CBT’s House of Cards?

applying the acid tests of the Cochrane Collaboration Tool and the GRADE Handbook for the quality of randomised controlled trials, studies of low intensity CBT fail to clear the methodological bar. Whilst only high intensity studies for depression and the anxiety disorders make a successful jump. This calls into question IAPT’s penchant for disseminating CBT for everything, with an imprimatur from BABCP, paying travel expenses of upto £100 for special interest group members to attend a pre-conference workshop Revolution in Mental Health Service Delivery: The Evolution of Low Intensity CBT on Tuesday 3rd September.

One of the seven domains highlighted by the Cochrane Collaboration tool for assessing bias is the blinding of outcome assessment. I have been unable to locate one outcome study of low intensity CBT that fulfills this criteria whilst there are a significant minority of studies of high intensity interventions for depression and the anxiety disorders that do.

The GRADE handbook for assessing the quality of trials comments in section 3.4 ‘not infrequently, outcomes most important to patients remain unexplored’, with regards to psychological interventions clients are rarely asked by someone independent of the study whether and if for how long they are back to their usual selves since treatment. Instead most commonly reliance is placed on a surrogate measure a client completed questionnaire, as opposed to an independent clinicians assessment using a standardised diagnostic interview to determine whether there has been a loss diagnostic status.

These concerns are crystallised in a study of CBT for Health Anxiety conducted by Cooper et al (2017), Behavioural and Cognitive Psychotherapy, 2017, 45, 110–123 doi:10.1017/S1352465816000527

whilst 10 of the 13 studies in a meta analysis used the DSM or ICD-10 to determine whether people should be admitted to the meta analysis, in no study was meeting these criteria used as an outcome measure. To be no longer suffering from the identified health anxiety at end of treatment/follow up would have been a client important outcome. Instead the self-report Health Anxiety Questionnaire was used as surrogate. Cooper et al (2017) attempted to rate studies using the Cochrane Collaboration tool using a summary score for the seven domains, but this bore no relation to outcome and as the authors admitted was a questionable procedure. Despite this CBT was claimed to be an effective treatment for health anxiety.

I am afraid I can’t join in the jamboree for IAPT services that takes place at the BABCP annual conference. I doubt that the ‘House of Cards’ will be discussed and it would likely be seen as banned literature on IAPT training courses.

Cochrane https://www.dropbox.com/s/bmr98o8z8fcfuzv/paths%20to%20mh%20Cochrane%20Risk%20of%20bias%20assessment%20tool.pdf?dl=0

GRADE handbook

https://www.dropbox.com/s/wudv2eu2oxw7qsr/GRADE%20handbook.html?dl=0

Dr Mike Scott

Leave a Reply