The draft scope for the NICE guideline on ME/CFS is now out for consultation, June 2018

Back to this again. This worries me Peter. Your comment suggests that when you speak of "reporting bias" you actually mean something very different to what everyone else means. You seem to be saying that the data itself does not hold any bias within it, that it is somehow pure and unbiased, "how it actually is". Which then makes me wonder what sort of reporting bias you are thinking of? Could you be saying the data is untainted, but the authors' reporting of their findings was biased, and you are referring to that as reporting bias? The reporting bias widely discussed is about self-reporting bias, which inevitably taints the data itself, so this cannot be what you are talking about, because you are saying it is "actual data", unbiased by any reporting effects.

Like I say, this worries me.

Edit: Or is it me getting confused between reporting bias and self reporting bias? The data is most certainly not free of the latter.

 

No one here thinks ME does or should include vaguely defined 'chronic fatigue' patients. The point is, though, that PACE claims to be covering both chronic fatigue and chronic fatigue syndrome patients (and ME patients too, since the authors claim to use the London criteria on a subset) - and yet still can't show any benefits.

They specifically say their findings will be useful to anyone with fatigue as their main complaint. They also claim the findings apply to ME patients because some of their patients met ME criteria too, though they don't separate any of these patients from each other when discussing their results. The mixing of patient cohorts is inherent to the PACE trial, which in effect is implying that ME patients are CFS patients are CF patients, and that they all respond alike.

But here's the rub: even by those overly inclusive standards, it still shows no real benefit of treatment. Not for ME or CFS or CF.

So we don't need to say 'well it might work for fatigue patients', because it doesn't work on anyone. Better to say: 'It doesn't work for ME or CFS patients, or even CF patients, so there's no evidence for any of it.'

Arguing that the trial looks at the wrong patients is irrelevant. Even if it was testing on orange and blue aliens, it still failed to show any benefit for CBT or GET. That they tested on orange and blue aliens just highlights how desperate they were.

 

I am not sure where you get this from @Peter Kemp.

This is a public forum where people express their own opinions. You are as much part of the group as I, and you are welcome to the debate. Because the forum had the chance of being a 'stakeholder' we have been trying to get agreement on comments to send to NICE. Our comments are about taking note of the uninterpretability of evidence based on flaws that are common to all the trials that have been done. I think pretty much everyone agrees that PACE does not refute anything about treatment of ME/CFS because it is likely to have ended up with an unrepresentative group of patients.

I think you are confusing issues, Peter. A trial of patients with chronic unexplained fatigue would be relevant to people with ME/CFS because in general people with ME/CFS have chronic unexplained fatigue - of a particular pattern. In the case of PACE there are problems with selection that make it likely that any conclusions it might come to were specifically not applicable to ME/CFS but that is a specific issue with this trial. Otherwise it is standard to carry over conclusions from wide samples to subsets.

Practical medicine does not work by putting people into diagnostic pigeonholes first and then wondering what to do. It works by identifying clinical problems that might or might not turn out to usefully fall under one or other diagnostic category and judging what to do on the basis of estimated probabilities. The problem this guideline deals with is most widely called chronic unexplained fatigue. That does not in any way stop the detail of the guideline making specific suggestions about what to do if the diagnostic label ME/CFS seems to fit well, but so far it is not clear that it actually matters, except in that it is sensible to avoid pushing exercise in people for whom this makes them feel worse.

I spent my research career studying another syndrome - rheumatoid arthritis. When I started it was just a syndrome, like ME/CFS. Nobody understood its mechanism so the diagnostic criteria were just based on a pattern. There was no sense in which they were 'right' or 'wrong' because nobody knew what they were supposed to be identifying. Forty years later we know almost exactly what is going wrong in RA and we know that the old criteria identified quite well an immune complex based disease arising from a broken feedback mechanism arising through genetic and stochastic factors (and smoking). We also know that the old criteria would have misidentified some patients who in fact have a T cell based disease. Similarly for diabetes we have discovered that there are two major diseases of quite different mechanism type 1 and type 2. For ME/CFS we have not got to this stage. When we have we will almost certainly divide it up differently.

 

Sorry but this is totally wrong; if you read any of the many, many threads here on the PACE trial and those directly and indirectly involved in it (and also those on PR where a lot of us were up until the end of last year) you would know that.

 

@Jonathan Edwards , this remark has been rattling around in the back of my brain for a while. I think that both you and I, comfortably retired from a secure job, financially stable and male, can easily underestimate the pressures put on more junior members. If you have a large mortgage hanging around your neck and family responsibilities, and if promotion lies in the hands of those around you, the pressures to conform are enormous. Add to that the fact that women are open to even greater pressures to conform, and I can fully appreciate how risky it would be for those who are not near the top to speak out.

I suspect the medical profession is very heavily weighted towards a hierarchical structure, where specialists rule the roost and junior doctors are expected to toe the line. Isn't that one of our major problems?

I get the impression that, in research, in science that is not really the case: newcomers are expected to challenge existing ideas. But is that the case in medical research?

I don't know what life is like in these professions, but I do know the tremendous power that headteachers have over their staff: a governor once said to me that if someone was unhappy, they could always leave and get a job elsewhere. He was quite surprised to find that references are taken up before decisions are made about who to interview, and it is always expected that a reference from the current head is included. If that is your first teaching post, then you can feel quite vulnerable.

