Transdiagnostic therapy for persistent physical symptoms: A mediation analysis of the PRINCE secondary trial 2022 Chalder et al

Abstract
The PRINCE secondary trial did not find any evidence that transdiagnostic cognitive behavioural therapy (TDT-CBT) plus standard medical care (SMC) was more efficacious than SMC for patients with Persistent Physical Symptoms (PPS) for the primary outcome Work and Social Adjustment Scale (WSAS) at final follow-up (52 weeks). There was a significant treatment effect for TDT-CBT plus CBT compared with SMC for two secondary outcomes: WSAS at the end of active treatment (20 weeks) and symptom severity (Patient Health Questionnaire, PHQ-15) at 52 weeks.

To understand mechanisms that lead to effects of TDT-CBT plus SMC versus SMC we performed a planned secondary mediation analysis. We investigated whether TDT-CBT treatment effects on these two secondary outcomes at the end of the treatment could be explained by effects on variables that were targeted by TDT-CBT during the initial phase of treatment. We pre-specified mediator variables measured at mid-treatment (9 weeks).

Reductions in catastrophising and symptom focusing were the strongest mediators of TDT-CBT treatment effects on WSAS at the end of treatment. Improvements in symptom focusing also mediated the effect of TDT-CBT on PHQ-15.

Future developments of the TDT-CBT intervention could benefit from targeting these mediators.

https://www.sciencedirect.com/science/article/pii/S0005796722001954?dgcid=rss_sd_all

 

Who forced Chalder & co to have that as the first sentence?!?!

 

IMO, to avoid conflict or an uncomfortable scenario, people agree with therapists: "Yes, I feel more positive, yes, my symptoms aren't such a big deal." But, what about objective outcomes?

 

https://www.s4me.info/threads/princ...physical-symptoms-2021-tack-and-tuller.21930/

?

 

This is a completely illegitimate way of doing things. Medicine has lots its damn mind, there is a growing disconnect between reality and whatever the hell this is all about. It's beyond clear that non-controlled randomized trials are nearly useless at giving useful answers without an actual theoretical basis. They can validate some straightforward questions but as fishing expeditions they are the equivalent of fishes-in-a-barrel, whatever one wants to find they will find it because it was put there to be found. The bias level is extreme.

Even controlled trials are a terrible form of evidence since they're only as good as how good the control is and it never even comes close to it when there is a specific and objective thing that can be measured. There are simply too many things that need to align perfectly, without any way of knowing whether they did, that basically amount to movie-plot level of "this has to work, we got nothing else".

Otherwise all this is is bias with a lot of obfuscation. It's as arbitrary as any political ideology and leaves everything down to rhetoric and politics.

 

IMO, any trial employing psychological therapies, or viewing issues and items to do with the human psyche, stands a good chance of being fraught with bias. Bias that cannot be satisfactorily controlled for.

Humans are not black boxes. There are too many variables to control for in psychological studies.

 

The disturbing thing here is 'transdiagnostic' - so great they are keeping pushing at this idea of having 'any old Bob' taking part and all this 'must have PEM' and criteria that can be questioned to check the sample has the disease you claim you are studying no longer an issue. Since when did anything work for people without diagnosing the cause?

The 'baseline demographics' not these people came from 5 different clinics from cardiology to gastro. And for some reason intrusively asked people whether they owned or rented their house and education level. Why is it OK to do baseline but not report the 'N' for each measure at each time period particularly the ones they report as being significant e.g. 52 weeks have no 'N'? - all we seem to have is:

"Of the 324 randomised participants 259 (80%) provided WSAS data at 20 weeks and 258 (80%) provided PHQ-15 data. Completion rates of the putative mediators at 9 weeks post randomisation ranged from 75% to 81%."

Their statistical analysis notes: "To deal with missing values in questionnaire items outcome and mediator scales were pro-rated when fewer than 20% of items were missing." - is this acceptable?

"There was a significant treatment effect for TDT-CBT plus CBT compared with SMC for two secondary outcomes: WSAS at the end of active treatment (20 weeks) and symptom severity (Patient Health Questionnaire, PHQ-15) at 52 weeks.

To understand mechanisms that lead to effects of TDT-CBT plus SMC versus SMC we performed a planned secondary mediation analysis."

