. Every primary care organization possesses its own cultural norms, practices and leadership. It is impossible to achieve change without adopting an approach that considers the individual and the team of providers, the organization setting and the greater system within which it is embedded [54]. Wolfson and colleagues attributed the success of QI in different primary care practices to facilitators in various levels of the organization including: presence of an initiative champion; physician, staff and patient cooperation; leadership investment; team practice and progress tracking [55]. The uptake of CCM elements in the studies required a primary care culture supporting willingness to change and quality improvement at the individual clinician, team and organizational levels. Implementation was most successful when there was a shared vision and a recognized need across the organization for new care change approaches to promote effective execution of the CCM [35,36,39,44,52].
Transforming care practices in an organization requires a supportive culture of change and learning [23]. Clinical provider beliefs and attitudes about evidencebased practice can influence the culture and learning environment, particularly when the provider perceives the evidence as unreflective of their day-to-day clinical decision making. This suggests the need to involve clinicians in early stages of intervention development and implementation [22]. Interventions that incorporated providers, patients and their experiences in the planning phase of the intervention were more successful in operationalizing the CCM [50,51]. This approach may bridge the cultural divide between leadership and clinical providers, which can hinder quality improvement efforts if left unaddressed. On the other hand, literature shows that lack of a group-oriented culture, as well as hierarchical relationships where the leadership is unsupportive of Table 4 Thematic analysis shows the barriers and facilitators identified by the studies mapped on to their corresponding CFIR domains and constructs (Continued) change, are negatively associated with implementation of care change processes [55]. Marshall and colleagues highlight the importance of culture and cultural change when implementing clinical governance in primary care. Cultural traits that support implementation efforts include commitment to accountability by the organization, willingness for collaborative work and learning, and ability to evaluate and reflect on mistakes [56].
Implementing and managing change processes in primary care can require time and flexibility. Organizational transformation can be slow and resistant to change, while spread of best-practice can be a challenge [57]. In some cases, even when an organization's culture is supportive of the CCM, the inner structures of the primary care organization, such as a lack of staff and financial resources or a lack of clinician expertise, can impede organizational readiness for implementation and cause unexpected setbacks [34,48,52]. A study evaluating the implementation of evidence-based practice revealed that the current primary care system is not adaptive to rapid change, or accommodating to the additional duties associated with adopting new interventions. What this suggests is the need to set realistic implementation goals that are reflective of the organization and staff capacity for changes associated with the CCM. This requires comprehensive planning at all stages of component adaptation, to mitigate impeding factors such as rigid bureaucracies and organizational policies.
On the other hand, clinical leaders and champions can be drivers of change by ensuring the availability of resources and providing adequate staff supports [58]. Indeed, leadership support for change has been shown to be positively associated with QI outcomes and sustainability in primary care [24]. Implementation of CCM in primary care requires tailoring interventions to the local context, as well as altering the context, for the process to be successful. The two can co-adapt and evolve during the implementation process, thereby ensuring sustainability [59]. The majority of the studies included in the review identified the impact of the CCM on existing routines, practices, and culture of the organization. There was variability in how each organization adapted the CCM, i.e., translating the framework's components into practice resulted in implementation heterogeneity. What became clear is that a standardized one-size-fits-all approach was difficult to put into practice when the components were conceptualized differently by each primary care organization.
Tailoring the intervention necessitates accounting for innovation-promoting and hindering factors at different levels of the organization, as well as reconfiguring aspects of the primary care setting itself [50].
This systematic review has several limitations. First, our search strategy meant that we did not assess: grey literature, studies that have not been published in peerreviewed journals or those published in languages other than English; therefore, articles that were relevant to our review may not have been included. The search may have excluded studies that implemented CCM-based interventions but which were not explicitly referred to as such in the articles. In addition to the challenge of consistently identifying such studies, it would be difficult to be certain that implementation issues were reflective of issues relevant to the CCM. Another limitation is that the articles that were included were selected and assessed by one reviewer, thus limiting the reliability of the selected studies. Given that the articles were abstracted qualitatively by a single data abstractor, there is a possibility of bias in how the attributive statements were mapped under CFIR constructs and domains. While abstraction and coding was carried out by one reviewer, extensive and continuous discussion took place between both authors occurred during the study selection and data abstraction process. While using the CFIR as a guiding framework is a strength of our review, the numerous construct and sub-constructs meant that areas with few facilitators and barriers identified received less consideration (although these were captured in Table 4).

Conclusion
These findings highlight the need to evaluate factors that influence successful implementation of CCM in primary care. The CFIR can be used to guide the formative evaluation processes of CCM interventions. Assessment of organizational capacity and needs is important prior to and during the implementation of the intervention, in order to gain a better understanding of health care providers and organizational perspective. The barriers and facilitators identified under the CIFR domains can be used to build knowledge on how to adapt the CCM to different primary care settings.