AVATAR Therapy for Auditory Hallucinations: If You Can't Beat the Voices, Join Them

Tom K J Craig, PhD, FRCP Tom K J Craig, PhD, FRCP

About 65% of patients with schizophrenia experience verbal auditory hallucinations, typically in the form of voices that emanate from perceived “others,” who tend to fit a common unsavory profile – they're domineering, derogatory and unremittingly hostile. Current therapies help to ease hallucinations in many patients, but for roughly a quarter of people with psychotic conditions, available treatments just aren’t enough.Enter AVATAR therapy. AVATAR, an acronym for Audio Visual Assisted Therapy Aid for Refractory auditory hallucinations, began garnering attention when it was pilot tested as a treatment for patients with auditory hallucinations between 2009 and 2011 by the UK’s National Institute for Health Research.The therapy allowed patients to create an avatar, or visual representation of the source of their perceived auditory hallucinations, known as the “persecutor,” whose speech closely matched the pitch and tone of the persecutory voice in their heads. Patients were then encouraged to engage in a dialogue with the avatar, who was controlled by a therapist. Instead of propagating a relationship where the persecutory voice dominates a submissive patient, the therapist could control the avatar so it would slowly yield control to the patient as time passed.Results of the pilot study were encouraging. Patients who underwent the novel AVATAR therapy showed mean reductions in total Psychotic Symptom Rating Scale (PSYRATS) auditory hallucinations of 8.75 (P = .003), and in the Omnipotence and Malevolence subscales of the Revised Beliefs About Voices Questionnaire (BAVQ-R) of 5.88 (P = .004). On the other hand, the control group experienced no changes during the study period.Researchers behind a recent follow up study published in The Lancet set out to recreate those results in a larger, powered, randomized controlled trial, and again found that patients experienced favorable outcomes, with 83% meeting the primary end point – a reduction in auditory verbal hallucinations at 12 weeks.For the study’s lead author Tom K J Craig, PhD, FRCP, emeritus professor of social psychiatry at King’s College London, the positive results came as a surprise. The most compelling evidence, he said, was the number of patients who clearly improved with therapy.“Most dramatic were the people for whom voices stopped entirely. Although Julian Leff had found this in his first pilot work, we did not really expect to see it repeated in our larger [powered, randomized] controlled study,” Craig told MD Magazine. “While that was the most striking outcome, the wider reductions in frequency and severity of voices reported by many people was also striking.”In the follow up study, Craig and colleagues described the significance behind the transition of power from persecutor to patient.“The operation of power within this relationship is viewed as crucial…the voice-hearer assum[es] a submissive role characterized by feelings of inferiority and powerlessness that can reflect social relationships more generally,” researchers wrote. “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."The results are especially encouraging because the trial involved a sample of people suffering from persistent psychoses who reported unremitting and distressing auditory hallucinations for at least the previous year, despite regular supervision and continuing pharmacological treatment. Moreover, more than a third of all patients across both therapy groups had a clinical record of treatment resistance and were prescribed clozapine before the start of the study.In the follow-up study, the reduction in PSYRATS total score at 12 weeks was significantly greater for AVATAR therapy than for supportive counseling, with a mean difference of -3.82 [SE 1.47], 95% CI -6.70 to -0.94; (P