'Recovery' from chronic fatigue syndrome after treatments given in the PACE trial

[user9876]

One way, but not the only way, to get problems with papers noted is to have letters to the editor published. I would encourage as many people as can, to try this. Even if your particular letter doesn't get published, the letter can be used on other occasions and could, for example, be highlighted on a thread collating all the responses here.

I assume if the editor gets a flood of letters pointing out the most basic mistakes it might also make them think twice about being more careful with any further publications.

 

[alex3619]

Alex, or anyone else, do you have any more good individual examples of this blatant deceptive trick of redefining words, in published psychogenic studies and literature (particularly from the ME/CFS literature).

If I can get say 5 good examples, I will try to get a letter published in a newspaper or science journal. And I may also set up a web site specifically exposing this fraud of deliberately misleading word redefinition that these psychologists have engaged in (I am good at getting web sites to appear in the top position of a Google search).

This is something that I am investigating and will have more to say on in time. However for PACE the two big ones are the obvious ones, "normal" and "recovery". In general though they misrepresent more obvious things, including "there are no abnormalities" or "there are no tests" or emphasizing "deconditioning" when it cannot ever explain sudden improvement or decline in long established patients.

 

[Firestormm]

I agree entirely, while I remain doubtful about CBT/GET being a source of significant harm, and there is some likely benefit to some M.E/CFS patients in some circumstances, the way that the PACE, FINE etc monolith is progressing is IMO profoundly harmful to a science referenced NHS.

Just so everyone is clear there is not a formal process of 'franchising' rather it is an effective franchising within the bastardised NHS 'internal market' process.

Buying into the stats is a part of this 'shadow franchising' although the commissioning level managers will not be required to demonstrate effectiveness of a treatment, that will remain the responsibility of the service level contractee. The local commissioning bodies will have little in way of resources to check effectiveness and will be reliant on the service providers own data (which in the case of a PACE model CBT/GET provision would very likely use the PACE etc, statistical analysis) with the primary arbiters of effectiveness being the NHS Commisioning Board and NICE (don't be fooled by Government claims of power devolving to local commissioners !).

For CBT/GET delivery, unless there is a very dynamic group of local commissioners, the development/expansion within NHS Trust hospitals of Liaison Psychiatry services or seperate private sector contractees, will be the most likely models. The only general limitation to this isthe massive cuts (sorry efficiency gains) the NHS is having to absorb, however this is leading to much more energetic competition so active selling of CBT/GET on the back of PACE is likely to be underway.

IVI

I am aware that within the NHS, 'targets' are driven by 'clinical outcome measures' or the latter is a consumer-friendly term for the former.

Whilst our local ME Service - now newly commissioned and with a more succinct 'business' model - does not prescribe GET and CBT a la PACE manuals - it must surely at some level be able to demonstrate some outcome measure.

I need/want to do some more work on this side of things. The outcomes are not equated with 'cure' or 'recovery' I know that much but there must be something substantive that they work towards even if only as a means of measuring quality of care.

But what can a clinical team do when they are only - on average - able to deliver six-eight sessions per client? Compare that to PACE intensity.

From what I have seen from my own home-delivered Occupational Therapy - there is no attempt (though it is very early days) to complete an assessment or measure outcome.

From what little I have thus far gleaned it is entirely focused on helping me to learn how best to manage my health. To try and tailor advice and support to personal and realistic goals - my own all stemming from a desire to live independently once more but with support.

I cannot yet - more generally - see why Clinical Commissioners would want the PACE recovery data. If you strip even the principles of CBT and GET or Graded Activity from clinical care - what have you got left?

These principals are so ingrained if not employed with precisely those nomens that I have to ask what difference this data will make.

If PACE was seen to have proved these therapies even if delivered solely a la PACE manuals - were ineffective - the ramifications across all of healthcare for all conditions would be... well... dire. Or would it?

This was a model. Using the principals. Applied in such a way that was felt relevant to CFS/ME by those who figured they knew what they were doing.

It could be argued by others outside of PACE that it was less effective than hoped because the model was wrong, not the principal behind the therapies themselves.

 

[Stukindawski]

It could be argued by others outside of PACE that it was less effective than hoped because the model was wrong, not the principal behind the therapies themselves.

It could be argued that pursuing CBT/GET research in an illness without a known cause is like licking a muddy puddle when there's a clean glass of water on the table. I think criticism of the recovery data is definitely better aimed at putting pressure on the MRC to fund investigative types of research.

