Understanding collaboration beyond contracting: the relationship between inter-organizational collaboration and hard/soft control

Abstract

Purpose

This paper aims to extend insights about the relationship between inter-organizational collaboration and approaches to control from the perspective of decision-makers. We investigate the relationship between approaches to control and intended forms of integration between actors responsible for solving the complex problem of integrated person-centered care for elderly with diverse and significant needs.

Design/methodology/approach

The empirical study is based on a content analysis of contractual agreements. We have analyzed a total of 118 collaboration agreements and associative documents between all Swedish regions and municipalities.

Findings

The study shows that intended integration is subject to remarkable variation in intended forms of inter-organizational collaboration in this Swedish case. The paper illustrates that decision-makers’ intentions with proposed collaboration in each given context are important for the chosen approach to control. Regardless of intended forms of integration, our study suggests that an imminent soft approach to control is expressed alongside limited signs of hard control. Various forms of intended integration can be managed by the two approaches simultaneously insofar as the agreements appear to have a two-sided purpose.

Originality/value

Our paper proposes an empirically driven taxonomy of intended forms of integration initiatives. The taxonomy provides resources for studies about how collaboration can be managed when it is stipulated by national legislation but local self-governance gives actors considerable freedom to decide on how to organize and manage services. By presenting the taxonomy and relating this to approaches of control, our iterative study builds on and adds to a recent stream of research arguing that the relationship between collaboration and approaches to control may by fuzzier and more complex than originally thought.

Citation

Frid, K., Funck, E.K. and Glenngård, A.H. (2024), "Understanding collaboration beyond contracting: the relationship between inter-organizational collaboration and hard/soft control", Journal of Public Budgeting, Accounting & Financial Management, Vol. 36 No. 6, pp. 47-71. https://doi.org/10.1108/JPBAFM-06-2023-0086

Introduction

During the past few decades, complex challenges in the public sector have been described as a result of the era of austerity, cutback management and efficiency initiatives (Bracci et al., 2015; Jacobs and Cuganesan, 2014). Some of these problems cannot be clustered within single organizations’ boundaries because they stem from multiple causes and involve many actors from different organizations (De Waele et al., 2021; Kokko and Laihonen, 2022). Thus, to tackle complex problems, scholars in public administration suggest that public service providers and departments need to come together and work across organizational and professional boundaries in collaborative and coordinated efforts based on a common purpose (Ansell and Gash, 2007; Bianchi et al., 2021; Klijn and Koppenjan, 2000; Osborne, 2010; Wang and Ran, 2023).

While collaboration across boundaries might increase the capacity to deal with a complex problem, it inheres several tensions and contradictions (Agranoff and McGuire, 2001; Berthod and Segato, 2019; Bryson et al., 2015; Evans et al., 2023). Firstly, inter-organizational collaboration is a matter of incorporating and balancing values from different settings as well as managing divergent and potentially conflicting goals (Axelsson and Bihari Axelsson, 2006; Bianchi et al., 2021; Skelcher and Smith, 2015; Vangen and Huxham, 2012). Secondly, many of the challenges associated with collaboration across boundaries involve issues of control. Evidence suggests that controls are significant in shaping boundaries for accountability (Lee, 2022). Control may encourage collaboration and co-ordination by measuring outcomes and evaluating efficiency and effectiveness (Cuganesan et al., 2012) and by providing management support (Axelsson and Bihari Axelsson, 2006; Fırtın and Karlsson, 2020).

When discussing control, one may distinguish between hard and soft control approaches (Karrbom Gustavsson and Hallin, 2014; Lartey et al., 2023; McGregor, 1960; Truss et al., 1997). Hard control defines roles, expectations and performance metrics. It also provides guidelines for the collaboration (Bauer et al., 2022; Gazley, 2008; Karrbom Gustavsson and Hallin, 2014). Hard control is thus structured, stable, goal-oriented and relies to a high extent on binding agreements (Guest, 1995). By contrast, soft control relies on verbal commitments, social capital and trust. It supports autonomy, learning and room for local action as it depends on information exchange, interactions and shared norms (Bauer et al., 2022; Gazley, 2008; Guest, 1995; Karrbom Gustavsson and Hallin, 2014). Some studies have compared the relative influence of hard versus soft control in inter-organizational collaborations. These studies suggest that whereas hard control undermines trust and relational governance (Adler, 2001; Romzek et al., 2014), a more supportive soft control may encourage collaboration, improvement initiatives, flexibility and innovation (Agostino and Arnaboldi, 2018; Romzek et al., 2014). Other studies indicate that both hard and soft control can lead to positive outcomes (Gazley, 2008). Yet, tensions may arise between the two, making them work against each other (Romzek et al., 2014).

While scholars agree that managing and balancing tensions is critical for the inter-organizational collaboration’s functioning and performance (Berthod and Segato, 2019; Wang and Ran, 2023), few empirical studies have focused on how this can be managed (Bryson et al., 2015), especially on system and organizational levels (Karlsson et al., 2020). The purpose of our study corresponds to this call by focusing on one specific tension: to extend insights about the relationship between inter-organizational collaboration and approaches to control from the perspective of decision-makers. Specifically, we investigate the relationship between approaches to control and intended forms of integration between actors responsible for solving a complex problem. The complex problem we address is integrated person-centered care for elderly with diverse and significant needs. This problem is highly relevant given the challenges facing healthcare systems stemming from ageing populations and associated morbidity in many countries. In Sweden, where the empirical study is conducted, the responsibility for these patients is shared between two local authorities – municipalities and regions. While these two can be assumed to share the common goal of delivering integrated person-centered care, they differ in terms of responsibilities and employ individuals with different professional backgrounds. Municipalities have the responsibility of care and treatment in ordinary and special housing for elderly, provided by, e.g. registered nurses, physiotherapists and assistant nurses. Regions, on the other hand, have the responsibility of providing care by general practitioners (GPs) to these patients. GPs are employed by public and private primary care centers (PCCs) under the responsibility of the regions.

According to the Swedish Health and Medical Services Act (2017:30, ch. 16)), regions and municipalities should collaborate around the delivery of services for individuals who receive care from both the region and the municipality where they reside. Regions and municipalities should establish contractual documents to clarify roles and responsibilities and to facilitate integration of services, hereafter referred to as collaboration agreements. While collaboration is stipulated by national law, the strong tradition of local self-governance in Sweden gives regions and municipalities considerable freedom to decide on how to organize and control services and collaborative working methods in accordance with local conditions. Hence, variation is expected across the country when it comes to inter-organizational collaboration and approaches to control of actors involved in the care of elderly. The variation provides an excellent context for enhancing our understanding of collaboration across organizational boundaries and approaches to control from the perspective of decision-makers by analyzing these collaboration agreements.

In this paper, we build on and add to a recent stream of research arguing that the relationship between collaboration and approaches to control may be fuzzier and more complex than originally thought (Evans et al., 2023). We do this by presenting a taxonomy of so called “integration initiatives” and relating this to approaches to control. By bringing together insights from inter-organizational collaboration (Axelsson and Bihari Axelsson, 2006, 2013) with those of hard and soft control (Bauer et al., 2022; Gazley, 2008), our study suggests two important contributions. First, by further nuancing forms of integration (Axelsson and Bihari Axelsson, 2006, 2013), we respond to recent calls for more research on how to conceptually grasp the content of different forms of collaboration (D'Amour et al., 2005). Second, our study adds to the discussion of tensions between control and collaboration (Evans et al., 2023) by empirically showing how the relationship between intended forms of integration and approaches to control is asymmetrical and that hard and soft control may be combined based on context and decision-makers’ intentions.

