Keys to unlocking service provider engagement in constrained co-production partnerships

Co-production partnerships

Co-production is defined as a process where service users and providers collaborate to improve, challenge, transform and innovate public service delivery (Osborne & Strokosch, 2013). It entails a different kind of relationship between service users and providers (Bovaird, 2007; Pestoff, 2012). A key challenge is to unlock the potential of genuine co-production partnerships (Osborne & Strokosch, 2013). In this article, we ¹ explore what may strengthen service providers’ engagement in co-production. We define the co-production partnership as a reciprocal ‘working alliance’ (Bodorkós & Pataki, 2009; Trevithick, 2003). This can be a dynamic partnership between service users and providers, where role migration (between ‘teachers’ and ‘learners’) and reciprocal empowerment are core qualities (Larsen & Sagvaag, 2018; Yorks et al., 2008).

Our study contributes to existing literature by presenting empirical records of how participative co-production (Osborne & Strokosch, 2013) was facilitated within a public specialised mental health and substance abuse service (SMHS). Participative co-production is the democratic involvement of service users and providers in service development on a strategic planning level. This article contributes to existing action research literature with a co-operative inquiry inspired approach that facilitated the participative co-production of a public SMHS. This framework (see Figure 1) may be regarded as a co-creative contribution to action research inquiry (Friedman et al., 2018).

OPEN IN VIEWER

Figure 1.

The four phases of inquiry.

Structural power imbalance

In health institutions (Smith et al., 2017) and total institutions (e.g. prisons) (Drake, 2014), an inherent structural power imbalance challenges the ideal of parity in action research. Neither staff nor their organisation operate in a vacuum, free to reframe all operating rules and principles: Elements like organisational policy and professional codes of conduct may constrain the relationship between service users and providers (Larsen & Sagvaag, 2018).

Although admission to treatment and participation in the inquiry was voluntary for patients, staff job descriptions support power imbalance: Patients cannot take possession of their own treatment completely, but they can participate in improving selected services. However, through reports and patient records, staff may contribute to the involuntary discharge of patients who do not accept or follow the rules of treatment. For staff, this would be equivalent to ‘being sacked on the spot’ – an exit that, for most, would be experienced as a degrading, unexpected shock. This dilemma arose in the inquiry, where the idea was for existing services – such as consultations with specialised professionals and certain therapeutic activities (occupational and milieu therapy, sports and more) – to be co-produced on an equal footing.

Context

A basic premise for the study was a consensus-based agreement to initiate an in-house project about increased patient involvement in an SMHS in Norway. A key objective was to improve treatment services through dialogue between stakeholders. The inquiry took place from 2010 to 2013, in a voluntary inpatient treatment unit.

At the outset of the study, the morning meeting was the only fixed shared meeting for patients, staff and leaders. Meetings were led by staff and were obligatory for patients; leaders were occasionally present. The most powerful professional decision-making body in this unit was the multidisciplinary treatment meeting. In these weekly meetings, final decisions about therapeutic approaches and patient discharge were made by leaders, based on dialogue with staff. Staff, leaders and patients participating in the inquiry appeared to have no prior knowledge of co-production conceptualisation or experience with involving patients systematically in service development.

Design

The cyclical principles of action research (Brydon-Miller et al., 2003; Reason & Bradbury, 2008) were applied as a single case study (Flyvbjerg, 2006; Mabry, 2008; Yin, 2009), starting with conceptualising and particularising the problem in collaboration with stakeholders, and moving through several interventions and evaluations (Heron, 1996; Heron & Reason, 2008). The inquiry was designed in line with co-operative inquiry principles – researching with rather than on people seemed an appropriate way to facilitate participative co-production (Heron & Reason, 2001; Osborne & Strokosch, 2013).

This design called for particular roles, i.e. patient co-researchers (PCs), staff co-researchers (SCs), leader co-researchers and the action researcher (i.e. first author). During the action research process, these individuals collaborated on interview guides, data collection, interpreting and disseminating findings, and proposing service changes. The researcher facilitated the full inquiry while conducting participatory observation (May, 2001; Savage, 2000); documented via minutes and reports (made accessible to contributors); attended all formal and most informal service-related meetings relevant to the study’s objectives; and provided training and supervision to qualify co-researchers to lead test interviews, 6 stages of multistage focus group interviews (Hummelvoll, 2008), 10 individual interviews (Silverman, 2006) and 4 dialogue seminars. Following the action research process, the action researcher withdrew from the action context to explore the data.

