UK-based online therapy provider Ieso Digital Health has applied deep learning to provide new insights into which aspects of psychotherapy are the most effective.

Using data from 90,000 hours of its internet-enabled cognitive behavioural therapy (ie-CBT) transcripts, Ieso has trained a deep learning AI to recognise the content of language used by therapists during CBT sessions.

Ieso Digital Health senior scientist Michael Ewbank said: “With our deep learning model, we can extract knowledge accumulated across thousands of hours of CBT in a way that would be impossible for a human to do.”

During ie-CBT, the patient and practitioner communicate using a real-time text-based system, allowing the conversations to be captured as transcripts.

The records of over 14,000 patients aged 18 to 94 undergoing ie-CBT were analysed. This collective of patients had a reliable improvement rate of 63.4% and an engagement rate of 87.3%.

The researchers identified 24 therapy feature categories, such as greetings, mood checking and setting goals. The mean number of words for each feature, averaged across all sessions, was calculated for each case.

The researchers found that when therapists focused on CBT behaviour change methods, patients were more likely to show an improvement in their symptoms and engagement with the therapy.

Conversely, patients were less likely to improve when session had an increased quantity of ‘non-therapy’ content, such as small talk. They would also be more likely to disengage from their treatment.

The study has been published in JAMA Psychiatry.

The findings support the key principles underlying CBT and provide validation for CBT as a treatment. Unlike other medical treatments, psychotherapy is comprised of a series of one-to-one discussions, meaning systematically measuring the efficacy of treatment is difficult.

Ewbank said: “What is exciting about this study is that it demonstrates the potential of Ieso’s data set, where we can understand more about what the active ingredients of therapy are, what works for whom, and develop new and more effective treatments for mental health disorders. Our work represents a first step towards a practicable approach for quality-controlled behavioural health care with the goal of improving the efficacy of psychotherapy.”