The Meaning of Voices in Understanding and Treating Psychosis: Moving Towards Intervention Informed by Collaborative Formulation

From a medical perspective, hearing voices is perceived as a symptom of mental illness and their content as largely irrelevant. The effectiveness of antipsychotic medication has made it central to the treatment of psychosis. However pharmacological treatment alone is rarely sufficient for this disabling condition. This review examined the feasibility of formulating an understanding of the meaning of voices in psychosis to inform intervention. Examination of the literature demonstrated the need for a paradigm shift to a recovery model, drawing on biopsychosocial factors in formulating an understanding of the meaning of voices in the context of a person’s life. Providing the opportunity to talk about their experiences may aid the development of an interpersonally coherent narrative representing opportunities for psychological growth. Findings have implications for treatment planning and assessment of outcome. Collaborative formulation regarding the subjective meaning of voices may aid in understanding their development and maintenance and guide intervention. Hearing voices with reduced negative effects on wellbeing and functioning may reduce distress and improve quality of life even in the presence of voices. CFT, CBT, Relating Therapy and Open Dialogue may be effective in applying these principles. Findings are limited by the lack of controlled studies. Further controlled studies and qualitative explorations of individual experiences are recommended.

The National Institute for Health Care Excellence (NICE, 2014) recommends that all individuals diagnosed with psychosis ought to be offered a form of talking therapy. Nevertheless, the majority of people with psychosis are not currently given the opportunity to access such interventions (Jolley et al., 2012). The medical model categorises psychosis as an illness requiring medication. The Division of Clinical Psychology (DCP, 2013) state that while antipsychotic medication has been found beneficial in the treatment of psychosis, its limitations in managing psychotic phenomena have increased calls for a paradigm shift that contextualizes human distress (DCP, 2013). A formulation approach, allows exploration of the meaning to the individual of the experiences, relationships and social circumstances of their life in shaping their current difficulties (BPS, 2014). Within this approach, the individual experiencing distress works collaboratively with the clinician to develop a shared understanding of their difficulties and develop an intervention plan (DCP, 2013).
A recovery approach adds to the medical model in that it's goal is to assist a person in their personal growth by building their self-esteem, sense of identify and developing a meaningful purpose in society (Allott & Loganathan, 2003). The recovery model suggests that while the condition is not 'cured', the recovery process can continue in the context of ongoing symptoms (Roberts & Wolfson, 2004). The move from a model which views voice hearing as a medical symptom alone, to an awareness of voice hearing as a meaningful experience, shaped by biopsychosocial factors, which may guide personal growth and recovery has garnered support from many who hear voices, their families and clinicians (Corstens, Longden, McCarthy-Jones, Waddingham, & Thomas 2014). In this regard, a recovery model may add to the medical approach by helping direct the most suitable form of intervention for each person. Therefore, this paper will review the feasibility of formulating an understanding of the meaning of voices in psychosis to inform choice of intervention and improve distress and quality of life.

