A cost effectiveness of the PACE trial

[Bob]

Has anyone got any insight into how the first bullet-point in the following quote fits in with all the reports that CBT and GET were the most cost effective?:

What were the basic results?

The researchers report the healthcare and societal findings of this study separately. The findings include the following:

• patients receiving specialist medical care alone had significantly lower healthcare costs than those receiving CBT, graded exercise therapy and adaptive pacing therapy
• the healthcare costs per QALY gained for CBT and graded exercise therapy were likely to be lower than the £30,000 threshold that, in England, has become a standard for the reasonable use of NHS resources, indicating they were cost-effective
• the healthcare costs per QALY for adaptive pacing therapy were found to be unlikely to be lower than this threshold and, therefore, thought unlikely to be a good use of NHS resources
• when all three were compared to specialist medical care alone, the incremental healthcare cost per QALY was £18,374 for CBT, £23,615 for graded exercise therapy and £55,235 for adaptive pacing therapy, meaning that CBT appeared to be the best value for money

http://www.nhs.uk/news/2012/08augus...onic-fatigue-syndrome-not-cost-effective.aspx

 

[Dolphin]

Has anyone got any insight into how the first bullet-point in the following quote fits in with all the reports that CBT and GET were the most cost effective?:
patients receiving specialist medical care alone had significantly lower healthcare costs than those receiving CBT, graded exercise therapy and adaptive pacing therapy

One can't bash them for that: any extra treatment is going to cost more money (all other things being equal i.e. it doesn't decrease other healthcare costs during the course of using it).

What they then looked at was how much it cost for any health gains got (in terms of QALYs, improvement of 2 or more in Chalder Fatigue score (likert scoring), improvement of 8 or more in SF-36 physical functioning).

 

[Esther12]

Ok, don't think Barts should get any credit for this. There is a woman on MEA FB complaining how her service in Essex won't write any letters for benefit applications. They claim it's some sort of "conflict of interest".

Some posts from her from the thread:

• 
  She mentions somewhere she may become homeless because of the policy.

I've heard the same about the East Sussex CFS service too.

The one useful thing they could do....!

 

[Purple]

Doctor Christian Jessen, a UK TV doctor who presents Embarrassing Bodies and Supersize v Superskinny, has tweeted twice about this study (he is a prolific tweeter with a large Twitter following of about 128k)

Friday 3 August 6:49 PM
Dr Christian Jessen @DoctorChristian

Graded exercise and CBT have been found most cost effective treatments for ME according to a new study. Are they always available on NHS?

Saturday 4 August 12:29 PM
Dr Christian Jessen @DoctorChristian

Many dispute the value of CBT and graded exercise recently 'found' to be cost effective for ME. Heres why:http://www.investinme.org/Article 432 - PACE Trial Update July 2012.htm

His Twitter page: https://twitter.com/DoctorChristian
And his Facebook page where his tweets also appear: https://www.facebook.com/pages/Doctor-Christian/112190008831149

Information about Doctor Christian:
http://en.wikipedia.org/wiki/Christian_Jessen
http://www.drchristianjessen.com/

 

[Bob]

Misleading quotes from the MRC press release:
http://www.meassociation.org.uk/?p=12343


"Two treatments found previously to be the most effective for patients with Chronic Fatigue Syndrome or Myalgic Encephalomyelitis (CFS/ME) have now been found to be the most cost-effective treatments evaluated in a large clinical trial."

The press release starts off with a load of nonsense. If they are going to treat the SMC control group as an extra therapy then, in fact, SMC was more effective than CBT and GET, by all primary measures. The PACE Trial wasn't set up to assess SMC as a separate therapy.


"Adaptive pacing therapy (APT) was not cost-effective."

This assertion is not strictly misleading in itself, untill they start making conclusions about 'pacing'.
The cost effectiveness of APT tells us nothing about the benefits of teaching patients about pacing.
The reason that APT was not cost-effective was because it was a failed novel attempt at a 'therapy'.


"There is now a strong case for the NHS to invest in providing these therapies."

If the therapies don't work for 87% of patients, and CBT is clinically ineffective at reducing physical disability, then I don't understand what the 'strong' case is.
And we still don't have the deterioration rates. Once those have been published, then the therapies will probably look even weaker than they do now.


