Therapeutic implications of counselling psychologists’ responses to client trauma: An interpretative phenomenological analysis

The past two decades have seen a surge of interest in the impact of working with trauma survivors on therapists’ psychological well-being. Existing literature assumes that therapists’ strong subjective responses to traumatic material adversely influence the therapeutic process. However, this has not yet been directly researched. Nine counselling psychologists were interviewed regarding the clinical impact of their responses to the disclosure of traumatic material. Interpretative Phenomenological Analysis showed that significant challenges are experienced in the use of self in therapy with trauma survivors, including making sense of horrific human actions, negotiating complex interpersonal dynamics and responding to ethical dilemmas in therapy. Results emphasised the importance that therapists attached to the development of their therapeutic use of self in therapy with trauma survivors, the value of learning from others and reaching a place of acceptance and hope when working with trauma survivors. Finally, specific training and development implications are proposed.


Introduction
Within counselling psychology, the reflective use of self in therapy and the therapists' appreciation of subjective and intersubjective factors are considered to be as important, if not more so, than the techniques that are employed. This is particularly the case in therapy with trauma survivors, where non-specific factors underpinned by the therapist's use of self are known to determine therapeutic outcome (Meichenbaum, 2013). Non-specific factors refer to aspects of therapy that are shared across modalities, including the therapeutic relationship, the therapist's skill and adherence to their modality. Solomon and Johnson's (2002) outcome review discusses the importance of non-specific factors in therapy with trauma survivors for establishing and maintaining trust and a good therapeutic relationship. The therapeutic relationship is thought to be particularly important to the therapeutic outcome for clients who have experienced interpersonal trauma (e.g. Cloitre, Stovall-McClough, Miranda, & Chemtob, 2004) and for those who have experienced extreme distress within a relationship (Briere & Scott, 2006).
A potential challenge to the therapeutic use of self in therapy with trauma survivors some researchers have observed is that therapists' strong subjective responses to traumatic material can have a detrimental impact on the therapeutic process (e.g. Neumann & Gamble, 1995). In this paper, we investigate the experiences of therapists working with clients who describe traumatic events, and how therapists consider their responses to the disclosure of traumatic material to impact upon the therapeutic process.

Impact of client trauma on the therapy relationship
Whilst the therapeutic relationship is thought to be central to therapeutic work, van der Kolk, McFarlance, and Weisaeth (1996) state that therapeutic relationships with trauma survivors can be highly complex in terms of potentially replaying aspects of damaging interpersonal dynamics that involve intense, previously avoided emotions that may be almost intolerable for both client and therapist (p. xvi). In addition, issues of moral complexity may arise for therapists when working with trauma survivors, for example, war veterans who have been traumatised by their own violent acts against civilians (e.g. McNally, 2010). A growing body of literature shows that therapists can develop forms of traumatisation variously termed compassion fatigue, secondary traumatic stress and vicarious trauma, involving intrusive and avoidant experiences, physiological arousal and feelings of helplessness and isolation (Sabin-Farrell & Turpin, 2003). These terms all describe the same fundamental concept; for the sake of consistency, this paper uses the term vicarious trauma.
In their review of the vicarious traumatisation literature, Sabin-Farrell and Turpin (2003) note that working with clients who describe traumatic events may elicit strong emotional, physical and behavioural responses in the therapist. This suggests that therapists' use of self may be more challenging in therapy with clients who describe traumatic experiences. Symptoms of posttraumatic stress disorder (PTSD) and psychological distress in therapists have been found to be associated with certain types of trauma work, including working with greater numbers of clients who have experienced interpersonal violence and abuse, particularly involving children (Creamer & Liddle, 2005;Cunningham, 2003) as well as working in disaster response teams (Holtz, Salama, Cardozo, & Gotway, 2002). Further, new and trainee therapists have been described as being particularly at risk of experiencing vicarious trauma symptoms in relation to their work (Adams & Riggs, 2008;Pearlman & MacIan, 1995). Neumann and Gamble (1995) describe Pearlman and McCann's (unpublished) survey of nearly 200 trauma therapists, which found that newer therapists reported more intrusive imagery from the work as well as more anxiety, depression and physical symptoms than did therapists with greater experience. This is consistent with research into the professional development of therapists, indicating that psychological distress is higher for therapists who are less experienced (Adams & Riggs, 2008;Hellman, Morrison, & Abramowitz, 1987;Rodolfa, Kraft, & Reilley, 1988).
It has been found that whilst skin conductance levels reduce in victims describing traumatic experiences, those listening become more aroused and their skin conductance levels become raised (Pennebaker, Barger, & Tiebout, 1989;Shortt & Pennebaker, 1992). In addition, findings indicate that working with a higher proportion of traumatised clients (Brady, Guy, Poelstra, & Brokaw, 2002) and cumulative exposure to traumatic material (Sabin-Farrell & Turpin, 2003) are associated with elevated symptoms of PTSD in therapists. Therapists' negative responses to traumatic imagery include frustration, shock, pain, anger, sadness and distress (Steed & Downing, 1998), as well as feeling horrified, angry, sad and nauseous during and after sessions (Iliffe & Steed, 2000). Adams, Boscarino, and Figley (2006) argue that in order to protect themselves psychologically, trauma workers may avoid empathic engagement with their clients. Pearlman and MacIan (1995) report that trauma therapists with greater experience demonstrate more disconnection from their inner experience and lower concern for others. They suggest that this may be a way of distancing themselves from the distressing pain of their clients. Arnold, Calhoun, Tedeschi, and Cann (2005) note that several clinicians in their sample discussed responses that seem to reflect a struggle with "empathic strain" or difficulty in maintaining empathic attunement as identified by Wilson and Lindy (1994). Despite the persuasive power of these arguments, the notion that so-called vicarious trauma impairs therapists' empathic responsiveness or negatively impacts on the therapeutic relationship has not been directly investigated.