Medical Model
Medical approaches to psychosis argue that conceptualising psychosis as a mental illness removes stigma from the individual (World Psychiatric Association, 2001), facilitates a shared language among professionals and promotes access to resources, treatment and care (Campbell, 2007;Pitt, Kilbride, Nothard, Welford, & Morrison, 2007). This illness conceptualization can provide a way of communicating about groups of people and can make people feel less alone in what they are experiencing (Pitt et al., 2007).
However, a diagnosis can have detrimental consequences for the individual's wellbeing, such as feelings of hopelessness and decreased confidence (BPS, 2014). Diagnosis may shift focus away from the potential meaning that voice hearing could possess for the individual and from possible triggering social-emotional difficulties that could be managed in a reparative manner (Cooke, 2008). The widespread acceptance of the concept that diagnostic labels, like schizophrenia are biological illnesses has meant that it is frequently believed that voice hearing always results from a brain based difficulty (BPS, 2014). This has contributed to a culture where the primary treatment tends to be medication alone (BPS, 2014).
The results of randomized controlled trials (RCTs) indicate that the use of antipsychotic medication during the initial incidence of psychosis can be effective in preventing the development of structural changes to the brain that are linked with psychosis (Lieberman et al., 2005;Nakamura et al., 2007). Furthermore, the common use of antipsychotic medication has reduced what can be the more impairing presentations of schizophrenia e.g. catatonia and highly disorganised symptoms (van der Heijden et al., 2005). This highlights the important role of medication in reducing the most severe impairments frequently associated with psychosis (van der Heijden et al., 2005).
However, there is an increasing acknowledgement that pharmacological treatment alone is rarely sufficient for this disabling condition (Pilling et al., 2002), with factors such as compliance, limited effectiveness of medication and side effects limiting the benefits of antipsychotic medication alone (Pilling et al., 2002). There is evidence that psychological interventions such as Cognitive Behaviour Therapy (CBT), in combination with medication can be effective in the management of psychosis (NICE, 2014). However, there is minimal evidence for CBT alone, without anti-psychotic medication (NICE, 2014). Developing different approaches for the treatment of psychosis is crucial frequent number of individuals who may; decline to take anti-psychotic medication, experience no benefits from it, or encounter negative side effects (NICE, 2014).
Mental health diagnoses such as psychosis can be critiqued for being ways of categorising experiences as a medical disorder, rather than explaining them or exploring where they fall on a continuum of normality (BPS, 2014). As with other psychological problems, experiences like hearing voices occur on a continuum where normality mixes with abnormality (David, 2010). For example, many people can occasionally experience voices during periods of high stress, while for other people they can be more extreme and upsetting (BPS, 2014).
Hearing voices is often associated with intense emotions and feelings e.g. anxiety, low mood or by difficult experiences (BPS, 2014). This highlights the importance of understanding the meaning of voices in the context of individual's life stress and underlying biological vulnerability.

Biopsychosocial Model
It is well established that stressful life experiences, such as childhood maltreatment increases the risk of developing psychosis (Varese et al., 2012). However, this is also the case for other psychiatric disorders (Barker, Gumley, Schwannauer, & Lawrie, 2015). Currently, the causes of a person's susceptibility to different psychiatric difficulties is uncertain due to comparable environments, however, a probable contributing factor is a genetic predisposition (Barker et al., 2015). Hypothalamic pituitary adrenal (HPA) axis activation takes place in reaction to stressors in the environment and has been associated with the progression from chronic stress to a variety of psychiatric disorders (Barker et al., 2015). Research indicates that people with schizophrenia possess a different HPA stress reaction and HPA axis hyperactivity may facilitate the connection between chronic stress and psychosis (Walker, Mittal, & Tessner, 2008). The stress-diathesis model proposes that it is the interaction between biological vulnerability and influences in the environment that result in a person's difficulties (Zubin & Spring, 1977). However, currently a more cohesive biopsychosocial framework is required in order to shed light on how these interacting factors develop into the presentation of psychosis (Barker et al., 2015).
As Auditory Voice Hallucination's (AVH) often persist when antipsychotic medication is administered (Shergill, Murray, & McGuire, 1998)  can impact on the quality of relationships with other people, mediated by psychological and brain adaptations to the original stressors (Read, Bentall, & Fosse, 2009). In a study of British children raised in economic deprivation, individuals with no family history of psychosis were found to be seven times more likely to develop schizophrenia (Harrison, Gunnell, Glazebrook, Page, & Kwiecinski, 2001). Other social risk factors which have been identified for psychosis are; maternal health, and stress during pregnancy, being an unwanted pregnancy, early loss of a parent, childhood physical, sexual and emotional abuse, childhood neglect, and trauma (Bentall & Fernyhough, 2008;Conus, Cotton, Schimmelmann, McGorry, & Lambert, 2010;Janssen et al., 2003;Larkin & Morrison, 2006;Read, Fink, Rudegeair, Felitti, & Whitefield, 2008;Welham, Isohanni, Jones, & McGrath, 2009).
Evidence suggests it is crucial for professionals to understand the meaning individuals attribute to their voices if they are to help alleviate distress (Lakeman, 2001) and promote a process of recovery (Royal College of Psychiatrists, Social Care Institute for Excellence and Care Services Improvement Partnership, 2007). One study found that mental health practitioners commonly felt that their clinical work may be adversely impacted if they gave attention to the content of psychotic phenomenon (Aschebrock, Gavey, McCreanor, & Tippett, 2003).
Understanding the experience of voice hearing from the viewpoint of the hearer is fundamental to developing useful approaches for formulation and clinical practice (Holt & Tickle, 2015). This is reflected in increasing grey literature and first person perspectives of voice hearing (Beavan, 2011;Romme, Escher, Dillon, Corstens, & Morris, 2009). Many genetic and brain researchers, have neglected the psychosocial causes of psychosis (Read et al., 2009) symptoms. The frequent number of repetitive AVHs identified indicates that trials of anti-obsessional medication may be worthwhile (Stephane, Polis, & Barton, 2001) and the CBT employed with OCD patients may also be beneficial (Whittal & McLean, 1999). McCarthy-Jones et al.'s (2014) finding that 64% of participants had a relationship with their voices, indicates that while psychotherapeutic approaches aiming to alter the hearervoice relationship (Whittal & McLean, 1999), can be helpful for some people, it will not necessarily be helpful for all.

