Trial and error,together: divergent thinking and collective learning in the implementation of integrated care networks

Abstract

Hybrid networks that link disparate professionals and organizations are a common approach to deliver integrated care to patients. Recent literature argues that successful implementation of these networks demands a socio-cognitive perspective in which stakeholder mental frames and thought processes are prioritized, investigated, and compared. The aims of this article are to identify where mindsets diverge among clinical and managerial stakeholders involved in the implementation of integrated care networks known as ‘Health Links’ (HLs) in Ontario, Canada, and to describe strategies to support stakeholders’ capacity to collectively learn and develop more convergent views. Drawing from shared mental model theory and practice-based learning theory, a secondary analysis was conducted of interview data with 55 healthcare professionals and managers involved in the implementation of HLs. We identified examples of divergences in stakeholders’ conceptualization of the HL design and approach (‘strategy mental model’) and their perceptions of each other and how they work together (‘relationship mental model’). We also identified four strategies that facilitate learning and possibly mental model convergence. The results of the study may help guide stakeholder dialogue towards collective learning and coordinated action for integrated care delivery.

Points for practitioners

The findings suggest that in the implementation of large-scale change involving multiple stakeholder groups, there are predictable areas where divergent views are likely to occur and may have a negative impact on coordinated action. An awareness of these potential divergences can guide practitioners to examine them explicitly and regularly, and to proactively develop strategies to support practice-based learning and the development of a convergent perspective.

Keywords

healthcare hybridity implementation integrated care learning networks policymaking shared mental models

Introduction

Healthcare delivery is largely organized by specialization and composed of independent professions and organizations that rarely function as a unified system (Glouberman and Mintzberg, 2001). This fragmentation is problematic for patients with complex healthcare needs such as those with multiple chronic conditions. These patients require services from numerous professionals across diverse care settings such as primary care clinics, hospitals, specialist clinics, long-term care facilities, and social service agencies (Bodenheimer, 2008). Frequent transitions across care settings contribute to inconsistent patient monitoring, duplicative tests, delays in diagnosis, and medication errors, which can threaten patient safety, quality of care, and health outcomes (Bodenheimer, 2008). Patients with complex healthcare needs thus require care that is integrated across professionals and settings over time (Singer et al., 2011).

While a fully integrated healthcare system where professionals and organizations share a single mission, resources, and patient information is ideal (e.g. Kaiser Permanente), it is often challenging or impossible to achieve (Denis et al., 2011). Instead, most healthcare systems introduce new policies, financial incentives, and one-time grants to stimulate the development of hybrid networks that link disparate healthcare professionals and organizations (Evans et al., 2013). The term ‘hybrid network’ refers to an entity that links different sectors, organizations, and/or stakeholders with diverse and often conflicting belief systems and practices (i.e. institutional logics) (Skelcher and Smith, 2015; Thornton and Ocasio, 2008). Prominent examples of hybrid integrated care networks include Accountable Care Organizations and Patient-Centered Medical Homes in the US, and Integrated Care Pilots in England. Hybrid integrated care networks have demonstrated a range of positive outcomes, including reduced emergency department (ED) visits and nursing home placements, lower institutional costs, higher patient satisfaction and improved health outcomes (Bardsley et al., 2013; Curry and Ham, 2010; Ouwens et al., 2005; Wodchis et al., 2015). However, there is considerable variability in the success of hybrid integrated care networks, and the literature continues to emphasize the challenges in cultivating collaboration across diverse stakeholder groups (Kreindler et al., 2012; Ling et al., 2012; Pate et al., 2010; Tsasis et al., 2013; Williams and Sullivan, 2009).

Delivering integrated care requires organizations and professionals to work together in a coordinated fashion, sharing relevant medical information and making care decisions that are consistent with and informed by the care delivered by other professionals (Singer et al., 2011). The literature suggests that integrated care delivery requires not only structural and process improvements, but also change in the mindsets and behaviours of stakeholders such as managers and healthcare professionals (Denis et al., 2011; Evans and Baker, 2012; Ferlie et al., 2005). In order to deliver integrated care, stakeholder perceptions of their roles, relationships, and practices must evolve alongside clinical, organizational and environmental changes (Denis et al., 2011). Perceptions play an integral role in the delivery of integrated care because they provide the underlying logic for action (Walsh, 1995). As such, the successful implementation of hybrid integrated care networks demands a socio-cognitive perspective in which stakeholder perceptions are prioritized, investigated, and compared (Evans and Baker, 2012). If stakeholders have conflicting ideas about the services to be integrated, the steps and partners involved, or the underlying purpose, they may be working towards different visions and their interactions may be disorganized and unproductive (Evans and Baker, 2012).