I wonder if some outcome measures only need to claim to have 'treated' a patient to meet a quota.

I had a day surgery cancelled recently because I pass out when I fast. The admins will not keep me in the system while I navigate my way through the GP and my specialist to pin point what's causing it. I will have to start the surgery referral process from the ground up.once I'm ready - which basically equates to repeating unnecessary appointments. It seems the highest value is placed on emptying queues.

 

[biophile]

It seems apparent that the authors of this PACE study actually want their completely redefined word "recovery" to be misinterpreted. Thus, this is not a mistake by the PACE study authors, but is in fact fraudulent.

In recent correspondence between Professors White/Wessely and the Countess of Mar, White refuted claims that he did nothing to correct the erroneously reporting of "normal range" as "recovery" back in 2011. He pointed out that the difference was clarified in their authors' reply in the Lancet (several months after the fact), while the Countess of Mar later accused him of double-standards (doing nothing to combat the claims of "recovery" in news articles, while rapidly writing in response to David Tuller's article which questioned the generalizability of the results).

Now, this so-called "normal range" in fatigue and physical function, which overlaps with trial eligibility for "disabling fatigue", is officially presented as a "recovery" in the latest paper anyway, with additional optional criteria bolted on. I have to agree with you that the confusion about recovery is not a mistake. The unnecessary ambiguity surrounding "recovery" is serving a purpose, just as the ambiguity surrounding use of the word "functional" is serving a purpose (noted in Wessely's paper).

So far the general defense of the paper can be simplistically paraphrased as: "Give the authors some slack, recovery is very difficult to define in CFS, they have done their best in a difficult situation, and CBT and GET (the only effective therapies known) outperformed APT and SMC anyway, so all criticisms are irrelevant."

I do not think they deserve much slack. The redefined recovery falls far short of their original conceptualization of it, does not include what many patients would deem to be important, and is contradicted by a bunch of other more objective outcomes (eg CBT and GET are being praised as leading to recovery, while the group average in welfare and insurance payments actually increased in these groups, etc). Furthermore, the redefined recovery is based almost exclusively not on major improvements but thresholds which require almost no improvement and which either overlap with or are on the border of what was defined as "disabling fatigue". When it comes to fatigue and physical function in particular, even the weakest thresholds in the original protocol, i.e. those used for a "positive outcome", are more strict than the strictest ones used in the latest redefinition of complete "recovery". It is rather questionable to call it "comprehensive and conservative".

So thoughtful of them to put us in the "depression" section of their website.

<sarcasm>Stop hating on mental illness and perpetuating the stigma! Did you not know that depression is real too and that mind-body dualism is unhelpful? Stop denying your psychological problems, how can you recover from a problem without first acknowledging it?</sarcasm>

It could be argued by others outside of PACE that it was less effective than hoped because the model was wrong, not the principal behind the therapies themselves.

Please elaborate the differences between the principle and the model.

I do have an open question to all you other "spicy" recalcitrants ...

What would the methodological requirements be for a trial on CBT/GET in order for it to be respected and lead to a reconsideration for the role of these therapies?

I guess that the answers will relate to using case definitions which better reflect them, stricter definitions for improvement and recovery based on actual healthy norms, consideration for other measures (employment, welfare, 6MWD, exercise test, welfare, actometer, etc), adequately defining safety, even designed and conducted and overseen by an independent authority which has not built their careers on CBT and GET (due to the lack of trust, and concerns about spin?).

 

[Stukindawski]

I do have an open question to all you spicy recalcitrants:

What would the methodological requirements be for a trial on CBT/GET in order for it to be respected and lead to a reconsideration for the role of these therapies?

I will guess that the answers will relate to case definitions which better reflect them, stricter definitions for improvement and recovery based on actual healthy norms, consideration for other measures (employment, welfare, 6MWD, exercise test, welfare, actometer, etc), designed and overseen by an independent authority which has not built their careers on CBT and GET (due to the lack of trust?).

I think there is always going to be something of a methodological black hole in CBT/GET research unless the effect of the treatment is substantial.

But I'd really like to see something massive.

OI symptom tracking
POTS tracking
VO2 max repeat exercise testing
Inflammatory response post exercise tests
Mental acuity testing.
3-5 years worth of follow up.
Employment/Welfare changes are a must (of course, a factor like the ESA/PIP debacle could actually impact research at this time)

I'd favour patient sign & symptom profiling over a strict definition. If there are responders it'd be interesting to see what symptoms they report and what signs are detected.

