The significance of interprofessional and interagency collaboration in reintegration after prison: A qualitative study exploring Norwegian frontline workers’ views

The Norwegian Correctional Service is well known for its focus on rehabilitation and the humane treatment of offenders. However, welfare issues and comorbidity are overrepresented among offenders, and recidivism rates remain unacceptably high. Mental health problems, substance abuse and a lack of housing suggest that offenders need support from a range of services in their reintegration processes. This calls for collaboration between frontline workers, welfare agencies and non-governmental organizations, especially in the transition from prison back into society. 
In the present study, we aim to explore frontline workers’ views of interprofessional and interagency collaboration among frontline workers working with offenders suffering from substance abuse issues in their reintegration after prison. 
Semi-structured interviews were conducted with nine frontline workers employed in welfare agencies and the correctional service, with workers directly engaged in supporting offenders´ reintegration after prison. 
Findings suggest that interprofessional collaboration is perceived as multifaceted. The participants in the study perceived the welfare needs of offenders as complex, and the transition phase from prison as particularly vulnerable. Finally, findings suggest that frontline workers’ individual values and engagement in the work, as well as a lack of shared knowledge and shared information among frontline workers, are perceived as important factors in how collaboration processes unfold. 
We further argue that there is a need for additional knowledge, such as theoretical frameworks and conceptual models, to increase the understanding of interprofessional collaboration in the interface between prison and welfare services. We discuss substance abusers’ transition from prison into society and interprofessional collaboration in this context, using relational coordination as a theoretical framework. 
This study shows that relational coordination contributes to a greater understanding of interprofessional collaboration in the prison-welfare context, but an understanding of this phenomenon may be further developed by expanding the theory of relational collaboration, and by using other relevant theories and models. New insights are presented and illustrated, combining the theoretical and practical aspects of interprofessional collaboration.

governmental organizations, especially in the transition from prison back into society.
In the present study, we aim to explore frontline workers' views of interprofessional and interagency collaboration among frontline workers working with offenders suffering from substance abuse issues in their reintegration after prison.
Semi-structured interviews were conducted with nine frontline workers employed in welfare agencies and the correctional service, with workers directly engaged in supporting offenders´ reintegration after prison.
Findings suggest that interprofessional collaboration is perceived as multifaceted.
The participants in the study perceived the welfare needs of offenders as complex, and the transition phase from prison as particularly vulnerable. Finally, findings suggest that frontline workers' individual values and engagement in the work, as well as a lack of shared knowledge and shared information among frontline workers, are perceived as important factors in how collaboration processes unfold.
We further argue that there is a need for additional knowledge, such as theoretical frameworks and conceptual models, to increase the understanding of interprofessional collaboration in the interface between prison and welfare services.
We discuss substance abusers' transition from prison into society and interprofessional collaboration in this context, using relational coordination as a theoretical framework.
This study shows that relational coordination contributes to a greater understanding of interprofessional collaboration in the prison-welfare context, but an understanding of this phenomenon may be further developed by expanding the theory of relational
Several researchers have highlighted the need for collaboration in supporting offenders who are being released from prison (Clairborne & Lawson, 2005;Friestad & Kjelsberg, 2009;Larsen et al., 2019;Travis, 2005). It has been suggested that effective interprofessional collaboration is a necessity to achieve sufficient service quality in the rehabilitation process (Denton, 2014;Gisler et al., 2018;Helgensen, 2019). However, Bond and Gittel (2010) demonstrated that welfare services and criminal justice services often operate independently of each other, competing for resources and operating within organizational and professional silos. It has been documented that professional boundaries often foster a logic of competition rather than collaboration, established along the lines of distinct theoretical and disciplinebased frameworks and professional jurisdictions (D'Amour et al., 2005). Some may even see interprofessional collaboration as a threat to traditional professional autonomy (Pihl, 2009).

