A qualitative process evaluation of social recovery therapy for enhancement of social recovery in first-episode psychosis (SUPEREDEN3)

Abstract Background: Many individuals with first-episode psychosis experience severe and persistent social disability despite receiving specialist early intervention. The SUPEREDEN3 trial assessed whether augmenting early intervention in psychosis services with Social Recovery Therapy (SRT) would lead to better social recovery. Aims: A qualitative process evaluation was conducted to explore implementation and mechanisms of SRT impact from the perspective of SUPEREDEN3 participants. Method: A subsample of SUPEREDEN3 trial participants (n = 19) took part in semi-structured interviews, which were transcribed verbatim and analysed thematically. Trial participants were early intervention service users aged 16–35 years with severe and persistent social disability. Both SRT plus early intervention and early intervention alone arm participants were interviewed to facilitate better understanding of the context in which SRT was delivered and to aid identification of mechanisms specific to SRT. Results: The six themes identified were used to generate an explanatory model of SRT’s enhancement of social recovery. Participant experiences highlight the importance of the therapist cultivating increased self-understanding and assertively encouraging clients to face feared situations in a way that is perceived as supportive, while managing ongoing symptoms. The sense of achievement generated by reaching targets linked to personally meaningful goals promotes increased self-agency, and generates hope and optimism. Conclusions: The findings suggest potentially important processes through which social recovery was enhanced in this trial, which will be valuable in ensuring the benefits observed can be replicated. Participant accounts provide hope that, with the right support, even clients who have persistent symptoms and the most severe disability can make a good social recovery.


Introduction
Despite considerable advances in psychosis treatment (Kingdon and Turkington, 2017;Nordentoft et al., 2014), functional outcomes of individuals with psychosis remain poor. Even with specialist early intervention services, less than 40% of people experiencing first-episode psychosis achieve social recovery (Chan et al., 2015;Hodgekins et al., 2015a;Morgan et al., 2014), defined as engagement in normative interpersonal and occupational activities (Hodgekins et al., 2015b). Such difficulties often become entrenched, resulting in long-term social disability (Revier et al., 2015;Strauss et al., 2010). Symptomatic remission was long considered the primary goal of psychosis treatment (Greenwood et al., 2010), but more recently there is increased focus on improving social recovery (Alvarez-Jimenez et al., 2016;Fowler et al., 2010;Santesteban-Echarri et al., 2017). This shift reflects the importance of socio-occupational functioning within patient constructions of recovery (Law and Morrison, 2014;Pitt et al., 2007), and the substantial economic costs of persistent social disability (Andrew et al., 2012;Park et al., 2016).
Social Recovery Therapy (SRT) is an individual psychosocial intervention designed to promote time spent in structured activity, through understanding the client's values and goals (Fowler et al., 2013;Fowler et al., 2019). Barriers to engagement in structured activity are formulated using a Cognitive Behavioural Therapy (CBT) approach and in vivo behavioural strategies used to overcome avoidance and promote meaningful behavioural change . Evidence gained from behavioural work is used to inspire hope and positive beliefs about self and others . Unlike conventional CBT, therapists take an assertive outreach approach and work with families and external agencies to facilitate structured activity engagement . SUPEREDEN3 (ISRCTN61621571) was a randomised controlled trial testing whether providing SRT to young people who had social disability, despite having received at least one year of early intervention in psychosis services (EIP), leads to better social recovery than EIP alone (Fowler et al., 2018). In addition to persistent social disability, participants had severe symptoms at trial entry; over 50% of participants had positive psychoisis symptoms and 94% negative symptoms, with high levels too of depression, social anxiety and hopelessness (Fowler et al., 2018).
Providing SRT alongside EIP resulted in large, significant and clinically important improvements in social recovery, with SRT participants engaging in an additional 8 hours per week of structured activity compared with those receiving EIP alone (Fowler et al., 2018). This gain reflects a full working day and is twice that identified by a consensus group of patients and clinicians as the minimum clinically important effect. Whilst SRT is a demonstrably effective intervention for this complex group, focusing on outcomes alone does not uncover why SRT is effective. As being able to answer the 'why' question is crucial to successfully implementing complex interventions (Moore et al., 2015), we conducted a qualitative process evaluation alongside the trial. Our aim was to explore how SRT was implemented and the mechanisms of impact specific to SRT from the perspective of trial participants.