From a marketing science background I can explain it:

only those who found they had an improvement (and wanted to tell someone but noone probably wants to listen since they got dumped in that bucket) replied to their 52 week questionnaire 30 weeks after follow-up - nevermind anyone completing the treatment in the first place once they realised it was all about changing their answers on the questionnaire.

And we all know why 'at the end of active treatment' (with therapist asking the questions?) 'works' - if we took the questions asked on said WSAS and compared the script people were given for their sessions then perhaps the answers might be in there in the form of 'rewording how they described their symptoms'


Anyway the appendix doc for 'data' has no N values on it, but does give a sense of how this is measure enough variables (9 measures many of which seem to be extensive questionnaires in themselves e.g. PHQ-9, GAD, Acceptance Scale, CBRQ) and 4 different points and maybe you'll find something that creeps into some sort of significance through sheer dice-rolling for the odd six. This meant 57 questions on repetitive scales four times over - 228 questions by week 52.

The sheer amount of questions these people must have answered over the space of this 'trial' is mind-boggling, and how many of these overlap with each other - this is surely attempting to benefit merely from the brainwashing effect of forcing people to repeatedly answer the same questions over and over in order to basically force them into people's thinking. It is utterly inappropriate to call it a measure rather than an answer-changing treatment if they used this many, 57 scaled questions is like a psychological coshing into submission,no way even well people could be 'conscious' and in control of their answers by the end.

The other section of gobsmacking note is the conflicts of interest section. Boy do these 2 authors have a lot of grants, payments, licenses, expenses listed under all of this from sooo many organisations. No wonder they've a 'credit authorship statement' to I assume be clear about dividing up any spoils that relate directly to this piece of 'research', who wrote what, and how it underpins each of these?

 

There is an endless list of competing interests they've had to list in that section to choose from?

 

The repetition illusory truth effect: https://cognitiveresearchjournal.springeropen.com/articles/10.1186/s41235-021-00301-5

The effect of repeated questioning in reliability of children's eyewitness testimony: https://calio.org/wp-content/upload...y-and-consistency-in-eyewitness-testimony.pdf

feedback given after one round of questions can influence the answers given in the next round: http://www.pc.rhul.ac.uk/sites/rheg/wp-content/uploads/2011/05/wysman-et-al-bsl-in-press-17feb14.pdf

surveys with repetitive questions yield bad data: https://www.sciencedaily.com/releases/2022/01/220128153553.htm

I strongly suspect the impact of this would be even greater on those who are cognitively exhausted and feeling a power differential. And I'm certain there are many more if my brain could think of the right words for the search terms

 

In the quote box is table 1 which shows all 57 questions on 9 different scales that participants would have had to slog through. I've removed the last bullet (total score) for brevity, but kept the others as I think it is also important to note that:
- these scales are on different likerts, some are 3 point, some 4, 5 or 7 point scales
- whilst if you design a questionnaire to not be leading you make sure you switch display of the scales around, but if you have lots of them you might not in order to reduce confusion.. this seems the worst of all worlds with all but one of these scales placing the higher choice on likert as 'more problem' except one - the acceptance one (third down).

Almost definitely meaning some congitively fatigued people won't notice higher score is more acceptance, or......... more likely, and naughtily... inferring that accepting your illness isn't a good thing (the old push through mindset again)

 

Spurious Claims For The Impact of Transdiagnostic CBT on Persistent Physical Symptoms

"In a just published study James et al (2022) extol the virtues of their transdiagnostic CBT therapy (plus standard medical care) for Persistent Physical Symptoms.
s. In an earlier study these authors found no difference in outcome at 52 weeks follow up on the Work and Social Adjustment Scale (WSAS: the declared primary outcome measure) between those undergoing the transdiagnostic CBT plus standard medical care and those having the latter alone. Not to be outdone they have gone on a fishing expedition in their latest paper James et al (2022) and focus on two secondary outcome measures, WSAS score at the end of active treatment and the PHQ-15 score at 52 weeks, claiming that there was a significant difference between the transdiagnostic CBT and standard medical care using both these secondary measures."

http://www.cbtwatch.com/spurious-cl...agnostic-cbt-on-persistent-physical-symptoms/

 

More from Mike Scott' blog post, copied in full with the author's permission:

In a just published study James et al (2022) extol the virtues of their transdiagnostic CBT therapy (plus standard medical care) for Persistent Physical Symptoms.

s. In an earlier study these authors found no difference in outcome at 52 weeks follow up on the Work and Social Adjustment Scale (WSAS: the declared primary outcome measure) between those undergoing the transdiagnostic CBT plus standard medical care and those having the latter alone. Not to be outdone they have gone on a fishing expedition in their latest paper James et al (2022) and focus on two secondary outcome measures, WSAS score at the end of active treatment and the PHQ-15 score at 52 weeks, claiming that there was a significant difference between the transdiagnostic CBT and standard medical care using both these secondary measures.