Research on interprofessional collaboration
The field of interprofessional collaboration is complex and hard to grasp, and touches several disciplines such as psychology and criminology (Willumsen & Ødegård, 2016). There is a significant amount of research implicitly describing collaboration, such as research concerning the restorative justice model (Robinson & Shapland, 2008) and desistance theory (Weaver & McNeil, 2014), although research regarding the phenomena of interprofessional collaboration explicitly is rather limited (Reeves et al., 2017). However, the quantity and quality of research on interprofessional collaboration, as well as the consequences it has for service quality, have grown in recent years (Körner et al., 2016;van Leijen-Zeelenberg et al., 2015).
Interprofessional collaboration may be defined as the process by which health and social care frontline workers work together positively to impact the provision of welfare service and care (D`Amour et al., 2005;Reeves et al., 2017;Willumsen & Ødegård, 2016). As a phenomenon, interprofessional collaboration often consists of factors on several levels (Ødegård, 2006), such as interpersonal relations between team members and organizational and structural factors (D`Amour et al., 2005).
Within the prison context, studies focusing on interprofessional collaboration are also increasing.
In their study exploring collaboration between criminal justice and substance abuse treatment agencies, Fletcher et al. (2009) identified a two-tier system of collaboration: less structured informal networking on the one hand, and structured, formalized levels of collaboration on the other. Their research also found that the size of collaborating organizations was influential, with smaller facilities having a positive effect on collaboration, bringing frontline workers into closer contact (Lehmann et al., 2009). Similarly, , measuring relational coordination (RC-scale) between mental health services and the Norwegian Correctional Service, found that even though nurses, social workers and other prison officers communicated with each other the most, prison officers seldom communicated with psychiatrists in mental health and drug services. In another study, Hean et al. (2018) also found logistical issues, limited resources and differences in professional judgments on referral and confidentiality as additional challenges that limited interprofessional collaboration in the criminal justice environment.
Despite a growing number of studies on interprofessional collaboration in the criminal justice and other different contexts, there is still a need to build theoretical frameworks and conceptual models to explain and articulate collaboration (Reeves et al., 2017).
Moreover, there is currently a lack of detailed understanding about how collaboration takes place in specific prison populations and at particular points in an offender's journey. There is also a lack of understanding about the views of interprofessional collaboration from the perspective of the frontline worker. These workers are the key to the successful implementation of national policy, and are the stakeholders likely to know the offender and the offender's needs the best (Bazemore & Erbe, 2004;Ward & Maruna, 2007).
In particular, there is a need to better understand the perspective of the frontline worker, and to explore the processes of collaboration as they play out within the collaborations regarding specific prison populations. In addition to their outcomes, knowledge about interprofessional collaboration processes in this context is essential if welfare services are to be further developed for this stigmatized and marginalized group of offenders. A qualitative in-depth study of these processes, and how they are experienced subjectively, is required, and is the main aim of this paper.

The Norwegian context
Norway is a useful context in which to explore collaboration in a prison context, as it is known for its strong welfare system characterized by universalism and egalitarian principles (Gisler et al., 2018;Hatland et al., 2018). Its strong commitment to these principles may explain the Norwegian Correctional Services' humane treatment of offenders and its focus on rehabilitation (Garland, 2001;Pratt, 2008aPratt, , 2008b. On the other hand, the Norwegian prison population shares many similarities with the international prison population, thus making findings relevant in an international setting as well. A total of 60% of Norwegian offenders are reported to be dependent on drugs; 60% are homeless and only 8% of the prison population show no sign of mental illness (Cramer, 2014;Friestad & Hansen, 2004;Revold, 2015). Although recidivism rates are among the lowest internationally, recidivism rates are as high as 75% among offenders between the ages of 25-44 convicted of theft, and 37% among offenders convicted for the use or sale of illegal substances (Fazel & Wolf, 2015;Graunbøl et al., 2010). A review of the literature, albeit in crime prevention in Norway and work involving the police, found that offenders with substance abuse issues are marginalized in prison, and do not have access to necessary welfare services.
Several projects have been initiated, but the problem is not primarily a lack of projects, but rather a lack of action and an implementation of specific measures regarding the reintegration of this offender group (Rundhovde & Skjevrak, 2018).
During the last decades, the reintegration of offenders with substance abuse issues has been approached through the following strategies in Norway: 1) a legally binding collaboration between agencies and professions (Law of Execution of Sentences, 2001. §4;Law of Labour and Welfare Administration, 2006, §8); 2) the employment of reintegration coordinators (Falck, 2015); 3) collaboration agreements between welfare agencies (Falck, 2015); 4) the establishment of departments in prison for substance abusers (Helgesen, 2019); and 5) the involvement of third-sector voluntary and mentor work (NGO), for example, the Red Cross Network after imprisonment and the Church's City Mission (The Norwegian Correctional Service, 2021).
Despite these measures, living conditions and recidivism rates remain stable and high for offenders with substance abuse issues (Fazel & Wolf, 2015;Revold, 2015).