Method
Participants SUPEREDEN3 trial participants were aged 16-35 years, had non-affective psychosis, had been accessing EIP in Birmingham, Lancashire, East Anglia or Sussex for 12-30 months, and had social disability; defined by less than 30 hours per week on average in structured activity (assessed using the Time Use Survey; Hodgekins et al., 2015b;Short, 2003) (Fowler et al., 2018). Participants were randomised to receive either SRT in addition to EIP or treatment as usual (TAU), which reflected EIP alone. TAU participants had access to the full range of psychological, social and medical interventions offered by specialist EIP services, which includes support aimed at increasing social and occupational functioning, but did not receive SRT. Further details of the trial interventions and procedures can be found in the main trial outcome paper (Fowler et al., 2018).
A sub-sample of SRT plus EIP and TAU trial participants were invited to participate in the qualitative process evaluation. Convenience sampling was used within a sampling frame that aimed to: (a) recruit SRT and TAU participants in a 2:1 ratio, reflecting the focus on providing insights into potential mechanisms of SRT, and (b) to ensure the number of participants from each of the four sites broadly reflected the composition of the overall trial sample.

Procedure
Interviews were conducted in-person by trained research assistants, in participants' own homes, or mental health service or community venues, at participant preference. Flexible semi-structured interview schedules elicited: experiences of trial participation and intervention received; future hopes and expectations (SRT plus EIP or TAU); and perceived SRT intervention mechanisms and outcomes in the SRT plus EIP arm. Interviewers attempted to elicit detailed accounts and probed for negative as well as positive views. Interviews were audio-recorded and transcribed verbatim.

Analysis
We conducted a reflexive thematic analysis (Braun and Clarke, 2006;Braun and Clarke, 2013;Notley et al., 2014). The themes were identified separately for SRT plus EIP versus TAU participants. This reflected the use of different versions of the interview topic guide, and our interest in exploring ways in which the experiences of TAU participants appeared similar to, and distinct from, those of SRT participants, i.e. to elucidate whether the current data provided any preliminary evidence as to whether SRT appears to be associated with distinct mechanisms and outcomes.
The first analysis phase was deductive; the first two authors developed a coding framework based on the SRT adherence checklist , which was refined after application to eight transcripts across both trial arms. The adherence checklist consists of the 14 key components of SRT as detailed in the therapy manual, including engagement, assessment and formulation, cognitive and behavioural strategies and involving other systems. Next, this coding structure was applied to the remaining transcripts and additional codes generated inductively to capture semantic units not present in the deductive framework. This approach was intended to facilitate both exploring the extent to which participants' experiences of the intervention aligned with the therapy manual  and additional unanticipated factors important to participants' experiences of SRT. We took a critical realist epistemological stance (Danermark et al., 2002), seeking to understand participants' individual views of reality through close engagement with their personal narratives.
The analysis team consisted of four individuals involved in the development and/or delivery of the SRT (J.H., M.F., K.P., D.F.) and four not involved in its development or delivery (B.G., C.B., A.L., C.N.). The analysis was led by research team members not involved in the initial development or delivery of SRT (B.G. and C.B.) and the initial deductive phase of the analysis was carried out by these two researchers. During the inductive phase of the analysis (again led by B.G. and C.B.), each transcript was independently coded by at least two analysis team members and meetings held to discuss coding and to reach consensus on themes identified and data interpretations.