Research across disciplines confirms the importance of shared views in the execution of coordinated action and has used a variety of terms to capture this phenomenon, including ‘strategic consensus’ (Kellermanns et al., 2011), ‘cross-understanding’ (Huber and Lewis, 2010), ‘common ground’ (Dewulf et al., 2011), ‘congruent understandings’ (Vlaar et al., 2006), and ‘shared framing’ (Fiol, 1994). In public administration, synthesizing the perspectives of diverse actors is a common challenge, particularly in the implementation of initiatives – like integrated care – that require coordinated action (Dewulf et al., 2011; Rein and Schon, 1996). Although numerous studies of integrated care involve comparing stakeholder perspectives, divergences in views have been identified and reported using varied methods and language (Jiwani and Fleury, 2011; Kreindler et al., 2012; Ling et al., 2012; Pate et al., 2010; Tsasis et al., 2013; van Wijngaarden et al., 2006; Williams and Sullivan, 2009). The lack of consistency across studies makes it challenging to communicate and accumulate knowledge regarding divergent mindsets. For example, we lack knowledge regarding which divergences in understanding are most common and impactful, and what strategies leaders can use to facilitate convergence of thinking across stakeholder groups involved in integrating care. The Integration Mindsets Framework was developed to support the conceptual and empirical advancement of this line of inquiry (Appendix A) (Evans et al., 2014a).

The Integration Mindsets Framework draws from theory, empirical evidence, and practice to identify key areas where a lack of shared understanding may significantly hamper efforts to integrate care (Evans et al., 2014a). The distinguishing feature of the framework is its roots in industrial psychology, specifically shared mental model theory. In the literature on integrated care, cultural differences are often offered as explanations for failed or suboptimal initiatives (e.g. Friedman and Goes, 2001; Pate et al., 2010; Suter et al., 2009). An alternative approach to the dominant focus on general cultural attributes is to examine individual and shared mental models specific to integrated care.

The aim of this article is twofold. First, we identify where mindsets diverge among clinical and managerial stakeholders involved in the implementation of hybrid integrated care networks known as ‘Health Links’ (HLs) in Ontario, Canada, and map these divergent mindsets on to the Integration Mindsets Framework. Second, we identify strategies from the HLs case that may enhance stakeholders’ capacity to learn and develop more convergent views.

Theories: shared mental models and practice-based learning

To guide this study, we draw from two theories: shared mental model theory (Mohammed et al., 2010) and practice-based learning theory (Raelin, 1997). Mental models are internal ways of thinking about specific tasks or situations. They develop over time through experience, direct communication and interaction with others, and vicarious learning (Fiske and Taylor, 1991). When multiple individuals develop a common psychological understanding of a task or situation this is referred to as a ‘shared mental model’ (SMM). SMMs fall into three broad categories: task-related (goals and performance requirements), team-related (interpersonal interaction requirements and skills of team members), and beliefs (preferences or expectations) (Mohammed et al., 2010). SMMs in these areas allow individuals to behave in ways that are consistent and coordinated with each other in the completion of interdependent tasks (Cannon-Bowers and Converse, 1993; Mohammed et al., 2010). Identical mental models are not necessary or feasible; rather, the aim is a level of consensus that is broad enough to accommodate differences (Mohammed et al., 2010). Empirical research over the past 20 years, including a meta-analysis of 65 studies, confirms a positive relationship between SMMs and team performance (DeChurch and Mesmer-Magnus, 2010; Mohammed et al., 2010).

Although SMMs are team-level phenomena, they can manifest and impact performance beyond the team level (Hysong et al., 2005; Kellermanns et al., 2011; Vlaar et al., 2006). For example, in a study examining healthcare personnel’s mental models of the organizational implementation of clinical practice guidelines, personnel in high-performing facilities exhibited SMMs, while those in lower performing facilities did not (Hysong et al., 2005). Constructs such as ‘industry mindset’ (Phillips, 1994) and ‘macrocultures’ (Abrahamson and Fombrum, 1994) suggest that SMMs also shape inter-organizational behaviour and system-level performance. In this vein, Evans and Baker (2012) propose that SMMs can influence inter-organizational or ‘health system’ performance, particularly in the implementation of large-scale change involving multiple stakeholder groups. The development of hybrid integrated care networks is a prime example of such a change, and previous work by the authors applies SMM theory to such integrated care efforts (Evans and Baker, 2012; Evans et al., 2014a; Evans et al., 2014b; Tsasis et al., 2012). For a thorough justification of the importance of congruence and the value of SMM theory to integrated care efforts, please refer to Evans and Baker (2012).

McComb (2007) proposed that mental model convergence occurs via three phases: orientation, differentiation, and integration (McComb, 2007). First, stakeholders orient themselves to the task at hand and those involved in its execution. Second, stakeholders use the information gathered during orientation to explore similarities and differences in their mental models. Finally, new information is integrated into their existing mental models, and as more integration occurs, stakeholders’ mental models may become more similar. This convergence process emphasizes the importance of identifying and understanding mental model divergences (the differentiation phase) as a means to understand how convergence does (or does not) unfold. Understanding the differentiation phase is particularly important in the development of hybrid integrated care networks where divergences are likely and can derail progress before opportunities for reconciliation emerge.