In the next section, we outline a conceptual framework based on inter-organizational collaboration as differentiated or integrated (Axelsson and Bihari Axelsson, 2006, 2013) in combination with the notion of approaches to control as hard or soft (Gazley, 2008; Karrbom Gustavsson and Hallin, 2014; Lartey et al., 2023; McGregor, 1960; Truss et al., 1997). Subsequently, we describe our approach to collecting and analyzing our empirical data, i.e. collaboration agreements and associative documents for care of elderly covering all Swedish regions and municipalities. We then synthetize and discuss our results with reference to our conceptual framework before presenting our empirical findings, conclusions and contributions.

Conceptual framework

We combine Axelsson and Bihari Axelsson’s (2006, 2013) framework about inter-organizational collaboration in public health with the concepts of hard and soft control (Gazley, 2008; Karrbom Gustavsson and Hallin, 2014; Lartey et al., 2023; McGregor, 1960; Truss et al., 1997) to explore intended forms of integration and approaches to control.

Differentiation and integration

Research about collaboration is described as progressive yet tainted by an older view, which is said to describe collaborative behavior in a flattening and simplistic way (Gazley, 2017). For example, some literature discuss how collaboration can be described as levels (Gray, 1989; Himmelman, 1996; Huxman, 1996; Mulford and Rogers, 1982) or illustrated as increasing activity from one stage to the next (Gazley, 2017). Instead of measuring categories or degrees of collaboration in this way, we agree with the understanding of collaboration as a dynamic matter, characterized by complex human collectivity in organizations to service the needs of professionals and clients (Axelsson and Bihari Axelsson, 2006; D'Amour et al., 2005; Gazley, 2017). This view on collaboration highlights the concept as complex (Axelsson and Bihari Axelsson, 2006; D'Amour et al., 2005; Gazley, 2017), dynamic (D'Amour et al., 2005) and as voluntary assistance to reach a common goal (Castañer and Oliveira, 2020). The same stream of literature mention nonetheless that a plethora of definitions for “collaboration” and associative concepts adds confusion to the theoretical understanding in the research field (Axelsson and Bihari Axelsson, 2006; Castañer and Oliveira, 2020; D'Amour et al., 2005).

With the aspiration to overcome a too simplistic view of collaboration, we select a framework about inter-organizational collaboration in public health based on differentiation and integration (Axelsson and Bihari Axelsson, 2006, 2013). We are inspired by a conceptual view on various forms of integration in-between organizations and organizational units, i.e. inter-organizational collaboration (Axelsson and Bihari Axelsson, 2006). In this framework, the core of what collaboration stands for is integration; to bring actors and activities together into a greater unit (Axelsson and Bihari Axelsson, 2013). The essence of integration is collaboration across organizational boundaries to respond to demands from the outside world (Axelsson and Bihari Axelsson, 2013). Axelsson and Bihari Axelsson (2006, 2013) distinguish between horizontal and vertical integration. The former takes place between organizations or units that are on the same hierarchical level or have the same status, whereas the latter takes place between organizations or units on different hierarchical levels.

Based on these two dimensions, the outcome is a conceptual scheme with four forms of integration, i.e. contracting, co-ordination, co-operation and collaboration (Axelsson and Bihari Axelsson, 2006, 2013). Contracting refers to a situation with a low degree of both horizontal and vertical integration. Contact between actors is almost out of the picture and competition becomes the guiding principle for integration, regulated by established contracts (Axelsson and Bihari Axelsson, 2013). Here we find division of responsibilities and separate accountabilities, i.e. differentiation. Co-ordination defines a situation with a high degree of vertical but a low degree of horizontal integration. Integration becomes a matter of exchanging services, informing and supervising. Integration is accomplished primarily through the existence of a common management hierarchy and decisions are made at the top of this system. Co-operation refers to a situation with a high degree of both horizontal and vertical integration. Decisions are made by the hierarchical management but leaves space for voluntary activities, mutual adaptations and more informal communication across organizational boundaries (Axelsson and Bihari Axelsson, 2006). The organization of work enables an exchange of comprehensive solutions in-between organizational units and organizations. Finally, collaboration defines a situation with a low degree of vertical but a high degree of horizontal integration. Lessons learnt and comprehensive solutions around complex problems are communicated. Integration is rooted in voluntary agreements, mutual adaption, a high level of communication and a willingness to work together across organizational boundaries (Axelsson and Bihari Axelsson, 2006). Integration is achieved through intense networking, oftentimes implemented as multi-professional teamwork (Axelsson and Bihari Axelsson, 2013).

Hard and soft control

Regardless of whether we discuss collaboration formality (Bauer et al., 2022), collaboration contract (Gazley, 2008; Poppo and Zenger, 2002) or interdisciplinary collaboration and accounting (Cuganesan et al., 2012; Jacobs and Cuganesan, 2014), it becomes evident that issues of control are significant in shaping and supporting collaborations (Evans et al., 2023; Jayasinghe et al., 2020). Collaboration agreements can be formulated as formal contracts or more informal commitments (Bauer et al., 2022; Malhotra and Lumineau, 2011). At the same time, control can be strictly formal or more informal within collaborative settings (Krause and Swiatczak, 2021). In this paper, we discuss formalization through the concepts of hard and soft control.

The distinction between hard and soft management control can be traced back to the work of McGregor (1960). McGregor’s argument was that our view of human nature (Theory X or Theory Y) ultimately influences management control approaches. In this paper, hard control is delineated as a form of structured, stable and goal-oriented management. It relies heavily on binding agreements while clearly defining roles and expectations, and providing explicit guidelines for collaboration. Hard control establishes precise performance metrics and goals, which are meticulously monitored and controlled (Bauer et al., 2022; Gazley, 2008; Guest, 1995; Truss et al., 1997). In contrast, soft control places emphasis on fostering commitment through trust in employees, investing in their training and development, and granting them autonomy. Soft control advocates for autonomy rather than control. It relies on information exchange, interactions, shared norms and ambiguously defined goals managed by negotiation and discussion rather than monitoring and evaluations (Bauer et al., 2022; Gazley, 2008; Karrbom Gustavsson and Hallin, 2014).

Control should however not be seen as a binary concept but rather as two endpoints – hard control and soft control. In this paper, we conceptualize formalization in the form of hard and soft control from two dimensions: the degree of formalization and the content of formalization. The first dimension encompasses the abundance and intensity of guidelines, checklists, routines and manuals, as well as the quantity of performance areas, the nature of performance measurements and the frequency of evaluations. Consequently, we assert that the rigidity of control increases with a higher number of guidelines, evaluations of more performance areas, measuring of more targets and higher frequency of evaluations. Conversely, a lower number of guidelines and performance areas, lesser measuring of targets and more infrequent evaluations indicate a softer approach to control.

Nevertheless, details and quantity are not the sole criteria for determining hard or soft control as the content of formalization must also be considered. The content of formalization pertains to the purpose of evaluation (accountability or learning and improvement) and the form of evaluation (documentation/report or dialogue). The purpose of evaluation holds significant importance since studies have emphasized the necessity to scrutinize how measures are employed and appraised – whether for the purpose of mere monitoring or for fostering learning (Guest, 1995; Karrbom Gustavsson and Hallin, 2014; Truss et al., 1997). In essence, this involves discerning whether the overarching goal of evaluation is geared towards accountability or learning and improvement. Evaluation systems take on a more stringent, or “hard”, approach when information is used in a targeted manner, whereas a softer approach is characterized by clear focus on learning and durability (Leeuw and Furubo, 2008). Consequently, we interpret formulations in the collaboration agreements to be indicative of hard control when accountability is prioritized over learning and improvement. Additionally, we interpret absence of dialogue in favor of documenting and reporting as a sign of hard control.