Participants
Nineteen SUPEREDEN3 participants (Table 1) were interviewed: 14 who received SRT in addition to EIP, and five who received EIP alone (TAU).
On average, participants were doing 9.24 hours (SD 7.72) of structured activity per week at baseline, which rose to 32.01 hours (SD 27.94) at 9 months, and 28.12 hours (SD 23.16) at 15 months. Fifteen study participants experienced a meaningful increase in their time spent in structured activity (defined as in excess of 4 hours gained). For 11 of the participants, this increase was sustained at both the 9-and 15-month follow-up point. With respect to clinical symptoms at baseline, participants scored a mean of 60.53 (SD 12.29) total Positive and Negative Syndrome Scale (PANSS) symptoms, 20.72 (SD 11.66) total Beck Depression Inventory (BDI) symptoms, and 44.22 (SD 14.67) total Social Interaction Anxiety Scale (SIAS) symptoms. The mean (normative population scaled) index of neurocognitive performance was on average -0.24 (SD 1.15). Participants scored on average 32.76 (SD 9.99) for the total Trait Hopelessness Scale (THS). Compared with the full trial sample (Fowler et al., 2018), therefore, participants involved in the process evaluation experienced slightly more social disability at baseline, but made greater gains at 9 months and maintained slightly greater gains at 15 months. Process evaluation participants had slightly reduced PANSS symptoms but experienced slightly greater depression and social anxiety at baseline. Process evaluation participants had very similar levels of hopefulness, and better neurocognitive performance, compared with the overall trial sample.
SRT participants completed an average of 19.29 (SD 8.75) sessions, which is slightly higher than the trial average of 16.49 sessions (SD 8.39); albeit similarly variable. Variability in number of therapy sessions is partly explained by the flexible, adaptable, and highly personalised nature of the intervention , but also by the representation in the present study of participants receiving full versus partial versus no dose of the intervention. Seven of the process evaluation participants had a full dose (50.0%), six (42.9%) had a partial dose, and one (7.1%) no dose. In the full trial, 57.3% of participants had full dose, 24.0% partial and 18.7% no dose, meaning that participants with a partial dose are slightly over-represented and those with no dose under-represented in the process evaluation sample.

Experiences of SRT
Six themes describing the implementation of SRT and perceived mechanisms were identified. These themes were used to develop a participant-perspective model of SRT as implemented in the trial and why it was effective in enhancing social recovery. Pseudonyms are used to protect participant anonymity.

