From the Editor

“The voices won’t leave me alone.”

A patient made this comment to me recently. And it’s an experience many patients have had. Despite medications and follow up, the voices continue to be problematic.

This week, we look at a new paper from The Lancet Psychiatry which considers a novel approach: AVATAR therapy.

In this study, patients who experienced auditory hallucinations created a computerized simulation (avatar) of the voice they most wanted to influence, including what the voice said, how it sounded, and how the “entity” with the voice looked like. Patients, working with a therapist who controlled the avatar, then had therapy sessions in which they could talk to it. Patients were compared on several measures to those who only received supportive therapy.

The paper has received significant media attention including CNN and BBC.

1124avatar2The Face of the Voice – and a Step Toward Healing?

Spoiler alert: the therapy helped – at least initially – but the results are complicated. (And, no, this isn’t “fake news,” to borrow a phrase from an American politician.)

In this Reading, we consider the paper and its findings.


Voices and Therapy

“AVATAR therapy for auditory verbal hallucinations in people with psychosis: a single-blind, randomised controlled trial

Tom KJ Craig, Mar Rus-Calafell, Thomas Ward, Julian P Leff, Mark Huckvale, Elizabeth Howarth, Richard Emsley, Philippa A Garety

The Lancet Psychiatry, 23 November 2017 Online First (Open Access)

Auditory verbal hallucinations, which are typically of a derogatory and threatening nature, are reported by approximately 60–70% of people with schizophrenia. Although pharmacological therapy is effective at reducing hallucinations in many people, approximately 25% of people with psychotic conditions continue to experience them. Cognitive behavioural therapy for psychosis is also helpful for many people, although average effect sizes are in the small to moderate range, and training and resource requirements mean that, in practice, therapy is delivered to only a fraction of those who might benefit. Consequently, there is considerable interest in the development of novel therapies that draw on the principles of cognitive behavioural therapy for psychosis but which are shorter, specifically targeted at auditory verbal hallucinations, and are capable of being delivered by a wider workforce.

Several novel therapies build on the perspective that auditory verbal hallucinations are experienced as coming from entities that have personal identities, speak with purpose, and with whom the hearer establishes a personal relationship. The operation of power within this relationship is viewed as crucial. The voice is typically experienced as dominant (even omnipotent), with the voice-hearer assuming a submissive role characterised by feelings of inferiority and powerlessness that can reflect social relationships more generally. In light of this finding, explicitly relational and interpersonal approaches have been developed that locate voices (and voice relationships) within the person’s biographical context and target key interpersonal dimensions such as power and proximity.

AVATAR therapy belongs to this new wave of relational approaches but, uniquely, the voice-hearer’s experiences are brought into therapy in a new way, allowing a face-to-face interaction with a digital representation (avatar) whose speech closely matches the pitch and tone of the persecutory voice. The therapist (switching between speaking as therapist and as avatar) facilitates a dialogue in which the voice-hearer gradually gains increased power and control within the relationship, with the initially omnipotent voice loosening its grip over the hearer by becoming more conciliatory over time.

A pilot study that compared AVATAR therapy with a treatment as usual in 26 patients who had a longstanding single or dominant persecutory voice found significant reductions in the frequency, distress, omnipotence, and malevolence of the voice. We report the results of a larger, randomised controlled trial that compared AVATAR therapy with an augmented supportive counselling intervention.

Tom Craig

Here’s what they did:

  • The study was a single-blind, randomized controlled trial done in South London and Maudsley NHS Trust.
  • Participants were referred to the study by their treating clinician between Nov 1, 2013, and Jan 28, 2016.
  • Inclusion criteria included diagnoses of schizophrenia or schizoaffective disorder, ongoing auditory hallucinations for at least 12 months despite medications, age 18 or older. Exclusion criteria included organic brain disease, enrolment in therapy at the time of the trial, medication refusal.
  • Participants were randomized to the AVATAR therapy or supportive counselling. In the AVATAR therapy, participants first created a computerized representation of the entity that they believed was the source of the main voice. They then received six weekly sessions of 50 minutes.
  • “Therapy proceeds through two phases. Phase one (typically sessions one to three) involved exposure to the avatar speaking the typical verbatim content of the participant’s voices while the therapist encouraged assertive responding—eg, that the person tell the avatar that they are no longer prepared to accept these threats and insults and to challenge any apparent misconceptions the avatar seems to have. In phase two (typically sessions four to six), the dialogue gradually evolved as the avatar conceded ground and acknowledged the strengths and good qualities of the participant. There is an explicit focus on self-esteem and acknowledgments of the participant’s strengths and capabilities.”
  • The primary outcome was to look at auditory hallucinations – that is, the total score on the Psychotic Symptom Rating Scales, auditory hallucinations subscale (PSYRATS–AH) at 12 weeks. They also looked at other subscales, other measures (like negative symptoms), and adverse events.
  • Statistical analyses were done, including a linear mixed-effects model for the primary outcome (PSYRATS-AH scores at 12 weeks).


King’s College has a youtube video with Craig explaining AVATAR therapy. The video can be found here:

Patient-created Avatars

Here’s what they found:

  • Of the 394 people referred, 150 were eligible and were randomly assigned to receive either AVATAR therapy (75) or supportive counselling (75).
  • Demographically: participants were male more than female (60% vs. 40%), White British (43%), unemployed (85%), had schizophrenia (77%), and with a mean age of 42.7 years.
  • “At 12 weeks, AVATAR therapy led to significantly greater reductions in auditory hallucinations than did supportive counselling, as assessed by the PSYRATS–AH total score (estimated mean difference −3·82, SE 1·47, 95% CI −6·70 to −0·94; p=0·009; d=0·8). There were also significant differences in reported frequency of voices and reduced distress at 12 weeks…”
  • “Nine participants reported a complete absence of voices during the preceding week at the week 12 assessment (seven in the AVATAR therapy group and 2 in the supportive counselling group).”
  • “At 24 weeks follow-up, the AVATAR group maintained the improvements on PSYRATS-AH, but the supportive counselling group continued to improve “such that there were no significant differences between the two groups by this point. No significant differences between the two groups were observed for any of the other secondary outcomes at either 12 or 24 weeks.” !! See figure below.

AVATAR therapy was feasible to deliver, acceptable to participants, and did not result in any adverse events that could be attributed to the therapy. The study corroborated the primary hypothesis concerning clinical efficacy by showing a rapid and sustained reduction in the severity of auditory verbal hallucinations by end of therapy at week 12 that was significantly superior to that achieved by supportive counselling.

A few thoughts:

  1. This is an interesting study in that it takes a common problem and tries a different approach.
  1. At 12 weeks, they had a big result.
  1. At 24 weeks, they didn’t. It raises good questions. First and foremost: why didn’t the AVATAR therapy outperform supportive counselling over time?
  2. Ben Alderson-Day and Nev Jones wrote a Comment that runs with the Craig et al. paper. “Understanding AVATAR therapy: who, or what, is changing?” is highly readable. You can find it here:

They offer a candid assessment of the results. “Although these results are encouraging, significant differences between the treatment and control groups were no longer evident at 24 weeks, and the authors note a roughly equivalent number of participants in both groups reporting no voices at the end of the trial.”

They also consider the nature of this type of therapy.

To understand the complex dynamics of this kind of treatment, we need to look both back and forward. If auditory verbal hallucinations can be dialogued with in this way, is this a possibility only for those with many years of developing voices, or might it be relevant for the young person using early intervention services for the first time?… If we think that the voices or their power relations have changed, does this actually persist beyond therapy, and why are gains apparently not maintained when compared with control interventions?

  1. Picking up on the points made by Alderson-Day and Jones, we can ask: is AVATAR therapy helpful, but to a sub-population? For the record, Alderson-Day and Jones close on a thoughtful note: “the question now is this—how does the conversation continue?”
  1. Interested in avatars? Evidence-based Mental Health has a paper on using avatars for people with borderline personality disorder. You can find the Falconer et al. paper (which is open access) here:


Reading of the Week. Every week I pick articles and papers from the world of Psychiatry.