Theories such as game theory, agency theory and transaction cost theory argue that hard control supports trust-like relationships (Sitkin and Stickel, 1996). Counter-positioning theories argue that hard control will undermine trust and even restrict effectiveness and efficiency of joint efforts (Adler, 2001; Chaserant, 2003). What is stated is that soft control trumps hard control when collaborations and synergistic effects are in focus at the same time as it improves flexibility, innovations and the performance of the collaboration (Ghoshal and Moran, 1996). Still other studies indicate that both hard and soft control can lead to positive outcomes and that the two approaches can complement each other (Gazley, 2008; Karrbom Gustavsson and Hallin, 2014; Poppo and Zenger, 2002). However, whereas hard control in the form of formal protocols, rules and structures may enable collaboration and reduce uncertainty, previous research suggests that soft control (e.g. facilitative local leadership, trust, commitments, shared understanding and values) is important for putting collaborative organizational settings into practice and for generating sustainable outcomes (Adler and Borys, 1996; Bianchi et al., 2021; Jordan and Messner, 2012). Consequently, control has the potential to provide benefits and to cause unintended consequences in organizational settings where collaboration should take place.

Synthesis – conceptualizing integration and approaches to control

The theoretical discussion above helps us to scrutinize inter-organizational collaboration beyond basic expectations of contractual agreements as such, i.e. beyond the form of integration referred to as contracting. By analyzing differences in approaches to control in relation to intended forms of integration between actors responsible for solving a complex problem, we can say something about the set-up for inter-organizational collaboration at the system level. Are the intended activities characterized by partnership, teamwork and common activities leading to synergistic effects (integrated), or are they typified by division of labor and separate responsibilities between actors (differentiated)? What are the characteristics of the approaches to control within this setting? Is the degree of formalization (expressed rules, regulations, performance metrics and monitoring) high or low? Are principles for evaluation and monitoring designed for the purpose of stimulating accountability or learning and improvement? In other words, are the described approaches to control hard or soft?

Previous research has expressed the importance of soft control to foster integration, while relating hard control to differentiation (Axelsson and Bihari Axelsson, 2006, 2013). Soft control supports autonomy and encompasses information exchange, interactions and shared norms (Bauer et al., 2022; Gazley, 2008; Guest, 1995; Karrbom Gustavsson and Hallin, 2014) – elements that are crucial to foster horizontal integration, i.e. collaboration (Axelsson and Bihari Axelsson, 2006, 2013; Karlsson et al., 2020). Hard control is associated with clear definitions of roles, expectations and performance, providing guidelines for the collaboration (Bauer et al., 2022; Gazley, 2008; Karrbom Gustavsson and Hallin, 2014) – elements that are associated with differentiation, i.e. contracting (Axelsson and Bihari Axelsson, 2006, 2013). More recent research suggests that multiple approaches to control may enhance collaboration (Evans et al., 2023) and that the context, including decision-makers’ intentions with the proposed collaboration, matters for the chosen approach to control of such activities (Karlsson et al., 2020). Our conceptual framework allows us to nuance forms of integration between the two endpoints of integration and differentiation by elaborating on the relationship between intentions with integration vis-à-vis approaches to control from the perspective of decision-makers. Moreover, such an elaboration can identify possible implications about collaboration beyond contracts as such (Gazley, 2008).

Research context, data and method

Context

The Swedish healthcare system is primarily tax funded and involves three levels of government. At the national level, the Ministry of Health and Social Affairs is responsible for overall health and healthcare policy. National government agencies provide guidelines and high-level oversight. The responsibility for funding and organization of healthcare rests with the twenty-one regions and 290 municipalities, whereof one (Gotland) is both a municipality and a region. The Swedish Health and Medical Services Act (2017:30) is designed to give local authorities considerable freedom in organizing their healthcare services. Of relevance for this empirical study, municipal health and medical care refers to care and treatment (including rehabilitation and technical aids) provided in ordinary or special housing for elderly by registered nurses, physiotherapists, occupational therapists, assistant nurses and care assistants. The regions, on the other side, are responsible for the provision of healthcare services in hospitals and PCCs and to provide care by GPs in home care (in ordinary and special housing). In 2018, about 340,000 individuals aged 65 and above received municipal health and medical care in Sweden. About two-thirds received care in ordinary housing and one-third in special housing, i.e. nursing homes (NBHW, 2020-3-6603).

The regional and municipal authorities are guided by both national regulation, guidelines and local priorities in their decisions. According to national legislation, regions and municipalities should establish collaboration agreements to clarify roles and responsibilities in the delivery of services to individuals in need of both regional and municipal healthcare. This includes routines for coordinated individual care planning (CIP) (Health and Medical Services Act, 2017:30; Social Services Act, 2001:453), access to adequate care by GPs and safe handling of pharmaceuticals (Health and Medical Services Act, 2017:30), and proper handling at discharge from hospitals (Act on Coordinated Discharge, 2017:612). However, the strong tradition of local self-governance in combination with variation in local conditions implies that local priorities lead to variation with regards to how provision of healthcare is organized and governed. While there is a common goal – integrated person-centered care for elderly with complex needs – there is local variation in arrangements towards reaching that goal. This variation makes an empirical study of differences and similarities concerning how Swedish regions and municipalities use written agreements to govern and manage collaborative working methods a relevant case to enhance our understanding of collaboration beyond contracting.

Data collection

The primary source of data is written contractual documents regarding collaboration in care of elderly with complex needs in twenty Swedish regions and 289 municipalities, valid at the time of collection (February–March 2022). The sample covers all of Sweden except for the exclusion of one (Gotland), which is both a region and a municipality. The documents concern various subjects in relation to care of elderly, such as home care, responsibility take-over in ordinary housing, discharge from hospitals and GP involvement (see Appendix). We also include available supplements, clarifications and reviews of such contracts and local contracts.

Our initial online search revealed that such documents are dispersed with different designs and titles across public homepages and headlines. Thus, we contacted key representatives in the regions to assure a continuous inclusion and exclusion of relevant material. Our combination of collecting material from homepages and key representatives resulted in 118 included documents, excluding attachments. The names of the twenty regions and the respective municipalities are collectively pseudonymized in alphabetical order, A-T (see Appendix).

The included documents cover some material other than contractual documents, namely mission statements, directions, guidelines, instructions and user manuals as well as additional content such as reports and action programs. We consider this material to be associative since these documents are listed alongside the contractual documents on homepages, presented as reference points in the contracts and/or sent to us by the key representatives. Hence, we find them necessary to include for a complete representation and understanding. Altogether, the wide scope of material is regarded useful because it contributes to a holistic understanding of how inter-organizational collaborations are governed and organized.

Data analysis

The data analysis of the documents takes the form of a directed content analysis – a structured research technique where coding can start directly with pre-determined codes and progress into further coding (Hsieh and Shannon, 2005). The first step of our content analysis was an initial understanding of the material based on underpinning theoretical assumptions. We created 32 preliminary codes within seven general themes about, inter alia, definitions of the documents and other formalities, collaboration within and beyond law-binding requirements, and monitoring and evaluation. This preliminary code system was tested on documents for three regions and the respective municipalities (F,J,L), before developing the system in terms of expansion, re-arrangements and clarifications.