In addition to service meetings and the scheduled inquiry (Figure 1), there were ad hoc inquiry meetings with leaders, staff and/or patients to address issues raised in the inquiry: The dialogue meeting was facilitated when staff and PCs encountered communication difficulties, to resolve conflict between leaders, staff and PCs. In this meeting, a new fixed meeting ² was established for patients, staff and leaders to discuss treatment quality and co-design services.

Inquiry phases

The project used the four phases of knowledge development in co-operative inquiry as a framework ³ (Heron, 1996; Reason, 1994). In phase 1, there were two co-research workgroups (one with staff, the other with patients), which then merged into one joint work group. In phase 2, one patient co-research team led multistage focus group interviews (with staff, leaders and patients), semi-structured individual interviews (with patients) and a dialogue seminar (with staff, leaders and patients). In phase 3, one staff co-research team led multistage focus group interviews (with staff, leaders and patients), semi-structured individual interviews (with staff) and a dialogue seminar (with staff, leaders and patients). In phase 4, a co-research group of leaders, staff and patients were supported by facilitators (a staff member and a former patient) in one dialogue seminar, and then led the final dialogue seminar (with staff, leaders and patients) themselves ⁴ (see Figure 1).

Documentation and analysis process

Over the three years of inquiry, 109 persons (66 m, 43 f) participated. Data for this article were chosen to illustrate specific circumstances regarding communication. These data consist of preliminary ⁵ and cited documentation, including reports, minutes, field minutes and journal notes from phases 1 and 2. In addition, some descriptions in the ‘Results’ section are condensed interpretations from field notes and other data.

The cited documentation in the ‘Results’ section was member-checked by the respective participants (with the exception of the journal notes) (Lincoln & Guba, 2007). Member-checked field notes are regarded as ‘field minutes’. Participants were urged to look for missing elements or misinterpretations; the texts were amended according to their recommendations, then accepted by the respective participants for external dissemination. The reports, minutes and journal notes were written in Norwegian and the selected findings were professionally translated. To ensure consistency, familiarity and readability, all documentation was written in accordance with the SMHS’s documentation tradition. Minutes were thus predominantly condensed descriptions of conversations, not verbatim transcriptions (Hammersley, 2010; Poland, 1995). Some participants were quoted verbatim – these instances are underlined.

All data were subjected to qualitative conventional and directed content analysis using NVivo 9 (Graneheim & Lundman, 2004; Hsieh & Shannon, 2005). The investigation was abductive (Timmermans & Tavory, 2012). Abduction jump-starts inductive and deductive reasoning, without engaging in purist dichotomies: The researcher’s theoretical position is acknowledged and there is no intention to prove or falsify a theory. Rather, in abduction, induction and deduction may be repeated in an inferential creative process: as new anomalous and surprising findings emerge, new hypotheses and theories are produced (Timmermans & Tavory, 2012).

The data were revisited repeatedly through a process of de-familiarisation (theoretical cultivation and sharing ideas among communities), and alternative casing (creating new cases with the help of theory) (Timmermans & Tavory, 2012). Three themes ⁶ concerning patient–staff communication emerged from the data once the theoretical lens was applied. These themes were reconstructed into three categories ⁷ that were condensed, decontextualised and compared in a table. The final analysis involved marking each condensed category in the table with the following theoretical concepts: ‘move’, ‘follow’, ‘oppose’ and ‘bystand’ (Isaacs, 1999); ‘genuineness’/‘authenticity’ (Isaacs, 1999; Newton & Goodman, 2009); ‘empathy’ (Newton & Goodman, 2009); and ‘power’ (Emerson, 1962). This produced the three themes described in the discussion.

The inquiry was approved by the Norwegian Centre for Research Data.

Encounters preceding the meeting

Direct communication between patients and staff seemed challenging to both parties. Several encounters preceding direct communication were common. One illustrative example occurred when a PC suggested changing the protocol for routine drug and alcohol searches. A staff member ‘adopted a very humble attitude’ (Journal) and ‘wanted to invite the unit leader to the next morning meeting’ to discuss the matter in more depth (Field minutes). However, the staff member then told the researcher that she ‘thought such matters might be uncomfortable if there was insufficient time to prepare’ (Field minutes). The researcher informed the PC about this concern, and he ‘concluded that the morning meeting was not the appropriate arena to address such a substantial matter; he would prefer to address it in the new forum’ (Journal). The PC, however, did not propose this to the staff member directly.

I met the PC in the corridor and he seemed rather upset (his hands were shaking). He said that the staff member had informed him that the unit leader was coming to the morning meeting tomorrow and that it was wrong now. (Journal)

With support from the researcher, the PC met with the staff member and informed her about his decision. Several encounters were thus necessary to establish a suitable meeting space in which a specific topic could be discussed with relevant parties – one of many examples of the challenges around facilitating direct communication between stakeholders.

