IAPT’s Training Fails But CBT Can Make a Real World Difference

three papers just published in the journal Cognitive Therapy and Research, tell contrasting stories: two are by leading lights in IAPT assessing the competence of trainees, in neither study did they demonstrate any real world outcome. By contrast in a study by Perrin et al (2019) of individual CBT for children (aged 10 to 18) suffering from generalised anxiety disorder 80% no longer had GAD by the end of 10 sessions of treatment compared to 0% in the waiting list.   These impressive results were maintained at 3 month follow up.

IAPT could learn from the Perrin et al (2019) study in that client’s diagnostic status was assessed using  a standardised diagnostic interview  and again at the end of treatment using blind assessors, further therapists followed an evidence based protocol for the identified disorder. Whilst it is costly to make such rigorous assessments and IAPT might fear having to explain to Clinical Commissioning Groups the necessary change in modus operandi, IAPT might then at last make a socially significant difference.

IAPT has been provisionally scheduled to be the focus of presentations on BBC TV and Radio on Wednesday, November 13th. 

Perrin et al (2019) Cognitive Therapy and Research (2019) 43:1051–1064

Liness et al (2019) Cognitive Therapy and Research (2019) 43:959–970

Liness et al (2019) Cognitive Therapy and Research (2019) 43:631–641

Dr Mike Scott

CBT Is Overeaching Itself – Clients and Therapists Are The Likely Casualties

A re-examination of the evidence base for CBT, using published guidelines for the evaluation of randomised controlled trials [ Guidi et al (2018)], suggests that low intensity interventions and interventions for ME, long term physical conditions and psychosis are not evidence based. Such studies lack credibility either because of the abscence of blind outcome assessment or when blind assessment has been conducted the results have been negative. Further the number of blind credible trials supporting the efficacy of CBT for depression and anxiety disorders is about half the number of studies usually considered as evidence. Dissemination of CBT beyond the boundaries of an evidence base hampers finding real world solutions to a clients difficulties and will likely result in demoralisation of the latter and therapists. This casts doubt not only on the wisdom of IAPT’s expansion beyond depression and the anxiety disorders but the ethics of its’ treatment of staff.

An international team of Experts [Guidi et al (2018) see link below] have developed evaluation guidelines stipulating the need for blind independent assessment of psychological interventions, particularly when psychometric tests are the outcome measure.


The PACE trial for chronic fatigue syndrome was heavily criticised [ Edwards (2017)] because it relied on self-report measures of outcome without blind assessment, a methodology that is unacceptable in medicine and in the evaluation of pharmacological products see https://journals.sagepub.com/doi/full/10.1177/1359105317700886

To my knowledge there are no blinded assessment of outcomes for any low intensity interventions. Efficacy has a way of disappearing when there is blinded assessment, for example Morrison et al (2018) conducted a blinded outcome assessment of CBT for schizophrenia and found no clinically meaningful difference, see link below:


One other stipulation of the Guidi et al (2018) guidelines is that studies of an intervention should involve an active placebo, in order to ensure that any impact of treatment is not just due to raised expectations and attention. But more than 80% of trials in the anxiety disorders have used waiting list control groups [Cuijpers (2016)] as opposed to active placebos .


Carpenter et al’s (2018) , study of anxiety disorders see link below found that there were only 41 studies using an active placebo and in only two thirds of them was there a low risk of bias because outcome assessment was blinded. Thus though CBT was still regarded as efficacious, this number of studies spread across all the anxiety disorders does not make the case for CBT being irrefutable.


As Zhu et al (2014), see link below, put it with regard to generalised anxiety disorder, the evidence for CBT is ‘strong but not definitive’. They point out that although the 12 randomised controlled trials they reviewed all had blind assessors, in 6 of them outcome was not based on the assessors assessment but on a self-report measure.


Of the 144 studies of depression, generalised anxiety disorder, panic disorder and social anxiety disorder reviewed by Cuijpers et al (2016) only half (48.6%) had a blind outcome assessment,


Further Cuijpers et al (2016) found that the effects of CBT are small to moderate when the comparison condition is usual care or active placebo compared to a large effect size when the comparison is a waiting list control condition.

In view of Guidi et al’s (2018) strictures around the evaluation of randomised controlled trials, it is wholly inappropriate for IAPT to admonish its therapists for ‘poor performance’ based solely on a psychometric test. There are surely grounds here for a therapist to claim constructive dismissal.

Dr Mike Scott

Better Than CBT?

‘Metacognitive therapy (MCT) is a new evidence based psychotherapy that is proving to be more effective than than CBT’ so runs the advert in the April 2019 issue of the Psychologist, promoting an MCT Conference at the end of next month. Inspection of the referenced supporting literature indicates that there is just one, to be published study, by Adrian Wells et al, on Generalised Anxiety Disorder, suggesting MCT outperforming CBT. In MCT their is allegedly a 70-80% recovery rate compared to average 50% in CBT.

But great care has to be taken in evaluating efficacy studies, those relating to GAD are an exemplar. Studies conducted only by the originator of a therapy (Adrian) are necessarily suspect, there needs to be at least one independent study by researchers without an allegiance to the therapy and in which there is blind assessment of outcome using a standardised diagnostic interview. Further the results should include blind rater assessments not merely self-report. Whilst Adrian’s work has not yet cleared this hurdle, a methodologically rigorous analysis of the CBT for GAD studies paints a less convincing picture than most CBT devotees would imagine. A review of CBT for GAD studies by Zhu and colleagues, found just 12 studies as worthy of consideration and commented:

‘Despite having blinded rater, in half the the studies the main outcome depended on the self-rating….The overall risk of bias was considered high in 8 of the 12 studies. And using the rigorous GRADE criteria the overall level of evidence was classified as ‘moderate’, which indicates that further research could change the widely accepted conclusion about the effectiveness of CBT. Thus the results in favor of CBT are strong, but not definitive’. Dropbox link to full article below:


When it comes to studies of CBT for long term physical conditions, the evidence is much weaker than that for GAD which raises the interesting question of ‘why IAPT is treating long term physical conditions’. This very question is to be addressed by a Psychological Welbeing Practioner at an IAPT PWP Conference on June 26th. Interestingly the Workshop is titled ‘Step 2 Support for long term conditions’. But there is surely a gross mismatch between a low intensity intervention and a long term physical condition! It rather looks like distinction between low and high intensity interventions is being blurred, not before time. However a colleague of mine working in high intensity has been trained in treating LTC’s but is restricted to 6 sessions! Despite none of the efficacy studies in this area offering just 6 sessions, I am off to a home for the bewildered and bemused.

Dr Mike Scott