"Professor Michael Sharpe from Oxford University and a co-author of the paper said:
In the PACE trial we found that the rehabilitative treatments CBT and GET improve the fatigue and disability of people with CFS/ME."

This is misleading, because CBT does not improve disability of people with CFS/ME. It failed to meet the threshold of a clinical useful outcome for SF-36 PF.


"These promising findings demonstrate how MRC funding can help to identify treatments that are not only effective, but are financially viable for the health service."

The treatments were not found to be 'effective', but only 'moderately effective' for the primary outcomes measures, and clinically ineffective for one primary outcome measure.


"In 2011, the first findings from the PACE trial showed that CBT and GET benefit around 60 per cent of patients with CFS/ME, for whom fatigue was the main symptom."

This is the biggest lie of all.

The proportion of participants who achieved a clinically useful outcome (clinically useful difference from SMC) were as follows:

CBT physical function 13% (NNT = 1 in 8)
CBT fatigue 11% (NNT = 1 in 9)

GET physical function 12% (NNT = 1 in 8)
GET fatigue 15% (NNT = 1 in 7)

By contrast, 58% (physical function) and 65% (fatigue) of the SMC control group achieved a clinically useful outcome.

 

[Bob]

Can anyone explain the following to me please?

"Specialist medical care was the cheapest option in terms of absolute cost to provide, but when the benefit of treatment on quality of life was taken into account, CBT became the most cost-effective option (likelihood of 62.7 per cent). There was a 26.8 per cent likelihood that GET was the most cost effective, while the likelihood for APT and standard care alone were 2.6 and 7.9 per cent, respectively."
http://www.meassociation.org.uk/?p=12343

If SMC was the cheapest option, in terms of absolute cost, and it vastly outperformed the other therapies, then how can it possibly be least cost effective? Surely that means that it was the most cost effective?

Have they completely ignored the primary outcome results for the cost benefit analysis?

 

[Dolphin]

Can anyone explain the following to me please?

"Specialist medical care was the cheapest option in terms of absolute cost to provide, but when the benefit of treatment on quality of life was taken into account, CBT became the most cost-effective option (likelihood of 62.7 per cent). There was a 26.8 per cent likelihood that GET was the most cost effective, while the likelihood for APT and standard care alone were 2.6 and 7.9 per cent, respectively."
http://www.meassociation.org.uk/?p=12343

If SMC was the cheapest option, in terms of absolute cost, and it vastly outperformed the other therapies, in terms of effectiveness, then how can it possibly be least cost effective? Surely that means that it was the most cost effective?

Have they completely ignored the primary outcome results for the cost benefit analysis?

SMC didn't "vastly outperformed the other therapies, in terms of effectiveness".

I don't think it's that hard to understand. Replace a drug for CBT or GET: it might cost more but if it improves quality of life, it can be considered "cost-effective" (if the cost required to give 1 QALY is less than £30,000 (say)). A health service is willing to spend money it collects from taxes to help people's quality of life, within certain limits.

 

[Dolphin]

I've heard the same about the East Sussex CFS service too.

The one useful thing they could do....!

What happens in Sussex seems to come from Peter White, just as it does in Essex. If I was in such areas, I'd be pretty annoyed at him.

 

[Sam Carter]

Can anyone explain the following to me please?

"Specialist medical care was the cheapest option in terms of absolute cost to provide, but when the benefit of treatment on quality of life was taken into account, CBT became the most cost-effective option (likelihood of 62.7 per cent). There was a 26.8 per cent likelihood that GET was the most cost effective, while the likelihood for APT and standard care alone were 2.6 and 7.9 per cent, respectively."
http://www.meassociation.org.uk/?p=12343

If SMC was the cheapest option, in terms of absolute cost, and it vastly outperformed the other therapies, in terms of effectiveness, then how can it possibly be least cost effective? Surely that means that it was the most cost effective?

Have they completely ignored the primary outcome results for the cost benefit analysis?

