Working inside the envelope – or pushing the boundaries
There is something very satisfying about reading a well-designed and beautifully written research paper. Not only can it produce some helpful answers (and usually pose multiple questions!), but the process that is followed means it’s pretty easy to see how the researchers arrive at their conclusions. This paper by White, Goldsmith, Johnson, Potts, Wolwyn, DeCesare, et al., is just one of those papers.

Why am I writing about research into four treatments for chronic fatigue syndrome?  Well, the treatment options used in this trial are exactly the same treatments used in chronic pain management (cognitive behavioural therapy, graded exercise therapy, adaptive pacing therapy and specialist medical care). And it poses the question: does ‘working inside the envelope’ of energy help people achieve better mood status and better functional status than challenging the assumption that ‘this is as good as it gets’?

On with the method.  This study examines four different treatments carried out in parallel, it’s an unblinded/masked randomised trial designed to establish the effectiveness of these treatments as well as the adverse events associated with each treatment, and reviews the outcomes up to one year after treatment ended.

Participants were carefully selected to meet the criteria for a diagnosis of chronic fatigue syndrome, and it’s interesting to see that only 28% of the people initially recruited actually met these criteria, and of this group, only 71% or 641 people actually progressed through to treatment.  Thankfully the CONSORT trial profile included in the paper shows very clearly who was ‘in’ and who was ‘out’ – and why!

Clinicians were carefully trained, supervised regularly, and treatments were monitored for adherence to the treatment type, the manual used, and that sufficient number of sessions were carried out.

A number of outcome measures were used – primary outcome measures of fatigue and the SF-36 physical subscale, with secondary outcome measures covering a range of variables including the ‘I think you look better’ clinical global impression scale, the six minute walk test, and the number of chronic fatigue symptoms remaining, along with a couple of others.

So what exactly were they testing in this study?

Adaptive pacing therapy is based on the idea that people with chronic fatigue have a limited amount of energy available, and that this is fairly static with limited potential for improvement.  People are asked to monitor their fatigue, and plan their days according to their energy levels, avoid ‘overdoing’ things, prioritise and therefore achieve what is important, and set the scene for natural recovery, if that occurs.  The strategies include looking at daily activities to identify links between activity and fatigue, becoming aware of ‘early warning signs’ of fatigue, regularly using rest and relaxation and generally attempting to ensure that no single activity exceeded ‘70% of the energy envelope’.  Participants could participate in more activities provided they felt they could and as long as their symptoms didn’t get worse.

Cognitive behavioural therapy is based on the idea that chronic fatigue is reversible, and that fear of activity (cognitions) and avoidance of activity (behaviours) influence physiology and perpetuate the problem.  By changing the appraisals of symptoms and introducing approach behaviour rather than avoidance within ‘behavioural experiments’, participants can test the usefulness of their beliefs and gradually, in a planned way, increase their activity level.  This is delivered within the context of identifying important activities and problem solving through obstacles that could get in the way of engaging in these activities.

Graded exercise therapy is based on the idea that people with chronic fatigue are, at least in part, deconditioned and intolerant of activity.  By gently extending the level of activity, participants can reverse the physiological changes that perpetuate chronic fatigue, and gradually return to participate in everyday activities.  Participants develop a baseline level of activity, then in planned increments of time rather than symptoms, start to increase the amount of time they spent in exercise.  Target heart rates and time exercising were set collaboratively, and most participants started with walking.

Specialist medical care consisted of education about chronic fatigue, prescription of symptom-alleviating medications (mainly for sleep, pain and mood), and general ‘self help’ advice.

What happened?

Well, surprisingly, both the specialist medical care and the adaptive pacing therapy achieved similar results across the outcome domains.  And both CBT and graded exercise therapy achieved similar results over the domains, with CBT improving mood slightly more than GET, and GET improving physical activity slightly more than CBT.  What was incredibly surprising is that both of these therapies were more effective than pacing and specialist medical advice.  And even more surprising was that the rate of adverse events was pretty similar – with CBT having the least number of events than those in adaptive pacing or graded exercise therapy.  None of these adverse events was considered, according to predetermined criteria, to be a ‘serious’ event, and adverse events generally were quite common.

How does this study finding apply to chronic pain management?

I think the underlying hypotheses of the treatments used in this study are pretty much the same as those used in chronic pain management.  The question is – do we help people ‘live within their limits’, or do we help people extend beyond those current restrictions and progress towards doing things they think they can’t?  It’s a question that some clinicians don’t even consider because the focus in chronic pain management has so often been on curing or fixing the pain.  And then there are clinicians like myself who want to help people live well despite still experiencing pain.

But in a large number of support groups, the whole idea of learning to do more seems invalidating.  To want to change from being seriously disabled to living pretty much normal lives means giving up some of the ‘being disabled’ identity.  It can mean no longer being given special consideration, or having the limitations of choosing what to do and when to do it acknowledged by others.

I’m not sure that I give enough time and attention to helping people mourn these losses as we try to push the boundaries.  At the same time as making space for mourning though, I want to give hope that life can be good, and better, than what is current.  Time to push the boundaries – but gently does it.

