Half of IAPT Clients Attend Only One Treatment Session

according to the UK Governments Improving Access to Psychological Therapies (IAPT)  annual report for 2019-2020 -https://files.digital.nhs.uk/B8/F973E1/psych-ther-2019-20-ann-rep.pdf,  489,547 people curtailed involvement after having one treatment appointment,  whilst 606,192 had two or more treatment sessions. If half of people never returned to a particular Cafe in their locality it would not be a good advertisement. But the IAPT Cafe is the only affordable one for most people, with a third of its’ clientele returning for further courses of ‘sustenance’, Hepgul et al (2016) https://doi.org/10.1186/s12888-016-0736-6.  Doesn’t sound like this Cafe confers health.

This haemorrhaging of clients is not a feature of the NICE recommended treatments, yet IAPT claims fidelity to such treatments! What then does one make of IAPT’s claim of a 50% recovery rate? This would be a fitting subject for Radio 4’s ‘More or Less’ statistical programme. To set the context, of 1.69 million referrals to IAPT in 2019-2020, 1.17 million left the starting gate, 30.77% (almost 1 in 3) were non-starters. Further only 1 in 3 (36.8%) got around the course (defined curiously by IAPT as attending 2 or more treatment sessions). IAPT’s terrain is much like that of the Grand National horse race. The ‘50% recovery rate’ refers to the significant minority who cross IAPT’s finishing line. Thus even using IAPT’s own yardstick  the true recovery rate is much less than 50%.

With regards to those who cross IAPT’s finishing line, there is no indication that their ‘success’ is lasting. It is not known what proportion of them ever ‘race’ again. Thus, the true recovery rate is likely to be far less than 50%. My own independent study of 90 IAPT clients, Scott (2018) https://doi.org/10.1177%2F1359105318755264  suggests that in fact only the tip of the iceberg, 10% , recover in any real world sense.

It serves all Organisation’s well to be vague about its’ criteria for success, they almost certainly operate with a confirmatory bias, cherry picking data that would justify its’ existence in the eyes of its paymaster. Biases or heuristics do not necessarily operate consciously but are nonetheless powerful and likely to be employed if survival is at stake. See for example Lilienfeld and May’s (2015) https://doi.org/10.1002/9781118625392.wbecp567 critique of reliance on self-report measures, a particular penchant of IAPT. These authors also point out that the only sure way of de-biasing is to have independent review, but IAPT has never been subjected to publicly funded independent review. 

Dr Mike Scott