Hi Bob,

Here are the percentage of participants who improved by at least 2 points for fatigue and 8 points for physical function (ie. who achieved an (apparent) 'clinically useful difference') at 52 weeks:

42% APT
59% CBT
61% GET
45% SMC

 

[Bob]

Hi Bob,

Here are the percentage of participants who improved by at least 2 points for fatigue and 8 points for physical function (ie. who achieved an (apparent) 'clinically useful difference') at 52 weeks:

42% APT
59% CBT
61% GET
45% SMC

Thank you Sam, they are the figures for the post-hoc analysis for improvements in both fatigue and physical function. My figures are for the improvements considered separately, for each primary outcome measure, as per the primary results in Table 3 of the original PACE Trial paper.

The figures that you give, Sam, are the ones that the authors prefer to quote, as they are slightly more favourable to CBT and GET. But I think that they add to the general obfuscation of the data, and they are just a quick post-hoc analysis, designed to make the results look better.

 

[Bob]

I don't think it's that hard to understand. Replace a drug for CBT or GET: it might cost more but if it improves quality of life, it can be considered "cost-effective" (if the cost required to give 1 QALY is less than £30,000 (say)). A health service is willing to spend money it collects from taxes to help people's quality of life, within certain limits.

So the cost benefit analysis is purely based on QALYs, and not on the primary outcomes of the main paper?

 

[Dolphin]

So the cost benefit analysis is purely based on QALYs, and not on the primary outcomes of the main paper?

They did one set of calculations based on QALYs (which just use the EQ5d).
They also did the same set of calculations using (i) whether an improvement of 2+ on the Chalder Fatigue Questionnaire was achieved and (separately) (ii) whether an improvement of 8+ on the SF-36 PF was achieved.

They didn't use the thresholds for the CFQ or SF-36 PF in the original protocol but these really weak/easy to achieve post-hoc ones.

They looked at each question in too ways: just looking at healthcare costs (GET/CBT/APT cost more - is it value for money in terms of the general philosophy that a health service will spend money to help patients) and then looking at societal costs (again therapies might cost more society wise but a health service can spend money to help patients within certain limits).

 

[user9876]

So the cost benefit analysis is purely based on QALYs, and not on the primary outcomes of the main paper?

I believe that is true there seems to be a few ways of converting EQ-5d into QALYs. Found this paper talking a bit about alternative techniques for generating QALYs.

http://www.ces-asso.org/sites/default/files/H5-Edwards.pdf

I'm still thinking about it and how errors propogate through. So far I have failed to understand the probability graphs in the paper.

It would be interesting to see how the QALYs values and improvement they come up with compare with other diseases and treatments. My guess is there is a relatively small community of health economists that deal with these things and it makes little sense outside of that group. Surprisingly for an economics paper it is not full of equations in fact I was feeling the need for an equation or 2 to explain what they are doing.

 

[Bob]

Dolphin, thanks for the info.
I've started reading it now, so I'll be better informed soon.

 

[Bob]

An attempt to summarise some of the main outcomes, excluding the QALY analysis.


Differences in Overall Costs and Totals


Table 3: Service costs at baseline and follow-up.

Total Health Costs:

"SMC patients had significantly lower healthcare costs than those receiving APT, CBT and GET."

"Controlling for baseline, healthcare costs after randomisation were significantly lower for SMC than for APT (difference £840, 95% CI £637 to £1117), CBT (difference £904, 95% CI £613 to £1205) and GET (difference £829, 95% CI £534 to £1165). The differences between the APT, CBT and GET groups were small and non-significant."

Total Societal costs (consists of: healthcare, informal care and lost employment/production costs)

Differences between CBT, GET and SMC "were not statistically significant."

"Societal costs (i.e. healthcare, informal care and lost production costs) were significantly lower for patients allocated to CBT compared to APT (difference £2607, 95% CI £432 to £5585). Other differences were not statistically significant."

There is, however, a difference between the totals, even if it is not significant. CBT and GET show (statistically insignificant) total societal savings. It seems that it is this (insignificant) difference from SMC that the cost benefit analysis is partly based on.

This is all very confusing because some of the cost benefit analysis is based primarily on societal costs/savings (the difference between CBT/GET and SMC.) So, even though health care costs are greater for CBT/GET than for SMC, and there is no 'significant' difference for societal costs between CBT/GET and SMC, the costs per QALY gained, and costs per patient improved, are considered worth the costs. (I'll try and get my head around this later.)

For CBT and GET there are overall health costs, not savings, and so the societal savings are made up of purely out of informal care and lost employment/production costs.