Theoretical framework
The theoretical framework underpinning our view of collaboration in this study primarily builds on the theory of relational coordination. Relational coordination (RC) is defined as "a mutually reinforcing process of interaction between communication and relationships carried out for the purpose of task integration" (Gittel, 2002, p. 301).
The theory of relational coordination has been developed and tested in the context of surgical care (Gittell et al., 2000), medical care (Gittell, Weinberg, Bennett, & Miller, 2008), long-term care (Gittell, Weinberg, Pfefferle, & Bishop, 2008) and the criminal justice system (Bond & Gittell, 2010;. According to Gittell (2001), relational coordination theory is useful in its ability to articulate the "work processes in which multiple providers are engaged in carrying out highly interdependent tasks under conditions of uncertainty and time constraints" (p.3). Relational coordination combines two dimensions of collaboration: a structural and a relational one. The relational dimension describes the extent to which different organizations/frontline workers share knowledge, hold common goals and have mutual respect for each other as collaborating individuals. For instance, a prison officer and social worker may share the goal of supporting a prisoner to find housing upon release from prison.
The communication dimension describes the structural features or systems that surround them, namely, the frequency with which the individuals meet/communicate, the timeliness of these interactions, the accuracy of the information shared and the problem-solving nature or focus of the communication .
Communication between frontline workers should be accurate and reliable, and of sufficient frequency if the reintegration process is to have a chance of being successful.
The relational and communication dimensions in relational coordination are interrelated, with each of them impacting the quality of the other (see Figure 1). (Gittell, 2011, 402, with permission from Gittell) Furthermore, RC is optimized if a variety of antecedents are in place that may promote this conceptualization of collaboration (e.g. working practices such as interagency meeting). Similarly, effective RC in turn impacts on service and client outcomes (e.g. quality and efficiency outcomes related to the quality of care) (Bond & Gittell, 2010;Gittell, 2011).

Figure 1: Dimensions of Relational Coordination
According to Gittell's (2011) viewpoint, relational coordination provides a counterpart to the traditional bureaucratic form of coordination and collaboration that often reinforces professional and organizational boundaries. Bond and Gittell (2010) argue that relational coordination is a growing theory which may enrich our understanding of interprofessional collaboration, a view supported by other researchers Otte-Trojel et al., 2017;Stühlinger, Schmutz, & Grote, 2019). Relational coordination offers a pragmatic, operational and bidimensional view of both collaboration and integration, and it is argued that it is significant in achieving an effective collaboration between welfare services in providing successful offender reintegration (Bond & Gittell, 2010). Relational coordination has been applied in the criminal justice context by Bond and Gittell (2010), and later by .
Both Bond and Gittel (2019) and  use surveys to assess levels of relational coordination. As quantitative studies can neither definitively explain the processes involved, nor why the relational coordination levels are as they are, there is scope for a more in-depth qualitative exploration of the reasons behind the levels of relational coordination in this context. This paper contributes to this perspective, by specifically exploring the views of frontline workers working with offenders upon their release, and exploring these views of collaboration through the lens of relational coordination.

