The Influence of Contextual Factors on the Process of Formulating Strategies to Improve the Adoption of Care Manager Activities by Primary Care Nurses

Research design

We conducted a qualitative multiple case study in three FMGs [2930] using an integrated knowledge translation approach to formulate implementation strategies with stakeholders. Integrated knowledge translation can be defined as “an ongoing relationship between researchers and decision-makers (clinicians, managers, policy-makers, etc.) for the purpose of engaging in a mutually beneficial research project or program of research to support decision-making” [31, p. 1]. To operationalize integrated knowledge translation in the planning stage of the implementation process, we followed steps one to four and part of step five of Grol & Wensing’s implementation of change model: 1) Development of a proposal for change; 2) Analysis of actual performance; 3) Problem analysis of target group and setting; 4) Development and selection of strategies; and 5) Development, testing and execution of an implementation plan [32]. We did not perform the two final steps of this model, i.e., integration of changes in routine care and continuous evaluation and adaptation of the plan.

We adopted a multiple case study methodology to reach the deep understanding of the context necessary to reach our aim [33]. We defined cases as the implementation planning process in each of the participating FMGs. Units of analysis within each case were: (1) current nursing and collaborative care practices for people with CMDs and long-term physical conditions and (2) the formulation of implementation strategies to optimize the adoption of care manager activities by PCNs.

We conducted the integrated knowledge transfer approach with regional and local stakeholders’ involvement, including an advisory committee, local working groups, clinicians, and patients. The advisory committee’s role was to counsel the research team regarding the provincial and regional contexts and support local working groups’ decision-making within FMGs. This committee included two patient partners with a history of CMD and at least one physical long-term condition, three researchers in the fields of collaborative mental health care and nursing, one nurse manager, one GP, one PCN, one psychiatrist, and one psychologist. Ethical approval was provided by the Centre intégré universitaire de santé et de services sociaux de l’Estrie-Centre hospitalier universitaire de Sherbrooke (CIUSSS-CHUS) ethics committee.

Recruitment

We recruited three FMGs based on convenience sampling and the lead GPs’ level of interest in this study. We contacted lead GPs by e-mail, providing a one-page summary of the project and describing the study’s aim and the expected staff involvement and timeline. We chose to target three FMGs to improve the results’ transferability to similar contexts while maintaining the study’s feasibility [30]. All three participating FMGs were located in medium-sized urban areas, but they varied in terms of mental health practices and resources (e.g., access to a consulting psychiatrist).

Within each FMG, a local working group was created, including at least one PCN and one GP. Local working groups were involved in discussing the gap between their practices and the CCM and in strategy formulation. In general, working group members were administrative leaders in their FMG or professionals whom their colleagues identified as mental health experts. Local working group members provided signed informed consent to participate in this study.

In each FMG, health care professionals involved in the care of people with CMDs (i.e., PCNs, GPs, psychologists, and social workers) were invited to participate in interviews. For patients recruitment, we asked participating PCNs to identify one adult patient diagnosed with at least one long-term physical condition and a depressive or anxiety disorder. Identified patients were then individually contacted by a research team member for participation in a qualitative interview and an observation session of a clinical encounter with their PCN. All participants to interviews and/or observation sessions provided signed informed consent before data collection.

Data collection

Data collection was carried out throughout the implementation planning process. From December 2018 to April 2019, data were collected through the following sources: (a) face-to-face interviews (≈45 to 60 minutes) with patients and clinicians (n = 32); (b) non-participatory observations of a nurse-patient encounter (n = 7); and (c) relevant documents on professional activities and collaborative care (documents on specific programs for people with long-term physical conditions or CMDs, educational materials, clinical nursing mental health assessment guides, and health care professionals’ referral forms). In addition to interviews and observations, each participant completed a questionnaire. Patients were asked sociodemographic and health-related questions, and clinicians were asked questions relating to their professional role. In addition, every advisory committee and local working group meeting was summarized in a document including attendees’ identity and the main points discussed (January 2019 to January 2020).