Intersubjectivity and the therapy relationship
Intersubjectivity, literally meaning "between subjects", is an interdisciplinary field of study spanning many areas of psychology as well as neuroscience, social science and philosophy. Within psychoanalytic literature, a number of writers have discussed a shift from an emphasis on intrapsychic drives to a primary focus on relationality. Intersubjective approaches focus on emotional experience as regulated or misregulated within continuing relational systems (Stolorow, 2007). Indeed, Mitchell (1988) notes that we live in a "relational matrix": "The person is comprehensible only within this tapestry of relationships, past and present" (p. 3). Rasmussen (2005) argues that a limitation of the research into vicarious trauma has been the "linear thrust of the investigations" (p. 22), in that most studies have attempted to measure the impact of providing trauma therapy on the therapist. However, based on intersubjectivity theory, Rasmussen (2005) argues that throughout the therapeutic process, both client and therapist continuously influence and mould the responses of the other, both consciously and unconsciously.
Intersubjectivity theory requires a focal shift to research questions that explore the "reciprocal and dynamic interplay of subjectivities of therapist and client and the ways in which they interact to help or hinder the therapeutic process" (Rasmussen, 2005, p. 27), including the methodological and even philosophical challenges that such an approach may well entail. The goal of the current study was to investigate the experiences of therapists working with clients who disclose traumatic material, and how therapists' responses impact on the therapeutic process. Our primary research question was: How do counselling psychologists view the impact of their responses to descriptions of traumatic experiences on the therapeutic process? Related to this primary question, the following areas of interest are explored: (1) How do counselling psychologists make sense of their subjective responses to working with trauma in their therapeutic relationships? (2) In what ways, if any, do counselling psychologists view their subjective responses to impact on the therapy? (3) What, if anything, has influenced\helped\hindered the way in which counselling psychologists work with their subjective responses to trauma in their therapeutic relationships? (4) How do counselling psychologists view the way they work with their subjective responses to trauma in therapeutic work to have changed\developed over time; and what, if anything, might have contributed to this?

Method
Interpretative Phenomenological Analysis (IPA) was developed to explore unique situations and lived experiences (Smith, 2004), and in particular how people make sense of their experiences (Smith, Flowers, & Larkin, 2009). IPA is an idiographic approach concerned with detailed analysis of the particular, as opposed to a nomothetic approach, which pertains to the study of universal laws. Smith and Osborn (2003) describe the interpretative elements of IPA as a two-stage interpretation process; "the researcher is trying to make sense of the participants trying to make sense of their world" (p. 51). The researchers selected IPA as an appropriate method to allow for an in-depth exploration of therapists' subjective experiences in response to clients' descriptions of traumatic events, and how they understand its impact on the therapeutic process.

Participants
In order that areas of convergence and divergence around the themes in participants' accounts can be explored in detail, a relatively homogeneous sample is required. Therefore, participants were recruited for the study from the British Psychological Society (BPS) Division of Counselling Psychology (DCoP). Since counselling psychology is broadly underpinned by a humanistic value base, and the therapists' appreciation of subjective and intersubjective factors is highly valued, it was anticipated that this would generate a level of homogeneity of perspective across the participants' accounts, such that variation within the sample could be explored in detail.
Advertisements for participation were circulated via the DCoP email list. Participants were also contacted directly through the DCoP practitioner database, using "interest in trauma" as the search criterion. In addition, the advertisement was sent to colleagues and supervisors working in trauma settings, to pass on to qualified counselling psychologists. Participants were required to have experience of working with clients who had experienced trauma, usually involving one or more of the following: sexual or physical abuse, experiences of military combat, terrorism, mass violence, natural disasters and accidents (Zimering, Munroe, & Gulliver, 2003).
Nine participants were recruited in total; two men and seven women. All participants had more than one year post-qualification experience of therapeutic work with trauma survivors. Participants were asked to describe their therapeutic orientation within the interview schedule. Three participants described themselves as working primarily from a CBT approach, two psychodynamic, one humanistic and three integrative. Participants had worked with trauma survivors in prisons, NHS outpatient and inpatient facilities, independent practice private clinics, voluntary sector therapy services and overseas outpatient clinics.
We have eliminated or disguised all identifying information to ensure anonymity and confidentiality.

