Evaluating co-production in an interactive webinar series concerning borderline personality disorder: a discourse and content analysis

Abstract

Purpose

This paper aims to evaluate the quality of co-production between lived experience practitioners (LXPs) and professionals in an interactive National Health Service webinar series aimed at supporting people who were diagnosed or identified with borderline personality disorder.

Design/methodology/approach

Transcripts from the webinars were subjected to mixed-method examination combining Foucauldian discourse analysis (FDA) and content analysis (CA).

Findings

FDA identified nine discursive objects: diagnosis beyond its medical context, diagnosis as a total explanation, being the other, universality, compassion, hope, faking it, mentalisation and co-production. CA demonstrated those nine discursive objects each corresponded with equalised airtime appropriated by professionals and lived experience practitioners.

Research limitations/implications

The sample was limited and if applied to other mental health settings might reveal different findings. More needs to be understood about the attitudes of professionals and LXPs that support discourse sharing. Although this study has offered evidence of the quality of co-production, it can say very little about whether the co-productive approach offers superior outcomes to other forms of treatment.

Practical implications

Further research could employ FDA and CA to further explore how co-production is being enacted in other situations, with different models, where comparable interventions are delivered. Future research could compare outcomes between co-productive and professional-only interventions.

Originality/value

This study examined naturalistic practice to build new theory in an under-researched area for a substantial mental health population.

Citation

Xenophontes, I. and Springham, N. (2024), "Evaluating co-production in an interactive webinar series concerning borderline personality disorder: a discourse and content analysis", Mental Health and Social Inclusion, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/MHSI-01-2024-0009

Introduction

Co-production methodology actively equalises participation between involved parties while valuing both consensus and differences between their perspectives. The UK National Health Service (NHS) long-term plan (NHS, 2019) enshrines an expectation for co-production in healthcare. Although that aspiration is laudable, the reality of equalising participation between professionals and non-professionals is not simple. Co-production methodology exists in recognition that where stakeholder groups have become distanced, or even antagonistic, deliberate mechanisms are needed to reconcile their perspectives. This demands traditional constructions of power and assumptions about what constitutes valid evidence all become subject to revision. After decades of provider-led service design and delivery, the application of such an approach at this point within the NHS inevitably positions co-production as disruptive. Unsurprisingly, such disruption is welcomed by some and less so by others, resulting in variable levels of involvement of different groups in projects. Therefore, there is a need to examine the sharing processes between participants within any claim of co-production.

As co-production becomes applied to the field of NHS mental health services, taken-for-granted assumptions are perhaps particularly challenged. Mental health service users’ perspectives have historically been undervalued as evidence on a presumption they lacked insight (Rose et al., 2006). Distinct barriers to inclusion also operate within the various mental health service specialisms in relation to the condition being treated. For example, the experience of those seeking help from NHS services designed for people who have been diagnosed, or have difficulties in association with, borderline personality disorder or emotionally unstable personality disorder in the UK (hereafter BPD), have historically been far from inclusive. The National Institute for Mental Health in England’s report entitled Personality Disorder, No longer a Diagnosis of Exclusion (NHME, 2003) detailed how those diagnosed were not routinely met with understanding or even with any service at all. The BPD diagnosis itself is controversial and can be stigmatising for some (Klein et al., 2022). Too many service providers still openly express contempt towards those with this diagnosis (McKenzie et al., 2022). Repeated experiences of being met without understanding can develop into a lifetime of disempowerment (Klein et al., 2021; Stalker et al., 2005). Given it is estimated that 50% of outpatients and 20% of inpatients in mental health services received a diagnosis of BPD (Beckwith et al., 2014; Chapman et al., 2022), such discrimination represents a significant challenge to the co-production of healthcare.

As practitioners attempting to improve NHS provision for BPD from within services through co-production from lived experience and professional backgrounds, respectively, we (the authors) routinely experience the impact of the troubled history of the field. All projects still require a search for narratives about BPD which can be trusted by both service users and providers. An ongoing lack of literature base exploring which discourses better equalise participation in co-production must ultimately limit progress in the field. The present study is aimed at addressing that knowledge gap. Given the question originates from practice, we reason that a practice-based research methodology is most likely to yield realistic answers. On that premise, it becomes important to detail the specifics of that practice setting.