Formulation and the Hearing Voices Movement
The DCP (2013) has suggested that one alternative for The international Hearing Voices Movement (HVM) is a mental health service-user approach that highlights the needs and views of the voice hearers and professionals working collaboratively to reframe medical understandings of voice-hearing and develop coping and recovery frameworks (Longden, Corstens, & Dillon, 2013). The main proposition of the HVM is that voice hearing is an experience filled with meaning (Corstens et al., 2014). Interventions that are focused on eradicating voice presence are typically not reliably achieved (Lepping, Sambhi, Whittington, Lane, & Poole, 2011). Consequently, therapeutic approaches have moved away from symptom eradication towards exploration of the personal meaning of the experience to reduce the distress (Romme & Escher, 2000). Voices hearing has frequently been found to be precipitated and maintained by difficult emotional experiences that flood the individual with voice content, often associated with more global difficulties in their life (Longden, Madill, & Waterman, 2012;Romme et al., 2009;Romme & Escher, 1993). Peer support and hearing voices groups have been developed as a result of the HVM (Corstens et al., 2014). Within the HVW, some voice hearers experiencing abusive voices, have been found to become progressively more positive or impartial while other people negotiate with their voices to clarify what it wants or gain more time alone (Jones & Shattell, 2013). However, arguments that have been key to the development of the HVM are grounded in limited evidence and require formal testing (Corstens et al., 2014).

Cognitive Behavioural Therapy
One intervention which attempts to help formulate an understanding of the voice hearer's thoughts, feelings, beliefs and response to voices is cognitive behavioural therapy (CBT). CBT for psychosis has been widely researched with studies demonstrating its efficacy as an intervention (Wykes, Steel, Everitt, & Tarrier, 2008).
The primary supposition in CBT is that distress is associated with the manner in which individuals interpret and react to events (Freeman & Garety, 2003). The therapist works with the person to develop a formulation, to develop a shared understanding of what may be occurring, such as any maladaptive cycles that may be present, and to map out possible avenues for change (BPS, 2014). NICE (2014) advises that CBT should be offered to all those diagnosed with schizophrenia. Studies examining the effectiveness of group CBT identified improvements in symptoms of psychosis, self-esteem, hopelessness, social functioning and perceived voice control and power (Chadwick, Sambrooke, Rasch, & Davies, 2000;Gledhill, Lobban, & Sellwood, 1998;Wykes, Parr, & Landau, 1999;Wykes et al., 2005). In contrast, two RCT studies of group CBT for voice hearers found no significant effect of treatment on severity of symptoms (Barrowclough et al., 2006;Wykes et al., 2005).
However, the primary aim of CBT for psychosis is not to reduce psychotic experience, rather it is to decrease the distress experienced in response to psychosis (Chadwick, Birchwood, & Trower, 1996). Owen, Sellwood, Kan, Murray, and Sarsam (2015) found that a four week group CBT for psychosis (CBTp) with inpatients demonstrated significant reductions in distress and increased confidence (Owen et al., 2015). While traditional CBTp focussed on recognising and altering maladaptive thoughts and beliefs, third wave acceptance based theory and practice (Hayes, Follette, & Linehan, 2004) build on this through a focus the way in which individuals relate to voices (Dannahy et al., 2011). Person based cognitive therapy (PBCT) builds upon the CBT formulation by explicitly emphasising acceptance of the voice hearing experience, identifying and testing positive self-beliefs, acceptance of self and a greater focus on relationship with voices (Chadwick et al., 1996;Chadwick & Birchwood, 1994;Dannahy et al., 2011;Hayward, 2003). Dannahy et al. (2011) examined PBCT groups for people hearing distressing, treatment-resistant voices and found significant improvements in wellbeing, voice distress and voice control. The findings of CBT for psychosis reviewed support the recovery model as CBT was found to increase participants' confidence more than it reduced symptoms, suggesting participants were developing their ability to accept and cope with challenging experiences, rather than necessarily reducing there frequency (Dannahy et al., 2011;Owen et al.'s., 2015).