Emerging codes then derived from discoveries of the actual presence of content in the documents. We formulated new codes accordingly (see examples in Tables 1 and 2) whilst developing our theoretical framework. This resulted in a final coding system of 100 distinctly formulated codes under the umbrella of seven themes: (1) the formalities of the documents, (2) expressed aims, goals and values, (3) law-binding requirements, (4) established collaboration beyond law-binding requirements, (5) goals, targets, monitoring and evaluation, (6) governance and organization of team-based care and (7) organization of multi-professional teams. We coded all the documents (A-T) in an iterative process of following the coding system, discussing theoretical implications and reading the documents for a solid understanding of broad patterns, similarities and differences.

Subsequently, we validated the first empirical findings by applying search words to assure coverage of all codes and by re-coding for the sake of confirmation. In addition, we conducted a complementary search of additional documents to reassure a complete inclusion of relevant material. Consequently, our data analysis progressed from the previous stage of pre-determined and refined codes into new categorization of data, which characterizes the process of a directed content analysis (Hsieh and Shannon, 2005).

Our methodological approach, all the way through the coding and construction of the theoretical framework, can be described as an ongoing movement between data and theory. The process for analyzing the data is therefore iterative in the sense that we constantly moved back and forth in-between text in the documents, literature and the emerging categorization of themes and codes (Ahrens and Chapman, 2006; Silverman, 2011). Throughout this iterative process, we concluded that the theoretical framework inter-organizational collaboration and the theoretical concepts soft and hard control are congruent and appropriate for our analytical framework. We constructed our analytical framework in the final step of the data analysis and will now move on to describe it in detail.

Analytical framework

The analytical framework includes the dimensions forms of integration (Table 1) and approaches to control (Table 2).

The first dimension, forms of integration, includes the indicative forms contracting, co-ordination, co-operation and collaboration (see Table 1). The operationalization starts with (1) contracting, i.e. very few signs of integration other than formal obligations and law-binding requirements. This notion stems from contracting as associative to separation of roles and responsibilities, i.e. differentiation (Axelsson and Bihari Axelsson, 2006). We operationalize signs of contracting in formulations in the documents concerning basic contract-related matters, for example law-binding responsibilities and specifications of payment liabilities.

The second step of the operationalization is (2) co-ordination, i.e. integration held together by coordinative efforts and common management decisions. We operationalize co-ordination in formulations which suggest a supervisory form of integration: common values and principles for the organizations to follow, precise goals for co-ordination and/or collaboration and appointment of local representatives in collaboration boards or similar. Our understanding of co-ordination is underpinned by co-ordination-related structures, a common management hierarchy and top-level decisions (Axelsson and Bihari Axelsson, 2006).

The third step of the operationalization is (3) co-operation, i.e. integration characterized by a combined form of hierarchical management and voluntary work along with some information sharing and communication across organizational borders. We operationalize co-operation in formulations which express shared information and feedback in evaluation activities, for example found in the content of what is evaluated and what evaluation results are used for. This comprehension is underpinned by characteristics of co-operation such as voluntary activities, mutual adaptations and more informal communication across organizational boundaries (Axelsson and Bihari Axelsson, 2006).

The fourth step is (4) collaboration, i.e. integration in the form of fusing actors together into one unit. We operationalize collaboration in formulations indicating efforts to turn talk into action in specific collaborative activities/projects and multi-professional teamwork. Examples of such indicative efforts are support systems for collaborations and denominations of multi-professional teams or equivalent. We base our understanding of collaboration on the idea of voluntary agreements, mutual adaption, a high level of communication and willingness to work together across organizational boundaries, oftentimes in multi-professional teams (Axelsson and Bihari Axelsson, 2013).

The second dimension of the analytical framework is approaches to control (see Table 2).

Control becomes significantly relevant in our empirical study since agreements can be formulated as formal contracts or more informal commitments (Malhotra and Lumineau, 2011). Evaluation can be formulated as institutionalized arrangements or more as knowledge-driven long-term activities (Gazley, 2008; Leeuw and Furubo, 2008). The keyword here is thus to what extent indications of formalization can be seen in the agreements according to degree and content of formalization. Control therefore gives a further appreciation of whether and how intended integration is surrounded by more or less formalization, i.e. indications of hard or soft control.

Our operationalization starts with (1) a hard approach to control, i.e. high degree of formalization and content of formalization in favor of accountability. We understand a high degree of formalization as associative to explicit guidelines, as well as precise performance metrics and goals which are meticulously monitored and controlled (e.g. Bauer et al., 2022; Gazley, 2008). Therefore, we operationalize signs of hard control in formulations which express a high number of guidelines, a high number of performance areas, measurements of targets and a high evaluation frequency. In addition, we find signs of hard control in content of formalization in favor of accountability. Formulations about evaluation activities for the purpose of accountability are especially relevant here, based on previous studies about using and evaluating measures for the sake of monitoring (Adler and Borys, 1996; Jordan and Messner, 2012). In extension, we interpret evaluations in the form of documentation as a sign of hard control. This interpretation is underpinned by the notion of evaluations as featured by a hard approach to control when information is used in a targeted manner (Leeuw and Furubo, 2008). Thus, we operationalize hard control in formulations expressing that evaluations are made for the purpose of accountability (for example to control compliance) and are documented (for example in the form of a report). Our operationalization is grounded in the view of hard control as a form of structured, stable and goal-oriented management.

The final step of the operationalization is (2) a soft approach to control. A soft approach to control is operationalized by us as low degree of formalization alongside content of formalization in favor of learning and improvement, contrary to our abovementioned interpretations of hard control. We operationalize signs of soft control in formulations expressing a low number of guidelines, a low number of performance areas, no measurements of targets and a low evaluation frequency. In addition, we find indications of soft control in formulations about evaluation activities for the purpose of learning and improvement. In addition, we interpret evaluations in the form of dialogue as a sign of soft control. We ground this understanding in previous research pinpointing evaluation systems as softer when characterized by clear focus on improvement, learning and durability (Leeuw and Furubo, 2008). The operationalization goes back to the view of soft control as a hallmark for commitment through trust, investment in training and development, autonomy and dialogue.

Empirical findings: analysis of intended integration and control

The following analysis focuses on intended integration and control in the studied documents of Swedish regions and municipalities. By analyzing these two dimensions, we can say something about patterns and tendencies in the organizational settings in terms of both inter-organizational collaboration and formalization. The analysis starts with formulated intentions for integration, continues with formulated intentions for control and ends with an inter-connection of the two dimensions.

Before elaborating on this, it should be mentioned that all the agreements express what is to be expected from contractual documents in terms of compliance with law-binding standards and requirements. There are extensive formulations about separate and shared responsibilities for the regions and municipalities in terms of obliged tasks, collaboration and communication, and payment liabilities. Required responsibilities and procedures are specifically described in relation to subjects such as routines for CIP (L #58), involvement of GPs (Q #97), discharge from hospitals (N #67) and person-centered collaboration solutions (J #47). These subjects are typically described alongside references to national laws. The common denominator of law-binding requirements indicates that all the agreements fulfill the basic function in terms of division of responsibilities and separate accountabilities, in other words integration in the form of differentiation. This implies that integration through “contracting” is fulfilled (see Table 1) and therefore needs no further analysis.

However, variation occurs in terms of the composition of documents. Some documents are collected in one place at the regions’ websites, but most of them are dispersed under different headlines, pages or in other ways not collected in one place. A similar discrepancy is observed in the various titles of the contractual documents (see Appendix). The notable differences in the compositions imply different interpretations of national legislation concerning collaboration, either as contract-binding or more lightly as agreements.

