Introduction
The need for adherent CBT for psychosis
Approximately one in ten of us in the general population report psychotic experiences such as hearing voices (Johns et al., Reference Johns, Kompus, Connell, Humpston, Lincoln and Longden2014), and one in a hundred experience clinical levels of psychosis (British Psychological Society/Cooke, 2017; Schizophrenia Commission, 2012). Psychosis can cause considerable distress to the person and their family, and have a devastating impact on relationships and work prospects (British Psychological Society/Cooke, 2017; NICE, 2014). In addition to the human impact, the economic costs of psychosis are estimated at £11.8 billion annually in the UK (Schizophrenia Commission, 2012). Ensuring the availability of effective treatments, including cognitive behavioural therapy (CBT), is a priority for people with psychosis, their families, and the healthcare economy (British Psychological Society/Cooke, 2017; NICE, 2014).
In the UK, everyone with psychosis should now be offered CBT (NICE, 2014). Whilst not unequivocal (cf. Jones et al., Reference Jones, Hacker, Meaden, Cormac, Irving, Xia, Zhao, Shi and Chen2018; Laws et al., Reference Laws, Darlington, Kondel, McKenna and Jauhar2018), the evidence indicates that formulation-based CBT is effective for people with psychosis (Turner et al., Reference Turner, van der Gaag, Karyotaki and Cuijpers2014; Turner et al., Reference Turner, Burger, Smit, Valmaggia and van der Gaag2020), and may attenuate symptoms or delay transition in vulnerable individuals (Hutton and Taylor, Reference Hutton and Taylor2014; Stafford et al., Reference Stafford, Jackson, Mayo-Wilson, Morrison and Kendall2013). Over the last five years, we have seen an increase in demand for CBT for psychosis (CBTp) in clinical services, and a corresponding increase in demand for nationally accredited training. Although we remain far from meeting best practice guidelines, routine clinical provision has started to improve in the UK, largely due to national auditing of services, most notably the Access and Waiting Times standards for people with first episode psychosis (Royal College of Psychiatrists, 2022). Training clinicians in high-quality CBTp is therefore essential to ongoing healthcare provision.
Formulating psychosis
Individualised CBTp involves working with the person to make sense of the psychological processes likely to be contributing to the maintenance of distress and disability associated with their psychosis. This case conceptualisation or formulation forms the basis for treatment planning, drawing on cognitive, behavioural and affective interventions likely to cultivate less-distressing appraisals, and behaviours aligned with the person’s values and goals, for example learning that the voices do not always tell the truth, or that the person is safe enough to visit friends and attend college even when feeling anxious or suspicious of others.
Making sense of unusual experiences affects both immediate and longer-term consequences. The very process of formulating psychosis can ease distress (Brabban et al., Reference Brabban, Byrne, Longden and Morrison2016), and Stainsby et al. (Reference Stainsby, Sapochnik, Bledin and Mason2010) found that people’s ability to make sense of their psychotic experiences predicted quality of life two years later. Formulation then forms the basis for individualised therapy (cf. Turner et al., Reference Turner, Burger, Smit, Valmaggia and van der Gaag2020).
Current theoretical models (on which formulations are based) include broad frameworks of psychosis, e.g. the model of positive symptoms (Garety et al., Reference Garety, Kuipers, Fowler, Freeman and Bebbington2001); psychosis as culturally unacceptable intrusions (Morrison, Reference Morrison2001); and the person-based or self-schema model of psychosis (Chadwick, Reference Chadwick2006). Others focus on specific symptoms, e.g. persecutory delusions (Freeman, Reference Freeman2007); paranoia as akin to social phobia (based on Clark and Wells, Reference Clark and Wells1995, Newman-Taylor and Stopa, Reference Newman-Taylor and Stopa2013); voices (Berry et al., Reference Berry, Varese and Bucci2017); command hallucinations (Byrne et al., Reference Byrne, Birchwood, Trower and Meaden2006); grandiose delusions (Isham et al., Reference Isham, Griffith, Boylan, Hicks, Wilson, Byrne, Sheaves, Bentall and Freeman2021; Knowles et al., Reference Knowles, McCarthy-Jones and Rowse2011); negative symptoms (Rector et al., Reference Rector, Beck and Stolar2005), and trauma in psychosis (Hardy, Reference Hardy2017; Hardy et al., Reference Hardy, van de Giessen and van den Berg2020). Finally, some models are stage specific, notably Gumley’s model of recovery (Gumley and MacBeth, Reference Gumley and MacBeth2006; Gumley and Schwannaeur, Reference Gumley and Schwannauer2006). In addition to differences in focus, these models vary considerably in complexity and quality of supporting evidence.
