Background
Many nations are arguing for the need to increase comprehensive and coordinated primary care services in light of an aging population that is living longer and managing multiple chronic conditions [
2]. Primary care is widely regarded as the locus for access and optimization of care for older adults, and many countries are engaged in primary care reforms to meet populations’ needs [
3]. Despite several decades of reform aimed at increasing access to and coordination of primary health care services, the Canadian health care system remains fragmented and primary care is largely detached from secondary and community care [
4]. Older adults also face barriers in accessing health and social care systems to address their needs resulting in higher use of emergency rooms and hospitalizations [
5]. Furthermore, the care that they receive from primary care is reactive rather than proactive, and joint goal setting, action planning, and care planning rarely occurs [
6,
7]. Given that older adults experience multiple co-morbidities [
8,
9] they can benefit from person-focused, wellness-oriented, tailored approaches to care [
10] to support them to age in place. Interventions to address older adults’ needs as well as manage chronic diseases typically require the adoption of models of care that support interprofessional team- and community-based primary care for older adults [
11,
12]. This paper examines the implementation of an innovative team-based primary care model to address the needs of this population.
A primary care intervention was developed through engagement of interprofessional primary care team members, volunteers, community service providers, and older adults [
13,
14]. Health Teams Advancing Patient Experience: Strengthening Quality [Health TAPESTRY (HT)] is an innovative, coordinated approach to care which seeks to build on existing healthcare system strengths while attempting to address its challenges [
15]. HT is centred on meeting people’s health goals and needs to help them live at home healthier and longer. It is argued that complex interventions are needed with multiple interacting components to produce change [
16]. As such, HT is comprised of four components: 1) interprofessional teams, 2) trained community volunteers, 3) e-health technologies, and 4) increased linkages between primary care and community organizations.
Pairs of volunteers visit older adults at home, build relationships, and act as connectors between the patient and their primary care team. Volunteers receive training related to role expectations and activities which can help to ensure that they are not assuming a health professional role. Roles included forming relationships with clients, gathering patient information around health needs and health-related goals, assisting clients to set up a personal health record, sharing community resource information and providing motivational support. Training is provided to support each of these roles [
14]. The volunteer pair consists of an older adult and a younger adult, who is often a university student. This team collects data on older adults’ health needs and goals using questionnaires on a tablet (via an online application). The surveys assess mobility, physical activity, nutrition, memory, medication, and also include open-ended questions about life and health goals.
HT reports completed by the volunteers are sent electronically to the patients’ electronic medical record. Interprofessional primary care teams meet in weekly interprofessional “huddles” (small team-level communication) [
17] to triage HT volunteer reports and support care coordination. Huddle teams of 5 to 6 providers engage in system navigation tasks as they facilitate older adults’ access to primary care and community-based programs and services. Communication occurs between the most responsible physician (MRP) and other clinicians beyond the huddle as needed. Huddles are not part of usual practice. However, interprofessional teams do participate in informal consultations among team members, and all primary care team members attend monthly academic rounds.
A randomized controlled trial was conducted to evaluate implementation and effects of the HT intervention [
15]. Guidance provided by the Medical Research Council indicates that it is important to understand both outcomes and processes of any new complex intervention [
16,
18]. Having a theoretical understanding is needed to explain how an intervention might be causing change, since effects can be influenced by implementation failure or success [
16]. Furthermore, without a deep understanding of implementation processes, effective interventions have small likelihood of being applied into real world settings [
19].
There is growing interest in applying theories and frameworks to explore implementation [
20]. We used Normalization Process Theory (NPT) to examine the implementation of HT’s four components [
21]. This theory has been widely used to explore implementation of health care interventions, specifically in primary care contexts [
19,
22,
23]. As a formal mid-range theory it is particularly appropriate for this study. It was developed to address observed challenges in integrating new interventions and reorganizing care delivery in health care settings [
21]. NPT posits that complex interventions are implemented via processes and integrated via structures in a professional and organizational context. Normalization is a result of the work people both individually and collectively undertake to implement and embed an innovation. NPT consists of four constructs [
21].
Coherence (i.e., sense making) refers to how everyone individually and collectively understands the intervention, its purpose, and potential value. It also includes how individuals and the team see the intervention as differing from usual care delivery.
Cognitive participation (i.e., enrolment) refers to individual and leadership buy-in of the intervention and agreement to work with it.