Mental model convergence inherently involves learning (McComb, 2007), defined as the development of new and diverse interpretations of events and situations (Fiol, 1994). Learning is triggered when two pieces of conflicting information are brought together or through sensemaking in response to ambiguous situations (Argyris, 1994; Collins, 1998; Weick, 1995). Collective learning involves developing enough consensus around diverse interpretations for organized action to result (Fiol, 1994). However, evidence suggests that the boundaries between professional groups and organizations produce strong social and cognitive barriers that facilitate learning and change within boundaries, but make knowledge ‘sticky’, or difficult to move, across boundaries (Carlile, 2004; Elwyn et al., 2007; Ferlie et al., 2005; van Wijngaarden et al., 2006). Cognitive disorder and ambiguity are thus common in the implementation of integrated care networks as individuals struggle to ‘make sense’ of other professionals and organizations, and of their current and future relationships (Denis et al., 2009; Evans and Baker, 2012; Williams and Sullivan, 2009).

The central premise of practice-based learning theory is that ‘learning by doing’ enables professionals to bridge the gap between explicit and tacit knowledge, and transcend ‘sticky’ organizational and professional boundaries (Raelin, 1997). Practice-based learning emphasizes context and is rooted in the coordinated activities of individuals and groups doing their ‘real work’ (Raelin, 1997). It is through this process that diverse professionals and organizations can learn to work together and build the SMMs necessary to deliver integrated care (Tsasis et al., 2013; van Wijngaarden et al., 2006).

Taken together, SMM theory and practice-based learning theory suggest that mental model convergence relies, in part, on the extent of opportunities and support for ‘learning in practice’. Building on these theories, Evans et al. (2014a) developed the Integration Mindsets Framework (Appendix A). An ‘integration mindset’ is an individual’s way of thinking (i.e. mental model) about integration that is based on knowledge and beliefs regarding (a) what is being integrated and how, why, and for whom it is being integrated (i.e. Strategy Mental Model) and (b) the stakeholders involved in integration and how they are connected (i.e. Relationships Mental Model) (Evans et al., 2014a). The Integration Mindsets Framework identifies and prioritizes specific mental models whose convergence or divergence across stakeholder groups are most likely to influence inter-professional and inter-organizational relations in integrated care efforts (Evans et al., 2014a). The Integration Mindsets Framework has not yet been applied in studies of integrated care.

Methods

We conducted a secondary analysis (Heaton, 2008) of semi-structured interview data from a study on the implementation of Health Link networks across Ontario, Canada (Evans et al., 2016). The Health Links (HLs) are a provincial initiative launched by the Ontario Ministry of Health and Long-Term Care (‘the Ministry’) in December 2012 with the aim of integrating care for patients with complex health and social care needs. An HL is a hybrid inter-organizational network consisting of health and social services organizations who voluntarily come together to plan and deliver integrated care for patients with complex needs. For more information on the HLs initiative, please see Appendix B.

The HLs provide an ideal opportunity to identify important divergences in mental models of integrated care. The Ministry gave few rules on how these voluntary networks should be designed and how they should operate, which increased the likelihood of divergent mental models across stakeholder groups (Angus and Greenberg, 2014; Grudniewicz et al., 2018). The HLs thus serve as an ‘extreme case’ or ‘critical case’ capable of generating rich data. Because of the ‘low rules’ nature of the HLs, inter-network meetings were convened regularly to bring together representatives from multiple HLs to exchange knowledge and coordinate approaches. These meetings provided opportunities for those involved to progress through the phases of mental model convergence (McComb, 2007).

Between August 2014 and February 2015, we interviewed healthcare professionals and managers from HL networks and Local Health Integration Networks (LHINs). The province of Ontario is divided into 14 geographic regions, each of which is serviced by an LHIN (a regional governance body that plans, funds, and coordinates care). The LHINs were tasked with supporting the implementation of HLs in their region.

We recruited LHIN employees using a Ministry mailing list of personnel responsible for the HLs. Using a semi-structured interview guide, LHIN representatives were asked to describe their role in implementing and supporting the HLs, success factors and challenges, and similarities and differences across their HLs. We asked LHIN representatives to identify healthcare professionals and managers actively working in HLs in their region. We stratified recommended individuals by HL and profession and purposefully sampled participants to ensure breadth of representation. HLs were purposefully sampled to maximize variation in geographic location, implementation stage, and lead organization. Recommended individuals were also purposefully sampled to ensure that multiple individuals per HL network were invited to participate, including administrative and clinical stakeholders from at least two of the partnering organizations in each HL. We asked HL representatives to describe implementation progress, success factors, and challenges. The interviews were conducted one-on-one or in small groups, and were audio recorded and transcribed verbatim.

We conducted a secondary analysis of interview data from the study on HLs described above and published elsewhere (Evans et al., 2016; Grudniewicz et al., 2018). Secondary analysis involves re-examining data collected in a previous study for a new or related purpose (Heaton, 2008). We analysed an issue that emerged, but was not fully addressed in the primary study by reviewing a previously coded set of data under the theme of ‘Commitment to Learning’, which captured values and practices that supported the ongoing development of new knowledge among participants. We also re-coded the entire dataset using the Integration Mindsets Framework. In the secondary analysis, we focused on identifying (a) strategies to support practice-based learning, and (b) divergent mental models (because HLs were still in early implementation, they had greater likelihood of being in the orientation and differentiation phases of mental model convergence, rather than the integration phase (McComb, 2007)).