Method
A constructivist philosophy underpins the exploratory and descriptive qualitative research approach (Caelli, Ray, & Mill, 2003) used in this study. This approach was relevant because the study aimed to explore the range of frontline workers' subjective views of collaboration and the processes involved in it.

Sample
The target population of the study was frontline workers employed in welfare agencies and the correctional service who handle offender reintegration after prison.
The selection of frontline workers was based on two criteria: (1) their having more than six months of experience working with offenders with substance abuse issues, and (2) their daily work being in the frontline service (rather than in a management position).
The sample was selected by sending out an inquiry to prisons, the Norwegian Labour and Welfare Administration (NAV), municipalities and substance abuse treatment centres. Those who were interested in participating in the study contacted the first author. Individuals from some welfare agencies were more unavailable than those from others. It was especially difficult to get a hold of participants from health care agencies and NAV. This may be explained by the ongoing pandemic, Covid-19, which created a lot of pressure, particularly on these welfare agencies.
In total, nine frontline workers participated in the study: three nurses (one working in the prison, a second from a substance abuse treatment centre, and the last, in the community), two prison officers working in the prison and one at a probation office, one therapist (from a substance abuse treatment centre) and two social workers (one from the volunteer sector -the Red Cross and one from NAV). The participants' experiences with the targeted group ranged from three to 25 years.

Materials
A semi-structured interview guide was developed to explore participants' views of interprofessional collaboration. The interview guide contained questions about the participants' backgrounds (e.g. years of work experience) and their experiences with this group of offenders (e.g. What are your experiences regarding reintegration after prison for offenders with substance abuse issues?). Specific probing questions were included that explored their experiences with interagency and interprofessional collaboration. These questions were underpinned by the theoretical framework of relational coordination (Bond & Gittel, 2010;Gittel, 2011). Hence, we asked participants to explore both the structural and relational components of their interagency and interprofessional collaborations, touching on themes in the interview that related to the means of sharing knowledge, the presence of common goals and mutual respect, and the nature of communications (e.g. frequency, timeliness, accuracy and constructiveness) (e.g. How do you experience communication with other frontline workers and agencies in the transition phase from prison to community? Do you receive enough and appropriate information from-and about other frontline workers?) The interview guide was critiqued by a panel of five to establish the trustworthiness of the instrument. The panel included researchers from academic institutions, trained social workers and prison officers from the correctional service.

Data collection and procedure
All interviews were conducted by the first author, and took place either in prison facilities or the respective welfare agency locations over a period of three months.
The interviews were audio recorded, and lasted approximately one hour. All interviews were conducted in Norwegian, and the relevant quotes later translated into English. Interviews were transcribed verbatim.

Analysis
The analysis employed a combination of inductive and deductive approaches (see Figure 2), as it has been suggested that gaining information using a combination of inductive and deductive approaches within a qualitative research design may provide new, and more in-depth, knowledge about the subject being studied (Bernard, 2017;Tjora, 2017).

Figure 2: An Illustration of Analysis
A five-step template analysis of the interview transcripts (King, 2012;Malterud, 2001) was conducted. This form of thematic analysis emphasized the use of hierarchical coding in the analysis but balanced this with a relatively high degree of structure provided by the RC framework and the research aim.
The five-step template analysis was conducted as follows (King, 2012): (1) The transcripts of each interview with frontline workers were read through in full several times to become familiar with the raw data material to be analysed.
At this stage, we had an open mind and restrained ourselves from doing any coding. The only purpose was to become familiar with the data.
(2) Second, a preliminary coding of the interviews was conducted. Although we had a set of a priori issues based on the theoretical framework of relational coordination and the former work experience of the first author as a probation officer, it was also important at this stage to maintain an open mind and not force the data to fit the a priori issues.
(3) Third, an initial template was developed. The theoretical framework of relational coordination informed the creation of this initial template, but themes that developed in this emerging template were also allowed to emerge from the data itself through an inductive approach.
(4) The initial template was then tested on the transcript to determine if a pattern in data material related to themes developed in the template was found, both in terms of the a priori themes from the relational coordination framework, and the preliminary themes developed from the inductive approach. The initial template was modified, and a "final" template was developed.
(5) The "final" template was applied to the remaining transcripts. It aided in the interpretation of these additional transcripts, although the template was allowed to develop as new data material entered the analysis.
The analysis was an iterative and reflexive process leading us to repeat the above steps several times, testing themes that were developing in the initial template as the analysis progressed (Tjora, 2017).