Communication ambience and power imbalance

Issues around communication and power were ongoing. Some patients explained how staff could make them feel subordinate:

The co-researchers from the patients report that the staff can overrule them and talk to them and set limits on them as if they were children. It is clear that the staff are above the patients and that they have the power. (Patients, joint report)

Although staff and patients were cautious about speaking openly in front of one another, patients could be explicit.

A co-researcher has observed that patients who have previously been critical in the unit have been discharged, saying that it is not necessarily advisable to shout too loud while on the ward. As a result, co-researchers also feel uncomfortable about participating in the research project, as there may be a threat that people who create conflict or ‘the one who shouts loudest’ may be discharged. (Patients, joint report)

This was arguably an invitation to explore power differences, trust and empathy between the stakeholders. The patients also urged staff not to take their feedback personally. However, in their response, the staff changed the topic, thus avoiding patients’ concerns about involuntary discharge:

SCs agree that the response should be perceived as constructive feedback and that focus should be placed on the services … Patient feedback is a good way for the staff to learn and improve. At the same time, constructive criticism from patients is a great way to ensure that we do not stagnate, but move on. (Staff, joint report)

As demonstrated above, challenges arose from the ambience surrounding communication between stakeholders. Staff were perceived as paternalistic, and it appeared that some issues risked being bypassed – however important they may have seemed.

Constraints and dialogue

Involuntarily discharge became a concrete risk in the subsequent phase. One day, the researcher was urged by staff to come to multidisciplinary treatment meetings, as they were concerned that PCs might be discharged due to their co-research activities. Staff and leaders both reported that it was difficult for them to reassure decision-makers who were not participating in the inquiry. A leader reasoned, ‘They don’t have same cool as we who know the project do’ (Leader, field minutes). He also confirmed that PCs could indeed be involuntary discharged. The researcher attended several meetings to observe developments and clear up misunderstandings about the nature of the co-research. A staff member also urged the researcher to intervene:

When [the PCs] ask questions about everything and demand change, the staff feels negatively towards them. She encouraged [the researcher] to guide the co-researchers regarding their role in the environment, given that they are often ‘going after’ the staff. … We agreed that I should talk with staff and leaders to facilitate a meeting where patients, staff and leaders can enter into dialogue about this challenge. (Staff and researcher, field minutes)

Staff and leaders were concerned with the communication ambience and staffs’ working environment, as PCs’ way of questioning existing practice was perceived as querulous and pushy. Though the PCs’ involvement was essential to resolving this conflict, the urgency of their situation was kept a secret – they were led to believe that communication difficulties were the main issue.

After several mediation meetings with the researcher, staff, PCs and leaders agreed to meet to resolve the conflict and engage in dialogue. Here, a staff member appealed to the PCs’ empathy:

It makes one feel inadequate as a staff member, and perhaps the patients get the feeling that they are not respected when the staff member cannot answer all their questions. He says that when the pressure becomes too much, patients might feel as though the staff is useless when unable to provide a proper response. (Staff, dialogue meeting)

In the meeting, a leader advised the PCs to handle their new role wisely:

The co-research also involves some trust on the part of the staff towards the patients; it would be uncomfortable if the research were used as a trump card to push things through. For example, patients threaten to use the research to get the final word in a discussion. … The staff must be included in the dialogue … Discussing changes is good milieu therapy, but you need to have good communication and not hide in the trenches. (Leader, dialogue meeting)

The PCs understood that their communication ‘can be perceived as querulous and that it may be excessive’, and expressed empathy for the staff by saying ‘that they will calm down a little’ (Patients, dialogue meeting). They suggested establishing a fixed meeting between patients, staff and leaders. The stakeholders agreed that this would be a meeting where they could resolve conflict, practise dialogue and develop the services.

After facilitating this dialogue and attending three multidisciplinary treatment meetings, the researcher was confident that the discharge risk had been mitigated. As the above shows, the ‘backstage’ activity preceding direct communication was evident. It is unclear whether the researcher’s guidance and several encounters preceding meetings mitigated the discharge risk; however, in such an imbalanced power hierarchy, mediation and facilitation of dialogue seemed imperative.