Results

We analysed data from interviews with 55 healthcare professionals and managers. Twenty-six participants were LHIN employees and the remaining 29 were directly involved in one or more HL networks. All 14 LHINs and 38 of the 56 then-active HLs (68 percent) were represented in our sample. HL participants came from a variety of organizations including primary care practices (48 percent), hospitals (35 percent), and community-based organizations (17 percent).

Overall, participants conceptualized the implementation of HLs as a learning process, involving trial and error and gradual adaptation until inter-professional and inter-organizational collaboration becomes ‘the new normal’ (HL Interview 6). However, while this process of learning was organic and energizing for some stakeholders, it was slow, contested, and challenging for others. Below, we discuss divergences in stakeholders’ conceptualization of the HL design and approach (‘strategy mental model’) and their perceptions of other stakeholders and how they work together (‘relationship mental model’). This is followed by a description of four key strategies from the HLs for fostering practice-based learning.

Strategy Mental Model

In the Integration Mindsets Framework, a Strategy Mental Model is a ‘conceptualization of what is being integrated and how, why and for whom it is being integrated’ consisting of six inter-connected knowledge elements and four belief elements (see Appendix A for details) (Evans et al., 2014a).

Response to the Ministry’s low-rules implementation strategy

Several divergences were evident in how stakeholders conceptualized the HLs design and approach, demonstrating divergent ‘strategy mental models’ (see Appendix C for a summary of all divergent mental models found in the data). This variation may be attributed to the Ministry’s ‘low-rules’ approach to the design and implementation of the HLs. To encourage local innovation, the Ministry allowed stakeholders to determine how to design and operate their HL (Angus and Greenberg, 2014; Grudniewicz et al., 2018). Participants were divided in their response to this approach, with many agreeing that it promotes innovation, but a significant minority arguing that it perpetuates duplication and confusion. The following two quotes illustrate this divide:

One of the things that we have really appreciated … is the low rules environment. This is really one of the first times that we’ve really felt that there is that opportunity to be creative, to be innovative, to provide some local solutions … we don't hesitate to think of things that are out of the box. (Administrator, LHIN stakeholder, Interview #1)

Our Health Link teams have found it difficult to work with this ambiguity … So we are taking somewhat of a unique approach in identifying what work needs to happen on a regional basis as opposed to 9 different solutions with 9 different Health Links. So we’ve adopted a strategy for planning on a regional basis. (Admin, LHIN3)

Aims and expected outcomes of the HLs

Participant perceptions also diverged regarding the aim of the HLs. For example, one participant said, ‘What is the outcome you’re trying to get? Maybe we don't know, and we’re just going to keep striving to drive costs down or control cost and improve care. Maybe that is the goal’ (Clinician, HL18). However, others noted that the HLs helped to establish common goals and expectations at a local level: ‘So the hospital, primary care, and CCAC [Community Care Access Centre, an organization that coordinates and delivers home and community services]. That’s the first time we all sat down at the table with a common purpose’ (Admin, HL19).

Participants also had varied views regarding measurement of desired outcomes, as the following two quotes demonstrate:

And when we start to look at what Health Links can look like in the future and where they need to go, we should start measuring that sooner as opposed to later. (Clinician, LHIN11)

We are so concerned about measurement, that it sometimes takes away from our capacity to be innovative. So if we can’t measure it, we won’t do it anymore. People’s understanding of what performance measurement is, is so varied. (Clinician, HL4)

Target patient population

As a result of the ‘low-rules’ policy, HLs independently identified their target patient population and, correspondingly, which sectors and organizations would be members of their network. Some HLs focused on ‘high-cost’ patients with high ED utilization. Other HLs identified patients with multiple co-morbidities or mental health and addiction issues. Still others focused on vulnerable or under-served populations in their regions such as Aboriginal, Francophone, the frail elderly, or the homeless population. One participant described their approach as ‘organic’ and said it involved asking ‘primary care providers “Who's keeping you up at night? Who do you think we need to support?”’ (Admin, LHIN11). Some participants disagreed with the Ministry’s focus on the top 5 percent of users of the healthcare system: ‘We’re actually missing out on who we really need to address. We need to hurry up and get further downstream and address the other 10 to 15 percent of the population. And if we don't do something with them urgently and dramatically right now, they will be tomorrow’s 1 to 5 percent of the population’ (Clinician, LHIN7).

Health link membership

HL network membership was influenced by characteristics of the patient population, as well as the extent to which participants adopted a medical versus social model of health. For example, one participant said:

We don’t apply a medical model. We look to social supports as a factor. Where you’re driven by a non-medical model, you’re getting more of that richness around broad community engagement and problem-solving. And not to take away from the great work that [HL name] is doing because it’s phenomenal. It’s just there’s been some active hospital participation, and with that is a more traditional view around the players of health. (Clinician, HL5)

Influence of resources

All participants identified resources as a challenge to HLs’ implementation and sustainability. However, they did not frame the resource issue in the same way. Some viewed a lack of resources as a barrier to integration. ‘You have to have the capacity and the resources to pull it off’ (Admin, HL23), and another added ‘Funding, stable funding. I mean that’s the biggest fear I have. I am afraid we’re going to promise a lot and end up delivering little’ (Clinician, HL3). However, other participants argued that a lack of resources was a driver and contributed to the need to integrate and innovate; this view was more common among rural HLs. For example, a physician leader recounted a conversation with peers, ‘I kept saying, guys, there’s no more money in the system. This is about working differently’ (Clinician, HL6). Another participant clarified that ‘It's not that everyone had extra money or capacity lying around. It’s been a leadership choice to realign because they saw the importance of this’ (Admin, HL7).