For Total Societal Costs, there would be very little cost benefit, if any, for CBT and GET, if the 'informal care' costs had not been included. So, in Table 3, I think the 'informal care' costs, alone, make GET better than SMC. CBT is still slightly better than SMC without informal care costs included.
It seems to me that including these 'informal care' costs was optional, just as excluding welfare benefits was optional, so they were very handily included by the authors to make their case.
If they hadn't been included, then I think they might not have been able to say that CBT and GET are significantly more cost effective than SMC, but I'm not certain about that.
So it looks like the whole basis of the paper, could rest entirely on the informal care costs.
Very clever.

Informal Care costs

"Informal care costs were substantial for each group and significantly lower after receiving CBT and GET when compared to APT and SMC."

Lost employment.

Lost employment 'days' are given in Table 2, and lost employment 'costs' are given in Table 3.

The paper says that CBT and GET did not improve employment prospects:
"There was no clear difference between treatments in terms of lost employment."

I don't understand this because the differences between the changes for SMC vs CBT and SMC vs GET for employment costs, look highly substantial. The unadjusted figures are a relative saving of £1157 for CBT and £711 for GET. (Savings compared to the SMC control group).

Note that there were lost employment improvements in the CBT, GET and SMC groups, but the differences between CBT/GET and SMC were not significant, so CBT/GET did not improve outcomes.

Lost Production costs

I think the paper says that lost production costs are the same as lost employment costs:
"The human capital approach was used with the value of lost work-time to society assumed to be reflected by national mean age and gender-specific wage rates and combined with the lost employment data to generate lost production costs."

I think there was no difference between groups, apart from between APT and CBT. All it says about production costs is:
"Lost production costs were significantly higher for APT compared to CBT (difference £1279, 95% CI £141 to £2772)."





Table 4: Welfare Benefits and Other Financial Payments

http://www.plosone.org/article/info...RI=info:doi/10.1371/journal.pone.0040808.t004

Note that these figures are not included in the cost effectiveness evaluations.
I think they are the only cost data in the paper that are excluded from the final analysis.

For the overall differences, for welfare and private 'benefits' taken as a whole, I think there were no significant difference between CBT/GET vs SMC, although it's not very clear what exactly they mean by 'benefits' in the following text:
"However, with the exception of a difference between CBT and APT, there were no significant differences in either lost work time or benefits between the treatments during follow up. In fact, benefits increased across all four treatments."


Income Benefits:
The proportion of participants claiming Income Benefits increased in every therapy group.
Looking at the unadjusted figures, compared with SMC, CBT made no significant difference to income benefits claims, and GET increased numbers on income benefits slightly, but I don't know if it is a significant difference. So there is little difference between SMC, CBT and GET.

The text says:
"Relatively few patients were in receipt of income-related benefits or payments from income protection schemes and differences between groups were not substantial."

Illness/disability benefits:
The proportion of participants claiming illness/disability benefits increased in each therapy group.
Compared with SMC, CBT/GET resulted in a relatively lower increase in numbers on illness/disability benefits, by my rough estimation using the unadjusted figures, of about 12/13 percentage points lower increase for CBT/GET than for SMC.

Although, this doesn't tally with the text of the paper, so maybe I've got something wrong here:
"Receipt of benefits due to illness or disability increased slightly from baseline to follow-up (Table 4). Patients in the SMC group had the lowest level of receipt at baseline but the figures at followup were similar between groups."

They seem to be looking at absolute numbers claiming benefits in each group, rather than the relative changes in numbers claiming benefits in each group over time. So they completely fail to comment on the relative changes in illness/disability benefits. I don't understand this. Maybe there's no statistical significance but that sort of minor detail doesn't usually stop them from making outlandish claims of success.

So for Illness/disability benefits, there were absolute increases for CBT and GET, but relative lower increases for CBT/GET than for SMC.

Income protection schemes or private pensions:
The proportion of participants claiming for income protection schemes or private pensions were higher in every therapy group.
CBT and GET resulted in relative increases, compared with SMC, (but I don't know if they are statistically significant increases), in payments from income protection schemes or private pensions. (Roughly 4 to 6 percentage points more participants made claims in the CBT and GET groups, compared with SMC, using the unadjusted figures.)