Learning to formulate psychosis
Given the range and complexity of current theoretical models, trainees can find the formulation process difficult and confusing. In the context of postgraduate training, we have often observed trainees developing complex formulations to satisfy assumed academic requirements, which are then put to one side to focus on familiar interventions. The absence of a guiding formulation is likely to result in therapeutic drift both from a clear focus on key maintenance processes, and from the person’s goals. Formulations need to be accessible, explanatory, and form a ‘road map’ for therapy; a failure to formulate effectively and then integrate the formulation into the treatment plan jeopardises the potential benefits of CBT (cf. Waller, Reference Waller2009; Waller and Turner, Reference Waller and Turner2016).
Clinical supervision is the primary means of facilitating trainees’ understanding and effective utilisation of psychological models (Fleming and Steen, Reference Fleming and Steen2004; Hall, Reference Hall2012; Milne, Reference Milne2007). In CBT training programmes, supervision typically involves discussion of prepared supervision questions, observation of linked video clips, and reflection on the content and process of therapeutic interactions (following Padesky, Reference Padesky and Salkovskis1996). Regular clinical supervision is a requirement for trainee and qualified psychological therapists to ensure safe and effective practice (BPS, 2017), and we now have a range of guidelines and frameworks to support this (e.g. Bennett-Levy, Reference Bennett-Levy2006; Gordon, Reference Gordon2012; Hawkins and Shohet, Reference Hawkins and Shohet2000; Milne, Reference Milne2009; Roth and Pilling, Reference Roth and Pilling2008/2015).
Importantly, the evidence suggests that CBT training is largely ineffective without the reflective process of supervision, and that the more reflection, the greater trainees’ learning (Bennett-Levy and Padesky, Reference Bennett-Levy and Padesky2014; Haarhoff and Thwaites, Reference Haarhoff and Thwaites2016). Supervision is therefore the place where trainees are most likely to be grappling with CBT models of psychosis, and where there may be the greatest opportunity to facilitate depth of understanding and skilful application.
Current study
We sought to explore CBT trainee and recently qualified therapists’ experience of formulating psychosis. With a greater understanding of what exactly people struggle with when using the underlying theoretical models, we can shape guidelines for their use in clinical practice – both for training courses and NHS supervisors supporting those new to CBTp. We also sought to elucidate barriers and facilitators to this process. To our knowledge, no previous research has examined trainees’ experience of formulating psychosis or drawn out the implications for training and clinical practice.
Method
Research team
The study was designed and run, and the analysis conducted by K.N.-T. and L.I. Both are consultant clinical psychologists, expert practitioners in CBTp, and course leads of postgraduate training in CBTp in the UK. Their a priori assumptions were that individualised formulation is a fundamental component of CBTp, with which trainees frequently struggle. Interviews were conducted by R.W., a senior clinical psychologist, expert practitioner in CBTp, and joint lead for the locality ‘psychosis pathway’ for psychological therapies in her NHS Trust. Interviews were recorded and transcribed by A.E., a psychology assistant supporting the project.
Participants
Participants were current or recent trainees completing postgraduate training in CBT for psychosis at the Universities of Southampton and Oxford, in the UK. A total of 17 trainees took part, including 11 women and six men, with an average age of 45.86 years (SD = 5.18) (five people did not report their age). All identified as White. Participants were healthcare clinicians, employed as clinical and forensic psychologists, CBT therapists, mental health nurses, social workers and occupational therapists.