Collective action (i.e., enactment) is operational; people have the needed resources to undertake new practices, and everyone knows who is doing what.
Reflexive monitoring (i.e., appraisal) includes feedback gathered on outcomes and impacts of the intervention, which can reinforce its continued application and adaptation.
A 2017 integrative review applied NPT to explore barriers and enablers in implementing interdisciplinary team working in primary care [
23]. The review showed that there was a paucity of research that examined all four NPT core constructs, other than collective action. It also showed that most implementation studies look at a few team members without considering the full team [
23]. The current study, therefore, aims to examine the implementation process of the HT intervention in relation to teamwork including all team members (providers, clinical managers, and volunteers) and patients. Results may inform others implementing complex interventions involving primary care teams.
Research questions
The primary care team was conceptualized as consisting of: 1) small interprofessional huddle teams, who received and reviewed the HT patient reports, 2) primary care team members outside the huddle teams, and 3) trained HT volunteers. The research questions were:
1.What are the perceptions of patients, volunteers and health care providers on the implementation of HT in relation to the work of interprofessional teams?
2.How has HT implementation affected existing structures and processes of primary care with the teams?
Methods
Design
This study used a descriptive qualitative approach [
1] to explore implementation processes of HT. The study was embedded in a mixed-methods, pragmatic randomized controlled trial [
15].
Setting
This study was situated in two primary care practice sites in a large urban setting in Ontario, Canada. McMaster Family Health Team, is a multidisciplinary team providing 7-day-a-week care, supported by an electronic medical record. Physicians are remunerated through blended capitation and affiliated with a partially provincially funded (salaried) interprofessional team [
24]. The family health team has approximately 37,000 patients, 21 family physicians (full-time equivalent), 28 family medicine residents, 8.5 nurse practitioners, 5.5 registered practical nurses, and 26 full- or part-time other health care professionals (occupational therapists, physiotherapists, social workers, dietitians, pharmacists, psychologist, system navigators, physician assistants, a lactation consultant, and a chiropodist).
Sample and recruitment
Participants included: health care providers, clinical managers, administrative assistants, HT volunteers, and a HT volunteer coordinator. Invitations to participate were emailed to all family health team members of the weekly interprofessional “huddle” teams, clinic managers, nurse practitioners, as well as physicians and medical residents at each site who had 10 or more patients participating in the Health TAPESTRY (HT) program. All HT volunteers and the coordinator were invited via email to participate in focus groups while the volunteer coordinator was invited to an individual interview. Initially, all patients who completed the HT intervention were invited to participate in an interview, with the goal of selecting 30 participants equally distributed by gender, clinic site, and across 2 age groups (70–79 and 80+). Selected HT patients were recruited by phone. We found this was resulting in a healthier sample of older adults based on their HT survey results, therefore, we selected from individuals identified with greater health needs based on their survey responses.
Data collection
Data were collected at two time points: 4 months (June and July 2015) following the start of the intervention and 12 months (February and March 2016) for all participants except for HT patients who were interviewed at the end of the 6 month intervention (November 2015 to February 2016). Phone interviews averaging 30 to 45 min were held with clinic managers, the volunteer coordinator, and physicians who were not part of the huddle team or could not attend a focus group. One-hour focus groups were conducted with huddle teams, physicians and medical residents at their workplace. Volunteer focus groups were held separately with older adult and student volunteers in an accessible room in the community. One-on-one interviews were held with patients in person at clinic or on the phone.
Data collection was completed by a nursing researcher and Department of Family Medicine research staff experienced in qualitative methods (LC, RV, JG, FP, MB, NF, DJ). All were female. A few providers were known to some research staff since they had interacted through previous studies. Participants were informed that the study was focused on learning about their experiences implementing the intervention, their thoughts on how well the program was working, impacts and outcomes, and program sustainability. Interview guides (available on request) were informed by NPT’s core constructs and questions were altered slightly based on participant type and divided into sections in keeping with the topics noted above. Patient interview questions focused on their understanding of HT, their experiences with the program, and perceived impacts on communication with the team, coordination of care, and sustainability. Participants received a $25 CAD gift card as a token of appreciation. Light refreshments were provided at focus groups. All interviews and focus groups were digitally recorded and transcribed verbatim.
In addition, a researcher (LC) took field notes during huddle meetings to track changes in the interprofessional team structure and processes and noted challenges and enablers in implementation. Any significant challenges were shared with the implementation team to provide them an opportunity to adjust their processes.