Ethics
The study was approved by the Norwegian Social Science Data Service on 30 January 2020, and by the Norwegian Correctional Service Region West on 26 February 2020 (201901020-34). Participants were recruited through a letter presented by managers at the agencies, and it was clearly emphasized that participating in the study was voluntary. Agreement to participate was declared in written form by the participants. Before the interview, it was made clear that the study was not about specific patients or offenders, but, rather, general experiences with interprofessional collaboration.

Findings
Three main and interrelated themes emerged from the analysis. Frontline workers working with offenders reintegrating back into society believed the following: 1. Offenders have multifaceted needs, and these offenders are particularly vulnerable when in the transition phase from prison back into society.

Frontline workers' individual values and their personal engagement in the lives
of "their clients" drive the collaboration processes. 3. Communication and relational dimensions influence collaboration negatively.
The themes were discrete, but interrelated. Participants describe the multifaceted needs of offenders, and the support needed from a variety of welfare professions and agencies; hence, these professions and agencies must then collaborate to provide the necessary quality in welfare service (Theme 1). According to the participants in this study, this collaboration is influenced by frontline workers' individual values and personal engagement (Theme 2), but also by the structures that mediate the quality and ease with which knowledge and information is shared (Theme 3). Implicitly, these factors affect the welfare service provided to the offenders.
Theme 1: Offenders have multifaceted needs. They are particularly vulnerable in the transition phase from prison back into society.
The initial template was built on codes related to frontline workers' experiences with the group in general, and their common experiences were related to the complexity in needs and vulnerability in reintegration. The participants revealed that they perceived the life situations of the offenders in the reintegration process as multifaceted. The multifaceted nature of offenders' needs is illustrated by some offenders needing, for example, both proper housing and economic support. Others lack a supportive social network and, in addition, need mental health treatment: They often lack education, have dyslexia, they`ve been bullied at school, often been sexually abused, lack of care in general. At least, many of them have personality disorders, anxiety, depression, no money, lack of resources surrounding them, no support. This is the majority of those (offenders) I`ve worked with. (nurse, substance abuse treatment centre) The multiple and bespoke nature of the permutations of needs of each individual offender, makes offenders, especially those with substance abuse issues, particularly vulnerable: All transition phases are vulnerable. Going from an isolated situation like prison and then entering community, without the necessary care, it gets highly challenging. The development of this theme started with factors viewed as being on an individual level in general. For example, many of the participants stated that collaboration is dependent on the frontline worker involved. A further analysis of these statements suggested that the individual factors primarily concern the values and attitudes held by the frontline workers, as well as the motivation and drive toward progression in the rehabilitation processes. Offenders, especially those with substance abuse issues, were described as not only having a low status in the community in general, but also among frontline workers in the welfare services. The participants claimed that offenders are often stigmatized because of how they have lived. One of the participants, himself an ex-substance abuser, stated: They have always been the last priority. There's a basic misunderstanding among people who have not had these problems themselves, that it is a choice. You have chosen to be a drug addict. I have been faced with that attitude myself. You are a bad person. (social worker, Red Cross) The low status of offenders with substance abuse issues, both among the general public and within welfare services, and their lifestyle, characterized by crime, lying and cheating, seems to activate the attitudes of the frontline worker. One of the participants claimed that among prison officers: … there are mixed attitudes in prison. Some have a mindset based on the importance of security, others have a mindset based on the importance of social work. (prison officer, closed facility) The above illustrates that individual values and mindsets among frontline workers may very well influence how collaboration unfolds, both between prison staff and between prison staff and frontline workers outside the prison.
However, according to the participants, professional background did not necessarily dictate the attitudes of individuals engaging with this population and their desire to collaborate. A nurse working in the community/municipality commented:

You have uninterested medical doctors, and you have interested medical doctors.
Another nurse working at a substance abuse treatment centre stated: The help you (offenders) get is totally dependent on who you meet in the welfare system. If you (offenders) meet someone who is engaged in the work and motivated it will make a difference. We know that.
Additionally, "personal chemistry" among the frontline workers was perceived as more important than formal education and formal communication.

Theme 3: Communication and relational dimensions influence collaboration negatively.
This theme captured the communication/structural and relational dimensions of the relational coordination framework that informed the initial template in the template analysis. In line with the RC communications dimension, a lack of information sharing is reported between the different agencies working with offenders with substance abuse issues, with participants describing how there are few opportunities to meet frontline workers from other services. This they see as a major challenge to collaboration: Restrictions on information sharing is a big challenge in collaboration. We often think, OK, for how long are we supposed to walk around not knowing? (nurse, municipality) I miss gathering more. Physically. I think it is a good idea to train on collaboration. Simulate it. Often. (therapist, substance abuse treatment centre) However, the quantity and quality of collaboration was not consistent across services.
A probation officer said that some agencies are more available than others, and that certain agencies are more or less impossible to reach. He seemed to feel highly frustrated by this, as some of the unreachable agencies are key stakeholders in the transition phase. Related to this, a social worker reported severe challenges in communication with the correctional service: Communication is slow. Things take way too much time. (social worker, Red Cross) The frontline workers interviewed reported mixed experiences when it came to the relational quality in the collaboration processes. For example, they reported challenges concerning shared knowledge.

Discussion
The offender's dependence on interprofessional and interagency collaboration in reintegration after prison The participants described the multifaceted needs of offenders, the support needed from a variety of welfare professions and agencies, and the importance of different professions and agencies collaborating to provide the necessary quality of welfare service during this vulnerable transition phase (Theme 1). These views are supported by other Norwegian studies (Friestad & Kjeldberg, 2009;Hansen, 2015;Larsen et al., 2019). Larsen et al. (2019) found in their small-scale study that offenders' experiences of support from the welfare services during vulnerable periods, such as reintegration after prison, varied, both within and between services. Moreover, in the Larsen et al. study (2019), the offenders reported that they found themselves floundering in the gap between prison and community. Another Norwegian study, by Friestad and Kjeldberg (2009), found that health promotion among offenders in prison is clearly a multidisciplinary task which requires close collaboration between different welfare service providers and systems of care. Hansen (2015) also highlighted the importance of collaboration, but found that collaboration had proven difficult to establish. He suggested that an explanation is that both the correctional service and the municipalities appear to view problems concerning offenders with substance abuse issues as problems that can be solved separately and sequentially.

Frontline workers' personal attitudes and characteristics influencing collaboration
The participants in this study emphasized the influence of personal aspects, such as the engagement, values and mind-set of the frontline worker involved, as well as "personal chemistry", among frontline workers, as being important in collaboration (Theme 2). This rather disturbing, though maybe not surprising finding, naturally raises the question of what impact these factors have on the welfare services being offered to offenders and the offenders' right to equally receive good quality help. To put things bluntly: Does the quality of service offered depend on whether the frontline workers like each other or not? Or on whether they have the same set of personal values and mindsets? Findings further reveal that the lifestyles of this group of offenders affect the moral views and basic values of the frontline workers (Theme 2). This interaction may influence collaboration. Frontline workers, having shared personal values and mind-sets, may create closer relations among them, but, naturally, the opposite situation can challenge collaboration. Furthermore, organizational culture is perceived as important, because organizational culture can either be supportive of interprofessional collaboration or not (Theme 2).
Consequently, one might come to the conclusion that the quality of the welfare services provided to offenders is based on chance: For instance, if you are lucky you can end up dealing with a welfare agency with a collaborative culture.
Furthermore, the participants highlighted factors such as having shared information and shared knowledge as important in collaboration (Theme 3). To some extent, these statements are in accordance with the dimensions of relational coordination, but these statements may be viewed as normative and hard to achieve, considering the multifaceted needs in the reintegration of this group of offenders, as well as legally restricted information sharing and welfare services' lack of resources (Hean et al., 2018).