Data collection
Semi-structured interviews of approximately one-hour duration were carried out by the first author in order to explore participants' understandings of their responses to descriptions of traumatic events, and how their responses impact on the therapeutic process. At the time of conducting the semi-structured interviews, the primary researcher was in the process of completing her doctorate in Counselling Psychology and had experience working with trauma survivors. Smith and Osborn (2003) argue that semi-structured interviews are generally the most successful way to collect data for qualitative studies in psychology, as the researcher's questions can be reconsidered and adapted through engaging with participant's ideas, and the researcher can spontaneously respond to interesting ideas that come up. A consent form and information sheet about the research were given to the participants to complete. These included information about how data would be used and steps to maintain confidentiality. The interview schedule was constructed to investigate how counselling psychologists view the impact of their responses to descriptions of traumatic experiences on the therapeutic process. The interview schedule (see Appendix 1) consisted of approximately eight questions that focussed on participants' experiences of trauma work and the therapeutic process. These questions were developed by the first author and research supervisor and were directly related to the research questions identified in the introduction. The following areas were included in the interview schedule: experiences of working therapeutically with trauma survivors, perspectives on how personal responses to trauma narratives may have impacted the therapeutic process, factors that helped\hindered the way in which counselling psychologists worked with their subjective responses to trauma in their therapeutic relationships and therapeutic orientations and therapy settings.
Questions were carefully worded to focus on participants' experiences and the interviewer also ensured any additional follow-up questions were focused on participants' experiences. The schedule was used flexibly in order to reflect areas that participants regarded as being important. Participants were invited to reflect upon and discuss therapeutic work across their client work rather than to discuss one specific client.
Interviews were tape-recorded and transcribed verbatim.

Data analysis
The first author conducted the analysis and the steps identified by Smith (2004) for IPA were used as guidelines. After an initial reading of the transcript, an overall summary impression was noted. Then the transcript was re-read; this time notes and reflections were made in the left-hand column of the transcript. On a third reading, themes were identified and noted in the right-hand column. These themes were created into a table of constituent themes based on the notations in the right-hand column, and then checked against the participant's words to ensure they were grounded in his or her own account. Smith, Jarman, and Osborn (1999) state that the table of themes should "capture most strongly the respondent's concerns on this particular topic" (p. 223), so this was used as a criterion to judge what was selected as constituent themes. The constituent themes were then clustered together, and those themes that appeared to the researcher to be central to the participant's experience were labelled as master themes. These master themes were checked against the original transcript to reflect on how they operated in the participant's account as a whole. In terms of the master list, themes were selected for their "richness" and their ability to "illuminate other aspects of the account" (Smith et al., 1999, p. 226). Another table containing these master themes and constituent themes was created. This became the basis for a development of the interpretation of the themes. The same process was carried out for all the transcripts. Following the analysis of all transcripts, areas of convergence and divergence between cases were explored in detail. This involved forming a "picture board" of connected themes for the group as a whole, and a final table of super-ordinate themes, together with their constituent themes (Smith et al., 2009).
To determine the prevalence of a theme, and thus the representativeness of subthemes and superordinate themes for the group as a whole, the researcher drew on the criteria provided by Smith (2011). The subtheme needed to be represented across a sufficient number of participants' accounts: according to Smith (2011), this would usually be half. In a sample of nine participants, that led us to set a minimum of four participants for each subtheme to ensure that all substantial themes were represented. In selecting extracts to support the theme, a minimum of three extracts from different participants was chosen, though usually more than this was selected in order to demonstrate the convergence and divergence within the theme (Smith, 2011). Extracts were selected on the basis of what the researcher judged to be most representative of a particular theme, and also extracts that provided the richest interpretative capacity to demonstrate both the variation and depth of the theme.
Credibility of analysis was sought in three ways. First, the research supervisor audited the analysis of one interview, which involved reading the transcript, evaluating the credibility of the reflections and themes and providing a forum for the discussion of emerging subthemes and superordinate themes. Second, the research was presented at a specialist group for IPA research within the University psychology department, where group members gave feedback on the analysis which helped to develop the richness of interpretations. Third, a peer-researcher conducted an analysis of one of the transcripts for comparison.