Keeping in mind that at least one of the authors works for an insurance company, the paper avoids commenting on the increase in payments from income protection schemes and private pensions, as a result of CBT and GET:
"Relatively few patients were in receipt of income-related benefits or payments from income protection schemes and differences between groups were not substantial."
(Note, that they do not say that the differences were not 'significant'. Crafty!)



So, in all benefits categories, there was an absolute increase in the proportion of participants making claims, in each of the therapy groups.

For both 'income-related benefits' and 'income protection schemes or private pensions', the increases in claimants for CBT/GET are higher (worse) than for SMC, but they just say that the differences were not 'substantial'. They don't say that the differences are not significant.

For illness/disability benefits, there was a lower increase in the CBT/GET groups in relation to the SMC group. The paper doesn't seem to comment about whether the differences are significant in this category.

For the overall differences, for all welfare and private 'benefits', taken as a whole, the paper seems to assert that there were no significant difference between CBT/GET vs SMC, although it's not very clear what exactly they mean by 'benefits' in the following text:
"However, with the exception of a difference between CBT and APT, there were no significant differences in either lost work time or benefits between the treatments during follow up. In fact, benefits increased across all four treatments."


Interestingly, there is no separate data specifically about 'private medical insurance' claims. I don't know if they collected it. If they did, then the data can't have been good because they didn't include it.




Well, that's my lot for now. When I look at it again, I might find that I've made some major mistakes.
I haven't yet worked out how QALYs relate to the above info.

 

[currer]

I'm sorry you're finding this so upsetting, Bob. I know you worked really hard to analyse the original Pace trial.
It is necessary to do this but - it is also, in a way dignifying a simple piece of propaganda to treat it as a scientific paper.

@ dolphin post 81
I remember when our CFS/ME service was taken over by the present psychiatrically led service. They wrote out the requirement to provide specialist reports from the clinic regulations. Normally in the NHS specialists have a duty to provide these reports. So this duty had to be specifically excluded before the clinic started up. It was clear at the time they were going to operate on a basis of no disability reports - after all they believe we are not ill - we just think we are.
Our previous clinic provided vary helpful disability reports for benefit purposes - until it was taken over by the current St Bartholomew's team.
This has been forseen and planned for a long time. The government itself is behind this lie. It is the government's lie.

We must reiterate the truth because eventually people will realise that it is the truth. many people are loosing faith in our governments because of this falsity and abuse of their own citizens.

If your specialist team will not give you a report on your medical condition, which it is their duty to provide under their NHS contract, you should go to your MP.

 

[Bob]

How can they say that any of their treatments are cost effective when they've got no control group (although SMC seemed to have originally been intended as a control group, they're now acting as if the improvements in questionnaire scores there are the result of the expert care provided)? Are they just ignoring the tendency for subjective questionnaire score to improve in RCTs regardless of treatment type? I wish they'd had a homoeopathy arm.

Yes, it's confusing isn't it.

 

[Esther12]

Just had another read through of biophiles post on their figures - thanks for that B.

Wait, what? Although lost employment does not really change much, for some unknown reason their cost analyses did not take into account the increasing use of welfare benefits or other financial payments. This could have significant impact on cost-effectiveness, as benefits/payments cost a lot of money per person. Notice how all groups in the trial reported improvements in symptoms but also had increased usage of welfare benefits and other financial payments? I think I already highlighted the potential irony here for the underlying rationale of CBT/GET if these are associated.

The cost-effectiveness of CBT and GET are minimal and fragile at best when it comes to overall service costs. I still haven't looked at the other cost-effectiveness analyses which AFAIK are not based around actual monetary exchange but based entirely around hypothetical QALY (EQ-5D), self-reported fatigue and physical function. As others have pointed out, there are potential flaws and biases in these measures as used in the PACE Trial.

So they didn't account for the cost of additional benefit payments in their calculations on social cost? It does seem strange that they want to take credit for all positive changes, but then ignore the negative changes!


Also: does anyone have any idea what a normal way of calculating cost effectiveness is? eg: If a drug kept someone alive for an extra year, would they also attempt to account for the additional medical and care costs this would lead to? Or any additional economic benefits this extra year of life would tend to bring?

These could be side issues. Really the validity of their methods for assessing their actual impact upon quality of life are more key, and I'm afraid that's an area I'm totally at sea with for the moment.