The limitations of relational coordination or limitation of time and resources?
When it came to articulating how collaboration takes place, participants seemed less able to provide in-depth descriptions of the process. Despite attempts to prompt them to explore relational (respect, trust and shared goals) issues and the communication (frequency, accuracy, problem solving and timeliness) dimensions of collaboration, they seemed only able to discuss the frequency of meeting up with other frontline workers and agencies. Why is this the case? It may be that the frequency of interaction is so limited that other aspects of collaboration cannot be discussed because contact is just not occurring in the first place. Participants suggested the training and simulation of interprofessional collaboration, with empirical studies supporting this suggestion Ødegård, 2006). This could be a way to raise the awareness of differences and develop a positive collaborative process, but it might also be a challenging solution for welfare services where resources are already cut to the bone. Within the field of social work, self-awareness and critical self-reflection among frontline workers has been central for decades (Taylor & White, 2000). What is needed is a critical self-reflection that goes deeper into-and asks critical questions about which underlying assumptions, attitudes and values, in addition to institutional and structural frameworks, affect practice and interprofessional collaboration (Taylor & White, 2000). Such a continuous selfreflection process might have a constructive effect on collaboration, and does not demand time or economic resources. Gittell et al. (2000 (Bond & Gittell, 2010) claim that the theory of relational coordination has been successfully tested in several contexts, including the criminal justice system. Gittell's theory of relational coordination (2011)  Nevertheless, Gittell (2011) proposes five potential directions for the further development of relational coordination. One of these five directions is to extend the theory from its focus on role and task relationships to include personal relationships, and to explore the interplay between them. This is a highly relevant development in light of the findings in this current study. Bolton, Logan and Gittell (2021) most recently suggested expanding the theory from a linear structure-process-outcomes model to a more dynamic model of change. This may also be considered an interesting development of the theory in this context, as the life situations and welfare needs of offenders change rapidly (Hansen, 2015;Larsen et al., 2019)

Methodological issues and limitations
Given the multifaceted problems involved in interprofessional collaboration, this study obviously does not provide a complete picture of the subjective views of collaboration in this context. First, this is a small-scale study presenting a snapshot in time of interprofessional collaboration based on a small number of participants. The ongoing Covid-19 pandemic may also have limited the recruitment process. Second, the first author has over 10 years of experience as a practitioner in the Norwegian Correctional Service. This experience creates a preunderstanding which could influence the researcher's perspective and, thus the interpretations and conclusions drawn in this study. However, the work experience may also have strengthened the study as the first author has knowledge and a deeper understanding about the information presented by the participants. Moreover, both the context in which the interviews took place, and the interaction between the researcher and participants, may have influenced the knowledge gleaned. Third, the complexity of interprofessional collaboration within a prison context may require the use of mixed methods that incorporate both quantitative and qualitative methodologies to capture the multifaceted nature of interprofessional collaboration (Mathieu et al., 2008;Ødegård & Bjørkly, 2012). Fourth, a theoretical framework may create a narrow approach, and factors outside of the framework may not be revealed despite an effort to have an open mind and conducting an additional inductive analysis.

Conclusion and implications
Based on frontline workers' subjective views, this qualitative study offers a range of tentative hypotheses about factors influencing collaboration in a prison context that now requires testing. Combined with other models and theories, the theory of relational coordination can be a contribution in further developing a theoretical framework for the field of interprofessional collaboration which has implications for welfare services offered to a stigmatized group of offenders in a vulnerable transition period.