Formulated intentions for integration

A summary of formulated intentions for integration is presented in Table 3.

First and foremost, we start by elaborating on formulations that imply intended integration in the form of “co-ordination” (see Table 3). We observe multiple signs in this direction across all the regions and municipalities, especially in terms of common management hierarchies and top-level decisions. A representative example is the comprehensive pattern of commonly expressed overarching values and principles for organizations to follow. A shared basis of watchwords is presented across all the regions and municipalities, with phrases such as “good quality, local health care” (D #25, p. 2), co-ordination (N #75) and a focus on the individual (J #47). The widespread formulations about all-embracing values and principles point at a first sign of common management hierarchies and decision-making at the top of these management systems.

Similarly, overarching goals in the organizations appear to be prominent. In addition, we find formulations about goals in the direction of co-ordination or similar. For instance, one goal statement starts with the sentence “the goal for co-ordination is to meet the individual’s need of continuity, co-ordination, and safety” (I #45, p. 5). Another statement contains the phrase “ … develop forms for co-ordination and collaboration …” (C #16, p. 5). Similarly, several statements concern coordinated health and social care (e.g. K #48; N #75). Other examples of phrases are “development goals” (J #47, p. 9) and “boundless local care” (S #108, p. 9). These types of fluent goals appear to form a pattern of management-led decisions, but only one goal stands out for its specific description:

The goal for a revised regional co-ordination structure is to achieve: clearer co-ordination structure, greater continuity in co-ordination, a more efficient organizational co-ordination and meeting structure, good co-ordination culture and strengthened relationships. (O #79, p. 9)

This goal formulation includes concrete specifications of goal achievements to reach co-ordination-related structures wherein services and information can be exchanged. Hence, it appears to be a precisely formulated goal for co-ordination. The remaining goals in the documents come across as more generally formulated, hence we can say that only one region and its municipalities appear to have a specific goal for co-ordination.

Furthermore, all the regions and municipalities have some form of platform for communication where common matters are discussed and processed. The majority takes it one step further by pinpointing boards or equivalent with appointed local representatives from the municipalities. Seemingly, they have common management hierarchies that specify roles and responsibilities for the purpose of inviting local actors to participate in coordinated forums for exchanging services and informing each other. Overall, yet apart from the general approach to goals, we can therefore see a widespread tendency towards the integration form “co-ordination”.

More variation is however found with regards to formulations that imply intended integration in the form of “co-operation” (see Table 3). On the one hand, shared information appears to be common in relation to evaluation activities. The apparent approach is to evaluate and develop agreements in a participatory way by means of interviewing managers and staff (C #13), implementing information efforts and practical guidelines (G #37), or presenting proposals (I #42). These forms of information sharing imply that there is space for informal communication and mutual adaptations. A further example is joint assessments for all local agreements about involvement of GPs (L #58). Some documents also emphasize communication across organizational boundaries, for instance patient surveys about CIP’s and comparison of results across municipalities and regions (e.g. E #29), or national developments and comparisons to develop future home care (P #88).

On the other hand, few of the documents bring up some form of feedback and other forms of more voluntary activities in relation to evaluations. In one such case, employees and managers from municipalities and PCCs have been interviewed as part of analyzing and evaluating how an agreement works in practice:

The ambition has been to show a genuine curiosity about the interviewees’ different dimensions and perspectives towards the questions that the mission involves. The bottom-line is to, by means of interviews, lay a foundation for the conclusions and suggestions for improvement which emerge during later stages of a mission. (C #13, p. 5)

The formulations “… to show a genuine curiosity …” and “… lay a foundation for the conclusions and suggestions for improvement …” speak for the initiative to ask for feedback in a way that enables exchange of comprehensive solutions in-between organizations to reach a common way forward with the agreement. In a similar manner, another region and its municipalities have assessed an agreement by means of sending a questionnaire to municipalities and districts (K #49). However, most of the remaining regions and municipalities appear to fall short of equivalent initiatives.

Formulations in the documents that imply intended integration in the form of “collaboration” are both congruent and different across the regions and municipalities (see Table 3). On the one hand, investments in collaborative activities and projects are seemingly rife. Most commonly, the documents depict skills-enhancing training such as information and training in a support system (A #1) or staff training regarding the provision of care to individuals with extensive needs of home-based healthcare (C #15). One region and its municipalities work with exemplifications of patient cases for various situations (B #7), whereas another has focus areas for development towards “good quality, local health care” (D #25, p. 2). Not least, digital support systems are integrated or under development in all the regions and municipalities. These activities and projects, mostly underpinned by training and learning, can be seen as examples of enabling communication about lessons learnt.

On the other hand, signs of intense networking through daily collaborative working methods, such as multi-professional teamwork, are inconsistent in the documents. Only half of the regions and municipalities include the denomination “multi-professional teams” or equivalent forms of teamwork in the documents. The formulations for teamwork vary, including phrases such as “multi-professional teamwork” (B #7, p. 1), “cross-border collaboration between professional categories” (C #11, attach. 1, p. 9) and “multi-professional working methods” (H #40, pp. 7–8). Other formulations are “team-based form of care” (J #47, p. 8), “mobile community-based teams” (M #63, p. 2), “team-based working methods” (P #88, attach. 1, p. 20), “common mobile forms of care” (R #104, p. 13) and “multi-professional team” (T #115, p. 8). A representative example is illustrated below:

The overarching developmental goal for all municipalities within [name of region] is that individuals with a need of specially organized team-based form of treatment shall have access to it. A form of treatment in which the individual, next of kin/relative and health professional sense a feeling of safety. (J #47, p. 22)

This formulation connects all the municipalities in the region and relates collaboration to development as well as a needed form of treatment, which indicates that teamwork is prioritized. Nevertheless, half of the regions and municipalities fall short of mentioning intense networking in the form of multi-professional teamwork or equivalent.

Our summary of formulated intentions for integration (see Table 3) can be seen as an overview of what the documents tell us about integration beyond contracting, i.e. besides competition and regulation through contracts (Axelsson and Bihari Axelsson, 2013). There appears to be several signs of general efforts for such integration – bringing actors and activities together into a greater unit (Axelsson and Bihari Axelsson, 2013) – yet numerous discrepancies with regards to what these efforts look like in relation to the integration forms co-ordination, co-operation and collaboration (Axelsson and Bihari Axelsson, 2006, 2013). The only consistent patterns for all the regions and municipalities are common values and principles alongside collaborative activities and projects. Common values and principles imply some co-ordination through common management hierarchies and top-level decisions, whereas the activities and projects lean towards some co-operation through informal communication and mutual adaptations (Axelsson and Bihari Axelsson, 2006). We find that the remaining intentions are subject to various extents and combinations across co-ordination, co-operation and collaboration (see Table 3). The results show that formulated intensions sometimes rhyme more with some characteristics in these three integration forms. Yet, the variation is too widespread for us to classify some regions and municipalities as mainly coordinative, co-operative and/or collaborative. This variation makes it relevant to divide our results into an emergent taxonomy to illustrate the formulated initiatives for integration that we can deduce, i.e. the intended forms of integration (see Figure 1).

Furthest to the left, we find the first group (A,F,D,N). We refer to this group as the Contract Connectors, as this is the group where we see integration mainly as contracting. We find that these regions and municipalities comply to contractual standards and formulate common values and principles, suggesting a clear connection to contracting and some efforts in a direction towards co-ordination. For the most part the initiatives stop here, ending with very little integration in total.

