The October issue of The Psychologist features a letter of mine and the Editor’s response:
‘In the September issue, Ross Harper – CEO of Limbic, providing AI software for mental healthcare – extolled the virtues of their collaboration with four IAPT services. The supposed gains were all in operational matters, e.g. reduced time for assessment, with no evidence that the AI has made a clinically relevant difference to client’s lives.
This is merely the most recent example of the biased reporting of The Psychologist with regards to IAPT. I raised these concerns in a 2014 submission, ‘IAPT – The Emperor Has No Clothes’, and the Editor rejected it with the line ‘I also think the topic of IAPT, at this time and in this form, is one that might struggle to truly engage and inform our large and diverse audience’. This response was breathtaking given that IAPT was/is the largest employer of psychologists.
Fast forward to 2018, and I had published a paper ‘IAPT – The Need for Radical Reform’ in the Journal of Health Psychology, presenting data that of 90 IAPT clients I assessed independently using a standardised diagnostic interview only 10 per cent ‘recovered’ (in the sense that they lost their diagnostic status). This contrasts with IAPT’s claimed 50 per cent recovery rate. The paper has received only a passing mention in the pages of The Psychologist [‘Letters’, Flatt and Lido, April 2021]. This year I wrote a rebuttal of an IAPT inspired paper that was published in the British Journal of Clinical Psychology, ‘Ensuring IAPT Does What It says On The Tin’, but again no mention of this debate in The Psychologist.
In my view the BPS is guilty of a total dereliction of duty to mental health service users in failing to facilitate a critique of IAPT. If psychologists cannot pick out the log in their own eye, how can they pick out the splinter in others?
Dr Mike Scott
Editor’s reply: Funnily enough, we’ve also received strong criticism over the years for being too critical of IAPT! While I stand by my ‘at this time and in this form’, those are ever-shifting considerations and I’m pleased to now be in contact with you about a potential contribution to the magazine’.
Further to the editor’s reply I have submitted an article ‘Spin In CBT’ see below, alas the Editor said ‘no’ . I am off to a home for the bemused and befuddled.
The Spinning of CBT
Michael J Scott, Joan S Crawford and Keith Geraghty
There has been a massive expansion of psychological therapy services since the inception of the UK Government’s, Improving Access to Psychological Therapies (IAPT) Service in 2008. Offering principally and allegedly, cognitive behaviour therapy (CBT), by 2023/24 the IAPT Service hopes to see 1.9 million people a year [IAPT Manual, August 2021]. This represents a quarter of the community prevalence of depression and anxiety disorders. The intent of IAPT is clearly laudable, but it is much less clear that it meets the needs of clients? Is it worth the money? Given that the typical IAPT therapist earns £35K a year (twice that of a Care Assistant), and with salary costs reaching over £0.5 billion per year by 2023/24, there is a pressing need for independent audit.
By Services Marking Their Own Homework
The first author was alerted that all may be not well at IAPT’s coalface, when as an Expert Witness to the Court he reviewed 90 cases, treated with alleged CBT, and found a recovery rate i.e loss of diagnostic status, in just 10% of cases, using a standardised diagnostic interview Scott (2018). This applied whether or not service users were treated before or after their personal injury. This ‘tip of the iceberg’ response, contrasts sharply with IAPT’s claimed recovery rate of 50%. Curiously there has been no publicly funded independent audit of IAPT which would help to settle matters.
To Mask A Fault Line In the Provision of Routine Psychological Therapy
There is a fault line in IAPT’s approach which we thought might be rectifiable Scott (2018) [IAPT- The need for radical reform, Journal of Health Psychology] and Scott (2021) [Ensuring IAPT does what it says on the tin. British Journal of Clinical Psychology], but which may in fact make its’ ‘building’ of services inherently unsafe. IAPT declares usage of the ICD-10 code (the World Health Organisation’s labelling system for all disorders). The recent IAPT Manual (August 2021) https://www.england.nhs.uk/wp-content/uploads/2018/06/the-iapt-manual-v5.pdf recommends that IAPT clinicians give at least one code to each client, to characterise their debility. But nowhere in the Manual does it suggest that IAPT clinicians make a diagnosis. An ICD-10 code is only as reliable as the diagnosis made. The Manual claims that NICE Guidelines are based on ICD-10 codes and that IAPT is therefore NICE compliant. However the treatments recommended by NICE are all diagnosis specific, it follows that if there is no diagnosis there can be no fidelity to a NICE protocol. A key part of IAPT’s narrative is to gloss over that IAPT’s interventions are based, not on diagnosis, but on ‘problem descriptors’. The silent assumptions are that: a) there would be reliable agreement (reliability) between clinicians about what would constitute a client’s main problem and b) there is a body of evidence that a problem descriptor acts as a key to unlock the door to a specific protocol. Further that the specific protocol has been demonstrated to confer an added value, over and above an active placebo, for the chosen problem descriptor. There is an assumption of clinical utility. But there is no empirical evidence for either a) the reliability or b) the clinical utility. Whilst IAPT interventions contain elements of protocols used in randomised controlled trials there is no evidence of fidelity to such protocols i.e of comprehensive coverage of treatment targets for a disorder and matching treatment strategies.