I could spend hours really thinking about this, but my head hurts, so I wont

 

[In Vitro Infidelium]

I am aware that within the NHS, 'targets' are driven by 'clinical outcome measures' or the latter is a consumer-friendly term for the former.

Whilst our local ME Service - now newly commissioned and with a more succinct 'business' model - does not prescribe GET and CBT a la PACE manuals - it must surely at some level be able to demonstrate some outcome measure.

I need/want to do some more work on this side of things. The outcomes are not equated with 'cure' or 'recovery' I know that much but there must be something substantive that they work towards even if only as a means of measuring quality of care.

But what can a clinical team do when they are only - on average - able to deliver six-eight sessions per client? Compare that to PACE intensity.

From what I have seen from my own home-delivered Occupational Therapy - there is no attempt (though it is very early days) to complete an assessment or measure outcome.

From what little I have thus far gleaned it is entirely focused on helping me to learn how best to manage my health. To try and tailor advice and support to personal and realistic goals - my own all stemming from a desire to live independently once more but with support.

I cannot yet - more generally - see why Clinical Commissioners would want the PACE recovery data. If you strip even the principles of CBT and GET or Graded Activity from clinical care - what have you got left?

These principals are so ingrained if not employed with precisely those nomens that I have to ask what difference this data will make.

My guess is that PACE - and all that has flowed from it, has been structured/restructured to have 'confluence' with these: http://www.iapt.nhs.uk/silo/files/iapt-data-handbook-v2.pdf and http://www.iapt.nhs.uk/silo/files/iapt-data-handbook-appendicies-v2.pdf - main page: Measuring Patient Outcomes - http://www.iapt.nhs.uk/data/measuring-outcomes/

Clinical Commissioners (more precisely the managers employed by the Commissioning Boards, as few Board members will have a day to day handle on the processes in operation) will not operate PACE derived data, but the data, as constucted in the current article will serve to provide headlines to 'sell' PACE defined services to the Commissioning Boards. Additionally although the IAPT set up closely defines what the CCBs have to report M.E/CFS could be in noman's land in that, although a service may be delivered from a Liaison Psychiatry base, there may not be a requirement to define M.E/CFS under the IAPT reporting criteria (not a mental health illness) but also it doesn't fit under any of the other described Clinical Datasets http://www.commissioningboard.nhs.uk/files/2012/12/clinical-datasets.pdf . In that case CCB's will (as far as I can tell) have freedom to choose appropriate metrics for which the PACE work could conveniently fill the gap.

PACE was seen to have proved these therapies even if delivered solely a la PACE manuals - were ineffective - the ramifications across all of healthcare for all conditions would be... well... dire. Or would it?

This was a model. Using the principals. Applied in such a way that was felt relevant to CFS/ME by those who figured they knew what they were doing.

It could be argued by others outside of PACE that it was less effective than hoped because the model was wrong, not the principal behind the therapies themselves.

Yes indeed. The test of plausibility is wholly missing from the 'evidence based' construct that lies behind the clinical data set system - it will be down to the CCBs to make plausibility judgements (i.e just because the numbers show 'X' doesn't mean that 'X' exists beyond a statistical artefact) when commissioning a service. How many will pursue that diligently, or will instead merely default to the safety of reported numbers remains to be seen.

Advocacy can clearly impact the process ( "Whilst our local ME Service - now newly commissioned and with a more succinct 'business' model " !!) but it will need to advanced pragmatically - the position of 'no psychiatry involvement under any circumstances' provides no basis for negotiation or the meeting of CCB level perspectives of M.E/CFS. The alternative is to leave the field open for PACE defined services to dominate the contractual landscape.

IVI

 

[Firestormm]

Speaking solely in regard to my local service. There was no 'model' in place previously for the service. No structure. The number of patients being seen was overwhelming, and still may prove to be. However, aside from the actual 'product' delivered, the clinicians are now delivering a structured service model.

The commissioners know in advance how much money the model will require. The deliverers know how many average consultations they can deliver etc. What is not part of this model is the actual product delivered and how the effectiveness of that product is measured. I don't believe that PACE will change anything in that respect at the delivery level.

