NHS Talking Therapies – No Better Than GP Care


So why is the UK Government spending a £1 billion a year on NHS Talking therapies? NHS Talking Therapies raison d’etre is that it has a 50% recovery rate [IAPT Manual (2019)], based principally, on a change of score on the PHQ-9 from above caseness, (a score of 10 or more) to below 10,  from 1st to last assessment. But in a study of a sample of GP patients  (n=100)    [ Gilbody et al (2015)] given usual care, 56% recovered within 4 months and 60 and 61% recovering by 12 and 24 months respectively, using the   PHQ-9 as the outcome metric. [Albeit that 13% accessed IAPT  (the predecessor of NHS Talking Therapies]. Similarly, Moore et al (2012) found 47% recovering within 3 months.  Whence the added value of NHS Talking Therapies?


The null hypothesis is that NHS Talking Therapies are no better than treatment as usual.  Funding of the Service cannot be justified without studies demonstrating the superiority of these interventions to a credible attention control condition. But no such studies have been forthcoming. NHS Talking Therapies Outcome studies have ignored the placebo response. Most recently, Strauss et al (2023) in a comparison of 2 forms of low intensity guided self-help, claiming the superiority of mindfulness guided self-help  over CBT guided self-help, despite reductions in mean PHQ-9 scores from 14-15 to 6-7  in 16 weeks in both arms of the study. No mention that similar results would likely be obtained with an attention placebo, nor comment that 25% dropped out of each form of self-help.   Instead, Strauss et al (2023)  call for an expansion of guided self-help beyond the 100,000 current recipients a year, preferably mindful because it was cheaper!

The Size of The Placebo Response

 In a review by Motta et al (2023) the average response and remission rates in placebo groups (across all anxiety disorders including PTSD and OCD) were 37% and 24% respectively.  [Motta LS, Gosmann NP, Costa MdA, et al. BMJ Ment Health 2023;26:1–8]. Those diagnosed with GAD and PTSD had larger placebo response estimates than those with PD, SAD and OCD. These figures were calculated by a within group Standardised Mean Difference, the average within subject placebo effect size was -1.1. By comparison the mean average placebo effect size in depression is 0.37. [Furukawa TA, Cipriani A, Atkinson LZ, et al.2016:3:1059-66} Placebo response rates in antidepressant trials: a systematic review of published and unpublished double-blind randomised controlled studies Lancet Psychiatry 2016;3:1059–66].

Recovery in NHS Talking Therapies

The service claims a 50% recovery rate but I found that only the tip of the iceberg recover, Scott (2018). My finding is consistent with the reported recovery rates in the above placebo studies, given that the NHS Talking Therapies population is most likely a mix of people suffering predominantly anxiety or depression.

The burden of proof is on NHS Talking Therapies to demonstrate that it produces a clinically relevant effects beyond placebo. But I do not think it will attempt this anytime soon, it would be like turkeys voting for Christmas. 

Dr Mike Scott