Sorry about the rest of this post... it's little more than my own notes, as I've started scrabbling about to try to better understand the of the EQ-5D, and what sort of results we should expect were the therapists nice, charming, caring and useless. It's getting late, so I've just collated stuff here... this was my day of work to get my head around it, and I've realised I'm only scratching the surface.

There's a 100 page report from NICE looking at criticisms that have been made of NICE's use of EQ-5d here (I didn't read much, so can't say whether it's likely to be helpful or not): http://www.nicedsu.org.uk/PDFs of reports/DSU EQ5D final report - submitted.pdf

I've started looking for EQ-5D results from placebo and alternative treatments. To me, a big potential concern is that the much of the reported improvement could be down to politeness, positive thinking, etc. As there was no placebo control group in PACE, it might be interesting if we could find our own (I've not found anything great yet).

This was for back pain and acupuncture (not ideal as there's some reason to think acupuncture could well have a genuinely positive affect on back pain):

The acupuncture service was found to be cost-effective at 24 months; the estimated cost per quality-adjusted (QALY) was 4241 pounds sterling (95% CI 191 pounds sterling to 28,026 pounds sterling) using the SF-6D scoring algorithm based on responses to the SF-36, and 3598 pounds sterling (95% CI 189 pounds sterling to 22,035 pounds sterling) using the EQ-5D health status instrument. The NHS costs were greater in the acupuncture care group than in the usual care group. However, the additional resource use was less than the costs of the acupuncture treatment itself, suggesting that some usual care resource use was offset.

http://www.ncbi.nlm.nih.gov/pubmed/16095547

This study looked at homeopathy and FMS, and used EQ-5d scores according to it's protocol, but I've not been able to find them yet:

http://apps.who.int/trialsearch/trial.aspx?TrialID=ISRCTN74040048
http://www.ncbi.nlm.nih.gov/pubmed/19358959

This looked at a number of results for an NHS acupuncture centre, and could suggest EQ-5D is prone to distortion:

http://acmac.net/box/Gateway Outcomes.pdf

A set of outcome questionnaires has been developed to measure the range of treatment effects of
traditional acupuncture. In this descriptive outcome study we validated these questionnaires in a busy
National Health Service funded acupuncture clinic serving communities of diverse socio-economic
circumstances. Some of the questionnaires performed better than others in this setting but EQ-5D and
MYMOP-qual both showed statistically and clinically significant improvement in physical and psychological
health after six weeks and six months, in all categories of disease and degrees of chronicity.

The EQ-5D utility score (range 0.59 to 1.0) improved by 0.18 (0.24;
0.13) and the EuroQol VAS score improved by 8.28 (11.59; 4.98).
These improvements were maintained in the smaller sample at 6
months.

3.4. Analysis of change according to duration of complaints and
type of problem
The mean change in MYMOP profile score after six treatments
varies very little according to the duration of the problem (Table 3)
or the type of problem (Table 4). Changes in EQ5D and PEI are also
stable across these subgroups.

The
EQ-5D utility score improved by 0.18 (0.24;0.13) that, for a scale
that has been criticized in the past for low responsiveness23 is an
unexpectedly large shift.

It also mentioned something which seemed particularly problematic if the CBT treatment was emphasising the extent to which fatigue is a normal part of human life... all people feel bad sometimes etc... if this brought down the patient's conception of what the 'best imaginable health state' was, then this would have a direct impact on EQ-5D scores.

EuroQol (EQ-5D).15 The EQ-5D has five domains (physical, selfcare,
ability to perform usual activities, pain/discomfort, and
anxiety/depression) and three levels of severity on each
domain. Scoring generates a single health state profile and
subsequently a single utility score. Patients also rate their own
health on a visual analogue scale, ranging from 0 (worst
imaginable health state) to 100 (best imaginable health state).