The second group (E,G,I,L,S,K) is referred to as the Management-Construct Coalitions. Apart from compliance to contractual standards, we can see how integration is mainly discussed as management initiatives. Common efforts in this group are the introduction of coordinated forums with local representation and information sharing among collaborative partners. Still, the regions and municipalities in this group do not make it all the way through the integration for inter-organizational collaboration as initiatives appear to stop with the construction of common management hierarchies rather than translating talk into action.

The third and the fourth group both show more efforts of intended integration in the form of activities, projects and intense networking, i.e. action rather than just talk. Still, the two groups differ in the vertical and horizontal integration. The third group (B,T) is referred to as the Decoupled Performing Collaborators. Apart from contracting, we find that these regions and municipalities relate consistently to two forms of integration: co-ordination and collaboration. We can read in pervasive formulations about common values and principles, local representation in collaboration boards, collaborative activities and projects, and multi-professional teamwork. Information sharing and feedback (co-operation) is however not expressed at all in all the documents, signaling somewhat less consistent integration in total. We might anticipate that this is taking place, given that shared projects and activities imply a certain degree of co-operation. This is still not evident in the documents, thereby our reference to decoupling between values and principals vis-à-vis activities. Co-operation is a crucial missing link since it incorporates both horizontal and vertical integration, creating space for informal contacts and communication across organizational boundaries, thereby promoting inter-organizational collaboration (Axelsson and Bihari Axelsson, 2006).

In the fourth group (H,M), the Informed Performing Collaborators, we can once again see how initiatives appear to flourish especially in the practical setting of organizational life through collaborative project, activities and multi-professional teamwork. This group takes matters a bit further than the previous group by mentioning initiatives for shared information, which implies that vertical and horizontal integration is taken into consideration, or that both the managerial level as well as the collaborative partners are informed about ongoing activities. However, some integration is missing overall given a complete lack of intentions to support local representation (co-ordination) and feedback (co-operation).

In the fifth group (J,P,Q,C,R,O), we find the Consistent Collaborators. Here we can see clear connections to all forms of integration: contracting, co-ordination, co-operation and collaboration. Beyond contracting, these regions and municipalities appear to strive for consistent collaborative efforts through top-level decisions, sharing of information and feedback, and collaborative activities and projects as well as multi-professional teamwork. Thus, integration shines through consistently.

These five groups represent the intended forms of integration in relation to inter-organizational collaboration, based on patterns and examples in the documents. Our taxonomy is a development of the framework about inter-organizational collaboration by Axelsson and Bihari Axelsson (2006, 2013). The taxonomy presents a new division of various context-based efforts for integration beyond contracting and relates the efforts to an outcome, i.e. inter-organizational collaboration. The analysis will now continue with formulated intentions for control based on degree and content of formalization.

Formulated intentions for control

A summary of formulated intentions for control is presented in Table 4.

First and foremost, we can see a generally high degree of formalization across all the regions and municipalities given considerable quantity and detail in formulations about explicit guidelines (see Table 4). Process maps with instructive descriptions are rife, alongside checklists as well as various forms of listed routines and manuals. The procedures include indicative patient cases at times, depicted as “… exemplifying patient cases which highlight division of responsibilities in different situations” (B #7, p. 1) and similar. Overall, guidelines are frequent, and concern mainly discharge from hospitals and CIP, i.e. matters that are regulated in national legislation. For example, routines for discharge appear in the form of an eight-step process map with color markings and descriptions all the way from “send registration message” to “follow up and finish coordinated individual plan” (K #51, pp. 8–21). Similarly, routines for CIP’s show up in a seven-step guide with instructions all the way from “assess the need” to “terminate CIP” (A #5, p. 2). The widespread existence of such guidelines implies structure, stability and clear definitions of expectations. This pattern represents a consistent hard approach to control in all the regions and municipalities.

On the other hand, the quantity and content of performance areas and measuring of targets seem to lean towards a low to moderate degree of formalization across the documents. We cannot see clear signs of a comprehensive hard approach to control in this regard due to few formulations expressing meticulous monitoring and controlling of precise performance metrics and goals. With a few exceptions, the total amount of performance areas comes down to maximum five per region (see Table 4). Development of care treatment as well as discharge are the two performance areas where most regions and municipalities have formulated targets. Several targets for discharge from hospitals emphasize a safe discharge process as perceived by the individual (e.g. A #3) and early discharge, i.e. the possibility for individuals to return to their home as soon as possible when they no longer need hospital care (e.g. K #50). Similarly, target formulations for involvement of GPs encircle safe handling of patients (D #23), experienced participation and safety (M #59), and competence (P #89). The nature of these expressions is, still, seemingly general rather than precise. Most of the regions and municipalities appear to measure their targets nonetheless, most commonly with regards to days until discharge, patients’ experiences and quality indicators found in national quality registers.

Similarly, the pattern concerning frequency of evaluations points at a generally low or moderate degree of formalization. The frequency of evaluation activities comes across as low in all the regions and municipalities (see Table 4). The most common formulations are “once every year” (e.g. C #11, p. 15), alternatively “yearly” (R #104, p. 13), or “at least once every year” (H #40, p. 11). Yet, formulations such as “continuously” and “when needed” occur among more than half of the regions and municipalities. The only activity that seems to be evaluated more frequently are the CIP’s. The great majority express that they should be evaluated in conjunction with the CIP. In addition, three regions and their municipalities use monthly evaluations of compensation amount in the discharge process (E #27, attach. 9; G #38; K #50). The common approach is low evaluation frequency, nonetheless. This pattern suggests more autonomy than control in general, in other words more of a soft rather than a hard approach to control.

With regards to content of formalization, the main purpose of most evaluation activities appears to be long-term learning, organizational development and improvement for the sake of performing satisfactory care to patients (see Table 4). We can sense a soft approach to control driven by this clear focus on learning and durability rather than accountability. For instance, one document describes evaluation of dementia in the following way: “The purpose of the evaluation is to assess the progress of the disease and its consequences, and thereby offer individualized support” (H #40, p. 10). Another description emphasizes knowledge-building and development:

The goals for co-ordination in [name of focus area] shall be followed to create prerequisites for safe and healthy ageing and a better life for elderly in [name of region] as well as a dignified end. The focal point shall be about health-enhancing and preventive activities, fewer cases of elderly people falling and hurting themselves as well as seamless care, good drug treatment and good care during the end of life. (N #75, p. 10)

The two examples represent common formulations in which the apparent purpose of evaluating something is to build knowledge, progress continuously, enhance quality and strive for efficient processes and routines. Other purposes for evaluation are to assess the collaboration, but also the ability to change or add something to an agreement. Some documents thus express more tendency towards a purpose of monitoring, which suggests a harder approach to control. Some of these examples are to evaluate the division of accountability (D,I,M,N), to control compliance and to analyze deviations or the amounts of compensations (E,I,O,R).

Lastly, the form for presenting evaluations is a matter of deviation within and across the regions and municipalities, including dialogues, meetings, inspections and reports. The trend appears to be somewhere in-between a hard approach to control in the form of documenting and reporting versus a soft approach to control in the form of interacting and discussing (see Table 4). For instance, one region and its municipalities repeat that some activities concerning planning and evaluations are to be documented whereas forms for other evaluation activities are unspecified (E #27, attach. 1,4,9). A great part of the documents nevertheless express requirements for some form of documentation, which indicate that information is sometimes used along the lines of a hard approach to control.