Results
Analysis of the nine interviews resulted in the identification of two superordinate themes with six subthemes (see Table 1).
Superordinate theme 1: Demands and challenges in the use of self in response to trauma All participants described the difficulty of bearing the pain and horror aroused in them, and for some participants, this was experienced as traumatising in response to their Subtheme 1a: The difficulty of hearing painful trauma histories All participants described the difficulty of experiencing painful emotions in response to traumatic stories. When Beth conveyed some particularly sadistic details of abuse, she noted the difficulty of putting her feelings into words: Those feelings of, how can I put it, of kind of horror.
The words "kind of" suggest that "horror" is the best approximation but perhaps doesn't, and never could, capture the actual experience. In addition, all participants expressed their sense of the emotionally demanding nature of working with trauma survivors as well as times of feeling "traumatised" by what they had heard. Lydia described a confusing physical synchronicity between herself and a client: My mouth was also going very dry, which is interesting. She said she had a dry mouth and wanted to drink, she was actually frothing at the mouth, so whether that was related or just down to my own anxiety? [] She just wanted to regurgitate all of this trauma that she'd witnessed [] In this extract, there is a sense of the trauma as highly indigestible, and at the same time of experiencing physical sensations and wondering how to make sense of this. Beth told me she found herself becoming distrustful of people in general. She linked this to working in a prison setting and a breakdown in the usual sense of boundaries between offenders and non-offenders: When you are in a prison environment, you realise this is your next-door neighbour, this is your Dad, this is your brother, this is the man down the street. And you start looking out for things.
In this extract, there is a sense of the loss of the illusion of separateness or "otherness" of those who carry out acts of violence and abuse.
Subtheme 1b: Negotiating complex interpersonal dynamics Seven of the participants described challenging dynamics in their relationships with clients who had experienced traumatic events. For example, Elisa said: I felt absolutely terrible, as though I'd totally betrayed her by allowing it to happen, as though this was down to me [laughs]. And I remember apologising to her and saying "I'm sorry this has happened".
Other interpersonal difficulties described included experiencing the self being drawn in to the relationship in non-therapeutic ways. Serra described being challenged by a relationship dynamic relating to a client's childhood where being abused was associated with being special: It was as if she was, with her, it never felt the talking was that therapeutic, it just felt like a bit of a test, a bit of manipulation.
The notably tentative, uncertain language here seems to convey the highly complex nature of this interaction. Where the client's childhood abuse was associated with being special, it appears that Serra perceived that the therapeutic relationship was being tested; perhaps to test out whether this relationship would in fact be safe. What appeared to be most problematic for Serra was that something was happening but was not made sense of in the therapeutic communication. Elisa described her sense of concern about discussing her client's sexuality, which was also striking in its complexity: I think [it was] probably to do with this communication, that's not all at the verbal level, and how exposing that would be of me and my own sexual inclination, and how crossed over that could get, possibly[]. That you're not quite sure where you finish and they start, so it feels a bit too unboundaried a topic to get into.
As in Serra's description, there is a sense here of the murkiness around trying to understand the interpersonal dynamics and as a result, Elisa delineated clear boundaries in terms of the areas she is prepared to explore in therapy. Lucy talked about the danger of communicating her internal responses when working with trauma: The healers really have the capacity to harm even more than the original perpetrators do, I'm very careful of giving them feedback on my feelings.
On the other hand, Michail emphasised the importance of the therapist having an emotional response to a client's traumatic narrative: I do not think that being there as a dead face in response to a traumatic account is useful at all. Because if you're not seen to be touched by the experience, then you become an onlooker. And of course with traumatic events, especially if they're sexual, with traumatic events, part of the trauma is that there are onlookers who refuse to intervene.
In this account, Michail argues that it is important that the therapist be "seen to be touched". He argues that the importance of the therapist being emotionally moved is a way of providing a different experience to the original trauma where onlookers remained untouched and did nothing. On the other hand, Timothy noted that as a male therapist working with a female client with a history of violent and abusive relationships with men, there was an opportunity to provide a new kind of experience: Working with a male therapist I think she was a bit unsure at first. But I think [] she trusted me, and maybe that's helped her see men in a different light.
Similarly, Serra described that the potential of the therapeutic relationship to offer the possibility of trust is of vital importance with trauma survivors "[]because when you're that traumatised, it's quite difficult to rebuild your trust in people". Indeed, this relational component is of central importance to Serra in terms of her professional identity as a counselling psychologist, where "the most important thing for us is the relationship". Beth conveys a sense of her humanising recognition of difference in this account, because she directly names her client's disability in the context of the client's trauma history, rather than skirting around the issue. Lucy articulated the importance of recognising the implications of trauma in terms of social roles: Lucy argues that the thoughtful use of language has an important part to play in terms of the client's relationship with themselves. Michail extended this idea to the notion of the importance of healing at a social level, as well as within individual therapy: With victims of abuse, therapy is useful but it's also useful when the person gets a prison sentence, [when] there's some visible sign that their suffering is recognised. I think we need to think of the social aspect of trauma, the larger relational context in which trauma happens, is allowed to happen.
This radically locates trauma within a wider social context where the traumatic experience is "allowed to happen", and the notion of a "visible sign" communicates the importance of the social recognition of trauma. It appears that Rachel prizes her perspective on the commonality of human distress, emphasising her healing role and separating herself as a practitioner from the wider social context and political meanings to allow this to happen. Timothy described taking a similar position in working with both "victims" and "perpetrators" of abuse/trauma: It's one of the things that happens for whatever reason that later on, they become people that do it themselves. []So I suppose it's just me accepting that these people I'm working with, this is what's going to happen, and rather than giving people labels, treating them as individuals.
This quote conveys the perception of a clear pattern that those who have been abused go on to abuse others. This aids Timothy in "treating them as individuals", which allows him to identify with a healing role. A number of participants indicated that working through moral dilemmas raised issues that they grappled with in a complex way. In contrast to Timothy, Beth experienced conflicting thoughts and feelings in working with people who have been abused and then go on to abuse others: They blame the fact that they've been abused for the reason why they're abusers. But actually there are millions of people out there who've been abused and don't go on to abuse themselves, does that make sense? So it's no excuse really.