Procedure
Current and recent trainees were invited to take part in the study. All who expressed an interest were provided with information and consented. Participants were asked to attend one of two focus groups run at each of the two universitiesFootnote 1 . Semi-structured interviewsFootnote 2 were conducted by a senior clinical psychologist and CBT therapist independent of the training programmes, supported by an assistant psychologist who recorded and transcribed the interviews. The interviewer used pre-agreed prompts to facilitate discussion, exploring concepts that appeared to have meaning and relevance to participants (Bertrand et al., Reference Bertrand, Brown and Ward1992; Kitzinger, Reference Kitzinger1994). The assistant psychologist made field notes while observing the interviews to identify speakers and aid later transcription.
Data analysis
We used thematic analysis to explore the data as we sought a rich thematic description of an under-researched area (Braun and Clarke, Reference Braun and Clarke2006). Specifically, we used a coding reliability approach whereby themes (or topic summaries) are developed early, and coding is used to identify evidence (or absence of evidence) for these in an iterative process (Braun and Clarke Reference Braun and Clarke2021). Two of the four transcripts (one from each of the sites) were initially coded in full by two members of the research team (K.N.-T. and L.I.) who independently recorded ideas for possible codes before discussing. These preliminary codes were informed both by the data and the researchers’ assumption that trainees often struggle to formulate psychosis. There was a large degree of overlap in the preliminary codes generated, although a difference was the extent to which a broad versus fine-grained coding approach was used. Discussion regarding the relative costs and benefits of each approach (cf. Guest et al., 1989) resulted in the adoption of a relatively broad coding framework consisting of three key themes (revised topic summaries) (see Fig. 1). A third transcript was then independently coded by the same raters using this framework to assess inter-rater reliability. There was moderate to good agreement (Burla et al., Reference Burla, Knierim, Barth, Liewald, Duetz and Abel2008) between raters for each code (Cohen’s kappa ranged from 0.63 to 0.84), and where there were differences, theme definitions were further refined to achieve agreement. All four transcripts were then coded by K.N.-T. using the final framework to gather evidence for each of the themes. The codebook collated verbatim material grouped by themes, with definitions. We used NVivo 12 software for all analyses.
Themes (topic summaries) generated in qualitative analysis.
Results
Three over-arching themes were generated: (1) purpose of formulation, (2) formulation in practice, and (3) (reflection on) learning to formulate (see Fig. 1).
Purpose of formulation
Trainees reflected on the function of formulation, to inform decisions about which models to use. This theme captured a combination of uncertainty and striving to determine the focus and purpose of formulation.
I think I went into it quite blindly that [formulation] was the thing to do and it was the right thing to do, without perhaps fully thinking about … what the purpose of it was. [Group 3, ppt A]
There are all different sorts of formulations. They’re all useful in making a decision about what to do with someone. That’s something I’ve got clarity about … what’s a formulation, but well you have to be more detailed on that, don’t you? Is it the formulation for the person? … But it’s like – where are we with the client? It’s about like what’s the formulation of what’s blocking you from your goal? So, what’s the problem? … I’ve got to start computing – what am I formulating here? Am I formulating the person as a whole or this strange symptom that they’re reporting? … So, I think formulation is linked really closely with goal planning and problem solving, assessment and… I think I got much better at doing that. [Group 1, ppt C]
Some highlighted how formulation helped to make sense of unusual and initially un-understandable behaviours, possibly meeting trainees’ own needs to make sense of psychosis, early in therapy.
I think it has [helped me make sense of their problems]. And then it kind of gives you [a] framework on which to think about what you’re going to do next and that. Something about that is quite containing. [Group 3, ppt C]
If you’ve got someone that’s sort of barricading their bedroom, like objectively that behaviour seems outrageous, you know – unusual, odd, weird, all of that, kind of concerning. But I guess working backwards from that… and trying to formulate it and trying to identify the individual’s appraisals of what’s happening. So, usually it’s threat based … ‘Someone is going to kill me, kill my family’, ‘I can prevent it’ or ‘I need to engage in these to keep me safe’. Then suddenly it becomes far more understanding that they’re barricading themselves in the bedroom. Because that’s what I’d do. [Group 1, ppt G]
Given differences in complexity of theoretical models and linked clinical heuristics, many expressed a preference for simple maintenance cycles, while others valued developmental formulation as a means of fostering a contextualised understanding of the person’s psychosis.