Altogether, we can see that the documents generally lean towards soft control due to clear signs of low to moderate degrees of formalization (e.g. Bauer et al., 2022) alongside content of formalization characterized by learning and durability rather than accountability (Leeuw and Furubo, 2008) (see Table 4). In Figure 2, we illustrate the control initiatives together with the intended form of integration to inter-connect the two dimensions.

We find that the general pattern for all the five groups is a close to moderate connection to a soft approach to control. Regions and municipalities from all the groups seem to evaluate mainly for the sake of learning and improvement, and to present evaluation results in both oral and written forms (A,G,L,S,K,B,H,Q,C). This approach stands out as the most common undertaking of evaluation activities. Regions/municipalities from three of the groups express learning and improvement as the main purpose for evaluations but do not say anything about the evaluation form (F,T,P). We can also conclude that low evaluation frequency is common.

The only consistent sign of a high degree of formalization for all the five groups is the apparent focus on explicit guidelines. Such detailed guidelines can be understood as a way of bringing efficiency to professional work (De Waele et al., 2021; Fırtın and Karlsson, 2020) and supports the performance metrics and monitoring approach that characterizes hard control (Bauer et al., 2022; Gazley, 2008). Some additional signs of hard control also appear across the five groups, mainly concerning evaluation for the purpose of monitoring (e.g. Karrbom Gustavsson and Hallin, 2014) and documentation over dialogue (Leeuw and Furubo, 2008). Regions D and R, from the first and the fifth groups, express that evaluations are made for accountability purposes (as well as learning and improvement) and show no signs of dialogue. In addition, we find formulations in favor of accountability evaluation purposes in seven regions and their municipalities across four of the groups (D,N,E,I,M,R,O). Amongst these, we can however see evaluations merely in the form of documentation (D,R), documentation together with dialogue (E,I,M,O) or no expressed evaluation form at all (N).

Our results point towards a soft rather than hard approach to control across the regions and municipalities, which is described in previous research as important for long-lasting collaboration (Adler and Borys, 1996; Bianchi et al., 2021; Jordan and Messner, 2012). Collaboration is, to a large extent, depicted as a matter of communication, learning and networking (Axelsson and Bihari Axelsson, 2006, 2013). These are, also, characteristic traits of a soft approach to control (Bauer et al., 2022; Gazley, 2008; Karrbom Gustavsson and Hallin, 2014). Therefore, based on the suggestive associations between integration and soft control in previous research, we would expect to see a symmetric relationship between more intended integration vis-à-vis more soft control in the five groups (see Figure 2). However, the illustration of our results highlights a surprising dissemination between the positioning of the five groups (the extent of integration) and approaches to control (the extent of soft or hard control). The more integration, the more tendencies we can see in favor of managing and supporting inter-organizational collaboration through the documents. Yet, the results demonstrate no difference in approaches to control between the five intended forms of integration. Regardless of the initiatives, signs of soft control are evidently clear in the constant focus on learning and dialogues while signs of hard control mainly flourish in the explicit guidelines.

Conclusions and contributions

The call for enhanced, integrated and sustainable service delivery has led to an increase in collaborative efforts among organizations within the public sector, especially in healthcare (Ansell and Gash, 2007; Bianchi et al., 2021; Kokko and Laihonen, 2022; Wang and Ran, 2023). While existing research has predominantly addressed the reasons behind and the nature of such collaborations, there is a need for deeper investigation into how collaboration can be managed, especially how challenges and tensions within collaboration is handled on system and organizational levels (Bryson et al., 2015; Evans et al., 2023; Gazley and Guo, 2015; Karlsson et al., 2020). Our study corresponds to this call by focusing on the relationship between intended forms of integration and approaches to control from the perspective of decision-makers. By combining Axelsson and Bihari Axelsson’s (2006, 2013) framework about inter-organizational collaboration as differentiated or integrated with the notion of approaches to control as hard or soft (Bauer et al., 2022; Gazley, 2008; Karrbom Gustavsson and Hallin, 2014; Lartey et al., 2023; McGregor, 1960; Truss et al., 1997), we analyze formulations in collaboration agreements and associative documents for care of elderly between Swedish regions and municipalities. By presenting a taxonomy of so called “integration initiatives” and relating this to approaches of control, our iterative study builds on and adds to a recent stream of research arguing that the relationship between collaboration and approaches to control may by fuzzier and more complex than originally thought (Evans et al., 2023).

Our study suggests two important contributions. First, we respond to recent calls for more research on how to conceptually grasp the content of different forms of collaboration (D'Amour et al., 2005). Addressing this call, our paper proposes an empirically driven taxonomy of intended forms of integration initiatives that provides resources for studies about how collaboration can be managed when it is stipulated by national legislation but local self-governance gives actors considerable freedom to decide on how to organize and manage services. Our taxonomy (see Figure 1) denotes that intended integration between Swedish regions and municipalities is subject to remarkable variation in intended forms of inter-organizational collaboration. The most common forms are Management-Construct Coalitions and Consistent Collaborators.

Previous research suggests that inter-organizational collaboration in public health is above all a matter of the integration forms co-operation and collaboration (Axelsson and Bihari Axelsson, 2006), for example voluntary work, informal information, networking and multi-professional teamwork. In support of Evans et al. (2023), our findings add to this understanding by illustrating that the relationship between collaboration and approaches to control may be even more complex in reality. From the perspective of decision-makers, we show that various combinations of the integration forms in previous research (contracting, co-ordination, co-operation and collaboration) (Axelsson and Bihari Axelsson, 2006, 2013) exist in intended management and support of inter-organizational collaboration. This is not only an addition to the notion of collaboration as a dynamic and evolving process involving humans (Axelsson and Bihari Axelsson, 2006; D'Amour et al., 2005; Gazley, 2017). It also says something about how little it takes to find new combinations and versions of integration within a given context. The taxonomy presents a division of collaborative efforts with new combinations of the integration forms because previous research about different integration forms (Axelsson and Bihari Axelsson, 2006, 2013) was insufficient to fully illustrate signs of intended integration in this Swedish case. Our findings, based on observations from only one sector in one country where actors are subject to the same regulations and overarching complex problem, illustrate that decision-makers’ intentions with proposed collaboration in each given context are important for the chosen approach to control (Karlsson et al., 2020). The seemingly important role of context is underpinned by the remarkable variation in the content of collaboration agreements between the regions and municipalities.

Second, and related to the discussion about tensions and support between control and collaboration (Evans et al., 2023; Jayasinghe et al., 2020), our study suggests that all the regions and municipalities express an imminent soft approach to control in the agreements alongside limited signs of hard control. The signs of hard control are mainly visible in the explicit guidelines (see Figure 2). Previous research imply that hard control may undermine trust and even restrict effectiveness and efficiency of joint efforts (Adler, 2001; Romzek et al., 2014). We may consequently expect that integration will rely on soft control for collaboration initiatives to be efficient. At large, our findings support these arguments. However, the findings are true regardless of the intended forms of integration. Whether the agreements stipulate multi-professional teamwork or only express common values, the approaches to control appear to stay primarily soft. As a result, we cannot assert that a certain intention is associated with a certain form of control (Romzek et al., 2014). This asymmetry stands in contrast to suggested connections between more soft control vis-à-vis more collaboration in previous research (e.g. Adler and Borys, 1996; Axelsson and Bihari Axelsson, 2013; Bianchi et al., 2021; Ghoshal and Moran, 1996; Jordan and Messner, 2012). As such, our findings show that the relationship between intended forms of integration and intended approaches to control is fuzzier and more complex than originally thought (Evans et al., 2023).