Relationship With Voices
Approaches that have attempted to identify what the psychological elements are that may impact the level of emotional distress caused by hearing voices have predominately emphasised cognitive factors, such as the individual's beliefs about the voices control and purpose (Birchwood & Chadwick, 1997;Chadwick & Birchwood, 1994). Recent developments in theory and research have built on this by highlighting the role of the relationship between the voice hearer and the voice (Hayward & Fuller, 2010). Hayward and Fuller (2010) qualitatively examined 'Relating Therapy' that aimed to modify distressing relationships with voices. The authors found evidence for a model that proposes to bring about positive change by altering relationships with voices (Hayward & Fuller, 2010). The voice hearing phenomenon occurs in a relational context (Hayward & Fuller, 2010). For example, phenomenological studies have found that voices are often linked with important people in the voice hearer's life (Garrett & Silva, 2003;Nayani & David, 1996). Within this framework for formulation, the feeling of helplessness experienced by voice hearers in response to upsetting voices is reflected within their social relationships (Birchwood et al., 2000). The paralleling of voice and social relationships may create opportunities to intervene therapeutically to promote change in one area with the goal of subsequent positive change in the other (Hayward & Fuller, 2010). Therapists highlighted the value of ongoing formulation with hearers to shape interventions based on their experience of voice hearing, and described the therapy as helping clients see connections between their voices and what is going on in their lives (Hayward & Fuller, 2010).
The process of exploring clients' interpretations of relationships has been found to improve voice hearers views of themselves, their relationship with their voices and self-esteem, and social relationships more generally (Birchwood et al., 2002;Hayward, 2003;Hayward & Fuller, 2010). These outcomes support the recent focus on a recovery model emphasising quality of life (Hayward & Slade, 2008). However, in their qualitative study Chin, Hayward, and Drinnan (2009) found that participants both accepted and rejected the concept of a relationship with their voice, with rejection most strong in those adopting illness/symptom narratives about their voices.
Research has indicated that while voices can reflect attributes of significant relationships, more discrete relational qualities differed across interactions between the hearer and the voice and the hearer and the significant people in their life (Beavan, 2011;Mawson, Berry, Murray, & Hayward, 2011;Mawson, Cohen, & Berry, 2010). A possible explanation for this difference is that when hearers thought they suffered stigma in relation to their voices, they demonstrated a greater likelihood to perceive their voices as upsetting events (Hayward, Berry, & Ashton, 2011). Therapeutic interventions focused on the relationship with the voice may need to be framed within the context of the voice hearers possible interpersonal history of low rank/power, bullying and high shame . Results of studies suggest that relationally based interventions, such as developing assertiveness and forming a more equal relationship with strong voices, may be helpful in reducing distress and improving the relationship between the voice hearer and the voice . A comprehensive formulation of the voice hearer experience may be supported by clinicians eliciting a detailed interpersonal history, allowing voices to be understood within the context and idiosyncrasies of the hearer's relational history . Forums such as hearing voices groups and family meetings have been found to assist voice hearers in the meaning-making processes through hearing others stories, sharing their own stories, exploring theirs and others reactions and the process of developing alternative understanding of themselves as voice hearers (McCarthy-Jones, 2012;Taylor & Murray, 2012). The most non-directive of family meeting approaches, 'Open Dialogue', offers an opportunity for individuals to listen to each other's understanding of what is occurring (BPS, 2014;Seikkula & Arnkil, 2006). This has been found effective in reducing symptoms and the frequency of service use, and increasing rates of return to the work place (Seikkula et al., 2006). However, the evidence base still requires more rigorous, controlled evaluation.