Increased self-knowledge
Participants described SRT as facilitating increased self-knowledge. Consistent with cognitive behavioural approaches for anxiety, depression, and psychosis, increased self-knowledge was centred on identifying the factors underpinning problems. For instance, Hannah spoke about how her therapist drew diagrams to help explain what maintained the symptoms she was experiencing: 'He explained a lot, like I really got into the cycles of why everything's happening and like he used to draw them on paper and it used to really help and that's the one thing that sort of stuck in my head really about it, is just he sort of understood how I was feeling.' Hannah, Lancashire For other participants, increased self-understanding came through uncovering latent goals and ambitions. For instance, Samir described realising that, through exploring his previous interests and valued activities with his SRT therapist, one of his main priorities was improved physical health: 'We were just talking about my interests and kind of things that I used to do, or would have done, or would like to do, and it turned out that the biggest thing that was on my agendaon my mindwas to lose weight, for example and have regular exercise.' Samir, Lancashire Another participant described not realising how little he was doing with his time until this was explored in therapy and coming to understand the impact of being 'stuck' in a pattern of low activity on his mental health: 'We worked together kind of trying to pinpoint exactly what the anxieties were, why my mood was so up and down, and from that we kind of realised that it's because I wasn't doing too much : : : I think I got stuck in a routine of doing not much. I didn't realise how little I was doing.' Stephan, Sussex Stephan's use of the pronoun 'we' in this extract conveys his sense that he and his therapist worked in collaboration to discover what might be causing and maintaining his symptoms. Participants' accounts point to the importance of the strong therapeutic relationships developed in facilitating the process of self-discovery. The importance of collaboration and need to develop shared understanding of the client's difficulties and strategies needed to overcome them, in line with recommendations for good CBT practice, is further highlighted by the experience of Ben (Birmingham) who disengaged from the therapy. He recounted finding the exercises he was asked to do 'bizarre' and 'a bit of a hotchpotch', suggesting he did not have a clear understanding of the rationale for the strategies employed.
Ben's experience contrasted with other SRT participants who perceived the variety of techniques used as indicative of the therapy being individualised. For instance, Nicky (Lancashire) stated that SRT had been more helpful than previously received therapy because 'it was personal and it was tailored around me'.
'Sometimes you just need that little bit of a push' The relationship with the therapist was central to participants' experiences of SRT. Participants described the therapeutic relationship as warm and supportive, but also spoke about the therapist assertively encouraging them to do things they felt initially uncomfortable or afraid of doing. Participants indicated that they experienced this assertive approach as helpful. Indeed, the therapists' firm and persistent encouragement seems to have been vital to facilitating change.
The perception that the therapist was motivated by the client's best interests appears to have been important to participants experiencing this assertive approach as helpful: 'Sometimes you do need a therapist or someone just to give you that little bit of a push, and also to give people the support as well, because they are kind of pushing you on to do it because they know it's gonna benefit you, and they know you're gonna feel amazing after you've done it, but then they dothey don't do it in an aggressive way, they do it in a supportive way, and I think that's really important.' Stephan, Sussex The therapist doing activities with participants also emerged as an important factor in the therapeutic relationship. Alison (East Anglia) contrasted her experience of SRT with a previous experience of CBT in which behavioural experiments were completed only as homework: 'I'd had like CBT and stuff and they'd be like "Right, this week, we want you to go outside and you're going to rate it" : : : you're going to have to do that all by yourself and motivate yourself to do that. Whereas with [SRT therapist] it was, alright she only came out once a week but she was like, come on we're going outside, we're going to do this. And alright she would leave me homework each week which admittedly was a bit of a drag : : : [but] because she'd done the first bit with me, it was a lot easier to continue doing it.' Alison, East Anglia Whilst Alison still found completing homework involving behavioural exercises challenging, it was made easier by her SRT therapist having started this work alongside her during sessions. She was therefore able to make greater progress than in previous CBT in which behavioural work was done solely as homework. Moreover, therapists' completion of activities with the client, and thus the sense of the therapist 'walking alongside' the client in their social recovery, appeared a powerful ingredient in a strong alliance. For instance, when asked whether he was satisfied by the level of support provided by his SRT therapist, Samir responded: 'Oh, it was excellent. I mean your psychologist takes you swimming, how good is that? You can't get any better than that.' Samir, Lancashire You have to face your fears to get over them The support of the therapist allowed participants to push themselves to do things they would not otherwise have felt able to, and in doing so, to overcome avoidance and ongoing symptoms in order to engage in activities. Participants described this process as challenging and, at times, unpleasant, but ultimately worthwhile. This theme is well illustrated by one participant's account of one SRT session: 'The most memorable point was when I was made to do karaoke : : : stone-cold sober, so yeah, that was the most memorable point for me, and that was quite hard. I did find that tough. There was a lot of anxiety beforehand and yeahbut obviously, finished that and thenso that was quite a good learning experience : : : if you think to yourself, "Well, if you can make that much of a fool of yourself", then other stuff's a lot easier [laughter]' Lee, Birmingham Being supported by the therapist to complete behavioural work during sessions appears to have allowed participants to challenge beliefs about the consequences of engaging in activities they would otherwise have avoided. Stephan expressed this point elegantly, saying: ' : : : that's the moral of the whole therapy, that's kind of the bottom line, is you'll never be as anxious as what you think you'll be'. It is thus through behavioural work in the community that participants developed a sense of their own hopefulness and of positive expectancies for the future. Participants spoke about being motivated to continue to push themselves to do things they found difficult by the sense of achievement they experienced during reaching goals in therapy. These goals were often described by participants as relatively modest, at least from an external perspective. The motivational effect of setting and achieving small goals within sessions is eloquently described by Alison: 'I think by having goals that small you do almost get that instant gratification nearly every time that, yes, I've done it and now I feel good about myself. And that pushes you to then do the next step because you've done it and you feel good about yourself you sort of have that confidence boost of actually yeah, I can go to the bottom of the road now.' Alison, East Anglia Alison recalled that her therapist ' : : : would constantly change the goals to me to make them obtainable', to ensure that Alison was left with a sense of accomplishment even when she was not able to complete the initial activity planned.

