Cognitive behavioural therapy for psychosis

What it is

Cognitive behavioural therapy for psychosis is a talking therapy for people with psychosis or schizophrenia. With a therapist, the person looks at distressing experiences, such as hearing voices or fearing harm from others, and at the beliefs attached to them.¹ It is given alongside medicine and other care. Treatment guidelines in the UK and the Netherlands recommend offering it to everyone with psychosis.^2,1

How it works

The therapy adapts methods from CBT for depression.³ Its starting point is that the way a person interprets an experience affects how distressing it is and what they do about it.⁴ Hearing a voice is one experience; believing that the voice can carry out its threats is another. CBT works mainly on the second.

With the therapist, the person links their thoughts, feelings and actions to their symptoms. Together they question the beliefs and perceptions involved, and weigh what supports them against what does not.^2,1 The aims include less distress, new ways of coping and better day-to-day functioning, not only fewer symptoms.²

History

In 1952, the American psychiatrist Aaron Beck described treating a patient’s long-standing delusion with talking therapy, outside hospital.⁵ Cognitive behavioural therapy itself arose in the 1960s, and a form for psychosis followed in the 1990s.¹ In 1991, the British psychiatrists David Kingdon and Douglas Turkington reported using it with 64 people with schizophrenia, alongside their usual treatment. They added a “normalising” explanation of how symptoms arise, to reduce stigma.⁶ In 1997, a trial in London and East Anglia became the first large randomised trial of the therapy for symptoms that medicine had not relieved.³

NICE in the UK recommends offering it to everyone with psychosis or schizophrenia, a recommendation dating from 2009.² In the Netherlands, the care standard for psychosis (Zorgstandaard Psychose, 2017) makes it standard care for every psychotic disorder, alongside information about the condition and antipsychotic medicine.¹

How well it works

Outcome	CBT and standard care	Standard care alone
Responded to treatment	50	31

That first large trial was small: 60 people took part.³ Later reviews, which pooled dozens of trials, found smaller benefits.

Across 34 trials, CBT eased symptoms by a small amount compared with the control groups. In trials where the people rating symptoms did not know who had had the therapy, the benefit was smaller still, and for hallucinations and delusions it was no longer clear.⁷ An earlier review found that when the raters knew, the benefit looked one and a half to two times as large. Even the most rigorous trials found some benefit.⁸

The largest comparison, of 53 trials of talking therapies with 4,068 people, found that CBT eased hallucinations and delusions more than usual care, inactive control treatments or supportive therapy. The benefit was small to medium, and confidence in these results ranged from moderate to very low.⁹

A review led by Dutch researchers found that CBT tailored to the person eased voices more consistently than delusions. Compared with other active treatments, the benefit for delusions was no longer clear.¹⁰

Against other talking therapies, the advantage of CBT is small. Across 48 trials that compared therapies directly, it eased hallucinations and delusions slightly more than the others.¹¹ A Cochrane review of 36 trials found no clear difference from other talking and social therapies in relapse or readmission to hospital, though the evidence was of low quality.⁴

In a trial of 74 people who had chosen not to take antipsychotic medicine, symptoms were consistently lower with CBT than with usual care alone, over 18 months. The researchers called for a larger trial.¹²

Possible downsides

The therapy means talking in detail about voices, fears and beliefs, over at least 16 sessions.² In trials, more people dropped out of CBT than of usual care alone,⁹ though fewer than dropped out of other talking therapies.⁴

In a German trial of 198 people, serious events such as suicidal crises or a severe worsening of symptoms occurred in 10 people having CBT and 5 having cognitive remediation, a training of attention and memory. The difference could have been chance, and there were no suicides. A small group had sharp rises in symptoms during therapy, and the researchers advised therapists to watch for this.¹³

Many people who could have it are never offered it. In one NHS trust in north-west England, about 7 in 100 people with schizophrenia were offered CBT over a year, and about 5 in 100 had it.¹⁴ In Dutch specialist mental health care, estimates range from 10 to 25 in 100, and fewer still have the recommended 16 sessions.¹

Format

One to one with a trained therapist, over at least 16 planned sessions. The therapist follows a treatment manual, preferably one tested in a trial, and is regularly supervised. Between sessions, the person may keep track of their thoughts, feelings and symptoms.² In the Netherlands, it can be part of outpatient care or be given during a hospital stay.¹ If it starts in hospital, NICE advises continuing the full course after discharge.² In the first large trial, therapy lasted 9 months.³

Contraindications and cautions

NICE advises offering it during an acute episode as well as later, including in hospital.²

For people who want talking therapy without antipsychotic medicine, NICE advises that psychological treatments work better alongside medicine. If the person still prefers to go without, it recommends offering CBT and family intervention, and reviewing the options, including medicine, within a month.²

Alternatives

AVATAR therapy (opens in a new tab) and Talking With Voices (opens in a new tab) are for people distressed by the voices they hear. The Maastricht interview (opens in a new tab) maps a person’s voices and links them to their life story, as a basis for further treatment. NICE also recommends family intervention for relatives who live with or are close to the person.²

What we don’t know yet

• How large the benefit is in trials where the people rating symptoms do not know who had the therapy.
• Whether it helps more than simpler, less costly talking therapies.⁴
• How well it works without antipsychotic medicine. Only small trials have tested this.¹²
• How to make it available to everyone the guidelines say should be offered it.