However, moments later Beth stated:
You have to put their offence to one side and just concentrate on their trauma [] we had to work within coming to accept the crime he'd done in order for us to, if you like, lower the PTSD symptoms and everything.
There is a sense of the effort it takes for Beth to try to put their offence aside and she conveys the idea that accepting the crime is therapeutically healing, and therefore that this is something she "had" to do as a counselling psychologist. Michail described taking a different position, arguing for the importance of highlighting power relations: I'm very interested as well about power relations. How would this person think it's OK to lash out in that way, or talk to you in that way? How's that organised in the family or in the system? This reflection on trauma within a context of power may offer the client the therapeutic opportunity to (re)consider their position within their social context.

Superordinate theme 2: Developing the therapeutic self in response to trauma
Participants conveyed the way in which their therapeutic selves were developed over time, through formative training and therapeutic experiences. This was presented as a challenging journey, with significant changes for some in terms of their sense of self and/or the world. Participants communicated a sense of development in terms of their ability to respond therapeutically to traumatic material, and this was brought about by reaching a point of some acceptance.

Subtheme 2a: Learning from others
In all of the participants' accounts, there were a number of ways in which peers and supervisors were involved in the development of participants' use of self when working with traumatic material. Serra highlighted that the development of her emotional strength to work with trauma survivors was gained through experience: I think you need to have a level of emotional strength to deal with it. That doesn't come with training; that comes more with experience and learning from others' experiences … successful therapists and through supervision.

Michail described:
It's not useful to tell the client "god, I wish your father was run over by a truck", or something [] So that needs to be managed somehow. [] It's managed through supervision, and it's managed because I guess eventually you become a bit more confident that you can manage.
In this extract, Michail highlights the value for him of supervision and experience in being able to manage intense emotions, rather than the emotions leaking into the therapy in an unhelpful way. Rachel's supervisor also had an important role in her development; because of her experience of working in an environment that was so violent, she described her supervisor observing that: [I was] most probably, dissociated from the emotional response, in order to stay strong. So, she said, right, could you bring the human back a bit? And it took some years, and I said to her, fair enough, but at the same time still being able to feel not flooded by the information.
Thus, in Rachel's account, there is a sense of negotiation in terms of being "human" in one's emotional responsiveness and yet still maintaining some distance in order not to feel "flooded". In this way there is a sense of an ever-present threat of her self becoming overwhelmed by traumatic material.
Subtheme 2b: Reaching a place of acceptance and hope Through the emotional difficulties of engaging with trauma survivors, five participants described reaching a point of acceptance in themselves in relation to traumatic material.
No matter what people say to me now, the chances are I've heard something similar, so the impact's already been and it's settled somewhere, really. So my own responses, I know them [] because, I suppose you work it out. You have to find a way to resolve that in yourself.
In this extract, Nadia conveys her experience of personal development in spatial and physical rather than temporal terms; that with experience, the impact of hearing about traumatic experiences has "settl[ed] somewhere". The notion of finding a way to "resolve that in yourself" indicates a sense of inner conflict that previously existed. Perhaps what was unacceptable to the self needed to be internalised as part of the therapeutic work. Beth also described a position of acceptance, and for her this is centred around her capacity to put trauma into context: It's about [] putting it into context. [] It's either you allow it to keep contaminating you and affect your behaviour and your life, or you accept that actually you need to kind of, it's hard to explain really. It's about seeing things for what they are, rather than reading into it too much.
Beth conveys her sense that this position is reached through both necessity and choice: that either one continues to allow oneself to be contaminated, or one chooses to accept "things for what they are". Rachel, through her training and experience, valued having reached an internal position of calmness and strength: There was like a calmness and peacefulness, knowing that you're with the person, but by the end of the therapeutic process, they would have calmness and the strongness.
In this extract, it appears that Rachel's sense of inner calm is brought about by her confidence in the therapeutic process and her belief that the client will develop these qualities themselves.