A Pandemic of Spin
Spin has been identified in half the abstracts of papers in psychiatry and psychology journals Jellison et al (2020). In this context spin referred to a claim that an experimental treatment was beneficial, despite a statistically nonsignificant difference for the primary outcome or to distract the reader from statistically nonsignificant results. IAPT claims its sojourn into the treatment of persistent physical symptoms (PPS), such as chronic-fatigue syndrome, is evidence-based, but it is an exemplar of precisely what Jellison et al (2020) identified. Chalder et al (2021) compared the effectiveness of transdiagnostic cognitive behavioural therapy (TDT-CBT) plus standard medical care (SMC) to SMC alone. The primary outcome measure was the Work and Social Adjustment Scale (WSAS). There was no significant difference in outcome on this measure but in the abstract Chalder et al (2021) proclaim their intervention ‘may be helpful with a range of PPS’, with an appeal to some outcomes on secondary measures.
But arguably there are other additional markers of spin a) when the primary outcome is not independently assessed b) when the primary outcome is not clinically relevant and c) when there is no prior specification of what would constitute a minimally important difference in the primary outcome measure. In this connection none of the studies used to justify IAPT’s low intensity interventions have involved an independent assessor using a standardised diagnostic interview. For example, the Stress Control (SC) Programme is the most commonly delivered first line group intervention in IAPT Dolan et al (2021). In the SC studies outcome was assessed purely with self-report measures without any guarantee that the measures related to the disorders that the clients were suffering from. Dolan et al (2021) made no attempt to explain a) what a change of X on these measures would mean as opposed to a change in Y and b) whether the changes of scores would be meaningful to a client. The study showed an effect size difference in outcome between SC and active comparison conditions and passive controls of 0.12-0.15, but this is so small as to be of doubtful clinical significance. Nevertheless under a heading of ‘Practitioner Points’ they declare ‘SC is appropriate and effective for mild to moderate anxiety and depression’. Dolan et al (2021) found that the SC studies had a mean quality score of 18.21 but fail to mention that this is much lower than the mean score of 27.8 [Ost (2008)] in CBT studies. Ost (2008) commenting on a series of studies that had a mean score of 19.6 declared that this set of studies could not therefore be considered an empirically supported treatment (EST). IAPT does not provide EST’s in their low intensity provision. Further there is no evidence of fidelity to ESTs in the high intensity interventions. Additionally the dosage of therapy routinely delivered in high intensity IAPT therapy falls far short of that advocated in randomised controlled trials of CBT.
The Genesis of Spin
Spin is often related to undeclared conflicts of interest. In the Dolan et al (2021) study all authors declared no conflict of interest. But the corresponding author for the Dolan et al (2021) study is a programme director of IAPT and another of the authors has IAPT involvement. Unfortunately this is not an isolated example, Scott (2021) challenged a similar non-disclosure by these authors earlier this year.
A National Failure to Address Spin
The National Audit Office (NAO) began an investigation into the IAPT service but then stopped it in June 2018 without publication of findings. Following a Freedom of Information request to the NAO, the first author was told in a communication dated February 17th 2020 that the investigation was halted because of variously, the collapse of Carillion and Brexit, with no intention of resuming its’ investigation. Further the NAO response added ‘The investigation was not intended to comment on clinical judgements or the extent to which services meet patient needs’! IAPT has successfully enlisted NHS England, Clinical Commissioning Groups (CCGs), the British Psychological Society (BPS) and the British Association for Behavioural and Cognitive Psychotherapy (BABCP) to proclaim its’ ‘world-beating’ [Dr Claire Murdoch, NHS England Mental Health Director, Health Business August 26TH 2021] status. IAPT is however eminence-based not evidence-based.
The Demise of The Psychologist’s Role
IAPT is the major employer of psychologists, who risk becoming deskilled by the climate change brought about by the Service. Psychologists tend to stay in IAPT for a few years before heading for the exit to secondary care of private work. This leaves an IAPT workforce bereft of the means of critical appraisal of their work. Staff are ill-equipped to challenge the edicts from on high and disagreement is seen as disloyalty resulting in burn out and worse.
- Michael J. Scott, Consultant Psychologist, Liverpool, UK
- Joan S. Crawford, Counselling Psychologist, Mersey Care NHS Foundation Trust, Chronic Pain Management Service (CPMS), St Helens. https://orcid.org/0000-0001-6400-1158
- Keith Geraghty, Centre for Primary Care, Division of Health Sciences and Population Health, University of Manchester, UK. https://orcid.org/0000-0001-5060-5022
Corresponding author: Michael J Scott email@example.com
Chalder, T., et al. (2021). Efficacy of therapist-delivered transdiagnostic CBT for patients with persistent physical symptoms in secondary care: a randomised controlled trial. Psychological Medicine, 1–11. https://doi.org/10.1017/S0033291721001793.
Dolan, N., Simmonds-Buckley, M., Kellett, S., Siddell, E., & Delgadillo, J. (2021). Effectiveness of stress control large group psychoeducation for anxiety and depression: Systematic review and meta-analysis. The British journal of clinical psychology, 60(3), 375–399. https://doi.org/10.1111/bjc.12288
Jellison S, Roberts W, Bowers A, et al. BMJ Evidence-Based Medicine 2020; Evaluation of spin in abstracts of papers in psychiatry and psychology journals 25:178–181.
Scott M. J. (2018). Improving Access to Psychological Therapies (IAPT) – The Need for Radical Reform. Journal of health psychology, 23(9), 1136–1147. https://doi.org/10.1177/1359105318755264
Scott M. J. (2021). Ensuring that the Improving Access to Psychological Therapies (IAPT) programme does what it says on the tin. The British journal of clinical psychology, 60(1), 38–41. https://doi.org/10.1111/bjc.12264