White PD, Goldsmith KA, Johnson AL, Potts L, Walwyn R, DeCesare JC, Baber HL, Burgess M, Clark LV, Cox DL, Bavinton J, Angus BJ, Murphy G, Murphy M, O’Dowd H, Wilks D, McCrone P, Chalder T, Sharpe M, & PACE trial management group (2011). Comparison of adaptive pacing therapy, cognitive behaviour therapy, graded exercise therapy, and specialist medical care for chronic fatigue syndrome (PACE): a randomised trial. Lancet, 377 (9768), 823-36 PMID: 21334061


  1. One thing about this post that struck me very forcibly in comparison with yesterday’s post. If you read the paper, the two therapies that are clearly the most effective (CBT and GET) are working from the same model: that patients are deconditioned and/or afraid of fatigue, and this needs to be addressed. This matches up to the “hedonistic goals” model mentioned yesterday. There is no acknowledgment anywhere that not every patient can fit this model. And looking at improvement scores, yes, the averages improved, and more people improved on CBT than on APT, but only about 62% of patients experienced a positive change after 52 weeks.

    This leaves a lot of people unaccounted for. And this is why I felt so strongly at the post yesterday: my experience is, the vast majority of doctors at the moment are working based on a single model, and are actually taking the results of studies like this as “everyone should do CBT/APT, that’s the best available treatment”. If doctors do not recognize that not every patient fits that model, then this could be a problem.

    Granted, the results of the paper show that the majority of the people probably do fit the model. But I’d be quite careful about the sweeping conclusions about what is “best” for everyone, and I would love to see the research that at least assesses people based on the models you presented yesterday, and attempts to determine how the different groups fare with different treatments. And I think it would be even more important for pain rather than fatigue, I guess, because (in my totally non-evidence-based opinion) the groups are likely to be more pronounced and different in different types of chronic pain.

    1. Some good points Mary. Yes, this study does compare models in broad generalisations, I accept that completely. I do think that chronic fatigue is likely to be slightly different from chronic pain in that the main presenting feature of chronic fatigue is low activity level, while in chronic pain, the main presenting feature is what I call ‘inappropriate activity level’. So in some cases, inappropriate activity levels can mean activity that is too low and the person can feel either frustrated at this or fail to see that life can be anything different; it can also mean activity levels that are too high, leading to increased vulnerability when ‘yet another stressor’ arrives on the doorstep (for example, coming down with the flu or having a change in work, or moving house etc); and it can also mean a ‘saw-tooth’ pattern of over activity (boom!) and under activity (bust!).
      The work of a good therapist is in establishing how well a clinical model fits the person they’re seeing. In your case, maybe the activity level was just right – but perhaps there are other aspects of having chronic pain that are troublesome, and the goals of therapy should be adjusted accordingly. In a non-research setting, this is what clinicians need to do. In a research setting this isn’t as possible because the aim is to identify a reasonably homogenous group of people, apply a number of treatments that are delivered in the same way, and then measure the outcomes across a range of variables. It’s only in this way that the ‘therapist’ factor is minimised so that the treatment can be seen as the active ingredient.

      I’d have loved to have seen this study identify two or three personal goals and, maybe using Goal Attainment Scaling, identify whether participants were able to achieve these individualsed goals.

      1. Oh, I know about the research study design, and issues that come with it, and I understand why things get designed the way they do, to avoid confounds and what you called a “therapist factor”

        When I get worried is the interpretation, and how this gets communicated to the doctors. If I were getting such results in my research area, I would be saying that, given the circumstances, and given that the “pacing” model and the CBT/GET models make opposite assumptions (one assumes that there is a fixed amount of energy; another assumes that there is no underlying limit and deconditioning is a problem), and both help between 40-60% of people, then we should consider how well the models fit, and look for the hidden variables that explain the effect, either in this data or in the follow-up studies.

        What I see in that paper, in the accompanying comment piece in the Lancet, and also in your blog report, is “CBT/GET is more affective on average; hence, APT should not be the first treatment of choice”.

        I actually like what you said about the “inappropriate activity level” a whole lot. This makes sense, and this is what my current physical therapist subscribes to, as well. The problem is, it took me 4 years to find a someone like that. Maybe I am unlucky, but every other doctor and physical therapist seemed completely fixated on the “too low activity” model, and, if they cited any research at all, then they cited those kinds of average comparisons. So yes, this is my personal frustration with how the research seems to be interpreted into practice, and that’s why I am wishing more of the research that you reported yesterday was actually out there and communicated to doctors, too.

      2. I hear your frustration loud and clear! I wish I could communicate more of the need to personalise the way in which treatments are applied, but it’s incredibly hard for so many clinicians to recognise that black and white rules don’t really exist in therapy. I think that’s why I really emphasise developing an individual case formulation that identifies hypotheses – hypotheses are meant to be tested to see whether they hold up in the light of empirical evidence. If the hypothesis is “the reason this person’s pain bothers them is because they are deactivated”, then the test of that is to guide the person into increased activity. If the hypothesis isn’t supported – the person doesn’t become less bothered with an increase in activity level, then it’s up to the clinician to generate another hypothesis, NOT blame the patient because they’re still bothered!
        I’ll write more about this in coming posts – it’s on my mind a lot.

      3. Your posting this week regarding ‘deconditioning’ has re-kindled my interest in this article. Absolutely agree that the lack of therapist’s guidance to set goals hindered APT efficacy (my buy-in is at ).

  2. Thanks to both of you! This post & subsequent discussion have been very valuable to me. I often think that the people who most think they need to push themselves need to rest more, and the people who most think they need to rest need to push themselves more. Though of course that’s a crude generalization. Both sorts, of course, will tend to seek out therapists who confirm their bias 🙂

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.