How do we make sure ME prevelance numbers are as accurate as possible?

Dr Carrot

Dr Carrot

Senior Member (Voting Rights)
Have been thinking about this a fair bit recently. One of the most common forms of advocacy with this illness is stressing how debilitating it is, but also how common, citing a greater prevelance than say, MS.

However, it’s difficult to know how accurate our numbers really are with this. I believe the range of numbers you get for patients in the US is between 800,000 and 2.5 million - a huge difference.

While in the UK, the figure of 250,000 seems to have been used for at least 20 years, possibly more. In their latest note on the update of the guidelines, NICE cite a figure of 190,000.

I suppose my general question is - how does one accurately estimate the prevelance of an illness, and also what people think about the figures we usually use? To me, they’re not as robust as they could be, but I’m not sure how to improve that.
 
Dr Carrot said:
Have been thinking about this a fair bit recently. One of the most common forms of advocacy with this illness is stressing how debilitating it is, but also how common, citing a greater prevelance than say, MS.

However, it’s difficult to know how accurate our numbers really are with this. I believe the range of numbers you get for patients in the US is between 800,000 and 2.5 million - a huge difference.

While in the UK, the figure of 250,000 seems to have been used for at least 20 years, possibly more. In their latest note on the update of the guidelines, NICE cite a figure of 190,000.

I suppose my general question is - how does one accurately estimate the prevelance of an illness, and also what people think about the figures we usually use? To me, they’re not as robust as they could be, but I’m not sure how to improve that.
Click to expand...
I’ve been wondering this too. I presume there is a reference document somewhere ? Tagging @Russell Fleming who may know
 
It seems to me that no-one really knows, but that if a figure is needed for political reasons guestimating a prevalence rate of around 0.5% seems sensible, and there are studies that could be cited to support it. While there's still so much uncertainty over exactly who should be classed as having ME there's going to be uncertainty over the figures.

For accurate prevalence figures you really want to assess a large random sample of the general population, but studies like this tend to use less rigorous approaches to assessing whether people suffer from ME, often being based on no more than a few self-report questionnaires - carefully examining a large random sample in person would be very expensive.
 
The estimate of 800,000 US cases is based on Leonard Jason's 1999 study. He found a 0.422% prevalence rate (95% confidence interval, 0.29%-0.56%) and applied it to the US adult population in 1997 (the actual figure was 836,000). However, the US adult population has increased since 1997, and so the low end figure would now be over a 1,000,000.

https://www.ncbi.nlm.nih.gov/pubmed/10527290


By 2006, the CDC estimated that 900,000 adults in the US had CFS and another 2.5 million had a "CFS-like illness," which would be a total of 3.5 million people. Subsequently, circa 2010, it was estimated that 1-4 million US adults had CFS, which was probably an expansion of the 2006 figure based on a larger adult population. The 4 million figure has been disputed as too large, presumably because it includes about 3,000,000 "CFS-like" cases.

https://www.researchgate.net/publication/236995875_The_Face_of_CFS_in_the_US

Using the current (2018) US adult population estimate, the number of US adult cases would be about 1,060,000 based on Jason's 0.422% rate. The number of "CFS-like" cases would be about 2,954,000 for a total of 4,014,000.

This all assumes that Jason's prevalence rate is accurate and has remained constant since 1999.
 
Last edited:
Dr Carrot said:
how does one accurately estimate the prevelance of an illness
Click to expand...

There is an article here about the work of the London School of Hygiene and Tropical Medicine on establishing ME/CFS prevalence.

Their estimated minimum prevalence rate of ME/CFS is 0.2%, which lines up with previous estimates. This 0.2% is patients satisfying the CDC 1994 Fukuda criteria, and/or the Canadian consensus definition, and/or another definition called the Canadian and epidemiological case definition.

About half of this 0.2% (that is to say, 0.11%) met the more stringent Canadian consensus definition.

So this seems to verify the long established figure of 0.2% for ME/CFS prevalence.

They also established an accurate figure for ME/CFS incidence (new cases of ME/CFS per year), which is 0.015%. So in the UK for example, which has a population of around 66 million, you will get 0.015% x 66,000,000 = 9900 new ME/CFS cases per year.