I just think if it’s really, really simple then I think it’s going to be quite easy for them to kind of think about that again in the future, or bring it out again and go ‘OK, that links to that, that links to that, that links to that’ … a more complicated model, they might look at in the future and go ‘Oh, crumbs, what was this all about?’ [Group 2, ppt A]
I mean, like the [developmental] model is much better for me to get my understanding outside of sessions. Just to play around with that, and … feeding back to the team … ‘This is where we’re at.’ [Group 2, ppt D]
Working with people with all their difficulties, tend to be driven by formulation, often use maintenance cycles. But then I … try and come up with a, a shared longitudinal formulation so they can see what might be driving it. [Group 1, ppt D]
Some trainees also noted that formulating can be a way of managing (and avoiding) the course of therapy.
I think it’s useful for, for keeping therapy on track … it’s easy to either, for a therapist or the client right throughout the course of therapy, to avoid dealing with things that are most problematic or uncomfortable in any way. [Group 1, ppt A]
I also think you can hide behind formulations well. And particularly if you’re not really confident in delivering interventions. So… you can just keep on formulating and that can be a mechanism for just hiding behind your kind of, I guess your lack of competence in delivering sort of interventions if you haven’t delivered some of them before or you’re quite new at doing it. [Group 1, ppt G]
Formulation in practice
This theme highlighted participants’ experiences of using formulation in clinical practice. Trainees reflected on the impact of formulating within therapy sessions, and ‘fit’ for the particular person.
What’s most helpful in the session … he said it was [the] actual realisation that it was understandable why he did what he did. That was the most helpful thing. [Group 2, ppt D]
It was I suppose a very simple interpret- interpretation of sort of some physical sensations and that lead to… a decision about what he did, then that sort of maintained things …and we looked at how that was maintaining. But by actually interpreting, considering his interpretation and an alternative and doing something different, thinking about impact … it’s sort of a lightbulb moment, isn’t it? [Group 1, ppt B]
Everybody is different though, aren’t they? I’ve had people who have had to slow down and pull back and really slow down. There is another new gentleman I worked with recently and he just looked at the maintenance cycle and he went ‘Ah, I need to change that. If I go out with my hoodie on …’ and you know ‘by that taking it down I will see, maybe I’ve got this wrong’. He really got it straight away. [Group 2, ppt B]
Trainees also highlighted the collaborative nature of formulating, and not knowing, as a key part of the therapeutic process.
I think having a written and shared formulation on the table in front of us is something that is really helpful and … it has been incredibly useful to contain, to keep referring back to things as you go through the course of therapy. And to have it as a live document that I keep adding to. [Group 1, ppt C]
I think they’re there with it and talk about it and then all of a sudden, I’m sitting there thinking ‘No, no, they don’t get it’ so, again … I will formulate just in a maintenance cycle. So, it’s helpful … to have something for them to look at to know whether they’re there with you. That’s where formulation is so helpful. [Group 2, ppt D]
There’s this sort of therapist dilemma … if you don’t know it, if you, or something doesn’t fit or match with your understanding of things, just say it. And I find that quite liberating … because it just doesn’t fit sometimes … just saying it ‘I can’t really understand’, that’s really helpful. [Group 1, ppt G]
In line with concerns about complexity in the previous theme, some noted that more complex formulations can be unhelpful in practice.
My experience has been that it has been quite hard to collaboratively use the psychosis specific formulations, because they are complex. [Group 3, ppt C]
I think it’s often too many arrows. I think that’s the problem. People can get lost in the arrows. [Group 4, ppt A]
Trainees also reflected on what facilitated impactful formulation. Given concerns about complexity, trainees highlighted simplicity, a focus on goals, and not holding the formulation too tightly.
I just keep it really, really simple … it works. Keeping it simple really works. [Group 2, ppt A]
I think it’s the goal really. I think where I got into trouble formulating is normally because the goal is unclear in the first place. So, better clarity on what the goal is enables you to formulate. Keep the goal in mind. [Group 1, ppt C]
Not holding it too tight as well. Just to, it’s not something that you need to chisel out in stone. [Group 1, ppt E]
I think formulation is never set in stone as such. I think it’s important to kind of keep an open mind … and revise it as you get more information, which might come through the change work – they may be like ‘Oh, this happened as well’. Add that in and perhaps see how things change. [Group 3, ppt C]
(Reflection on) learning to formulate
Having grappled with the purpose of formulation, in order to determine which theoretical models or heuristics to draw on, and formulation in practice, to ensure goodness of fit and impact for the person, trainees reflected on the process of learning to formulate, some describing this as daunting.