Increased self-agency
Increased self-agency featured prominently in the accounts of participants who had a positive experience of SRT. The participants spoke about realising that they had control over how they felt and had the power to make positive changes in their lives: 'It all came down to me having control over how I felt and what I did, and about choosing the things to do to keep myself feeling good.' Stephan, Sussex 'I always thought that it was everything else and I kind of realised no, that it were the stuff in my head that was stopping me from, from being able to do that. I think once you realise that it's yourself that's the key.' Nicky, Lancashire SRT participants felt confident of their ability to work towards a positive future. Alison explained how she felt SRT had helped empower her to continue to pursue her goals: 'Because of the way I think the therapy worked I know I can push myself. And if I push myself, yes I feel uncomfortable while I'm doing it : : : but actually it's not that unobtainable, you can do it and once you do it you feel really good.' Alison, East Anglia Unlike other SRT participants, however, Kieran did not exhibit a strong sense of agency. Kieran did not describe SRT as helpful and spoke about how he ' : : : played the role of the patient' during therapy sessions so that the therapist would not feel he was wasting his time. Explaining why he felt SRT had not been effective for him, Kieran commented: 'I think I was at the stage where I was kind of complacent. So if I'm saying I want to get better or more confident, I'm not going to seek it out. I'm not really going to do it.' Kieran, Birmingham This highlights the importance of the client being ready to make changes for the therapy to be perceived as helpful.
'More of a cup half-full than a cup half-empty' Participants who received SRT spoke about seeing things in a more positive way as a result of participating in the trial and described hope and optimism for the future. Nicky expressed this change in her outlook by saying she has become ' : : : more of a cup half-full than a cup half-empty' person since taking part in the trial. This increased optimism and hopefulness appeared to result from unexpected positive changes occurring during the course of SRT, prompting the expectation that further positive change is possible: 'Just going back over 12 months, I've passed extra qualifications, and me NVQ2 : : : I'm in the running for [a] : : : driving licence. And 12 months ago, they weren't in the pipeline, so in another 12 months, who knows?' Rob, Lancashire 'Now I can see myself being successful. I can see myself being happy, going on adventures. I can just see a good future for myself : : : if I compare myself today to what I was 12 months ago, that is a massive change, so the next 12 months, 12 years, could be completely mind-blowing.' Stephan, Sussex

Experiences of TAU
Participants allocated to the TAU arm also reported experiences of being supported to increase their activity during the period in which they were enrolled in the study. Several participants talked about benefiting from being supported by their EIP team to attend social groups or take part in sports. For instance, Aaron (Birmingham) mentioned that attending a local youth group supported by his early intervention worker had increased his social skills and helped him to feel more confident. However, there were several ways in which the experiences of TAU participants were distinct from those of SRT participants, helping to clarify the features of participants' experiences that may be specific to SRT.
TAU participants spoke about past experiences of therapy being largely clinic-based, with limited opportunities to implement their learning. For instance, Niall (Lancashire), talking about his experience of receiving CBT, commented that ' : : : the main issue was that I didn't have a lot of opportunities to put everything we were talking about into practice'. Niall was clear that he benefited from CBT in the longer-term, but felt that his progress was slower than it might otherwise have been because his lack of social interaction limited his opportunity to ' : : : build up that evidence base that not everybody is a negative person'.
Several TAU participants' accounts conveyed a sense of lack of agency in their future wellbeing. For instance, when asked about how he saw his future, Jack answered: '[My future is] very unclear. Could be under a bridge, could be living abroad, you know. Just depends on if I'm not quite changed. I'm pretty much stuck here until I get lucky.' Jack, Lancashire This sense of the future being unpredictable and uncontrollable was a salient feature of TAU participants' accounts. Relatedly, TAU participants expressed largely pessimistic attitudes towards the future. For instance, Megan spoke movingly about her worries that she would never be capable of employment and would continue to be reliant on welfare benefits: 'I don't really see anything bright [in my future] but I'm trying not to dwell on it : : : one day I'm alright, the next day, well I just sit there and I cry and I don't even know why I'm crying and I think well can I actually function within a job? And then I think well I'm just a doledosser then, because physically there's nothing wrong with me.' Megan, Lancashire Her comments also powerfully convey the personal impact of the societal stigma associated with needing to claim benefits due to health difficulties that are less visible or less readily understood by others.