Dialogic leadership and power

Although we identified difficulties with direct communication, complementary roles in dialogic leadership also emerged. As co-researchers, patients often proposed inquiry pathways in a ‘mover’ role (addressing the staff’s paternalism, requiring change of search procedures, questioning existing practice, demanding change). However, the staff, when in the presence of patients, appeared hesitant to provide perspective (bystander), challenge validity (opposer) or offer new directions (mover) (Isaacs, 1999). Some staff instead took on a ‘follower’ role: For example, when PCs feared involuntary discharge, staff supported their initiative, taking patients’ contributions as constructive feedback but avoiding addressing the patients’ fears. Staff also avoided engaging in dialogue about search procedures without leaders present.

Another avoidance strategy for staff involved taking on an opposing role in ‘backstage’ discussions without patients present. Here, the co-researcher approach was questioned, and staff and leaders advocated a different path for patients’ behaviour. This backstage opposition was one reason the dialogue meeting was facilitated, so stakeholders could cultivate a practice of direct communication. Even so, serious issues still remained hidden from the patients.

The power relationships in this setting benefit from a simplified analytical dichotomy regarding movers and followers or opposers. On the one hand, staff and leaders may have struggled with their power disadvantage. As the staff felt unable to respond constructively to patient initiatives, PCs seemed to have a power advantage. Further, co-research activity was considered a ‘trump card’ that forced compliance with patients’ demands for change. With no previous experience with participative co-production, engaging in co-production on an equal footing with patients was a new experience for staff and leaders. On the other hand, also patients were dependent on the staff’s ability to respond sufficiently. After all, in this relationship, staff and leaders held the power to ‘grant or deny, facilitate or hinder’ fulfilment of the inquiry’s aim (Emerson, 1962, p. 32). For one, the PCs’ fear of involuntary discharge may have been mitigated at such an early stage because it was a difficult topic for staff to explore with patients. Confidentiality requirements may have made dialogue challenging, as staff could not discuss the involuntary discharge of former patients.

However, staff avoided exploring the issue even on a more general level (as bystanders). Here, exclusion was at the core of this power relationship, making the stakes in this collaboration uneven: The staff’s professional status protected them against exclusion, while patients were obliged to follow established rules to remain in treatment. Such a power-laden topic may have been difficult for staff to address in the presence of patients without preparation and support from leaders and colleagues. Moreover, later in the inquiry it seemed that avoiding confrontation by excluding PCs from the conversation/information loop was crucial to balancing their power advantage. Under these constraining circumstances, patients were dependent on staff, leaders and the researcher to direct patients’ behaviour (and thus have the power advantage), without revealing to them what was at stake.

Dialogue facilitation and co-production

Constructive dialogical roles seemed to be difficult for staff to enact, even if they complemented the patients’ role in the co-research process. Even so, one leader demonstrated several roles in the dialogue meeting. Having observed the process, he inquired into the co-researcher role and provided perspective (bystander) when confronting the PCs about how they managed their power (opposer). Also, while completing the initiated topic of ‘communication difficulty’ by supporting it (follower), he advocated a different path out from the trenches (opposer) and urged patients to include staff in dialogue. This suggests that, although he did not disclose the discharge risk, the leader role modelled authentic communication.

It is possible to balance and enhance the quality of dialogue with awareness, reflection and cultivation of communicative practices and, as the power balance tilts in favour of professionals who are more securely positioned than service users, they have greater responsibility to promote inclusion through equal and authentic communication. However, the staff’s avoidance and backstage opposition suggests that they were in need of training and guidance concerning inclusion and balancing dialogue and power in patients’ presence. We now turn to some related opportunities for future learning-in-experience among staff.

Learning in the presence of each other

Patients appeared fairly direct about topics that seemed uncomfortable for staff; they also appeared courageous and genuine when revealing their fear of discharge. However, it seemed that patients’ voices were not truly heard, as the staff appeared to engage in defensive organisational routines to avoid empathetic inquiry: They changed the topic, and, together with the leaders and researcher, pretended that communication difficulties were the reason for the dialogue meeting. The staff’s unwillingness to discuss certain topics with patients without a leader present may represent another defensive routine.

Staff members may thus have protected themselves against self-inquiry about emotions triggered in meetings with patients. For instance, staff members were concerned about patients’ reactions when they were unable to respond adequately to patient questions or demands. Because of this, staff may have felt the need to shield themselves, through avoidance tactics and only discussing issues in patients’ absence. Staff may also have empathised with the patients’ expressed needs but were unable to accept the affectivity that resonated within themselves. Either way, genuine listening would have been difficult in this context.

