The experience of hearing voices (verbal ‘auditory hallucinations’) is one of the prominent features of schizophrenia in current systems of diagnostic classification . Voice hearing is reported to occur in approximately 70% of patients with this diagnosis , although the experience is also common in other mental health conditions such as post-traumatic stress disorder and borderline personality disorder . National Institute for Health and Care Excellence (NICE) guidance in the UK recommends that psychological therapy in the form of cognitive behavioral therapy (CBT) should be offered to all patients who have a diagnosis of schizophrenia . Although this is appropriate in the case of psychosis more broadly, where an evidence base exists for moderate effects , the impact of CBT for psychosis specifically on distressing voices is less convincing . Therefore, consistent with a symptom-based approach , a voice-specific approach may be warranted. A recent review  was conducted of 16 published randomized controlled trials (RCTs) of CBT for psychosis that have reported at least one psychometrically validated outcome measure specifically related to voice hearing. The review found that the majority of studies failed to show a significant effect of CBT on voice hearing, and most of those that reported significant effects did so in the context of methodological weaknesses. Only one study reported an effect using a robust methodology. Interestingly, this was the only study that focused exclusively on voice hearing . A factor that may have limited the effect of CBT upon voice hearing concerns outcome measurement - early trials used measures of voice frequency and severity, despite CBT not focusing upon the eradication of voices. The authors conclude by calling for more robustly designed RCTs of CBT aimed specifically at distressing voices. They draw attention to the potential for the integration of other psychotherapeutic techniques that have shown early promise for assisting voice hearers. In particular, the approach of trying to understand and adapt the interpersonal-like relationships that can develop with voices (see [9, 10] for a review).
There is an evolving body of literature that has explored the experience of hearing voices from an interpersonal perspective, examining the interaction that can occur between the hearer and the voice(s) that is heard. Findings suggest that hearers, regardless of diagnosis, can have integrated, personally coherent relationships with their voices . Two studies using social rank theory found marked differentials of power and social rank between the hearer and the voice, which place the voice in a dominant position, and mirror perceptions of self in relation to significant social others [12, 13]. However, whilst there has been a focus on dominance, research has paid less attention to the role of intimacy and closeness within relationships with voices. If relationships with voices do mirror relationships in the social world, they are likely to be imbued with all the complexity and idiosyncrasy of social relationships. Relating theory  proposes that relating occurs on two axes: a vertical ‘power’ axis characterized at each polar end by ‘upper’ and ‘lower’ (analogous to power differentials), and a horizontal ‘proximity’ axis characterized at each polar end by ‘close’ and ‘distant’. Studies that have explored voice hearing experiences through the lens of relating theory report that distressing voices are perceived as relating dominantly and intrusively [15–17], and are responded to through distant relating from the hearer. Mirroring has also been found between proximity styles of relating to voices and those of the hearers within their social relationships .
Collectively, this body of research suggests that voices can be understood within interpersonal frameworks. This conceptualization has given rise to a new generation of therapeutic approaches that seek to modify the hearer-voice relationship. A pilot RCT  that looked at a therapeutic approach to address the power dynamic with voices that issue commands (cognitive therapy for command hallucinations [CTCH]) found significant reductions in compliance behaviour within the therapy group, and this finding has been replicated in a definitive trial . A therapeutic focus upon closeness and intimacy has been developed by Hayward et al.  in the form of relating therapy (RT) which aims to re-balance the hearer-voice relationship with regard to both power and proximity. Findings from a case series suggested that RT was acceptable to hearers who experienced persistent and distressing voices [16, 20]. Consistent with the call for evaluations of effectiveness to be methodologically rigorous , RT should now be evaluated within an RCT design.
Research objectives and hypotheses
The main objective of the study will be to inform the development of a phase III definitive trial . Specifically, this external pilot RCT will establish recruitment, retention, and follow-up rates to the trial, assess levels of adherence with the treatment, and establish treatment effect size relative to treatment-as-usual. This information will be used to finalize the design of the therapy protocol and the study protocol for the phase III trial.
This external pilot RCT is, by definition, underpowered to detect statistically significant effects. Voice-related distress, rather than voice activity (such as loudness or frequency) or voice attributions (such as beliefs about the origin of voices), is the primary target of the therapy. Therefore, the primary hypothesis for the definitive trial is that RT added to treatment-as-usual will reduce the distress associated with voice hearing compared with treatment-as-usual on its own.
Secondary hypotheses for the definitive trial are that RT will lead to: 1) Reductions in voice-related distress relative to the control condition that will be maintained over at least three months; 2) Reductions in voice-related distress that will be mediated by improvements in relating to and/or by voices (improvements in relating defined as reductions in voice dominance and intrusiveness, and hearer distance - measured by the Voice and You [VAY]questionnaire); 3) Improvements in recovery (measured by the Choice of outcome in CBT for psychoses [CHOICE] questionnaire), relating to people in the social world (measured by the shorter version of the Persons Relating to Others Questionnaire [PROQ3]), and mood and anxiety (measured by the Hospital Anxiety and Depression Scale [HADS]); and 4) Improvements in relating to people in the social world will be associated with improvements in relating to and/or by voices (improvements in relating defined as reductions in voice dominance and intrusiveness, and hearer distance - measured by the Voice and You [VAY] questionnaire).