Discussion
This qualitative process evaluation was designed to explore how social recovery therapy (SRT) and treatment as usual (TAU), comprising early intervention in psychosis services (EIP), were experienced by a subsample of participants in the SUPEREDEN3 trial. The aim of this process evaluation was to provide insights into why SRT was more effective in improving social recovery than EIP alone. Six themes were identified, characterising the implementation of SRT and the mechanisms through which it enhanced social recovery from the perspective of trial participants. The findings were used to generate an explanatory model of the effectiveness of SRT for enhancing social recovery following first-episode psychosis.
The findings of this study suggest the importance of the SRT therapist cultivating increased self-knowledge in order to identify meaningful goals and stimulate initial motivation pursuing them. Participant experiences of this aspect of the intervention align with the SRT manual (Fowler et al., 2013;Fowler et al., 2019) which directs the therapist to carry out a detailed assessment of the client's personal goals, and develop a formulation focused on the individual's unique barriers to social recovery. Participants highlighted the importance of the therapist being able to assertively encourage clients to face feared situations in a way that is perceived as supportive and collaborative. Participants' accounts suggest that the therapists' use of assertive outreach strategies to encourage them to push themselves beyond entrenched patterns of avoidance, and readiness to undertake behavioural work with clients during sessionsto 'walk alongside' themwas key to facilitating change.
Participants in the SUPEREDEN3 trial presented with extreme withdrawal and inactivity, copresent with complex ongoing symptoms of psychosis, anxiety and depression. It could be that, with such clients, therapists need to accompany clients on their initial behavioural experiments to engender engagement and success rather than solely suggesting activity as homework. This approach appears to additionally build a sense of meaningful connection between the therapist and client. The ability of therapists to link personally meaningful values and goals to achievable targets also appears to have been important to participants' willingness to engage in challenging activity.
Participants reported that by encouraging and supporting them to complete activities they would not otherwise have felt able to do, and breaking down larger goals into smaller more achievable goals, therapists scaffolded a sense of accomplishment. This generated motivation to continue to set and work towards more ambitious goals, initiating a virtuous cycle. By gaining increased motivation through experiencing positive change, participants often succeeded in achieving goals they had thought unreachable. Again, it appeared to be the experience of in vivo change in their day-to-day life that led to increased self-agency. This is in line with the theory underpinning SRT which suggests that judicious use of behavioural work can instil hope and promote positive beliefs about self. Participants' experiences additionally echo the cognitive hope model, in which hope comprises self-agency and pathways-thinking oriented towards meaningful goals (Snyder, 2000). This model espouses the positive feedback loop observed by SRT participants, in which hope increases goal attainment, which generates positive emotion and, in turn, greater hopefulness (Snyder, 2000). Importantly, people who are high in hope experience positive emotion even when encountering goal barriers (Snyder, 2000). SRT therapists appear to have effectively supported the development of hopeful thinking and demonstrated how flexible goal re-setting can ensure goal pursuit remains active even when barriers are encountered.
The experiences of SRT recounted by participants echo some of the themes identified in research exploring experiences of other psychological therapies for psychosis (Holding et al., 2016). For instance, the centrality of the therapeutic relationship, the role of improved selfunderstanding, and increases in optimism were identified themes within patient accounts of CBT for psychosis (Kilbride et al., 2013). However, some aspects of the experience of SRT appear to distinguish it from other such interventions; the assertive nature of the therapeutic relationship, the sense of the therapist 'walking alongside', the opportunity to face feared situations with the therapist, and the rich articulation of in vivo engendered hopeful thinking.
Quantitative investigation of the potential mechanisms identified is now needed to provide empirical support for the model generated by this study. For instance, research needs to investigate the extent to which increased hopefulness during the intervention mediates the effect of the intervention on social recovery.