We developed an interview guide for each type of participants (patients, PCNs, other primary care providers) to investigate current practices, challenges to nursing and collaborative care practices, and perceptions regarding an eventual change in practices. Interview guides were developed based on the results of previous studies on challenges to nursing and collaborative care practices for people with CMDs and long-term physical conditions in primary care clinics [1534], and studies on barriers and enablers to implementing the CCM [171819]. We reviewed the three interview guides respectively with a patient partner, a PCN, and a psychiatrist who were not otherwise involved in the study. The first author conducted interviews and observations and took field notes after each data collection session to record first impressions and propose links to previous interviews/observations. Interviews and observation sessions were audio-recorded and transcribed verbatim. Two additional researchers (EE, PR) also listened to two to three recordings to validate whether interviews captured all relevant elements of the practice and context. Finally, the first meeting with each local working group was audio-recorded while a professional research assistant took additional notes.

Implementation planning process

To operationalize the implementation planning process, we used CCM implementation tools available free-of-charge on the University of Washington Advancing Integrated Mental Health Solutions (AIMS) Center website. Thus, two implementation tools were adapted with permission from the AIMS Center: (1) Patient-Centered Integrated Behavioral Health Care Principles & Task Checklist, and (2) CoCM Behavioral Health Care Manager: Sample Job Description, Typical Workload & Resource Requirements [1135]. From those implementation tools, we constructed analysis tables to assess and compare each FMG’s nursing and collaborative care activities with professional activities described in the CCM (see Additional files 1–2). Table 1 presents the implementation planning process as conducted based on the first steps of the Grol & Wensing implementation of change model.

Data analysis

We first conducted an intra-case analysis, followed by inter-case analysis [3036]. For the intra-case analysis, meetings with local working groups served to build a case history regarding the formulation of implementation strategies. A systematic list of specifications was used to describe the case history for each FMG: the number of meetings conducted, the goal and attendance of each meeting, the main problem identified, stakeholders’ main concerns regarding the change in practices, and a summary of the implementation plan. We first conducted a thematic analysis of the case histories with an inductive approach, leading to the emergence of themes understood to have influenced the formulation of strategies to promote the adoption of care manager activities by PCNs. Field notes and results from the analysis of current practices were used to validate and enrich the initial case histories. Thereafter, we related the emerging themes to the contextual factors represented in the Consolidated Framework for Implementation Research (CFIR) [25]. One author (JBH) performed a thematic analysis for every transcribed document using the NVivo software. The interviewer (AG) reviewed the whole analysis process for each document, and both researchers discussed the main emerging themes in relation to the case histories until there were no remaining discrepancies in the interpretation of the CFIR’s contextual factors involved. For each FMG, we also constructed a schematic representation of the processes leading to care manager activities by PCNs to highlight the logical impact of specific contextual factors on those processes.

Regarding inter-case analysis, we created matrices to explore differences and similarities among emerging themes across FMGs. We also synthesized the relationship between contextual factors and the formulation of strategies to promote the adoption of care manager activities in a schematic form. The results’ interpretation was verified throughout the analysis through peer validation involving all the authors and by comparison with published theories and models. In particular, Weiner’s theory on organizational readiness for change was instrumental in clarifying the links between contextual factors and the process of formulating implementation strategies [37].

Characteristics of recruited FMGs and participants’ profile

Table 2 summarizes the main characteristics of each FMG. FMG01 and FMG03 were both university-affiliated, with a mandate and additional staff to train family medicine residents and primary care nurse practitioners. Nurses’ involvement in decision-making regarding their practices differed between FMGs. In FMG01 and FMG02, PCNs attended monthly team meetings to discuss their practices and to propose quality improvement projects to the medical director. In FMG03, PCNs met two to three times a year with their manager from the regional health center to discuss their practices and receive information about upcoming changes or programs.

Table 3 summarizes the profile of study participants. In addition to the members of each local working group, a total of 33 participants were recruited for interviews and/or observations (25 clinicians and eight patients), only five of which were male (two clinicians, three patients). Working group members could also participate in interviews and/or observations. Detailed information on each clinician (gender, age, years of experience) and patient (gender, age, mental and physical conditions diagnoses) are reported in a related paper [38].

Case descriptions

Influence of contextual factors on the formulation of strategies

Throughout the process of formulating strategies to improve the adoption of care manager activities by PCNs, working group members took various contextual factors into account either explicitly or implicitly. From the thematic analysis of data, we were able to highlight four perspectives from which those contextual factors were considered: (1) the gap between patient needs and the accessibility of external mental health services; (2) the PCNs’ work environment and the gap between nursing activities and care manager activities; (3) The PCNs’ perception of their competencies and of their local leaders’ interest in supporting them; and (4) the perception of the ability of the FMG to implement change in nursing practices for people with CMDs.