Trying to find your feet and get your head around [formulation] and then trying to do that with a patient in a productive and helpful manner, collaborative manner, can be a challenge. [Group 4, ppt A]
Even though I’m at the end of this diploma I still feel quite daunted by, you know, I can sit back and think of formulation, but doing it with a client in a useful way, I still feel really daunted to make it feel useful and simple enough for a client. [Group 1, ppt F]
Key learning and recommendations to others included the value of keeping the formulation simple (‘good enough’), prioritising the person’s goals, and practising.
When it’s complex, keep it simple. [Group 1 ppt G]
You can have a good enough formulation, it doesn’t have to be the best formulation. Because that could still be really helpful for the person. Really, having all the information, I don’t think it particularly matters as long as there are some poignant links for them. I think it’s about good enough and I don’t think that’s a bad way to look at it. You’ve got time to get to that realisation. [Group 2, ppt A]
Take your time, practise, really practise, use supervision to guide your knowledge. [Group 4, ppt A]
Trainees also noted helpful aspects of training, and made recommendations for trainers/supervisors. This included close supervision, illustrative material, and encouragement to ‘get it wrong.’
Supervision has been so helpful, trying to unpack these questions. [Group 3, ppt A]
And supervision, use supervision ... it takes practice and practice and practice, doesn’t it? Just know that you’re going in with the right direction. So, yeah. Just keep trying, keep doing it. [Group 2, ppt D]
A: Maybe, watching videos of clinicians ...
B: Yeah.
A: ... sharing formulations with clients...
D: That’s a good idea.
A: ... could have been helpful. [Group 1]
I think one thing that I’d say to carry on doing is that kind of permission to get things wrong and then trying stuff in supervision ... sometimes we would just be asked to present someone, and then say you’re a bit unsure, uncertain, the supervisor would say ‘Well start and we will do it together’ and that encouragement of, because you’re learning, I think that was a huge part of this course. [Group 2, ppt D]
Finally, trainees recognised their ongoing learning beyond training.
Using that knowledge usefully with the client, so that bigger formulation and applying that with the client in a simpler way, is still where I’m learning and training. [Group 1, ppt F]
It is a real skill to really get ‘Well I’m not quite there yet, but I’m getting there’. [Group 2, ppt D]
Discussion
Many CBT models of psychosis are theoretically complex. Having seen how trainee CBT therapists often struggle to draw on these to develop individualised and useful formulations, we sought to explore their experiences to help us understand how best to guide training and clinical practice.
A thematic analysis of interview transcripts yielded three over-arching themes: (1) purpose of formulation, (2) formulation in practice, and (3) (reflection on) learning to formulate. Trainees grappled with the purpose of formulation, seeking to clarify the function both for themselves and people with psychosis. In formulation in practice, trainees focused on both what and how to formulate effectively – highlighting goodness of fit and genuine collaboration in the process. Finally, trainees were able to reflect on the process of learning to formulate, noting that this could feel daunting, and identifying key learning for themselves and trainers (including supervisors).
The tension between theoretical complexity and the need for clear and useful formulations emerged across themes, with trainees emphasising the value of simplicity where possible and in line with the person’s goals. Others highlighted the need for more complex formulations when the person is likely to benefit from a developmentally contextualised understanding of their psychosis. This places responsibility on CBT training programmes to ensure trainees have the opportunity to practise formulating psychosis simply, and drawing on more complex models, and to use supervision to reflect on their clinical decision-making regarding these choices.
As we continue to work towards parity of esteem for people with psychosis, we need to ensure that adherent ‘full dose’ recommended therapies are available to all (cf. British Psychological Society/Cooke, 2017; NICE, 2014). This includes formulation-based CBT (cf. Turner et al., Reference Turner, van der Gaag, Karyotaki and Cuijpers2014; Turner et al., Reference Turner, Burger, Smit, Valmaggia and van der Gaag2020). We know that CBTp formulation can reduce distress (Brabban et al., Reference Brabban, Byrne, Longden and Morrison2016), and predict subsequent quality of life (Stainsby et al., Reference Stainsby, Sapochnik, Bledin and Mason2010), yet trainees, and possibly experienced therapists, often find this hard.