The Common Foundations Online intervention

The practice investigated in the present study occurred within Oxleas NHS Foundation Trust (hereafter Oxleas) in London, UK. The intervention was an interactive online support webinar entitled Common Foundations Online (CFOL) which was co-produced during the Covid-19 pandemic lockdown. The CFOL team comprised the following:

• Four lived experience practitioners (LXP) who had undergone an 18-month mentalisation-based treatment (MBT), following a diagnosis of BPD as per Bateman and Fonagy (2016). They also had developed skills through membership of Oxleas co-production network and/or the lived experience practitioner program.

• Six professional team members from art therapy, clinical psychology and medical psychotherapy professions, all of whom worked within the MBT service.

The CFOL intervention was co-developed through three distinct iterations to become the specific format delivered in the present study.

The first iteration was a co-delivered mentalisation-based treatment-introductory (MBT-I) psychoeducational group which aimed to prepare people for the 18-month MBT as described by Bateman and Fonagy (2016). Although successful in that aim, this model served a small number of the people who might benefit. Therefore, the second iteration restructured the intervention to be less exclusively MBT focused through a co-designed process that included professional and user-valued subject areas. By agreement between LXPs and professionals, it retained mentalisation at its core as a common mechanism of change (Bateman and Fonagy, 2016), as this added compatibility with other interventions offered within the service. This format was offered more broadly as preparatory support for other types of interventions or self-help; hence, the name “common foundations” was chosen.

The third iteration was a response to the need for mental health services to rapidly develop remote delivery formats in the Covid-19 lockdown (Patel et al., 2021). At Oxleas, this not only affected those within community mental health teams but also 600 Service User Network (Miller and Crawford, 2010) members. The situation was unprecedented, with limited knowledge available to develop the safe and effective online interventions needed at scale. However, it was known that people with an experience of, or self-identifying with, BPD frequently used social media platforms to increase social connectivity and build community (Woloshyn and Savage, 2020). Studies indicated that first-person vlogging from the same demographic promoted awareness of self, others and the wider world (Sangeorzan et al., 2019). CFOL drew on these insights to offer a live webinar format which involved a co-production team of LXP and professional facilitators in equal balance responding to questions emailed by viewers. The CFOL was launched to all its relevant service users and interested members of the public via social media within four weeks of lockdown.

The initial CFOL prototype used the same psychoeducational content as common foundations within a 90-min webinar format, delivered weekly through the Oxleas website. Based on viewer feedback, the preset psychoeducational curriculum was quickly replaced by topics suggested by viewers and responded to in the form of a live discussion between the CFOL hosts. Viewers contributed to that discussion via emails which were read out live (having been anonymised). Sessions were also recorded, archived and made available to view via several social media platforms. CFOL achieved between 200–300 views per week in this format. It ceased in July 2021 when face-to-face services resumed.

The aim of CFOL was to co-produce an online form of support for the BPD space. The research methodology we selected was an attempt to learn if we had done so and is now described.

Research methodology

The first author conducted the study, and two research questions were formulated:

RQ1.

What discourses are drawn on in response to viewer questions?

RQ2.

Do those discourses facilitate co-production between clinical professionals and LXPs in response?

The research used a mixed-methods design comprising Foucauldian discourse analysis (FDA) enhanced by content analysis (CA). FDA examined the discrete discourses and identified subject positionings operating within the CFOL webinar. FDA analysis defines discourse as “sets of statements that construct objects and an array of subject positionings” (Parker, 1994, p. 245). The study used the six stages of FDA as proposed by Willig (2013) to identify the following: discursive constructions, discourses, action orientation, positionings, practice and subjectivity within the material examined. Discourse analysis has been used as an enhanced form of reflective practice by practitioners wishing to understand the social constructs they draw on (Rossiter, 2005). Given the history of exclusion and stigma for BPD, this method was chosen to have the necessary sensitivity to detect and trace where power sharing occurred. CA then provided a nomothetic summary of the qualitative results from FDA. Typically, CA is used to complement analysis of naturalistic conversations, spoken words and written text as a psychometric measure (Feltham-King and Macleod, 2016). Within this study, CA was used to compare airtime (in seconds) appropriated by either LXPs or professions during each webinar. This provided a quantitative account of how webinar input between LXPs and professions was shared.