Results are organized according to NPT constructs. Quotes are used to support themes marked by the source as follows: Huddle team member = [Hud-1], [Hud-2], etc.; primary care team members other than the huddle team = [PC-1]; Volunteer = [Vol-1], etc.; and patient = [Pat-1], etcetera.
Analysis
Authors RV and LC have extensive experience in qualitative analysis. The coding structure was created based on interview questions and the intervention components (interprofessional teams [huddles and the larger primary care team]; volunteers; technology; and system navigation). Once the initial structure was developed, it was discussed amongst coders (FP, JG, NF) who then individually and inductively coded transcripts using NVivo Version 10 software [
25]. RV and LC supervised coding of the initial transcripts. They met over 4 to 5 half-day meetings held monthly during the active coding period to discuss meanings of codes and ensure consistency. After all transcripts were coded, RV and LC checked random selections of transcripts. Using constant comparison, they further refined the coding structure and developed themes and their elements (sub-themes). In the last phase of analysis, RV and LC met over multiple meetings to move themes and/or sub-themes under the constructs of NPT [
21] which served as sensitizing concepts to guide the final analysis [
26,
27]. The full research team met on 3 occasions to review the coding structure and reach consensus on final themes and their elements.
Discussion
This study used NPT to examine how an innovative primary care intervention aimed at older adults was implemented by an interdisciplinary team enhanced by trained volunteers. Using NPT increased our understanding about how providers and patients individually saw the HT intervention in comparison to usual care in general and in relation to teamwork (Coherence), how the team collectively bought into the new model of care (Cognitive Participation), how providers put the intervention into action (Collective Action), and how providers and patients appraised it (Reflexive Monitoring).
In relation to Coherence, participants in our study understood a purpose of the HT program was to strengthen interprofessional care delivery including volunteers as an expansion of the team as well as aiming to improve care coordination by the team. Their understanding of the intervention may be explained by HT clinical managers’ (i.e., lead physicians) who demonstrated support of the HT intervention through regular staff discussions in clinical rounds and other meetings. An integrated review of interdisciplinary working in primary care found that a barrier to collective teamwork was a physician training which is focused on the relationship of the doctor-patient dyad as well as the physicians’ overall sense of responsibility for patient care rather than sharing it with the team [
23]. However, similar to our results, the review also found that having senior physicians or local champions facilitate interprofessional team working enabled buy-in for interdisciplinary work [
23] and should be encouraged.
Under the construct of Cognitive Participation, huddle team members had a better understanding of each other’s roles and saw benefits in working as a team. Having dedicated meeting time to review patient charts and plan care for patient care in huddles appeared to contribute to this understanding. Fiscella and colleagues supported the notion of huddles to increase teamwork, as they found that these “teamlets” flourish when they have a shared cognitive model leading to an increased understanding of roles [
28]. This requires trust, cooperation, communication, and feedback, resulting in better coordination of tasks. Case studies of innovative primary care practices in the Unites States also found that huddles helped team members coordinate tasks more efficiently [
29]. Although huddles improved interdisciplinary practice, more work is needed to understand how to draw in team members outside of huddles.
The current study found that providers outside of the huddle team, namely physicians and medical residents, had difficulties understanding their roles in the HT intervention. This was mainly related to workflows which were not conducive to MDs (versus select MD champions) attending huddle meetings and ensuring that the huddle team regularly and effectively communicated with the MRPs regarding patient care planning. It appears that HT intervention developed a
provider huddle team rather than the
patient’s team. An effective
patient’s team would need to consistently include the MRP for case conferencing, as well as the patient and/or their caregiver, and the volunteers. An effective
patient’s team would engage all of these team members in care planning communication and follow up. Huddle teams were an effective strategy to get teams to work together; however, they were not necessarily effective in planning an individual’s care in the absence of all providers and family members who are members of the individual’s care team. Given the current focus for health systems to improve person-centred and more recently people-centered care [
30], this was significant oversight in the HT intervention and requires enhancement in future HT implementation.