Regular reviews will now take place to ensure that the model is appropriate and as a patient group we will be trying to ensure that patients are receiving an adequate, appropriate and timely service.

Even with this paper we are still left largely with the status-quo we had before. The claims made in NICE about CBT and GET being (from memory) the only proven methods of help remain.

Until such time as something comes along that can 'beat' the now published 'recovery' data - despite the valid criticisms of this paper and of the Trial itself - nothing has changed.

Even if a drug is trialled and shown effective and applicable for all or some with our diagnosis I do not see a time when CBT or GET will disappear from the NICE Guideline.

GET and CBT are 'safe' in so far as the regulators/advisor's are concerned and from personal experience GET is more akin with Graded Activity than a programme of progressively increased - regardless of symptom exacerbation - physical exercise.

GET and CBT will still be seen as more effective than 'pacing' or the APT that was applied in the Trial. If you can do something structured it is better than doing very little - which is how pacing is viewed.

However, again speaking personally, local delivery does include pacing. What I have found is that GET/CBT and pacing are very often delivered as a package as part of your tailored approach to a sustainable 'base-line'.

Not everyone needs or wants this level of help. Not everyone can get it of course.

I suppose the concern is that PACE and it's manuals will (perhaps are - I forget) being used verbatim in clinical service delivery.

On the plus side, we now know that even the authors have had to admit to the less than what I would deem 'moderate' effectiveness of their Trial.

At 22% clinical deliverers of the manuals will need to be far more upfront about their proposed regimen. No more talk of 'cure'. Or if there is then patients can throw this paper at them

 

[Dolphin]

I think there is always going to be something of a methodological black hole in CBT/GET research unless the effect of the treatment is substantial.

But I'd really like to see something massive.

OI symptom tracking
POTS tracking
VO2 max repeat exercise testing
Inflammatory response post exercise tests
Mental acuity testing.
3-5 years worth of follow up.
Employment/Welfare changes are a must (of course, a factor like the ESA/PIP debacle could actually impact research at this time)

I'd favour patient sign & symptom profiling over a strict definition. If there are responders it'd be interesting to see what symptoms they report and what signs are detected.

I could spend hours really thinking about this, but my head hurts, so I wont

A list like that could make a nice letter to the editor, if anyone was so inclined.

 

[Dolphin]

Until such time as something comes along that can 'beat' the now published 'recovery' data - despite the valid criticisms of this paper and of the Trial itself - nothing has changed.

Which is a very high bar to set for blinded trials which will probably use more rigorous definitions of recovery.

 

[SOC]

No: once people have finally cottoned on to the truth, the PACE team will take their skills into investment banking and cure the ills of the economy.

Do ya think they'll help us fix the economy by thinking it right? Perhaps the government just has false poverty beliefs and could actually pay for everything society needs without raising taxes. All we have to do is gather all our politicians together for a PACE-style seminar to cure their false belief in economic problems and everything will be fine.

 

[Dolphin]

At 22% clinical deliverers of the manuals will need to be far more upfront about their proposed regimen. No more talk of 'cure'. Or if there is then patients can throw this paper at them

Not sure why you say that. They found 21-22% recovery using their most strict criteria for recovery. This in a £5m-pound MRC/DoH/ScottishChiefScientistOffice/DWP-funded multi-centre trial.

 

[Stukindawski]

A list like that could make a nice letter to the editor, if anyone was so inclined.

Hmm, I've never actually written a letter to a journal before. I've drafted a letter, if you think it's good enough, let me know how I would go about sending it for publication:-


As a patient, I have yet to see substantial data that would make me feel confident in committing to such a treatment for the third time. There is a large amount of obtainable data in respect of the growing body of research in ME/CFS, which could make the case for this treatment much more compelling:-

Autonomic dysfunction tracking, including Orthostatic Intolerance and POTS (Based on Professor Julia Newton's work) [1][2]
VO2 max repeat exercise testing (Based on work at the Pacific Fatigue Laboratory) [3]
Inflammatory response testing following exercise (Based on Dr Alan Light's work) [4]
Mental acuity testing (Has the classic symptom of Brain fog been abated?)
A focus on welfare and employment changes which take note of external factors, I.E the migration from IB to ESA and DLA to PIP and the current state of the employment market (some attention to the availability of less demanding work and the type of work taken up by patients following treatment would also be valuable data).