In the study itself, they say: "the number of people with chronic fatigue who do not meet specific criteria for ME/CFS is higher still". So clearly when performing any study on ME/CFS prevalence, you have to be careful to apply the ME/CFS criteria rigorously, to include ME/CFS patients but exclude those who have only chronic fatigue.
 
Last edited:
Well, in Canada the diagnosis is still being made by ruling out other possible illnesses with overlapping symptoms.
Then add to that the number of "self reported" surveyed.

This brings the number up over 500,000 in a population of less than 37 million.

EDIT to add- MS rates are only at 100,000.
 
Last edited:
alDrifter, post: 229609, member: 605"]According to this paper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2379748/?page=2, the ME Association said there were 150,000 people in the UK who had ME back in 1993. Now they say 250,000 people have it. Where did the extra 100,000 come from?[/QUOTE]


There's a figure for new diagnoses/ year ( sorry I can't remember UK figure, 6,000/ yr quoted for Australia by Dr Guthridge in his radio interview recently).

So if death rate is lower than diagnosis rate, and we accept the low recovery rates, there will be an annual increase.
100,000 over 26 years isn't a large annual net figure.
 
In the UK, home of the delightful Oxford criteria, the PACE trial had 640 participants who satisfied the Oxford criteria for CFS, but only a little over a half of those satisfied London criteria for ME or "international" criteria for CFS. It's hard to know how many of the 250,000 said to have ME/CFS in the UK have been diagnosed with the sloppy criteria.

The study mentioned by @Hip seems pretty good, but a minimal yearly incidence rate of 0.015% puzzles me.

The spread of ages of folk taking part in the PACE trial was pretty flat, suggesting that the incidence of ME is pretty much age independent (I'm ignoring suggestions that teens are different). In other words, a build up of people with ME.

We generally reckon that very few people with ME recover. If the average lifespan is 80 years, and we say that ME strikes from 10 years old onwards (despite the fact that my son went down with it aged 8), wouldn't that mean a steadily increasing proportion of people having ME, with 70 x 0.015 = 1.05% of folk aged 80 having ME?

Working on the idea that 0.015% of 10 year olds would have it, then by the next year a further 0.015% would be added to that number, we would have a ramp of people with ME, giving an average figure of (0.015 + 1.05)/2 = 0.53 as a prevalence figure.

For that to drop down to 0.2 as a prevalence figure, an awful lot of patients would need to recover.

Has senility made me mathematically incompetent? What am I missing here?
 
I've been thinking more about this. Well, something has to be going on in my mind when I'm waiting to fall asleep.

My rough calculation is based on a number of simplifications.

1. I have ignored weird teenage effects.

2. I have assumed a negligible recovery rate. It would have to be quite high for it to have a significant effect: is it worth trying to do the sums for that?

3. The incidence rate is constant across ages. This is supported by the age distribution of PACE participants.

4. The incidence rate has been constant across time. There's a feeling that it has increased of late. In the UK the oft-quoted figure of 250,00 people with ME represents 0.38% of the population: let's go with the figure of 0.2%. For that to be right, we would need an average incidence rate of 0.2/35 = 0.0057% – a little over a third of the quoted figure.

So the big question is, assuming that the figures quoted by Hip's study are "in the ball park", why has the incidence suddenly trebled?

Alternatively, if we assume that the incidence rate has been reasonably consistent across time, then this would lend weight to the argument that there are very, very many more in the UK with ME than the quoted figure suggests. Are many of them simply putting their symptoms down to age?

Is it that doctors are simply recognizing the possibility of ME much more readily?
 
The paper, https://bmcmedicine.biomedcentral.com/articles/10.1186/1741-7015-9-91#Sec2, says
The majority of ME/CFS cases complied with the CDC-1994 criteria (97.1%), 52.5% conformed to the Canadian criteria and 16.9% to the ECD. However, there was a considerable overlap among case definitions, as shown in Table 3. For example, 270 patients met the CDC-1994 criteria. Of these, 127 also met the Canadian criteria, while 18 met both the Canadian and ECD definitions. A total of 103 fulfilled the CDC-1994 definition but neither of the others.
Click to expand...