Sampling strategy

The CFOL archive provided an opportunistic data source representing naturalistic co-production practice. In their recorded format, the interactions were amenable to the mixed method micro-analysis proposed. Importantly, the content was already within the public domain and sensitive information was pre-anonymised. On this basis ethical approval was granted from Goldsmiths University of London Ethics Committee.

Eight archived webinars livestreamed between June to December 2020 were selected, as this period is after the final CFOL format was established. Webinars were selected to include different combinations of facilitators and excluded when there were not equal numbers of LXPs and professionals involved. Webinar conversation was transcribed verbatim as suggested by Gardner (2001). Speakers were identified and classed as either LXP or professional, respectively.

Analytic strategy

To facilitate FDA, Byrne et al.’s (2021) model of semantic coding was used to group segments of text through explicit or surfaced meanings in data as communicated by the respondent. Semantically coherent segments were identified from the text while respecting the need for those to represent a realistic account of facilitators as responses to viewers’ emailed input. Once completed, the researcher conducted several rounds of secondary coding to identify, categorise and condense meaning units which would serve Willig’s (2013) requirements for FDA. Experienced research supervisors were consulted to position the researcher’s own subjectivity and identify likely biases for the purpose of bracketing as described by Tufford and Newman (2010). NVIVO 12 Pro software (Qsrinternational.com, 2022) was used to support coding. CA used an independent t-test to calculate the time (in seconds) appropriated by professionals and LXPs, per episode and total airtime covering the eight webinars examined. Prior to running the primary analysis, the data distributions were compared across both levels of independent variables in relation to the dependent variables.

Results

FDA resulted in nine discursive objects: diagnosis beyond its medical context, diagnosis as a total explanation, being the other, universality, compassion, hope, faking it, mentalisation and co-production. A comprehensive exposition of the full six-staged justification for these discursive objects can be found in Table 1.

Table 1 describes nine discursive objects in relation to discursive constructions, discourses, action orientation, positionings, practice and subjectivity.

CA revealed the clinical professionals and LXPs distributions were sufficiently normal for the purposes of conducting a t-test (Figure 1). After testing the assumption of homogeneity of variances via Levene’s F(14) = 0.396, p = 0.539, the sample t-test showed there was no significant difference in airtime appropriated between clinical professionals (M = 1,234.38, SD = 480.69) compared to LXPs (M = 1,249.50, SD = 345.30); t(14) = −0.0723, p = 0.943 (Figure 2).

When FDA and CA results are compared the nine discursive objects each corresponded with equalised airtime appropriated by professionals and LXPs. An example drawn from the “universality” discourse shows that both LXP and professional are given rights to comment and contribute equally:

It makes a difference to feel understood, not feel like a flake or that your problems aren’t real or that you are, I used to refer to myself as a nutter. This is just me that does this. And it turns out it wasn’t just me that did some of these things. And that was really important for me to stop feeling so much like an outsider. (LXP)

I think one of the things we’ve learnt over the month or two that we’ve been doing this is that people continuously feel that it’s only them who are having this problem and nothing could be further from the truth. It’s really important to try and combat this sense of isolation that people are feeling. (professional)

Another example can be drawn from the discourse of compassion:

I think in terms of the, the, idea of the wider things going on in the world, I think, once you learn that it’s OK to feel how you feel and you are just as deserving as, say, your best friend or your family or the people closest to you to express yourself and get the help that you need, you are just as valid. It doesn’t matter what’s going on in the outside world. (LXP)

I think there was also something about the mask that people often find themselves having to put up. And I think that for me, makes me think about something that discussed last week around compassion and how much, can be, erm, how easy it is to provide to others, but difficult to turn inwards, erm, and, there’s something about then not, not allowing your needs to be seen because you’re not viewed as important. (Professional)

Discussion

The present study sought to discover if we found enough common ground between LXPs and professionals to co-produce the webinar on the difficult subject of BPD. Considering the correspondence between discursive objects identified by FDA and shared airtime demonstrated through CA, it is reasonable to assume there was. FDA methodology was particularly valuable in illuminating the discourses that could be shared. Given the subjects discussed by hosts were initiated by viewers, and involved difficult and potentially divisive issues, FDA offers some insight that co-produced responding was not limited to superficial or anodyne issues.