Relationship Mental Model

In the Integration Mindsets Framework, a Relationships Mental Model is a ‘conceptualization of the organizations, groups, and individuals involved in integration and how they are connected’, consisting of five inter-connected knowledge elements and four belief elements (see Appendix A) (Evans et al., 2014a).

Selection of lead organizations

Each HL selected a lead organization, and in some cases co-leads, for their network. The rationale for who should lead differed across HLs, with some indicating that HLs should be led by: (a) hospitals for their capacity and infrastructure, (b) primary care practices to facilitate access to patients and enhance primary care engagement, and (c) Community Care Access Centres for their experience in intense case management and care coordination. Some participants argued that any organization could aptly lead an HL as long as they had a history of change and innovation: ‘The lead organization [must be] willing to try new things, challenge former beliefs and assumptions, and be open to creativity. That’s what makes the bigger difference versus who the lead organization is’ (Admin, LHIN2).

LHIN role

Another point of contention was the role of the LHINs in the implementation of the HLs. Some LHINs viewed themselves as ‘leaders’ in the HL initiative and thus had a hands-on approach involving standardization of structures or processes in the HL networks. Other LHINs took a backseat and saw themselves as ‘catalysts’ providing support as needed. In general, the role and accountability of the LHINs was unclear, as this participant described: ‘We weren’t quite sure who was doing what. There was a lot of communication coming out initially from the Ministry, directly to providers out in the field, not necessarily inclusive of the LHIN. It made it a bit confusing for everyone. I don’t think initially there was enough clarity around the role of the LHIN and using the expertise and relationships that the LHIN already has [on integrated care]’ (Admin, LHIN4).

Patient role

Patient engagement was a cornerstone of the HLs model, embedded in the Ministry requirement that all HL patients have ‘individualized’ coordinated care plans. However, the degree of comfort among healthcare professionals regarding patient involvement and patient empowerment varied, with some feeling that it was fundamental to success and others feeling that it was resource-intensive, inappropriate, or threatening. This theme was best captured by the following quote: ‘I’m a big champion of person-centred care and letting patients be in the driver’s seat of their healthcare. But healthcare providers have varying tolerance of that. Some find it very threatening that the patient’s chart and their coordinated care plan is theirs, not mine. Getting them to hear patients’ stories and talk to patients in a different way has been fun when it works and struggling when people won’t even try’ (Admin, HL6).

Strategies to support practice-based learning in hybrid integrated care networks

We identified four key strategies that facilitate practice-based learning and may support mental model convergence in the HLs. These varied in nature and presence across HLs and LHINs.

Select lead and member organizations with a learning culture

Many of the organizations that took on the role of lead or co-lead of a HL were described using terms indicative of a learning culture. These organizations were tolerant of ambiguity, open to change, and flexible or nimble enough to experiment. One lead organization was described as a ‘learning organization’ by a participant who went on to say, ‘That’s not new since Health Links. It’s been an ongoing journey for years. And so it sets the culture up for a willingness to change and a willingness to try new things, and to work together and learn together through different changes’ (Admin, HL11). When asked what is the most important characteristic of an organization leading or participating in an HL, the same descriptors were used, including ‘forward-thinking’, ‘can-do attitude’, ‘comfortable with ambiguity’, ‘creative’, ‘progressive’, ‘visionary’, ‘thinks outside the box,’ and ‘risk-taker’.

Create opportunities for interaction and communication

Regular interaction and communication between professionals and organizations fuelled collective learning as the nature and value of others’ knowledge was recognized. Two examples of this are provided below from the LHIN and HL perspectives, respectively:

A lot of providers had never really sat down at the table and talked about individual patients … A lot of times it wasn’t even having to do a complicated medical intervention to improve the [patient’s] pathway. It was just having that connection, with another agency that they were able to resolve some issues using simple solutions. (Admin, LHIN5)

As things emerge then we bring people in to have a conversation with them to make those relationships … We explain to them what we do. They explain to us what they do. And now we’ve elevated our own knowledge. (Clinician, HL11)

Develop formal structures for knowledge exchange

Several formal governance and knowledge exchange structures were developed to support learning across HLs and LHINs. These structures included ‘committees’, ‘secretariats’, ‘councils’, ‘communities of practice’, and ‘collaboratives’, and were often spearheaded by the LHINs, many of whom saw themselves as ‘knowledge brokers’. These diverse forums allowed for discussion of common challenges, identification of opportunities for standardization, and sharing of experiences and best practices. One participant noted, ‘Now each provider is going to come and present at our advisory council on how they are working differently so that we can all learn from ideas. And it’s like you hear an idea and you think, oh, we could do that, we’ll try that’ (Admin, HL6).