Discussion
Our analyses yielded two main findings. Firstly, participants described a number of struggles in terms of their use of self in response to traumatic material. Secondly, participants conveyed the way in which their therapeutic selves were developed over time, through formative training and therapeutic experiences.
In terms of making links between the different aspects of participants' accounts, Benjamin's (1990Benjamin's ( , 1999Benjamin's ( , 2000 work on the intersubjective concept of "mutual recognition" offers an illuminating framework, in that it allows for meaningful links to be made between participants' accounts. Recognition refers to the capacity to recognise the otherness of the other, in other words to uphold and cherish the difference of the other. According to Benjamin (1990) the development of our subjectivity, our self-understanding and our self-acceptance is dependent on the experience of recognition and mutual impact with care-givers in early life. Building on this, Pollock and Slavin (1998) state that with early interpersonal trauma the abused person is treated as an object whose needs are not valuable or important, which can impair the development of their agency. Benjamin's concept of recognition may be valuable to make sense of the current results and to think about why the participants considered their use of self in work with trauma survivors to be so vitally important and at the same time so highly challenging.
For instance, many of the participants tried to limit the extent to which the client was aware of the therapist's feelings and subjective responses. Indeed, the current findings indicate something of a taboo in relation to therapist self-disclosure, and particularly the sharing of strong feelings. However, a minority of participants argued for the value of therapeutic disclosure in order to acknowledge the "understandability" of clients' responses, and also to resist the position of a "neutral bystander", which might replay features of the original traumatic situation. Nevertheless, even participants who argued for the value of disclosure were highly cautious about how they used it, and were very wary of how it might impact on the therapy.
The literature on recognition (e.g. Benjamin, 1990), indicates that it is precisely the experience of having an impact on the other that can contribute to a sense of agency for the client. Paradoxically, whilst there is powerful therapeutic potential in the therapist sharing the impact, the client has on them, for more fragile clients, this could be experienced as highly threatening, particularly when they have experienced relationships in which they were overwhelmed by the impact of the abuser. This theory may make some sense of the ambivalence with which participants approached this issue of sharing the mutually impacting process of therapy and disclosure of the therapist's response. Previous research has not explored the issue of this disclosure in relation to working with traumatic material specifically, and the current study provides some initial insights on the issue and also highlights the need for more research in this area.
All of the participants described negative emotional responses to traumatic stories, including feelings of horror, anger, sadness and fear, and feeling overwhelmed by these emotions. Participants also described looking out for potential dangers and being mistrustful of other people, and experiencing greater distress themselves. This corresponds to McCann and Pearlman's (1990) concept of vicarious trauma whereby in therapeutic work with traumatic material, the therapist may become more aware of the experiences of powerlessness, lack of safety and betrayals of trust that people may experience, resulting in distress for the therapist. Further, participants described that when working with high numbers of people with trauma histories, they could become desensitised and might use defensive practices such as "switching off" in order to maintain their well-being. Thus, participants discussed the importance of limiting the amount of trauma in their caseload, as well as the amount of work with any one type of trauma, in order to prevent "saturation".
It was striking that such different positions were taken by participants in terms of how they described their moral stance. Some argued for taking a position of moral neutrality, whilst others argued for the importance, and value, of taking up a moral stance in relation to trauma. Lucy emphasised the role of language (e.g. "trauma survivors") in order to offer clients a more empowered perspective on their situation. Michail took this further, stating that it is not possible to take a neutral position, and he described focusing on power relations as a therapeutic aspect of his work to offer clients the opportunity to think about the way that power operates in their social contexts. Herman (2001) and De Zulueta (2006) strongly advocate for the therapist standing in moral solidarity with the client, and they link this with the importance of social recognition of the impact of trauma. However, as McNally (2010) observes, trauma does not always have clear victim and perpetrator distinctions, and participants in the current research indicated their difficulties in relation to taking a moral stance when this was the case. For example, Beth described a conflict at times between her healing role and her own moral compass in working with a client who had been abused and gone on to abuse others. Cooper (2009) states that "for many counselling psychologists, the essence of our profession … is that it is embedded in a particular set of values and ethics … counselling psychology is 'ethics-in-action'" (p. 120). However, the tradition within psychology of moral neutrality on the part of the therapist has arguably undermined due consideration of the moral stance of the therapist. Reinkraut (2008) argues that the "silence on matters relating to moral responsibility, justice and injustice tacitly collude [s] with an ahistorical, acontextual framing of individual suffering" (p. 8), and this results in a lack of consideration of social factors in therapy. Thus, the therapeutic use of self should explicitly include one's moral sensibility. However, the current study highlights the considerable complexity raised by these issues and the struggles of the participants in working with them.
The issue of one's moral position as a therapist was clearly important to participants, and it is interesting that it is not given greater prominence in the training, supervision and personal development of counselling psychologists. It was striking that participants appeared to be grappling with these issues entirely in isolation. Compared to other issues, none of the participants described discussing ethics in supervision, training, personal therapy or with colleagues. Perhaps, if this is the case, it may make some sense of the vastly different positions taken by participants in terms of an absence of collaborative thinking upon which to draw.
Nevertheless, perhaps a commonality in the participants' approaches was the attempt to offer their clients an experience of recognition. In the participants' accounts, thinking with clients about their experiences in terms of the social context and meanings included how the wider social context may allow trauma to happen, naming acts of violence and destruction as such, delineating how power works within the client's social context, and considering the implications of disability for ongoing experiences of abuse. In this way, clients were not treated in an acontextual way, but rather in a way that recognised their experience in meaningful social terms.
Participants described a number of difficult interpersonal dynamics with clients who had experienced traumatic events such as the relationship pulling them to relate in a particular way that they felt was non-therapeutic. In these relationships, there was considerable fear about what might be unintentionally communicated at the non-verbal level in the therapeutic relationship. Participants' experiences appear to relate to the theory on "enactments", where therapist and client experience themselves becoming stuck in emotional positions (Ginot, 2009). For example, participants described their concerns that their communications would be viewed as seductive. Serra understood this to be because the client's early experiences of relationship intimacy had been sexualised, and she described her fear that the communication between them was "untherapeutic". Participants appeared to be struggling with how to make themselves "recognisable" to the client, that their actions were being misconstrued by the client, and that the participants' sense of agency as a therapist was compromised. According to Benjamin (2000), where a person's agency, self-understanding and self-acceptance has been impaired, the capacity to recognise the other is also impaired. Ginot (2009) argues that whilst enactments inevitably involve the blurring of self/other boundaries, they are a valuable opportunity to gain an "unmediated" experience with what the client cannot yet verbalise. Indeed, the experiences of misrecognition and the lack of agency experienced by the participants appear to correlate with the early experiences of their clients. In this way, the experience perhaps provides the participants with direct access to relational patterns experienced by the client in earlier life. Ginot (2009) states that by becoming aware of and reflecting on this experience with the client, the therapist offers the client the opportunity to recognise and even integrate what the client could previously only enact. However, this is not an easy task. Certainly, in the current study, these relational dynamics were experienced as extremely difficult to make sense of and to work with, and therapists addressed them by becoming highly boundaried and avoiding discussions about some topics that were considered problematic. The current study thus highlights the considerable demands posed by enactments for the therapist, and the participants' lack of confidence in their capacity to work therapeutically with them. This area therefore warrants considerably more discussion and attention in therapeutic literature, and research to inform training and supervision practices that may help therapists in working in this complex area.
On the other hand, healing relationship dynamics were also viewed as a valuable part of the therapy. Examples of these included building up a sense of trust within the relationship where before, people were viewed as untrustworthy and, for female clients, the therapist providing an opportunity to view relationships with men in a different light through a new experience of a relationship with a male therapist.
A significant contribution to the development of the self was experienced in supervision. The value of supervision, for most participants, was less for the technical input and more for an experience characterised by recognition. The participants described the need to be witnessed, to share the experience, to be "held", and for supervisors to communicate their belief in participants' abilities. This supports existing theoretical literature (e.g. Hawkins, 2005;Herman, 2001) that highlights the value and importance of supportive contexts for therapists when working with trauma survivors.
In terms of participants' descriptions of their use of self, there is an overall sense of a developmental progression towards being able to "take in" traumatic material and then be able to think about it with the client, without experiencing this as unduly "contaminating", and hence needing to defend the self from this. This is consistent with research into the professional development of therapists, indicating that psychological distress decreases as therapists become more experienced (Adams & Riggs, 2008;Hellman et al., 1987;Rodolfa et al., 1988). Participants conveyed that through their work with trauma survivors, they had reached a position of acceptance in relation to traumatic occurrences. For one participant reaching acceptance was described as having a place within herself where she could hold dreadful things. Here, acceptance was presented as an embodied experience rather than an abstract idea. The position of acceptance was associated with the ability to bring a sense of hopefulness into their work with clients.