This one mention eq-5d, but one needs to sign up to access the paper - looks like it could be a dodgy journal (I've not yet) - it's a placebo RCT on homeopathy for some sinus thing:

http://www.explorejournal.com/article/S1550-8307(06)00563-5/abstract

Homeopathy for skin conditions (not seen full paper yet):

http://www.sciencedirect.com/science/article/pii/S1876382009000766

Here's a paper with EQ-5d scores for placebo RCT onankylosing spondylitis (?), it also includes figures for a measure of fatigue:

http://rheumatology.oxfordjournals.org/content/46/6/999.full

Here are graph of results, with placebo group being the least good performer:

EQ-5d utility:

http://rheumatology.oxfordjournals.org/content/46/6/999/F2.expansion.html

Eq-5d vas:

http://rheumatology.oxfordjournals.org/content/46/6/999/F1.expansion.html

BASDAI fatigue item:

http://rheumatology.oxfordjournals.org/content/46/6/999/F5.expansion.html

My googling also led me to this thread from Dolphin, about a paper critical of using EQ-5d for CFS:

http://forums.phoenixrising.me/inde...ported-outcome-measures-used-in-cfs-me.10592/

 

[Bob]

The cost-effectiveness of CBT and GET are minimal and fragile at best when it comes to overall service costs.

Yes, this is what I have taken from it all. The health costs are more for CBT and GET than for SMC. But the health costs are considered worth it for the QALYs gained per person and societal savings per person.

 

[Dolphin]

I've read it now, once. This is a really complicated paper, to get to grips with all the details.

I think that this is an even worse case of smoke and mirrors than the original PACE Trial paper, which hardly seems possible.

This paper seems to be almost purely a case of pre-deciding the results that you want to present, and then having free choice over the data you can use in order to present your intended result.

I expect that I have got some of this wrong, and some of it might have already been discussed, but here are my initial thoughts...


Table 3: Service costs at baseline and follow-up.

Total Health Costs:
Looking at table 3, (increases in) Total Health Costs are about twice as much for CBT and GET than they are for SMC.

Total Societal Costs:
For Total Societal Costs, there would be very little difference between SMC and CBT and GET, if the 'informal care' costs had not been included. So, in this table, it is purely the 'informal care' costs that makes CBT and GET look better than SMC.
It seems to me that including these 'informal care' costs was optional, so they were very handily included by the authors to make their case.
If they hadn't been included, then I think they wouldn't be able to say that CBT and GET are more cost effective than SMC, but I'm not certain about that.
So it looks like the whole basis of the paper, could rest entirely on the informal care costs.
Very clever.

Overall, there isn't much difference between the other costs, in Table 3, apart from the 'therapy' and 'SMC' costs... So apart from 'therapy' and 'SMC', it's all a bit of give-and-take, and the overall costs are all pretty much the same for CBT/GET/SMC.


Welfare Benefits and Other Financial Payments

Income Benefits:
Compared with SMC, CBT made no difference to income benefits, and GET increased numbers on income benefits slightly.

Illness/disability benefits:
Compared with SMC, CBT and GET reduced numbers on illness/disability benefits, by my very rough estimation, of about 14 percentage points for each therapy group.

Income protection schemes or private pensions:
CBT and GET resulted in significant (I think) increases in payments from income protection schemes or private pensions. (Increases of roughly 4 to 6 percentage points.)

Keeping in mind that at least one of the authors works for an insurance company, the paper avoids commenting on the increase in payments from income protection schemes and private pensions, as a result of CBT and GET:
"Patients in the SMC group had the lowest level of receipt at baseline but the figures at follow-up were similar between groups. Relatively few patients were in receipt of income-related benefits or payments from income protection schemes and differences between groups were not substantial." (Note, that they do not say that the differences were not 'significant'. Crafty!)

Interestingly, there is no data about private medical insurance claims. I don't know if they collected it. If they did, then the data can't have been good.


Lost employment.

"There was no clear difference between treatments in terms of lost employment."

So CBT and GET did not improve employment prospects.


Lost Production costs

I think there was no difference between groups.


Societal costs (i.e. healthcare, informal care and lost production costs)

Differences between CBT, GET and SMC "were not statistically significant."

I'm not sure how they come to this conclusion, because healthcare costs and informal care costs had considerable differences between CBT/GET and SMC.



Well, that's my lot for now. When I look at it again, I might find that I've made some major mistakes.
I haven't yet worked out how QALYs relate to the above info.

I think you make some useful points.
On Lost Production costs
"Lost production costs were significantly higher for APT compared to CBT (difference £1279, 95% CI £141 to £2772)."

The paper doesn't mention what significance is set at. I imagine 0.05. However, there are an awful lot of comparisons so many would adjust for these multiple comparisons.