FDA also added value in illuminating the process of co-production. Some discourses facilitated equal access to both LXPs and professionals, whereas others offered more barriers to one or other host. For example, “compassion”, “hope”, “universality” and “co-production” discourses tended to be shared equally by both LXPs and professionals. One observation about this set of discourses might be that they might be thematically grouped as deriving from folk psychology. Folk psychology refers to the common ways people understand and help each other in everyday life as distinct from specialist, distinctly clinical or research approaches (Sellars, 1956). Folk psychology discourses were highly relatable and could be adequately discussed in lay, non-professional language. The discourse of “mentalisation” operated similarly and has been cited as drawing on folk psychology by its originators (Bateman and Fonagy, 2016). It is possible though that “mentalisation” only operated as a shared discourse because all CFOL team members were steeped in MBT either as providers or users of those services. The shared length of experience of MBT may have accumulated enough shared understanding between hosts to provide the means for talking about the concept equally together. Yet, Ditlefsen et al. (2021) similarly found mentalisation-based psychoeducation groups facilitated shared discussion with professionals and peers in a very short time frame. This suggests mentalisation, despite its technical name, is a readily graspable concept which people can apply to their lives.

Other discourses tended to oblige either LXP or professional webinar hosts to respond first, depending on how the discourse initially positioned them. The discursive objects “diagnosis beyond its medical context” and “diagnosis as a total explanation” tended to be addressed to professionals who outlined the frame of reference used for diagnosis in treatment and theoretical aims for approaching people’s distress this way. This was accompanied by open critiques about historical and current problems with the term BPD and the limits to the explanatory power of diagnosis. This stance gave space for those concepts to be translated into the experience of being diagnosed and life outside the diagnosis to be coherently contributed by LXPs. Conversely, discourses of “faking it” and “being the other” resulted in LXPs-led responses, who could offer credible examples closely matching those described by viewers. Having expanded on how such processes made them feel and how they coped, LXPs could offer a link for professionals to discuss the same from their clinical perspective. Importantly, even though professionals did occasionally give examples of their own emotional responses at other points in the webinar, they made no attempt to draw any equivalence between their own experiences and those of LXP and participants in these discursive objects. LXPs and professionals shared the space to counter fears that such negative lived experiences were necessarily indicators of people being broken, bad or unable to recover.

Although the present study identified nine co-produced discursive objects, the results cannot suggest that those subjects are somehow inherently amenable to co-production. If other actors engaged in the same topics, it is possible each might be discussed in an exclusive, non–co-productive way. Indeed, the troubled history of NHS BPD services offers ample evidence to show that is precisely that. So, although the combination of FDA and CA methodologies can identify discourses where sharing did occur, they have some limitations in explaining how it occurred. It is possible the level of equalisation in co-production may well be determined by attitudes held by the participants engaging the discourse. On reflection, the fact that the CFOL team in the present study was so deeply rooted in MBT services is highly likely to be significant in this regard. In addition to identifying “mentalisation” as a shared discourse, perhaps the mentalising stance itself, epitomised as “I am taking a genuine stance of not knowing and attempting to find out” (Bateman and Fonagy, 2004, p. 315), may represent an attitude that invites equal participation from all. The whole team spoke explicitly about the intention to mentalise throughout the webinars. Therapeutic modalities other than mentalisation may also achieve this effect, but it is beyond the scope of the present study to comment meaningfully.

Limitations

The sample was limited and if applied to other mental health settings might reveal different findings. More needs to be understood about the attitudes of professionals and LXPs that support discourse sharing. Although this study has offered evidence of the quality of co-production, it can say very little about whether the co-productive approach offers superior outcomes to other forms of treatment.

Implications for future research

Further research could use FDA and CA to further explore how co-production is being enacted in other situations, with different models, where comparable interventions are delivered. Future research could compare outcomes between co-productive and professional-only interventions.

Conclusion

The troubled legacy of borderline personality disorder in NHS mental health services can be divisive and may reduce the potential common ground needed for co-production. In this study, nine discourses allowed input from webinar audience, LXPs and professionals to share participation collaboratively. This suggests distinct discourses can be identified where measurably equalised participation can be achieved to support co-production.