Limitations
A limitation of this study was the use of convenience sampling. While no qualitative study can claim generalisability, the use of convenience sampling increases the chance that the views of included participants did not reflect the breadth of experiences of all trial participants. A methodological concern could be that only those with more positive experiences of trial participation would volunteer to take part in the qualitative component. This concern does not appear to have been borne out, however: of the 14 SRT arm participants in the current study, two clearly reported that the intervention had not been helpful and thus at least some variation in participant views of SRT was captured. However, in comparison with the full trial sample, on average participants in the process evaluation received more intervention sessions and experienced slightly greater gains in structured activity, which may limit transferability of the findings to some extent. Furthermore, fewer TAU participants were recruited than originally intended due to fewer TAU participants accepting the invitation to be interviewed. This is in line with the greater attrition of TAU participants in the trial as a whole (Fowler et al., 2018) which we believe is due to the effect of SRT on participants' level of engagement.
A further potential limitation is that members of the analysis team were involved in the development of SRT and/or delivered therapy in SUPEREDEN3. We argue that the involvement of researchers with expertise in SRT and understanding of its implementation in the context of the trial is a strength; however, we acknowledge that we could have inadvertently minimised negative views or prioritised our own understandings of SRT over those of participants. We designed the analysis plan mindful of these potential biases; incorporating a deductive phase conducted by researchers who were not trial therapists to explicitly identify the extent to which participant experiences of the intervention aligned with the therapy manual, and ensuring all transcripts were independently coded by at least one researcher not involved in the initial development or delivery of SRT.

Conclusions
The SUPEREDEN3 trial demonstrated that offering SRT in addition to specialist EIP leads to clinically important gains in functioning relative to early intervention alone. This process evaluation contributes to explaining this finding by providing a model of the mechanistic impact of SRT as implemented in the trial, from the perspective of trial participants. This model could be used to inform quantitative investigations of the factors that moderate and mediate the effect of SRT on social functioning. Such an explanatory model will be valuable in guiding the development and implementation of SRT by providing insight into how, why, and in what contexts SRT facilitates enhanced social recovery, and thus how these effects can be successfully replicated. The findings of this study are additionally of wider relevance to professionals working with people presenting with persistent social disability following a first episode of psychosis. Participants' accounts provide hope that, with tailored intervention, even those with the most severe social disability and persistent symptoms can achieve ambitious personal goals.
Author contributions. All authors were involved in formulating the aims and designing this study. Many authors were involved in the SUPEREDEN3 trial for which the present study is a process evaluation, in designing the research programme and obtaining funding (D. Financial support. The present work was conducted as part of the SUPEREDEN3 trial, which was funded by the National Institute for Health Research (NIHR) under the Programme Grants for Applied Research programme (RP-PG-0109-10074). The funder had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The first, second, third and last authors had full access to all the data in the study and all authors had final responsibility for the decision to submit for publication.
Conflicts of interest. The authors declare none.
Ethical standards. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. All procedures involving human participants were approved by the Black County NHS Research Ethics Committee (12/WM0097). Written informed consent was obtained from all participants. All names used are pseudonyms.