It must be even worse if you are in some form of medical line, and you have ME. The prejudice against the condition is widespread.

 

This may have been answered elsewhere, but why are references to other documents in the Guideline Scope not actually referenced?
eg p2, line 10 mentions a "UK primary care cross-sectional study" which it doesn't give any further information about!!!!

Maybe I've missed something here. I'm a bit late to this.

 

It's interesting you understood it like that. My impression was this is not the opinion of s4me. (It's not my opinion.) But did I misunderstand you?

 

Welcome, @Peter Kemp! Sorry, I've only just caught up with this thread. I don't log on as often as I'd like.

I wanted to make one more point about the issue of case definition and the PACE trial. If you make the claim that the improvements reported by patients in the PACE trial reflect genuine improvement in a subgroup that aren't "real" ME patients - people who have CF but not PEM - then how will you explain the following?

The figure below is from the Lancet paper (White et al, 2011). It shows the means and confidence intervals for the two main continuous outcome variables (SF36 physical function scores and Chalder fatigue scores) across all trial arms at all time points.

The top two panels show the data for the entire cohort (all participants meeting Oxford criteria). The next four panels show data for patients that not only met Oxford criteria but also met additional criteria - the International consensus criteria and the London ME criteria respectively.

Notice that the patterns do not change when those more restrictive case definitions are used.

Now look at the bottom two panels, which shows the subset who also met criteria for depression. Notice that this subgroup also shows the same patterns overall (except perhaps that APT seemed, paradoxically, a little more effective at relieving fatigue in that group).

The paper also reports some statistics to back up these observations (which are questionable, and I certainly didn't accept them at face value, but I did some stats of my own). My stats and the figures in the table both bear out the overall point: the apparent benefits of CBT and GET are consistent irrespective of whether you define the condition broadly or more restrictively.

And of course they are! These effects are most likely artefacts due to the way people attended to and recalled their symptoms in the different trial arms. Those sorts of reporting biases are common to all of humanity, and it is totally to be expected that they would occur for all participant groups, irrespective of diagnostic criteria.

 

I wanted to add that in the above figure, there are different Ns (sample sizes) for the different analyses. Because, well, obviously the number of people that meet ICC or London criteria in addition to Oxford criteria is smaller than the entire cohort.

This means that not all the analyses will have the same statistical power. Specifically, its likely that confidence intervals will be wider as the size of your sample gets smaller. Because of this power confound, you can't read a whole lot into the details of the different figures - like how much the confidence intervals overlap in each subgroup.

Edit: one more thing to point out. Notice that APT and SMC (basic medical care) groups generate similar outcomes for self-rated fatigue, but APT fares worse on the self-rated physical function measure. This is obviously an artefact: if you tell your APT participants that they should place much greater limits on what they do, and then you ask them what they do, those that have complied with your treatment ought to report that they now do less. This is because its what you told them to do!

 

I don't think medicine is very different from science or teaching. In medicine newcomers are certainly expected to challenge ideas as much as in any other science. But juniors have to be aware of the power of seniors as in most walks of life.

I think I added a caveat that junior medical professionals who are also patients might wish to remain anonymous. But in the context of forming what seems to be a small (?a dozen) group that seems to engage in putting out opinion as advocacy I find it hard to see the justification of anonymity. It seems particularly odd that the webpage for VIRAS states that its members have these professional qualifications without saying who they are. The impression is that the statement is intended to give some sort of authority. If this is intended as information of value to patients then I think medical professionals are obliged to say who they are. The implication is that what is being said is contentious. That would be no surprise in the field of vector borne infection. In my view anyone who puts out information of that sort should be prepared to defend themselves in public.

 

I think that claiming expert support in anything, but failing to name those experts, is not a good idea. Any work should stand or fall on its own content, and only named proponents should "count" towards validating opinions there. Glibly stating that you have important people in the background behind you smacks of school playground arguments.

But we have always benefited from "expert" help here, from folk who wish to remain anonymous, for entirely understandable reasons. I always feel uncomfortable that their major contributions have gone unrecognized, but fully support their decisions.

You did add that caveat, but I wanted to make sure that the overall tone reflected the high value that many of us in the know have in our anonymous contributors, and a sympathetic understanding of their circumstances.

As far as the medical hierarchy is concerned, I get the impression from friends whose children have gone there, that the "specialist is God" syndrome is much more of a problem than it is in science or teaching though. In all of them, senior figures wield power, but my impression is that in the medical world there is an added God complex. The three of us have seen a variety of specialists over our very many years on this planet, and I must say that only a few could be described as approachable.

 

Have you looked at the VIRAS site, Graham? It is very different from the sort of discussion we have here.

 

Yes, I have, and in that context I agree with you. As I said, I was just trying to ensure that those folk here who have been of major help to us, but who have wanted to remain anonymous, knew that we supported and understood that decision. You didn't say anything that suggested otherwise, but equally, I thought it was important to make the point explicit.

 

Just read VIRAS submission. Fantastic! Good to see you went straight for the jugular.

 

How does that get made? A specialty, sub-specialty, or certificate program, or whathaveyou, for ME?
We need that.
Although we need more stuff for them to know, but even if we started with what we can demonstrate about the disease existing, how to diagnose, and some case histories, and just good basic supportive care (which most of us don’t presently get), and a special emphasis on evidence and critical reading of literature, and whatever else has a good evidence base, wouldn’t that be a place to start?