However, some participants argued that more needs to be done to support knowledge exchange and learning. One participant said, ‘It feels like a lot of digging and scratching to figure out where people are at. There’s not as much opportunity to learn across Health Links’ (Admin, HL19). Another participant added that, ‘What needs to happen is a centre for knowledge exchange which enables people and ideas to be brought together using web-based platforms. It’s the knowledge broker activity that I think is needed both at the clinical level as well as at the developmental level’ (Clinician, HL17).

Apply quality improvement methodology

Quality improvement (QI) methodology was adopted as a means to identify the most appropriate and effective HL design based on each HL’s unique context and target population(s). The most prominent QI principle and tool applied was ‘iterative testing’, using frameworks such as the Plan-Do-Study-Act (PDSA) cycle, as noted in the quote below:

We’re really working with an understanding that this is an evolving process. So we’re sitting there as a partner saying, okay, this program has changed twice already. We’re quite happy to change it a third time just to be able to investigate a better way of doing things. (Admin, HL15)

Discussion

This study examined the socio-cognitive challenges inherent in bringing together traditionally disparate providers and organizations to create hybrid integrated care networks known as HLs. The results provide conceptual validation of the relevance and importance of the content in the Integration Mindsets Framework. Examples of divergent thinking in the HLs data mapped on to all ten elements of the Strategy Mental Model and six of nine elements of the Relationship Mental Model (Appendix C). This finding may reflect the HLs’ early stage of implementation. Many of the HLs in the study had only just begun to see patients. It is not surprising, then, that most of the divergences in thinking centred on high-level design issues, rather than the minutiae of identities, roles, and relationships. Evidence to date suggests that the highest performance outcomes are achieved when SMMs are exhibited in relation to both task-work (i.e. strategy mental model) and teamwork (i.e. relationship mental model) (Smith-Jentsch et al., 2005). However, according to a meta-analysis, team-related SMMs typically have an indirect effect on performance, while task-related SMMs have a direct effect on performance (DeChurch and Mesmer-Magnus, 2010). As such, in hybrid integrated care networks, particular attention must be paid to Strategy Mental Model convergence in the early stages of implementation.

Many of the key factors identified in the Integration Mindsets Framework, and in our data, are well cited in the literature; we know the importance of a common vision, resources, appropriate network leadership and membership, and role clarity, among other factors (e.g. Suter et al., 2009). Although many of these factors are structural in nature, an SMM lens helps surface underlying social cognitions which determine how these factors shape integrated care efforts. In other words, the message is not, for example, that resource scarcity is an important factor for implementing integrated care networks. Rather, the message is that resource scarcity can be framed and understood in different ways, and the extent of convergence between stakeholders around the issue of resource scarcity shapes coordinated action.

Other scholars have drawn from the construct of ‘Institutional Logics’ (IL) to examine hybridity in the public administration field (Skelcher and Smith, 2015). We argue that an SMM lens complements and extends the IL perspective. IL are ‘the belief systems and related practices that predominate in an organizational field’ (Scott, 2001). IL are rooted in sociology, while SMMs are rooted in industrial psychology. Although the literature on IL includes brief references to cognition, IL are more closely linked with the concept of culture, and scholars have thus called for more attention to cognition (Thornton and Ocasio, 2008). In combination, IL and SMMs may enable a robust, multi-level study of social, cognitive, and institutional factors influencing hybrid networks.

This study has some limitations. First, inherent to the nature of secondary analysis, the data were not collected to address our specific research question. However, two of the authors (JME and AG) collected the data and were therefore aware of nuances in the context and methods that may be important to data interpretation. Second, the data were collected from networks in one Canadian province. Although the HL networks are diverse, the results may have limited generalizability. Third, we examined the data at a ‘macro level’, looking for divergent mental models across networks. We did not divide the data by network or professional group to examine mental model divergences within network or professional boundaries. This macro-level focus aligns with the proposition by Evans and Baker (2012) that SMMs can manifest and influence performance at an inter-organizational or system level, not only at the team level. Furthermore, the data were collected at one point in time early in the implementation of HLs. The results may thus be more applicable to the implementation, rather than the ongoing management and sustainability, of integrated care networks. Finally, given the early stage of implementation, we did not seek to identify shared mental models due to the low likelihood of convergence across diverse stakeholders province-wide.

Future research should examine both convergence and divergence of mental models within and across network boundaries at multiple points in time. Examining the extent to which SMMs exist across high- and low-performing integrated care networks (akin to Hysong et al.’s (2005) study) would provide the empirical evidence needed to drive future research on this topic. The Integration Mindsets Framework helps establish a common language for discourse and future research. Application of the framework across multiple studies and settings over time will enable us to draw conclusions regarding which mental model divergences are most common and impactful, and what strategies support convergence and coordinated action across diverse stakeholder groups.

Conclusion

The Integration Mindsets Framework and the results of this study can be used by healthcare policymakers, leaders and care providers to help surface and reconcile divergent mental models. For example, the Framework can be used as a checklist to guide discussion during the planning stages of an initiative, or to inform the development of education and training materials. If data on stakeholders’ mental models are collected, formally or informally, the results can be used to inform change management strategies. Finally, SMMs may be used as an indicator of readiness to integrate, or post-implementation convergence as a sign of successful integration (Evans et al., 2014a). In addition to using the Framework as a tool for improvement, stakeholders may also apply the four strategies identified in this study to facilitate practice-based learning and mental model convergence.