It may also have been challenging for staff to practise authentic communication in patients’ presence, without engaging in the emotional labour necessary to understand their own feelings and needs. If so, meeting without patients could have been a ‘cooler’ path that enabled staff to meet with patients in the dialogue meeting. When this meeting finally occurred, it appeared to be a space where stakeholders could think, feel and develop a joint understanding. Here, staff and leaders could air their fears and PCs could empathise and agree to adjust their communication. However, this communicative space was not equal for all stakeholders. By avoiding certain issues, staff may have indirectly limited patients’ contributions in the dialogue, as they lacked knowledge about the whole situation.

Secrecy may thus have prevented the direct affective exchanges that can enable deeper interpersonal relationships, communication and learning, and genuine meeting and learning on an equal footing. Further, co-production may have been delayed by organisational defence routines that allowed ‘beating around the bush’. Profound directness is often listened to, as it encourages honesty and helps put things in perspective, which motivates participants to search for a solution (Isaacs, 1999). However, speaking with a genuine voice is a main challenge of dialogic leadership, and it did seem that staff needed some time before they could reveal their subjective truth in front of patients.

Further, though discussions occurring in patients’ absence may have reduced conflict, they may also have postponed staffs’ professional development. The patients appeared to pressure staff to inquire into their own ability to respond and change, but staff seemed to focus on efforts to change patient behaviour when experiencing uncomfortable emotions. Little was therefore revealed about staff engagement in their own ‘coming to know’ journey. Bypassing learning-in-experience opportunities may have hindered or postponed their emotional and cognitive transition and learning. This may in turn have kept them in a single-loop learning process, i.e. repeating attempts to solve the same problem (uncomfortable confrontations with patients), with no variation of method (avoidance) and without questioning the end goal (service quality) (Argyris, 1995).

With sufficient skills, however, the parties may have been able to support each other’s double-loop learning processes and mutual growth. Informed by learning from being emotionally present in dialogue, the parties could have redesigned their actions and made more persistent changes within themselves and the services.

Methodological considerations and limitations

A participative perspective (Gayá Wicks & Reason, 2009) guided the relationship between the researcher and co-researchers. Although PCs were vulnerable due to the structural imbalance and delicacy of the themes raised in the inquiry, the project’s resilience was rooted in reciprocal trust and voluntary equal participation. The project respected ethical principles of participatory action research (Winter, 1996) with the SMHS granting authorisation for its participation. These principles required extensive consultations between participants, including transparency and access. In addition to full confidentiality, anonymity and access in the information exchange were guaranteed. Also, decisions regarding the direction and expected outcomes of the research were collective.

An action researcher is in effect a practitioner – an interventionist seeking to help client systems by making theory relevant to action (Argyris, Putnam, et al., 1985). The value of applying this perspective in co-production studies lies in the idea that knowledge should not be limited to its theoretical implications but to the practical value of the theory in use (Argyris & Schön, 1974). As an action research project, our study inherits the strengths of this kind of approach: it addressed a practical challenge (to increase patient involvement), generates new knowledge (co-production), enacted change (a fixed co-production meeting), was participatory (power sharing) and relied on a cyclical process (four phases of inquiry).

However, some have argued that a single case research design can contain pitfalls concerning methodological rigour, researcher subjectivity and external validity (Yin, 2009). Others have argued that by clarifying and developing the methodological techniques and epistemological grounding of single case studies, these issues are of little concern (Bennett & Elman, 2010; Flyvbjerg, 2006). Regarding the issue of generalisability, this is of limited relevance when the aim is one of particularisation, as in our action research project. As such, the trustworthiness of the study is not compromised by combining action research with a single case design (Lincoln & Guba, 2007). The use of an action research framework further ensured the rigour of our study. Attention to researcher subjectivity was also part of this framework and was promoted through continuous self-reflection around the ways in which emotions, reactions and preconceived notions may influence how participants and situations are perceived and recorded. Here, the steps in the abductive research cycle proved imperative across the action research process (Timmermans & Tavory, 2012).

Possible keys to unlocking balanced co-production

Co-production entails a relational shift between the traditionally power advantaged and power disadvantaged. To establish genuine co-production partnerships and facilitate organisational learning, service providers need tools to unseat unproductive defensive patterns. To ensure reciprocal working alliances between the parties, the norm must be that mutual learning happens in the presence of the other.

We urge the development of a mutual agreement. This should be co-developed and signed by service users and providers (including decision-makers), and should reveal agreed-upon principles for how to approach issues of power, exclusion/inclusion, confidentiality/transparency; and expectations around trust/openness, communication, collaboration, decision-making and the scope of co-production. Such formalisation may build trust and equalise the power between the participants.

One avenue for further research is to explore what service users and providers regard as necessary supervision, support and/or training to optimise the co-production relationship and the potential for co-innovation. Comparative cross-country action research may be needed to assess different conditions for organising co-production processes.