So doesn't it mean that the incidence rate of 0.015% is based on the NHS definition of ME? The incidence rate based on CCC would be 52.5% of 0.015%, so 0.008% (rounding up)?

Also if you look at
Screenshot 2020-02-12 at 13.43.57.png
there is a huge difference in the incidence in the London and East Anglia areas when compared to the East Yorkshire area. Does that mean that London, in particular, is skewing the figures as it often does (in general)? Or is East Yorkshire the oddity?

Also, as I understand it, the incidence figures only cover new cases diagnosed in one particular year, so that could skew things one way or another - a viral outbreak in London that travels to East Anglia but not to East Yorkshire in that particular year is just one scenario that comes to mind.

Ultimately, I'm not supplying any answers, just chucking more potential variables into the pot - no need to thank me ;)
 
Interesting considerations, @Graham.

One thought: perhaps the incidence figures might be exaggerated because of the phenomenon of post-viral fatigue, which can last one or two years before clearing up on its own.

This paper on CFS after glandular fever in adolescents showed that one year after the acute EBV infection, 7% met the criteria for CFS; but at the two year stage after infection, the number meeting the CFS criteria went down to 4%. So in the second year, quite a few with post-viral fatigue recovered.

If those patients with long-lasting post-viral fatigue are included in the incidence figures for ME/CFS, but then recover after a year or so, that would throw the incidence figures out a bit.
 
This distinction between PVFS and ME/CFS seems another underresearched area.

I often wonder how much ME actually does exist on a wider sliding scale than we assume, a wider spectrum of health degradation, with 1) PVFS (and maybe other idiopathic fatigue triggers) at one end and 2) very severe ME at the other end. Maybe in a way that is vaguely analagous to the HIV slide into AIDS.

Of course I'm only talking conceptually and not in terms of the phsyiology, but the idea of starting with X and ending up with "full blown" y makes intuitive axiomatic sense to me.

As much as we rightly draw distinction between being worn out and ME, I wonder how much being worn out sets a person up to fall, vulnerable to being PVFSed/MEed. That physical consumption by lack of rest/stress, if found to be measurable and relevant, would be part of the mix of congenital and environmental that would breed vulnerability in this idea.

If I'm not spouting rubbish, which I'm sure others have considered better, that would mess prevalence numbers much further than PVFS alone.
 
  • Like
Reactions: Hip
I'm fully aware of the Oxford criteria etc being a dangerous proxy for this, but I'm talking about more granularity in the epidemiology leading to better interrelated cohorts for research.

Please shoot me down.
 
Hip said:
One thought: perhaps the incidence figures might be exaggerated because of the phenomenon of post-viral fatigue, which can last one or two years before clearing up on its own.
Click to expand...
Good point.
What I don't understand is why the researchers who come up with these figures don't try to put them together as I have just to see if they are roughly consistent. But then, my experience with PACE leads me to think that perhaps many of them just don't think mathematically/scientifically.

InfiniteRubix said:
As much as we rightly draw distinction between being worn out and ME, I wonder how much being worn out sets a person up to fall, vulnerable to being PVFSed/MEed.
Click to expand...
It's hard to tell, isn't it? But the recent coronovirus scare reminds me that everyone has a slightly different immune system, and if the type of immune system that lays you open to ME was useless, it would have died out over the years, not spread to quarter of a million folk in the UK. After all, if we all had the same immune system, one specific disease could wipe us all out.

In the Spanish Flu outbreak, it was those with strong immune reactions that died. Sickle cell anaemia protects folk from malaria. I wonder whether the sort of immune system that lays us open to ME is protective from some other nasty condition: perhaps we are the ones that will survive the next bubonic plague. Mind you, it will be a bit of a struggle to repopulate the world.

Mind you, I also gather that there is more genetic variety in a troop of chimps than there is in the whole human population – evidence that our numbers dwindled drastically at one time.
 
You must log in or register to reply here.
Back
Top Bottom