Ensuring IAPT Makes A Real-World Difference - Scott: Dec 2019

Sly Saint

Senior Member (Voting Rights)
Key Messages

Over the last decade over £3billion has been spent on the UK
Government’s Improving Access to Psychological Treatment
programme, without any independent assessment of outcome.

IAPT claims a 50% recovery rate but other evidence suggests
that only the tip of the iceberg recover.

Expansion of IAPT beyond its remit of depression and anxiety
disorders should be halted, until it has been demonstrated
that it adequately performs its’ core task.
c) IAPT should not be allowed to expand its work, as it
is doing to medically unexplained symptoms and long-term
physical conditions until it has at least demonstrated that 50%
of patients with depression and the anxiety disorders lose their
diagnostic status for a period of at least 8 weeks.
e) Outcomes should be adopted that make sense to the
patient such as no longer suffering from the disorder for an
extended period of time. Point assessments using psychometric
tests should be seen as an unreliable metric for discharge.

http://mhfmjournal.com/pdf/ensuring-IAPT-makes-a-real-world-difference.pdf

see also
£3 Billion Spent On Talking Therapies For No Clear Benefit
http://www.cbtwatch.com/3-billion-spent-on-talking-therapies-for-no-clear-benefit/
 
IAPT claims a 50% recovery rate
I find it important to note that this was a target, is merely claimed as such because that was the target and not meeting the target would have been so bad for the continuation of the program that it's seen as preferable to lie.

So in essence IAPT has been "claiming" 50% "recovery" rates since before it even started. Just like PACE, where the numbers were fudged to align better with the initial target. There is no substance behind the claims of 50% recovery, it's merely aspirational and has pretty much been debunked.
Expansion of IAPT beyond its remit of depression and anxiety disorders should be halted, until it has been demonstrated that it adequately performs its’ core task
That would be devastating to the BPS model.
IAPT should not be allowed to expand its work, as it is doing to medically unexplained symptoms and long-term physical conditions until it has at least demonstrated that 50% of patients with depression and the anxiety disorders lose their diagnostic status for a period of at least 8 weeks
That would be extremely devastating to the BPS model.
Outcomes should be adopted that make sense to the patient such as no longer suffering from the disorder for an extended period of time. Point assessments using psychometric tests should be seen as an unreliable metric for discharge
That would be the end of the BPS model.

Though louder for those in the back:
Point assessments using psychometric tests should be seen as an unreliable metric for discharge
That would be a damning indictment of the entire BPS model and mark it down as a failure. Which it has been.

This is a very relevant observation:
But completion of a questionnaire for a therapist introduces demand characteristics - a wish to please therapist and to feel time has not been wasted, resulting in a possible artificial lowering of post-test scores. Further patients tend to present at their worst, often in crisis and are therefore likely to regress to the mean with the passage of time.
Stuff we have been saying for years but still feels good to see others realizing:
These psychological therapies are not however the mainstay of IAPT’s service provision. Indeed, in IAPT there is no attempt to gauge treatment fidelity, that is whether appropriate treatment targets have been identified and a matching treatment strategy deployed.
This would be devastating to Sharpe's career:
Serfaty [10] compared the efficacy of CBT delivered by High Intensity IAPT therapists with treatment as usual in patients with advanced cancer and found no difference. These authors concluded that IAPT-delivered CBT should not be considered as a first-line treatment for depression in advanced cancer and that the use of such services in patients with severe comorbid illness deserved careful scrutiny.
and literally dismantles his own argument about us being irrational since CBT is used to "help" cancer patients. Doing random nonsense doesn't formally count as helping, and anyway "helping" is literally the main argument for alternative medicine, without evidence it does not matter.

I hope people look at this and compare to the the US initiative called No Child Left Behind, enacted in the early 2000's with the aim of improving education by obsessively focusing on standardized testing and punitive measures for low-performing schools, which in practice lead to many of those schools outright cheating, with the teachers' and administration's oversight, to meet the arbitrary targets. People were essentially placed in situations in which they either lied or lost their jobs. Exactly what has been found in IAPT. Brilliant.

This program is widely seen as a failure and largely for the same reasons: ignorance and wishful thinking. Except IAPT is worse in every respect and was so obviously pie-in-the-sky. And as with every disaster there will be clean-up costs.
 
What worries me about Scott's analysis is that it seems to be the same as that given to me by Simon Wessely - that sausage machine CBT is not as good as proper CBT.

That is not a particularly helpful message for PWME.

The figures that Scott extracts may be helpful but ...
 
What worries me about Scott's analysis is that it seems to be the same as that given to me by Simon Wessely - that sausage machine CBT is not as good as proper CBT.

Do you know, is there anything very precisely different that separates the sausage machine variety from 'proper' CBT. I expect training would be one feature but how do they view this training or other differences so as to cause 'proper' CBT to be effective?

Or to put it differently perhaps, what is it about the sausage machine that is deficient in SW POV?
 
Do you know, is there anything very precisely different that separates the sausage machine variety from 'proper' CBT. I expect training would be one feature but how do they view this training or other differences so as to cause 'proper' CBT to be effective?