This looks a little fishy as they are the originators of the treatment. No mention of the need for independent replication of their findings. Further the comparison group was standard medical care, but this design does not control for attention or the presentation of a credible treatment rationale. Scepticism seems in order. This intensifies when the outcome measures used are considered.

The WSAS (Mundt et al 2002) was designed to assess the impact of mental health on functioning in 5 domains work, home management, social leisure activities, private leisure activities and close relationships. As such there is clear separation between the causal agent (mental health) and effect (functioning). In the initial validation study the WSAS was administered to a sample of depressed patients and a sample of OCD patients, the WSAS scores correlated with the severity of each disorder.

Mental Health can affect one’s ability to do certain day-to-day tasks in their lives. Please read each item below and respond based on how much your mental health impairs your ability to carry out the activity.

Table

Description automatically generated


If however the term ‘mental health’ in the above table is replaced with “Persistent Physical Symptoms’ the distinction between causal agent and effect is lost. Consider, a person who is terminally ill with cancer or a progressive neurological disorder, they would likely score high in each domain. But it would not be anticipated that any psychological intervention is going to attenuate their physical symptoms, because there is a fusion of physical symptoms and impairment. [In the original validation study the term ‘disorder’ was used because they were known to have either depression or OCD].

The burden of proof is on those who would reword the WSAS so that it related to PPS to demonstrate a meaningful distinction between say fibromyalgia and impairment in functioning or irritable bowel syndrome and functioning or between chronic fatigue syndrome and functioning. The WSAS was not validated for PPS but that has not stopped James et al (2022) using it as their primary outcome measure in their study of transdiagnotic CBT for PPS compared to standard medical care.



The PHQ-15 is a measure of symptom severity and is rated on a 3- point Likert scale. This brief validated measure looks at a range of symptoms and asks patients to report to what extent their symptoms bothered them in the past four weeks. The PHQ-15 contains 15 items and the total score can range from 0 to 30 where a higher score indicates greater symptom severity (Kroenke et al., 2010). The items include several related to pain, (stomach pain, back pain, pain in arms and legs or other joints, headaches, chest pain), fatigue, as well as other symp- toms related to the different systems of the body i.e. shortness of breath, dizziness, bowel symptoms.


But there is no clear mechanism by which a psychological intervention impacts such an array of symptoms. The authors considered 9 possible mediating variables (7 of which were cognitive), but only 2, catastrophising and symptom focusing appeared operative (Type 2 error). Suggesting that the supposed mediating variables may be chance findings. One of the suggested variables related to depression (PHQ-9) and the other anxiety (GAD-7) but for the supposed cognitive mediation model evinced by James et al (2022), they cannot be mediators, depression is known to covary with WSAS as such it cannot be an explanatory variable.

Assessment in the study is entirely by self-report measures but the authors totally ignore that they are subject to demand characteristics e.g wanting to please the therapist by reporting improvement or to convince oneself that time has not been wasted. It would have been more meaningful for an independent assessor to ask since ‘x’ have you felt the same, a little better, a little worse, much better, much worse.

James et al (2022) justify their transdiagnostic therapy on the basis that ‘people with different PPS share some cognitive and behavioural responses to symptoms, including catastrophising, symptom focusing, fear avoidance beliefs, avoidance behaviour and lack of acceptance (Deary, Chalder, & Sharpe, 2007)’. This is a sweeping statement certainly catastrophising and avoidance behaviour have been implicated in the development of chronic pain, but this is not to say that they are germane to the category of PPS (MUS). The use of the term ‘some’ makes the Deary, Chalder, & Sharpe, 2007 model incapable of falsification.

The term PPS as used byJames et al (2022), is a smokescreen for the construct Medically Unexplained Symptoms whose validity Keith Geraghty and I have challenged.

Dr Mike Scott