Methodological considerations and research limitations
It is important to be tentative in considering implications from qualitative studies, which involve small samples. We recognise that different researchers might have focused on other aspects of the participant's experiences. Further, the analysis does not cover all aspects of the participants' accounts, and focuses on material that is relevant to the research question. Finally, it is acknowledged that different researchers may have made different interpretations of the interview transcripts. Nevertheless, the reader may be able to draw on the current study in terms of theoretical generalisations and applicability to their own professional practice and understandings (Smith et al., 2009). For both logistical and ethical reasons, it was not possible to interview both clients and therapists. Thus, it is important to recognise that the participants' accounts do not represent the whole picture in terms of the therapeutic interactions they describe. The emerging literature on listening to client and user perspectives in the psychological therapies (e.g. Bates, 2005;House, 2003) perhaps means that in future, new, ethically sensitive methodological developments will enable the client voice to be heard more strongly when considering the kinds of issues being researched in this study.
A limitation of the study is that it was not possible to know how our findings might be influenced by the different backgrounds of each member of our sample. It would have been helpful to have collected additional demographic information to contextualise the findings within the range of experiences of the sample, for example, in relation to their use of supervision and years of experience. We would encourage new research with therapists to explore whether some themes may be more representative of therapists with particular backgrounds of experience and training. A further limitation is that although the sample was homogenous in the sense that they were all counselling psychologists, their theoretical orientations did differ. It may be that there are interesting differences between counselling psychologists' experiences depending on their orientation that we were not able to detect, and that further qualitative research could focus on groups of therapists that are similar in orientation.