In a qualitative study of CBTp trainees’ needs in clinical supervision, Harris et al. (Reference Harris, Maguire and Newman-Taylor2022) found that supervisors should prioritise knowledge, procedural learning and active experimentation. The current study indicates that in addition to learning the theoretical models in the area, we should be facilitating procedural skills in formulating, prompting trainees to practise (and get this wrong) in supervision. Modelling formulation is likely to be particularly useful. In an expert review of CBT training methods, modelling was rated as one of the most effective ways to facilitate procedural skills development (Bennett-Levy et al., Reference Bennett-Levy, McManus, Westling and Fennell2009). As an example, we have found that brief role play (of just a few minutes) in supervision can be a useful way to demonstrate skills. Once trainees have observed us and reflected on moment-to-moment clinical decision-making, they can be encouraged to repeat the role play taking the therapist role, with us coaching live to shape therapeutic interactions. The use of both ‘live supervision’ and linked reflection is likely to facilitate trainees’ ability to formulate psychosis effectively (Bennett-Levy and Padesky, Reference Bennett-Levy and Padesky2014; Haarhoff and Thwaites, Reference Haarhoff and Thwaites2016; Padesky, Reference Padesky and Salkovskis1996).
As a small, exploratory study with participants from just two universities, the generalisability of our conclusions is limited. Nevertheless, trainees’ reflections on the process of learning to develop individualised formulations for and with people with psychosis yield a number of implications for future trainees, training programmes and clinical supervisors (see Table 1).
Implications for formulating psychosis
We recommend that future research focuses on how formulation can be used most effectively for people with psychosis – to inform decision-making (and therefore training) about readiness for formulation, when to draw on simple or more complex models, how formulation is best delivered (e.g. whether narrative formulation might strengthen impact and optimism regarding next steps), and any adverse effects.
Conclusion
Training programmes and clinical supervisors should prioritise clarity of function, simplicity where possible, and collaboration in formulating psychosis. Modelling, practice and reflection on clinical decision-making are likely to strengthen procedural learning. This will support the development of CBT therapists who are clear minded about what they are seeking to formulate, and skilled in fostering a sense of joint endeavour. Individualised formulations that clearly articulate the processes maintaining distress or blocking valued goals are likely to facilitate joint understanding and decision-making, thereby providing the basis for effective CBTp.
Key practice points
-
(1) CBTp therapists should prioritise clear, targeted, collaborative formulation.
-
(2) CBT models for psychosis vary in focus, complexity and empirical support.
-
(3) Individualised formulations draw on maintenance and developmental models of psychosis, depending on the person’s needs and goals.
-
(4) Training programmes and clinical supervisors should prioritise opportunities for procedural learning – encouraging trainees to practise formulation and get it wrong in order to become skilled therapists over time.
Supplementary material
To view supplementary material for this article, please visit: https://doi.org/10.1017/S1754470X22000459
Data availability statement
The data that support this paper are available from the corresponding author upon reasonable request.
Acknowledgements
We thank our CBT trainees who help us understand how better to teach and supervise CBT for psychosis. We would also like to thank Micha Exton and Rosemole Vinny who drew on this data for their MSc projects and raised interesting questions in supervision.
Author contributions
Katherine Newman-Taylor: Conceptualization (equal), Formal analysis (equal), Writing – original draft (lead), Writing – review & editing (equal); Rachael Wood: Data curation (equal); Abbie Ellis: Data curation (equal); Louise Isham: Conceptualization (equal), Formal analysis (equal), Writing – review & editing (equal).
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conflicts of interest
Katherine Newman-Taylor is an Associate Editor of the Cognitive Behaviour Therapist. She was not involved in the review or editorial process for this paper, on which she is listed as an author. The other authors have no declarations.
Ethical standards
The authors have abided by the Ethical Principles of Psychologists and Code of Conduct as set out by the BABCP and BPS. Ethical approval was given by the University of Southampton (reference number: 48690.A1) and University of Oxford (reference number: EQ CIA 19 036).
Open access
Comments
No Comments have been published for this article.