Although this article focused on the healthcare industry, SMM theory has implications for professionals in public policy and administration seeking a theory-informed approach to facilitating stakeholder dialogue towards collective learning and coordinated action. The content of the Integration Mindsets Framework as well as the learning strategies proposed can be translated for application in diverse industries and settings.

Supplemental Material

Supplemental material for Trial and error, together: divergent thinking and collective learning in the implementation of integrated care networks

Supplemental material for Trial and error, together: divergent thinking and collective learning in the implementation of integrated care networks by Jenna M. Evans, Agnes Grudniewicz and Peter Tsasis in International Review of Administrative Sciences

Footnotes

Acknowledgements

Drs Walter Wodchis and G. Ross Baker contributed to the broader evaluation of the Health Links, upon which this study is based.

Funding

This study is supported by grants from the Canadian Institutes of Health Research and from the Ontario Ministry of Health and Long-Term Care Health System Performance Research Network. The views expressed in this article are the views of the authors and do not necessarily reflect those of the funders.

Ethical approval

Ethical approval for this study was granted by the University of Toronto Research Ethics Board (protocol #29787).

Supplemental data

Supplemental data for this article can be found at journals.sagepub.com/home/ras.

Jenna M. Evans, Scientist, Cancer Care Ontario and Assistant Professor, University of Toronto; expertise in organizational behaviour, health system performance, integrated care, and complexity science.

Agnes Grudniewicz, Assistant Professor, University of Ottawa; expertise in primary care, care coordination, and patients with complex health and social needs.

Peter Tsasis, Associate Professor, York University; expertise in interorganizational collaboration, complexity science, and organizational change.

References

Abrahamson

Fombrum

(1994) Macrocultures: Determinants and consequences. Academy of Management Review 14(4): 728–755.

Angus

Greenberg

(2014) Can better care for complex patients transform the health system? Healthcare Papers 14: 1–19.

Argyris

(1994) On Organizational Learning. Oxford: Blackwell.

Bardsley

Steventon

Smith

et al. (2013) Evaluating integrated and community-based care. Nuffield Trust.

Bodenheimer

(2008) Cooridnating care – A perilous journey through the health care system. The New England Journal of Medicine 358: 1064–1071.

Cannon-Bowers

Converse

(1993) Shared mental models in expert team decision making. In: Castellan

(ed.) Individual and Group Decision Making: Current Issues. Hillsdale, NJ: Erlbaum.

Carlile

(2004) Transferring, translating, and transforming: An integrative framework for managing knowledge across boundaries. Organization Science 15: 555–568.

Collins

(1998) Organizational Change: Sociological Perspectives. London: Routledge.

Curry

Ham

(2010) Clinical and Service Integration: The Route to Improved Outcomes. London: The King’s Fund.

10.

DeChurch

Mesmer-Magnus

(2010) The cognitive underpinnings of effective teamwork: A meta-analysis. Journal of Applied Psychology 95: 32–53.

11.

Denis

J-L

Lamothe

Langley

et al . (2009) The reciprocal dynamics of organizing and sense-making in the implementation of major public-sector reforms. Canadian Public Administration 52: 225–248.

12.

Denis

Davies

Ferlie

et al. (2011) Assessing initiatives to transform health care systems: Lessons for the Canadian health care system. CHSRF Series on Health Care Transformation. CHSRF.

13.

Dewulf

Mancero

Cárdenas

et al . (2011) Fragmentation and connection of frames in collaborative water governance: A case study of river catchment management in Southern Ecuador. International Review of Administrative Sciences 77: 50–75.

14.

Elwyn

Taubert

Kowalczuk

(2007) Sticky knowledge: A possible model for investigating implementation in healthcare contexts. Implementation Science 2: 44.

15.

Evans

Baker

(2012) Shared mental models of integrated care: Aligning multiple stakeholder perspectives. Journal of Health Organization and Management 26: 713–736.

16.

Evans

Baker

Berta

et al. (2013) The evolution of integrated health care strategies. In: Goes

Savage

Friedman

(eds) Annual Review of Health Care Management: Revisiting the Evolution of Health Systems Organization. Bingley: Emerald Group Publishing, pp. 125–161.

17.

Evans

Baker

Berta

et al . (2014a) A cognitive perspective on health systems integration: Results of a Canadian Delphi study. BMC Health Services Research 14: 222.

18.

Evans

Grudniewicz

Wodchis

et al . (2014b) Leading the implementation of Health Links in Ontario. Healthcare Papers 14: 21–25.

19.

Evans

Grudniewicz

Baker

et al . (2016) Organizational context and capabilities for integrating care: A framework for improvement. International Journal of Integrated Care 16: 15.

20.

Ferlie

Fitzgerald

Wood

et al . (2005) The nonspread of innovations: The mediating role of professionals. Academy of Management Journal 48: 117–134.

21.

Fiol

(1994) Consensus, diversity, and learning in organizations. Organization Science 5: 403–420.

22.