Illustration of contextual factors’ influence on strategy formulation

The four perspectives through which stakeholders considered contextual factors appeared interconnected in their impact on strategy formulation. Specifically, the first three perspectives appeared to create tension for changing nursing practices. This tension for change refers to the way PCNs and leaders perceived the situation as intolerable and the value they gave to solving the problem. The tension for changing nursing practices thereafter interacted with the perceived ability of the FMG to implement possible strategies to improve the adoption of care manager activities by PCNs. Together, tension for change and the perceived organizational ability to implement change appeared to influence stakeholders’ perception of the relative priority of implementation, which we refer to as “organizational readiness for change” as proposed by Weiner (2009) [37]. Organizational readiness for change is conceptually different from the “readiness for implementation” construct in the CFIR [25], which is similar to the perceived ability of the FMG to implement strategies. Indeed, we found that it is through organizational readiness for change that contextual factors ultimately drive the overall process of formulating and prioritizing implementation strategies. The influence of contextual factors on the formulation of strategies through organizational readiness for change is illustrated in Figure 1. Far from being static, organizational readiness for change evolved throughout implementation planning and strategy formulation.

Figure 1

Strengths and limitations

An important strength of this study resides in its originality in investigating a critical phase of the implementation process, i.e., implementation planning, which has received insufficient attention to date. Studying implementation planning helps shed light on limitations in the scientific literature regarding effective methods to support implementation in primary healthcare organizations, especially by providing an accurate assessment of their actual professional practices. In turn, practices analysis allows stakeholders to understand the gap between their practices and evidence-based models such as the CCM, facilitating the mobilization of organizational readiness for change and the formulation of contextually appropriate strategies. Our original perspective allowed us to balance promoting CCM-oriented evidence-based practices and mobilizing organizational capacity for change through a bottom-up approach to strategy formulation. To support the validity of our results, we referred to credible models from the research literature throughout the research process: Grol & Wensing’s implementation of change model to contextualize how the research team supported stakeholders in implementation planning; the CFIR to understand the contextual factors that would be expected to influence strategy formulation; and recommendations issued from the ERIC study to categorize the implementation strategies that were considered and prioritized in each FMG. The use of those models will facilitate comparison with results from other studies in other contexts, including when research methodology may differ.

Despite its strengths, some limitations of this study should be highlighted. First, when evaluating actual practices, we could not entirely benefit from the same analysis techniques and tools as a more traditional implementation study. Indeed, this study’s focus was not implementing the CCM or the role of care manager per se, but instead implementing strategies to enable clinicians to improve their practices and bring them closer to care management activities. When assessing actual practices, we did not work with stakeholders who explicitly understood their practices as CCM-related; the link between their practices and the model depended on the research team’s accurate assessment of their activities. While this contributes to our results’ transferability to quality improvement initiatives beyond the field of implementation science, we may have missed or misinterpreted some relevant activities during analysis. Second, although we provide an overview of study participants and a detailed description of the context, including only three FMGs may have limited the results’ direct transferability even though it was necessary to allow for in-depth analysis of current practices feasibly. In future studies, adopting a more quantitative approach to assess and compare participating organizations’ contextual characteristics would be a convenient way to involve more cases and rigorously test the relationships between contextual factors and strategy formulation that we observed in this study. One way to achieve this would be to exploit clinical registries and existing mental health indicators [44]. Third, our open-ended approach to strategy formulation through integrated knowledge transfer has led individual FMGs to consider strategies that did not necessarily involve techniques and tools that have previously been validated or shown reliable in clinical trials. In that regard, our study shares some epistemological principles with participatory research, although more directive regarding the objective of adopting evidence-based practices belonging to the CCM. Fourth, the fact that the research project ended before the proposed strategies were actually implemented leaves several questions unanswered, most notably whether the strategies for which there was sufficient readiness for change were likely to produce the improvement in practices that was sought. Therefore, this study should not be seen as proposing an alternative to previously published CCM implementation studies, for example. Conducting the whole implementation process remains as crucial as ever. However, we believe that our results should convince stakeholders involved in future studies to pre-emptively include a well-structured approach to implementation planning.