I'd expect to see at least 3-5 years worth of extensive follow up to account for symptom fluctuation and ensure that the treatment does not suffer from a 'plateau effect'. Some reconditioning may be of real benefit, which could later be lost when the patient's limit is reached (I.E relapse, being one of the most commonly reported patient experiences).

Sign and symptom profiling in each patient, as opposed to using stricter definitions might be key in identifying likely responders. I have yet to see solid evidence of homogeneity in this patient population which would be a reasonable competing explanation for the rather modest response reported by this paper.

[1] http://www.sciencedirect.com/science/article/pii/S2213158212000484
[2] http://onlinelibrary.wiley.com/doi/10.1111/joim.12022/abstract
[3] http://www.cfids-cab.org/MESA/VanNess.pdf
[4] http://www.ncbi.nlm.nih.gov/pubmed/19647494

 

[Dolphin]

Hmm, I've never actually written a letter to a journal before. I've drafted a letter, if you think it's good enough, let me know how I would go about sending it for publication:-


As a patient, I have yet to see substantial data that would make me feel confident in committing to such a treatment for the third time. There is a large amount of obtainable data in respect of the growing body of research in ME/CFS, which could make the case for this treatment much more compelling:-

Autonomic dysfunction tracking, including Orthostatic Intolerance and POTS (Based on Professor Julia Newton's work) [1][2]
VO2 max repeat exercise testing (Based on work at the Pacific Fatigue Laboratory) [3]
Inflammatory response testing following exercise (Based on Dr Alan Light's work) [4]
Mental acuity testing (Has the classic symptom of Brain fog been abated?)
A focus on welfare and employment changes which take note of external factors, I.E the migration from IB to ESA and DLA to PIP and the current state of the employment market (some attention to the availability of less demanding work and the type of work taken up by patients following treatment would also be valuable data).

I'd expect to see at least 3-5 years worth of extensive follow up to account for symptom fluctuation and ensure that the treatment does not suffer from a 'plateau effect'. Some reconditioning may be of real benefit, which could later be lost when the patient's limit is reached (I.E relapse, being one of the most commonly reported patient experiences).

Sign and symptom profiling in each patient, as opposed to using stricter definitions might be key in identifying likely responders. I have yet to see solid evidence of homogeneity in this patient population which would be a reasonable competing explanation for the rather modest response reported by this paper.

[1] http://www.sciencedirect.com/science/article/pii/S2213158212000484
[2] http://onlinelibrary.wiley.com/doi/10.1111/joim.12022/abstract
[3] http://www.cfids-cab.org/MESA/VanNess.pdf
[4] http://www.ncbi.nlm.nih.gov/pubmed/19647494

Well done on that. I've sent you a private message on it.

 

[Firestormm]

Not sure why you say that. They found 21-22% recovery using their most strict criteria for recovery. This in a £5m-pound MRC/DoH/ScottishChiefScientistOffice/DWP-funded multi-centre trial.

Delivers of this/these therapies will I think have to be more realistic. 'There is a chance of some recovery' and not 'This will cure you dude!' NICE of course never spoke of cures but anecdotal evidence from patients attending assessments have inferred that some clinicians are 'selling' CBT and GET as 'cures'. I no longer believe that they can do this. Some might try but informed patients will at least now know the chances are scant. And that whilst CBT and/or GET might help you with this episode you may well need them again for the next one.

22% may indeed be setting the bar high but perhaps only for other treatments addressing the whole disease - and how realistic are such treatments likely to be (outside of psychology) for such a heterogeneous mishmash of patients? I think what we are seeing are more focused efforts addressing either specific symptoms, aspects of the disease, or sub-categories of patients. Indeed, I think it perfectly reasonable to consider CBT and GET as pertinent perhaps to only some people with the disease or indeed some (associated?) symptoms. Neither therapy is recommended for 'severe' patients for example. And clinical delivery varies in terms of how they are delivered assuming of course that they are. Perhaps it is more the principals of CBT and GET that are built into the delivery. Not saying it hasn't happened but I have yet to hear of anyone using the PACE manuals as delivery mechanisms outside of the Trial.

 

[Firestormm]

From the paper:

Creating criteria for recovery from domains that are measured on a continuum requires the setting of operational thresholds based on population studies or trial eligibility criteria (Powell et al. 2004 ; Knoop et al. 2007 ; Malouff et al. 2008).