In addition to the above, our results illustrate that various forms of intended integration can be managed by soft and hard approaches to control simultaneously (Gazley, 2008; Romzek et al., 2014) insofar as the studied collaboration agreements appear to have a two-sided purpose. Whereas soft control permeates, the law-binding parts of the agreements are restricted by hard control. In this Swedish case, national legislation for collaboration (e.g. Act on Coordinated Discharge, 2017:612; Health and Medical Services Act, 2017:30) appears to necessitate a certain extent of hard control to add structure, stability and goal-orientation. At the same time, soft control appears to regulate collaboration beyond contracting, i.e. the parts focusing on long-term learning, information sharing, development projects and similar. This finding brings nuance to previous research about soft control as the significant approach for sustainable collaboration (Adler and Borys, 1996; Bianchi et al., 2021; Ghoshal and Moran, 1996) and aligns with previous research stating that hard and soft control can complement each other (Gazley, 2008; Karrbom Gustavsson and Hallin, 2014; Poppo and Zenger, 2002). Consequently, we suggest that the two-sided purpose of both promoting adherence to law-binding standards and facilitating collaboration is why hard and soft control exist side-by-side. The finding that actual signs of soft and hard control depend on the studied context adds to our understanding of the complex relationship between inter-organizational collaboration and control.

Our study also has practical implications for decision-makers. Our empirical results suggests that contractual documents are more than mere associations to competition between organizations and differentiation (Axelsson and Bihari Axelsson, 2006), i.e. they stretch far beyond compliance with contractual standards. When reading such documents, one gets impressions of intended approaches to inter-organizational collaboration and control in the organizations. To set the right expectations for decision-makers and managers, it is therefore highly relevant to aim for explicit formulations and continuous evaluations of the documents. The observed variety in this empirical study makes it just as relevant to consider the role that contracts play in action, i.e. in managing and supporting inter-organizational collaboration. Our suggested two-sided purpose of agreements implies a potential conflict of goals between managing and supporting inter-organizational collaboration according to hard and soft control simultaneously. A successful combination of control approaches is therefore important to understand and consider within each organizational context.

Our empirical study comes with both strengths and limitations. Despite our approach to include all the relevant documents in the chosen setting, the selection criteria were partly contingent upon presentation of documents as is in various and sometimes sprawling public listings online. The outcome is therefore limited to a slightly diffuse localization of data in combination with the interdisciplinary nature of the topic. It is appropriate to question whether the same empirical findings and conclusions would be realized by studying the same topic in other contexts. Nevertheless, we have assured a constant focus on validation, reviews and complementary searches of documents to mitigate bias. Here, we can say something about what is expressed in these specific documents in the given context. Further research in other contexts would be needed to extend knowledge about the relationship between inter-organizational collaboration and control. Would the nature of inter-organizational collaboration change in another geographical context or sector? Would the relationship between inter-organizational collaboration and control differ? We encourage further analyses of contractual documents about collaboration in other settings to test and compare our empirical findings for the purpose of generating more knowledge about tackling complex problems in society from a decision-making perspective. Finally, further research is needed about how intentions are translated into practice and what happens with collaboration at the actors’ level. What hindering and facilitating factors exist at the actors’ level? Answers to such questions presuppose multiple cases, dialogues with key representatives and longitudinal approaches in future research.

Figures

Figure 1

Emergent taxonomy of integration initiatives

Figure 2

Control initiatives among the different integration initiatives

Table 1

First part of analytical framework: four forms of integration

Integration forms	Content of integration forms	Examples of themes and codes (−) in categorization of data
(1) Contracting	Compliance with law-binding standards and requirements, division of responsibilities and accountabilities	Law-binding requirements - law-binding responsibilities for regions versus municipalities - specification of payment liabilities for regions versus municipalities
(2) Co-ordination	Co-ordination, exchange of services, information, supervision. Reference of patients between each other. Mainly common management hierarchy, top-level decisions - Values/principles - Precise goal - Local representation	Established collaboration beyond law-binding requirements - pronounced initiatives for collaboration in terms of quality/development/innovation - established structure for collaboration Goals, targets, monitoring and evaluation - goals in relation to collaboration and formulation of goals
(3) Co-operation	Hierarchical management hierarchy in combination with voluntary work. Some communication, informal information sharing, exchange of comprehensive solutions - Shared information - Feedback	Goals, targets, monitoring and evaluation - content of evaluation (what is evaluated?) - formulations of what the evaluation results are used for
(4) Collaboration	Fusion or coalition between actors into one unit. Voluntary work, high level of communication, networking, multi-professional teamwork - Activities/projects - Teamwork	Established collaboration beyond law-binding requirements - support for collaboration-related work and forms of support - support system for collaboration and definition of systems Organization of multi-professional teams - denominations of multi-professional team or equivalent

Source(s): Authors’ own work

Table 2

Second part of analytical framework: approaches to control

Approaches to control	Degree of formalization (high/low)	Content of formalization (accountability/improvement)	Examples of themes and codes (−) in reverse categorization of data
(1) Hard	- High number of guidelines - High number of performance areas - Measuring of targets - High evaluation frequency	- Evaluation purpose: accountability - Evaluation form: documentation/report	Established collaboration beyond law-binding requirements - support for collaboration-related work and forms of support
(2) Soft	- Low number of guidelines - Low number of performance areas - No measuring of targets - Low evaluation frequency	- Evaluation purpose: improvements/learning - Evaluation form: dialogue	Goals, targets, monitoring and evaluation - goals in relation to collaboration and formulation of goals - existence and definitions of measurements - evaluation frequencies - formulations of what the evaluation results are used for - forms for evaluation outcomes (dialogue/report)

Source(s): Authors’ own work

Table 3

Summary of formulated intentions for integration

Region/ Intentions	A	B	C	D	E	F	G	H	I	J	K	L	M	N	O	P	Q	R	S	T
Co-ordination
1. Values/principles	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
2. Precise goal															X
3. Local representation		X	X	X	X		X		X	X	X	X		X	X	X	X	X	X	X
Co-operation
4. Shared information			X		X		X	X	X	X	X	X	X		X	X	X	X	X
5. Feedback			X								X				X			X
Collaboration
6. Activities/projects	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
7. Teamwork		X	X					X		X			X		X	X	X	X		X

Source(s): Authors’ own work

Table 4

Summary of formulated intentions for control

Region/ Intentions	A	B	C	D	E	F	G	H	I	J	K	L	M	N	O	P	Q	R	S	T
Degree of formalization
1. No. of guidelines high (H)	H	H	H	H	H	H	H	H	H	H	H	H	H	H	H	H	H	H	H	H
2. No. of performance areas	1	1	2	2	1	1	3	0	3	2	11	1	3	2	5	4	3	3	4	3
3. Are targets measured?	No	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	No	No
4. Evaluation frequency low (L)/moderate (M)	L	L	L	L	M	L	M	L	L	L	M	L	L	L	L	L	L	L	L	L
Content of formalization
5. Evaluation purpose improvements (I) or accountability (A)	I	I	I	I/A	I/A	I	I	I	I/A	I	I	I	I/A	I/A	I/A	I	I	I/A	I	I
6. Evaluation form documentation/report (R) or dialogue (D)	R/D	R/D	D/R	R	R/D	–	R/D	R/D	R/D	R	R/D	R/D	R/D	–	R/D	–	R/D	R	R/D	–

Source(s): Authors’ own work

Appendix

The supplementary material for this article can be found online.