Researchers have found mixed levels of buy-in for interdisciplinary working by certain team members (i.e., physicians, pharmacists and nurses) with a tendency of physicians to resist collaborative teamwork, giving preference to working independently [
23]. Primary care researchers explored contextual factors influencing team performance in five jurisdictions in three countries and found that despite moving to collaborative teams, primary care clinicians still tended to work in parallel, infrequently using team approaches to solve clinical problems [
12]. They also found that team-based primary care was enhanced when direct physician involvement was not required in care provision. The huddle teams aimed to overcome these challenges with some success. However, similar to the challenges experienced by physicians and medical residents who were typically working outside of the huddle in our study, volunteers also largely worked on the periphery of the interprofessional team, expressing the desire to be more connected to the primary care team. Fully integrating lay community health workers into the primary care team has been shown to be a challenge by others [
31]. To improve future HT implementation, strategies are needed to strengthen multi-directional communication and engagement among all primary care team members including physicians and volunteers.
Under the Collective Action construct, our results showed that there were a number of changes in team processes that occurred to support implementation of the HT intervention in relation to teamwork. These included changes in care coordination and case management processes, a shift to more proactive approaches in care provision by the team and refining patient information flow and content (e.g., more comprehensive patient information and focus on priorities based on patient’s goals). Based on field notes, these processes were refined over time requiring extensive teamwork. Researchers have found that establishing protocols can help team-work [
23,
32]. One example in the current study was through the huddle’s creation of a decision path to support the triage process (Fig.
1). Despite these achievements, multi-directional information flow among team members needed improvement, which may have been strengthened through the development and full implementation of protocols to clarify roles and communication processes among all team members.
As seen in the current study, time and space resources were required to ensure team collaboration. This has long been supported by research that shows financial support for team activities has a positive effect on team performance [
12,
23]. Despite the need for resources to support team activities, results showed that teams and patients perceived that HT was valuable, embedded well into the clinic environment, and worth maintaining. HT improved team communication within huddle teams and patients were satisfied overall with the intervention. Having a researcher (LC) attend huddle meetings and share her reflections with the team about implementation challenges supported reflective monitoring by the team. Given challenges in linking the providers with volunteers in the circle of care, there may be value in holding regular full team meetings to reflect on work processes and enhance role clarity. An international study found that a) outreach facilitators charged with driving change in primary care by “creating ‘peer pressure’, modelling good communication, encouraging reflection, supporting momentum and providing accountability” (p. 283) in the United States, b) facilitators who worked on quality assurance in Australia, and c) leadership training for practice leads and additional resources provided for clinical administration and meetings in Canada supported primary care reform processes that fostered team-oriented family practice [
12]. These strategies may be valuable in spreading the HT model in other primary care settings.
This study had a number of strengths and limitations. Data was gathered from various sources (patients, providers, volunteers) and via various types (interviews, focus groups, and field observations) to support triangulation. Data was gathered prospectively over the course of the intervention. A robust coding strategy was developed over many meetings and reached agreement on final results within the large interprofessional research team. Primary care clinician scientists on the team supported peer debriefing and interpretation of results from a clinical vantage point. However, this could have introduced biases. The team attempted to prevent this by ensuring that a strong audit trail of decisions was maintained and having non-clinician researchers and research staff develop the initial themes supported by quotes. This intervention was tested in two sites of a Family Health Team in one urban area limiting generalizability.
Conclusions
This study provides important insights about factors to consider when implementing interprofessional primary care team processes within a new model of care such as HT. The application of NPT constructs provides guidance for future implementation of primary care innovations involving interprofessional teams. Since this study was launched, NPT has become available as a Toolkit with a 23-item survey for use by teams implementing and evaluating complex interventions (
http://www.normalizationprocess.org/nomad-study/). In this study, application of NPT constructs enriches our understanding of phases and timelines required for implementation of new approaches in care by interprofessional teams. Our data was collected at 4 months and 12 months into the implementation of a complex intervention, demonstrating that the normalization of new practices and processes were well underway. Given that timeframes for adoption of healthcare innovations may take years, this timeframe is relatively short to show evidence of normalization.
This study indicates that greater attention should be paid to the importance of the Cognitive Participation construct and focus on implementation strategies that would increase involvement of MRPs in interventions as a member of the patient care team as well as strengthening linkages of MRPs in huddle teams that may not directly involve them. This study also points to the important role of volunteers in primary care interventions, to suggest that these roles can also be integral to team-based and community-based care of older adults. A greater emphasis on clarity of roles, mechanisms for enhanced communication and strategies to ease workflow challenges should be carefully considered, planned, and implemented in normalizing primary care interventions among team members.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.