In this context it is important to note that recovery does not mean being free of all symptoms; population studies show that the average person in the UK reports a mean of four symptoms in any 2-week period (McAteer et al. 2011).

The three most common symptoms reported were fatigue, headache and joint pain; symptoms consistent with CFS (McAteer et al. 2011).

Anyone care to explain? Should I feel enraged or is this reasonable? I haven't read the papers cited. Thanks.

 

[Dolphin]

Anyone care to explain? Should I feel enraged or is this reasonable? I haven't read the papers cited. Thanks.

I doubt it's reasonable but haven't read the paper referenced yet (it's free http://europepmc.org/articles/PMC3020067). It would be interesting to compare the average number of symptoms recorded by the "recovered" CFS patients versus the average number of these symptoms reported by the public. My guess is the figure for the CFS patients would be higher esp. if they were age-matched (and didn't involve those with any of the exclusions, although the paper might not have those details). Also, it would be interesting to check the wordings of the questions to see if they are similar. Also, CBT and GET patients have been told the symptoms are normal so might potentially be less likely to report them.

Similar to the first point I make above, I imagine it's the case for a lot of the other measures used i.e. if they listed the mean SF-36 PF scores for the "recovered" patients, it would be lower than the mean for the general population, which would leading to rejecting the hypothesis that the "recovered" patients are like the general population, which is what would expect from a recovered group.

 

[biophile]

In this context it is important to note that recovery does not mean being free of all symptoms; population studies show that the average person in the UK reports a mean of four symptoms in any 2-week period (McAteer et al. 2011).

The three most common symptoms reported were fatigue, headache and joint pain; symptoms consistent with CFS (McAteer et al. 2011).

Anyone care to explain? Should I feel enraged or is this reasonable? I haven't read the papers cited. Thanks.

I had a very brief look : http://europepmc.org/articles/PMC3020067 .

Surveys sent to people drawn from general medical practices. Low response rate of 1/3. 45% had a chronic condition.

For the whole population, overall mean of 3.66 (SD = 3.47) symptoms over the previous 2 weeks, but mean number of symptoms was higher in those with a chronic condition than those without. Notice how White et al rounded up the symptom count from the whole population figures which included 45% of those with a chronic condition. Many conditions are excluded for a CFS diagnosis, and CFS patients would have had a recent medical assessment. The authors state that "Presence of a chronic condition, age, and employment status were the three factors most commonly associated with the 2-week prevalence of different symptoms."

 

[Dolphin]

I had a very brief look : http://europepmc.org/articles/PMC3020067 .

Surveys sent to people drawn from general medical practices. Low response rate of 1/3. 45% had a chronic condition.

For the whole population, overall mean of 3.66 (SD = 3.47) symptoms over the previous 2 weeks, but mean number of symptoms was higher in those with a chronic condition than those without. Notice how White et al rounded up the symptom count from the whole population figures which included 45% of those with a chronic condition. Many conditions are excluded for a CFS diagnosis. The authors state that "Presence of a chronic condition, age, and employment status were the three factors most commonly associated with the 2-week prevalence of different symptoms."

Out of 25 symptoms from what I see at a quick glance.

 

[biophile]

Return to work is not, however, an appropriate measure of recovery if the participant was not working before their illness and is influenced by other factors such as the job market.

This argument sounds like an avoidance of dealing with the implications of the poor employment and welfare data that was published last year. The average age of CFS onset in the PACE Trial participants was roughly about 35 years old and about 3 years before recruitment, so the vast majority of these people should have been employed before illness.

On the other hand, the "job market" argument is worth considering.

We recruited 641 participants from consecutive new out patients attending six specialist chronic fatigue syndrome clinics in the UK National Health Service between March 18, 2005, and Nov 28, 2008, and completed outcome data collection in January, 2010.

UK unemployment during most of the 2000's was generally low and relatively stable. The economic downturn started in June 2008. The 3rd patient newsletter gives a time graph of the cumulative recruitment into the trial, which is helpful for examining this job market claim : http://www.pacetrial.org/docs/participantsnewsletter3.pdf .

Assuming that there is about a year plus a few months at most between recruitment and collecting followup data, it appears that about half of the participants were recruited and followed up during the "good times".

Even if the bad times affected the data for some of the remaining participants, surely if the "completely recovered" rates were 22% for CBT/GET vs 7% for SMC, there should still be some positive effect on employment and welfare outcomes, unless the study was statistically underpowered?