Figures

Figure 1

Distributions of appropriated airtime

Figure 2

Homogeneity of variance descriptive plots

Table 1

Exposition for the full six-staged justification for discursive objects

Stage 1: Discursive constructions	Stage 2: discourses	Stage 3: action orientation	Stage 4: positionings	Stage 5: practice	Stage 6: subjectivity
Diagnosis beyond its medical context	Identifies diagnosis as a function of services and disputes its utility beyond that context. Also highlights how diagnosis traditionally underestimated individual history. Also acknowledges how powerful the term diagnosis is in our culture. Applies a psychosocial paradigm	Locate diagnosis as a function of mental health system, shows how that can be useful up to a point, but there is reduced utility outside of that system. Identifies self-stigma when the paradigm has been internalised	Professionals devolves authority outside of the scope of treatment, back to the service user to define identity. The service user has choice to differentiate a medical opinion from their own sense of self and understanding of emotional life	Show scepticism towards the diagnosis and have a more equal alliance with professional which balances both views	Validation with personal responsibility. Hope and freedom to apply non-medical interventions
Diagnosis as a total explanation	Professional tells service user about their experiences as symptoms. Symptoms ignoring trauma and other context. Although diagnosis is meant to define treatment needed in a service context, it extends to becoming a description of the person’s whole identity. No other explanation becomes pertinent, locating the problem in the individual, resulting in high blame	Putting a maker down about the historical power of the diagnostic process in relation to how an individual describes themselves	The professional imposes assumptions and the service user passively receives expert wisdom. In exchange of that, the service user is granted access to treatment intervention	The service user becomes dependent on services and professionals because they are the only ones who can understand emotional life	Subjective experience can either be stigma, but also validation. Although validation comes with helplessness, there is a hope that services will provide solutions and when services don’t meet expectations, we can become hopeless
Being the other	Recognises the experience of “othering” as dehumanising way of relating. Explicitly focuses on identifying with the subjective experience of having been “othered”	Left questioning the validity of their point of view. A default assumption that emotional reactions are always a mental illness issue	The individual can never claim legitimacy of their experience in certain areas. Those around the individual need never question their own perceptions	The individual must quarantine aspects of emotional life and engage in pretence. Those around the person doing the othering, never need to look at their part in things	The individual feels alone, cut off from social support. It’s only them that feels this way. People identify themselves acting differently to everyone else. Failure located in the self
Compassion	The compassion discourse has increasingly come into the professional realm to modify the emotional tone of the way people relate to themselves	Compassion can be applied as a deliberate reframing action	To right to show kindness and to mitigate actions and reactions	Gain some agency and unfreeze people from paralysing levels of self-blame	A sense of loss that compassion was not automatic, but also hope that it can be applied and bring relief from guilt and shame
Universality	The opposite of othering; this reframes all emotional experience as being part of a shared humanity. Previously defined symptoms are reframed as ordinary everyday reactions but with varying magnitudes	Solidarity between professional and service user	Normalises reactions otherwise pathologised giving service users a right to talk about their experiences outside of services	Seek support outside of mental health services	Individuals no longer feel like a “service user”; they feel connected and seen
Faking it	Individuals feel the need to selectively pretend and constantly perform to fit in	Imitating feelings or appearance of competence as defined by other people	Unable to communicate certain thoughts and feelings. Unable to challenge familial culture and social norms	Suffer in silence and isolate themselves	Isolated and exhausted
Hope	Belief based on experience that situations can improve. That there is good in the world	Deliberately talking about hope increases hope itself as an agent of change	Gives individuals the right to imagine beyond the current feeling or circumstance	Consider positive change through comparative experience	Individuals won’t always feel this way. Thoughts, feelings and reactions can change. A sense of autonomy
Co-production	A contemporary model for developing and delivering mental health services based on equal but differentiated contributors from users and providers	Triangulating “lived” and “learned” perspectives	Everyone has the right to comment on the difficulty described	Move beyond roles ascribed to individuals by a medical model	Experiences of decentring from your own perspectives, professionals and services users alike. A feeling of collegiality
Mentalisation	A therapeutic approach based on common mechanisms of change and triangulates attachment theory, neuroscience and psychotherapy practice. It requires service user and therapist to act as partners in thought	Don’t expect people to read our mind, use communication instead. Responses are not trying to “fix” people but encourage curiosity about assumptions and mental states	No one is an expert. Assumption that everyone has the capacity to engage in mentalising	Individuals can question everything without pathologising	Freedom from critical self and gives choice in perception and therefore action