Implications and suggestions for future research
The current study indicates that therapists vary in the type and amount of trauma histories they feel able to work with, and that participants valued very highly the ability to use their own judgement about their capacities and limitations. Some therapists may feel sufficiently supported within supervision, but the current study indicates the importance of therapists having access to additional support, such as peer supervision as one participant described, when working with high numbers of trauma survivors. Further qualitative research specifically investigating the role of supervision in working with trauma survivors would be helpful to inform practice in this area. In addition, future research could usefully explore the area of clinician experience and investigate the impact of number of years of clinical experience on the therapeutic use of self in therapy with trauma survivors.
The study highlights the considerable demands placed on therapists' use of self in therapy with trauma survivors, including making sense of horrific human actions and negotiating complex interpersonal dynamics. It therefore appears that this area warrants considerably more discussion and attention in the therapeutic literature, and research, to inform training and supervision practices that may, in turn, help therapists in working in this complex area. Further, the current study highlights the considerable complexity raised by moral and ethical issues and the lack of a forum available to participants to engage with these issues. One's moral position as a therapist was clearly important to participants, and this highlights the value of giving the development of one's "ethical sensibility", if it may be described in this way, much greater prominence in the training, supervision and personal development of counselling psychologists. In training this could involve discussion, debate, role-play and so on, the aim of which would be to give trainees a space to think through and develop their positions in relation to complex moral issues, such as the participants in the current study have described, and how they might respond to such issues.

Conclusions
The study is the first qualitative investigation that we are aware of that explores the experiences of therapists working with clients who disclose traumatic material, and how therapists view their responses to impact on the therapeutic process. It has provided rich, in-depth knowledge about the experiences of counselling psychologists working with trauma. Findings indicate the emotional and interpersonal demands posed by therapy with trauma survivors, as well as the challenge for therapists in responding to complex ethical dilemmas. These may be areas of therapy with trauma survivors where further input and support through supervision and training would be of value. Additionally, the research emphasises the particular importance that therapists attach to the development of their therapeutic use of self in therapy with trauma survivors, and the value of learning from others when working with trauma.

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributors
Olivia Merriman is a chartered psychologist (BPS), an HCPC-registered counselling psychologist and a BABCP accredited CBT therapist. She has a particular interest in working with trauma and is a trained EMDR practitioner. Following her counselling psychology doctorate at Roehampton University, she completed the postgraduate diploma in CBT at Kings College London. Olivia is currently the Improving Access lead for Wandsworth Improving Access to Psychological Therapies service.
Stephen Joseph is a professor of Psychology, Health & Social Care in the School of Education at the University of Nottingham where he is convenor of the counselling and psychotherapy cluster. He is a senior practitioner member of the British Psychological Society's Register of Psychologists specialising in Psychotherapy and an HCPC-registered counselling psychologist. Stephen's research interests are in positive psychological applications. He has published over 200 academic peer-reviewed papers, book chapters and books on trauma, posttraumatic growth and related topics in counselling and psychotherapy including What Doesn't Kill Us: The new psychology of posttraumatic growth (Basic books, 2011). He is also editor of the handbook Positive Psychology in Practice (Wiley, 2015). His latest book is Authentic. How to be yourself and why it matters (Basic Books, 2011).
(4) In what ways, if any, do you think your response, overt or "internal", to a specific client's description of traumatic experiences might have had an impact on the therapy? (5) Prompt: Your answer can encompass both conscious and unconscious levels of experience, if relevant/appropriate. (6) How would you describe the therapeutic relationship with this client\these clients changing over time? (7) What do you think was your most positive experience in working with clients who disclosed trauma and how do you make sense of this? (8) What do you think was your most negative experience in working with clients who disclosed trauma and how do you account for this? (9) Is there anything that has helped or hindered how you work with your subjective responses in therapy with trauma? (10) Prompts: Training? Supervision? (11) In what ways, if any, do you consider the way you work with your subjective responses in therapy with trauma to have changed over time? (12) Prompt: What if anything, has contributed to this? (13) Are there any other comments you'd like to make in relation to the research theme?