Or to put it differently perhaps, what is it about the sausage machine that is deficient in SW POV?

You would need to ask them but the impression I got was that IAPT was bypassing proper training in the wisdom of true psychotherapy. I guess it is all to do with a delusion of wisdom and importance and the desire to be seen as a high priest who can anoint the diligent and faithful. In simpler terms you and your buddies might be out of a job.

To be fair, it might be reasonable to criticise a service that uses people with three weeks training rather than three years. But not because of learning deep theories in three years, just getting some experience of people's problems.
 
that sausage machine CBT is not as good as proper CBT.
Which is where the old military saying that no battle plan survives contact with the enemy comes into play. That and the fact that his "highly trained" PACE therapists made no better than some rando with a CBT certificate anyway.

The only way IAPT could ever have unfolded was through fast-food type, there literally never was any other possibility (well, other than even worse fast-food). It was pie-in-the-sky and never made economic sense.

Reminds me of Edison's plan for a DC electrical grid. It never made any sense either, was a disaster in small-scale installations and would have been an even bigger disaster in full implementation, massively more expensive than AC and worse on every level. But it was his thing and that's all that mattered to him, he had an ideological battle to win with Westinghouse and Tesla and tiny details like entire towns engulfed in flames because of frequent electrical fires was just a small price to pay for civilization, even if it was massively more expensive and less reliable than the alternative.

Whether Wessely of any of the other IAPT architects did understand or not the most predictable thing about their big idea makes no difference, there never was any other possible outcome.
 
Do you know, is there anything very precisely different that separates the sausage machine variety from 'proper' CBT. I expect training would be one feature but how do they view this training or other differences so as to cause 'proper' CBT to be effective?

Or to put it differently perhaps, what is it about the sausage machine that is deficient in SW POV?
Wessely avoids precision at all costs, so you'll be very lucky to get a clear answer to that. The difference between the sausage machine and proper CBT is [insert self-deprecating bonhomie smokescreen here].
 
This is from the NHS Recovery Rate page:

This indicator shows how many people have shown a real movement in symptoms large enough to warrant the judgement that the person has recovered.
Recovery is measured according to a person's score on the PHQ-9 and ADSM questionnaires at the start and end of treatment. Each questionnaire has a clinical cut-off point, which is the score at which a person's symptoms are considered a clinically significant case.

For people who initially score above this "caseness" point on one or more of the questionnaires, recovery is achieved if they then score below this threshold on completion of treatment. This change in score shows a reduction in symptoms to a point where their condition is no longer considered a clinical case.

This indicator shows the proportion of people who completed a course of treatment who were considered to be a clinical case at the start of treatment (above caseness), and whose symptoms had improved to below the clinical threshold upon completing treatment.
This indicator differs from the reliable recovery indicator as "recovery" in this case simply means a change in symptoms from above to below the threshold for being considered a clinical case, regardless of the actual amount of change this represents. For example, someone who initially scored just above the clinical threshold for depression on the PHQ-9 at the start of treatment, and just below this threshold at the end of treatment, will have "recovered" in the sense of no longer having clinically significant symptoms.
However, such small changes in questionnaire scores may not be statistically reliable. Recovery is judged to occur if patients fall below the clinical/non-clinical cut-off for BOTH anxiety and depression at the end of treatment.

https://www.nhs.uk/Scorecard/Pages/IndicatorFacts.aspx?MetricId=6228
 
Uuuuuh....
Interpretation of Results


This indicator is not scored, as many factors such as casemix can affect a service's recovery rate. The term casemix is applied to reflect the complexity (mix) of the care provided to a patient.

Numerator: number of people recovered
Denominator: number of people completing a course of treatment – number of people not initially at caseness
So it's a metric that isn't scored, therefore not quantifiable, but it has a target of 50%. OK. I guess half is more feeling than number in this context.

It looks like the data is rolled into this report: https://digital.nhs.uk/data-and-inf...hub/mental-health-services-monthly-statistics, but it seems only to be a record of people going into the service, not about outcomes, which aren't even scored. And since we know people in the IAPT services have reported being told to lie and exaggerate outcomes...

So it seems to follow the same approach as with CBT/GET in ME: you completed the treatment, that means you're "recovered" (which we can't measure), now move along.
 
Last edited:
Rely Solely On A Self-Report Measure To Hike Up Funding and Fudge Outcomes
Jan 10 2020
the routine audit of mental health services such as IAPT, is based on client self-report measures such as the PHQ9. This carries the implicit assumption that the cut offs by themselves meaningfully distinguish cases from non-cases. Correspondence in this months British Journal of Psychiatry highlights how misleading reliance on a single self-report measure can be.

One study using this methodology claimed two fifths of 11-15 year olds had mental health problems but when in another study assessment was conducted using standardised diagnostic interviews and diagnostic criteria the figure was just 13.6%!. doi:10.1192/bjp.2019.225

Whilst claims of high prevalence rates might be good for funding purposes and placing mental health on the public agenda there is no real world change for clients, the powerholders are the only beneficiaries.

full post here
http://www.cbtwatch.com/rely-solely-on-a-self-report-measure-to-hike-up-funding-and-fudge-outcomes/
 
Back
Top Bottom