Compassion Focused Therapy
Compassion Focused Therapy (CFT) was developed for individuals who experience enduring and complex mental health difficulties associated with high shame and self-criticism, often within the context of difficult backgrounds (Leaviss & Uttley, 2015). It offers imagery-based interventions aimed at activating the emotion regulation system linked with self-soothing, the counterpart to the threat-focused emotion regulation system that is related to anxiety reactions (Gilbert, 2010). A sense of threat can develop from changes in feelings/ sensations and subsequent fear related to their possible meaning (Gumley, Braehler, Laithwaite, MacBeth, & Gilbert, 2010). There may be activation of trauma memories from previous episodes and voices arising can further increase internal self-attacking and threat processing (Gumley et al., 2010;Longe et al., 2010). One of the primary goals of CFT is to support people in developing a compassionate, reflective view of their experiences, to foster greater awareness of their interactions with self and others, and to improve emotional and interpersonal problem solving skills (Gumley et al., 2010). Lincoln, Hohenhaus, and Hartmann (2013) examined the effectiveness of a CFT group for people with sub-clinical symptoms of psychosis and found it reduced participant's paranoia, negative emotions, and increased self-esteem. The effectiveness of imagery on paranoid beliefs was facilitated through a reduction in anger, shame, anxiety and sadness (Lincoln et al., 2013). Beavan and Read (2010) examined the impact of voice hearer's auditory hallucinations and found that the content of the voice's was the only significant predictor of emotional distress and the strongest predictor of mental health service use (Beavan & Read, 2010). The authors found that individuals who had an unhelpful emotional response were significantly more likely to experience voices that commented on them or argued/ conversed with each other (Beavan & Read, 2010). Furthermore, studies have found that the processing and regulation of threat are areas of difficulty in psychosis (Braehler et al., 2013) as threat emotions, such as fear, anxiety and anger have been found to contribute to hallucinations (Freeman & Garety, 2003). Mounting evidence suggests that the content of voices may be key to understanding and supporting voice hearers (Beavan & Read, 2010). These findings highlight the need for a formulation approach which gives attention to the content and meaning of voices and their impact on emotional well-being.  (Corstens et al., 2014). This could aid in the development of a helpful and interpersonally coherent narrative that values voices as messengers representing opportunities for psychological growth (Corstens et al., 2014). Rather than symptom eradication, experiencing voice hearing with reduced negative impact on subjective wellbeing may reduce distress and improve quality of life (BPS, 2014;Corstens et al., 2014). These findings have implications for treatment planning and assessment of outcome. CFT, CBT, Relating Therapy and Open Dialogue are some of the approaches examined which may be effective in drawing on these principles to support individuals who hear voices. However, findings are limited by the lack of controlled studies exploring the meaning of voices and their treatment in psychosis. This qualitative review is limited by a non-systematic search strategy, meaning the possibility of selection bias cannot be ruled out. There is currently considerable research examining the biological factors influencing psychosis and its treatment. As a result, this review focused on highlighting the psycho-social factors which may benefit from greater integration into a biopsychosocial model of psychosis.
Consequently, this review is limited by the lack of detailed examination of the biological factors underlying psychosis. Further controlled studies examining the effectiveness of psychotherapeutic interventions for voice hearing is needed. This could be done by gearing therapeutic models to service users based on the formulation of their voice hearing experiences to inform the most suitable intervention e.g. CFT, CBT, Open dialogue. In this way, approaches are sufficiently tailored to each case to meet individual needs, while maintaining the core components of the most appropriate therapeutic model. This effectiveness of these approaches could then be compared to treatment as usual groups. Further qualitative explorations of individual voice hearing experiences and systematic reviews are also recommended.

Funding
The author has no funding to report.

Competing Interests
The author has declared that no competing interests exist.