Fiske

Taylor

(1991) Social Cognition. New York: McGraw-Hill.

23.

Friedman

Goes

(2001) Why integrated health networks have failed. Frontiers of Health Services Management 17: 3–28.

24.

Glouberman

Mintzberg

(2001) Managing the care of health and the cure of disease—Part I: Differentiation. Health Care Management Review 26: 56–69.

25.

Grudniewicz

Tenbensel

Evans

et al. (2018) ‘Complexity-compatible’ policy for integrated care? Lessons from the implementation of Ontario's Health Links. Social Science & Medicine 198: 95–102.

26.

Heaton

(2008) Secondary analysis of qualitative data: An overview. Historical Social Research 33: 33–45.

27.

Huber

Lewis

(2010) Cross-understanding: Implications for group cognition and performance. Academy of Management Annals 35: 6–26.

28.

Hysong

Best

Pugh

et al . (2005) Not of one mind: Mental models of clinical practice guidelines in the veterans health administration. Health Services Research 40: 829–848.

29.

Jiwani

Fleury

M-J

(2011) Divergent modes of integration: The Canadian way. International Journal of Integrated Care 11: e018.

30.

Kellermanns

Walter

Floyd

et al . (2011) To agree or not to agree? A meta-analytical review of strategic consensus and organizational performance. Journal of Business Research 64: 126–133.

31.

Kreindler

Larson

et al . (2012) Interpretations of integration in early accountable care organizations. Milbank Quarterly 90: 457–483.

32.

Ling

Brereton

Conklin

et al . (2012) Barriers and facilitators to integrating care: Experiences from the English Integrated Care Pilots. International Journal of Integrated Care 12: 1–12.

33.

McComb

(2007) Mental model convergence: The shift from being an individual to being a team member. In: Yammarino

Dansereau

(eds) Multi-level Issues in Organizations and Time. Bingley: Emerald Group Publishing, pp. 95–147.

34.

Mohammed

Ferzandi

Hamilton

(2010) Metaphor no more: A 15-year review of the team mental model construct. Journal of Management 36: 876–910.

35.

Ouwens

Wollersheim

Hermens

et al . (2005) Integrated care programmes for chronically ill patients: A review of systematic reviews. International Journal for Quality in Health Care 17: 141–146.

36.

Pate

Fischbacher

Mackinnon

(2010) Health improvement: Countervailing pillars of partnership and profession. Journal of Health Organization and Management 24: 200–217.

37.

Phillips

(1994) Industry mind-sets: Exploring cultures of two macro-organizational settings. Organization Science 5: 384–402.

38.

Raelin

(1997) A model of work-based learning. Organization Science 8: 563–578.

39.

Rein

Schon

(1996) Frame-critical policy analysis and frame-reflective policy practice. Knowledge and Policy: The International Journal of Knowledge Transfer and Utilization 9: 85–104.

40.

Scott WR (2001) Institutions and Organizations. Thousand Oaks, CA: SAGE Publications.

41.

Singer

Burgers

Friedberg

et al . (2011) Defining and measuring integrated patient care: Promoting the next frontier in health care delivery. Medical Care Research and Review 68: 112–127.

42.

Skelcher

Smith

(2015) Theorizing hybridity: Institutional logics, complex organizations, and actor identitites: The case of nonprofits. Public Administration 93: 433–448.

43.

Smith-Jentsch

Mathieu

Kraiger

(2005) Investigating linear and interactive effects of shared mental models on safety and efficiency in a field setting. Journal of Applied Psychology 90: 523–535.

44.

Suter

Oelke

Adair

et al . (2009) Ten key principles for successful health systems integration. Healthcare Quarterly 13: 16–23.

45.

Thornton

Ocasio

(2008) Institutional logics. In: Greenwood

Oliver

Suddaby

et al. (eds) The SAGE Handbook of Organizational Institutionalism. Newbury Park, CA: Sage Publications.

46.

Tsasis

Evans

Owen

(2012) Reframing the challenges to integrated care: A complex adaptive-systems perspective. International Journal of Integrated Care 12: 1–11.

47.

Tsasis

Evans

Rush

et al . (2013) Learning to learn: Towards a relational and transformational model of learning for improved integrated care delivery. Administrative Sciences 3: 9.

48.

van Wijngaarden

JDH

de Bont

Huijsman

(2006) Learning to cross boundaries: The integration of a health network to deliver seamless care. Health Policy 79: 203–213.

49.

Vlaar

PWL

Van den Bosch

FAJ

Volberda

(2006) Coping with problems of understanding in interorganizational relationships: Using formalization as a means to make sense. Organization Studies 27: 1617–1638.

50.

Walsh

(1995) Managerial and organizational cognition: Notes from a trip down memory lane. Organization Science 6: 280–321.

51.

Weick

(1995) Sensemaking in Organizations. Thousand Oaks, CA: Sage Publications.

52.

Williams

Sullivan

(2009) Faces of integration. International Journal of Integrated Care 9: 1–13.

53.

Wodchis

Dixon

Anderson

et al . (2015) Integrating care for older people with complex needs: Key insights and lessons from a seven-country cross-case analysis International Journal of Integrated Care 15: 1–15.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.58 MB