The course of the illness for ME patients in Norway, 2021, Schei and Angelsen

The Norwegian ME Association just published an impressive survey on 5.822 ME patients in Norway.

An English summary is available here: https://me-foreningen.no/wp-content/uploads/2021/03/Norwegian-ME-Association-2021-Report-on-the-course-of-illness-English-summary.pdf

 

Twitter summary: https://twitter.com/user/status/1369593779674611712


2) One has to take the results with a grain of salt because respondents to such surveys are not necessarily a representative reflection of ME/CFS patients as a whole. Nonetheless, almost 6000 respondents is huge: this might be approximately 1/3 of all ME patient in Norway

3) Some interesting results: “Many actors in the healthcare system describe a course of the illness with initial fluctuations and a gradual improvement. Such a course is possible but is only characterizing a small minority (12%) of the respondents.”

4) The survey criticizes unrealistic expectations of imminent improvement, especially in adolescents: “When there is no improvement, the patient is often blamed.”

5) The answers to the survey show a large increase in the number of new cases of ME in the period up to 2009: “we see a peak that coincides with the swine flu epidemic” A study by Magnus et al. suggested a 2-fold increased risk of ME following H1N1

6) Also interesting: “more than half of those who are very severely ill reported that they became ill before they turned 20. Thus we find a statistically significant correlation between an early onset and having a severe or very severe degree of ME”

7) The surveys also shows that many patients think activity levels have an influence on the course of their illness and severity level. “Almost 2 800 free text responses state that regulation of activity levels contributes positively to the course of their illness”

8) “Only 6 (0.1%) of the respondents report that they have become worse due to inactivity.” Coping strategies such as activity pacing and energy management were seen as the most important measures for a good outcome.

9) Also remarkable: “The survey had 56% more responses stating that the healthcare services have been a cause of health deterioration rather than improvement.”

10) The survey notes: “It would be difficult to find any other patient group where patients consider the healthcare service to have contributed more to the continuation or deterioration of the illness than to the improvements or recovery.”

11) Reasons for why the healthcare service was seen as a negative factor include mistrust of patients and their description of symptoms, misguided advice on activity levels and an unwillingness of GPs to provide symptom relief for sleep problems and pain.

 

That is staggering. It deserves to be retweeted widely!

 

The remarkable onset graph - thoughts?

The report includes a graph showing year of onset for 5800 people. This is a very large sample and the first big study I have seen that measures year of onset, as opposed to the much-less-useful year of diagnosis.



The most striking feature of the graph to me is that it shows such a strong pattern. That pattern might reveal something very important about the disease, though it might also be an artefact resulting from an unrepresentative sample. That's why I'm keen to discuss it here.

This is just the kind of puzzle that might interest @Jonathan Edwards

First, two important observations made in the study report itself.

1. The red bar for 2009 coincides with a swine flu outbreak in Norway, and the study authors note that a separate study had found that swine flu doubled the risk of getting ME/CFS. @Lucibee noted on Twitter that the high point in 2000, coincides with a bird flu outbreak in 99/00. Both peaks indicate substantial post-infectious onset for the illness.

2. The authors suggested that the decline in recent years could be an artefact resulting from the many years it takes to get an ME/CFS diagnosis. If you haven't had a diagnosis, and so don't know you have ME/CFS, you are unlikely to be taking part in this survey.

Also, the Norwegian population has grown 27% since 1990, but that would only explain a small proportion of the increase in new cases since then.

 

I can think of a number of possible explanations as to why the study finds a steady increase in the onset of ME/CFS, since the mid-80s to around 2015. But I am keen to hear others.

My starting point is that I expected to see a fairly flat graph. Apart from any epidemics that produced the spike in cases, I'd assumed that there would be fairly constant number of new cases every year. (While it's likely that people who have recovered wouldn't take part in the survey, the Cairns review of recovery found that it is rare.)

1. Increase in diagnoses.
Certainly in the UK, CFS has become much more common diagnosis since around 1990, when proponents of psychosocial theories encouraged its diagnosis, and "treatment". Perhaps there was a similar trend in Norway, but it would have to be a steady trend of increasing diagnosis (perhaps as awareness slowly increased among Norwegian doctors).

2. Underrepresentation of older people (who will generally have onset earlier in time).
Those with an onset earlier in time will, on average, be older than those with a more recent onset. If older people are less likely to use social media/access the survey, that would lead to increasing underrepresentation of people whose onset is further in the past. And so produce the effect of fewer cases further back in time.

3. A fairly steady rate of recovery from ME/CFS over time.
This strikes me as the simplest explanation. However, it doesn't fit with what we know about the illness (either from the Cairns study or anecdote).

The second point is testable. If we look at the 1990s, for instance, and assume that there is underrepresentation of older people who got ill in that decade, then the average age of onset (of those represented in the study) would be younger than for recent decades. (Someone who got ill aged 40 in the 1990s will be considerably older than someone who got ill aged 20 in the same decade.)

 

Looks like they did ask about participants age, so this should be easy enough to determine.

"Most were between 20 and 60 years old, with those between 40 and 50 years making up 30% of the respondents. 7% were below 20 years."

 

If people think this is useful information, I wonder if year of onset, age at onset, year of diagnosis, and age at diagnosis questions could be part of DecodeME – if it isn't already?

 

also discussed here:
https://www.s4me.info/threads/news-from-scandinavia.647/page-63#post-330317

 

This had me scratching my head, too. In addition to the possible explanations you came up with, @Simon M, I had the following thoughts.

1. Increase in diagnosis (similar to your point 1)

Respondents may have interpreted the question regarding onset as asking about date of diagnosis and there may have been an increase in doctors from the 1990s (re)diagnosing long-term ill people.

And maybe older people, as well as possibly being underrepresented in the survey, are simply unaware they have ME/CFS because prior to the 1990s they were diagnosed with something else and they have retained that different diagnosis.

Both could be the result of "fashion", not on the part of the patients as often claimed, but on the part of doctors who, becoming more aware of the CFS diagnosis around that time courtesy of certain BPS publications, chose it over burnout or depression or whatever they might have chosen previously for their mystery patients.

2. Increase in incidence

There could be a true increase in incidence for some reason. Maybe the chemical cocktail we all live in these days - or some other unknown factor more prevalent than it used to be - makes people already genetically predisposed even more vulnerable. I think this is a less likely explanation but if true it would be rather scary.

Adjusting for population shouldn't be too difficult, if somewhat tedious? Dig up the population of Norway for each year and calculate the onset per 100000 or whatever for each year. Though it may not be worth the effort as it wouldn't change the overall pattern.

 

The only factor that came to my mind to explain the rise in cases was the outbreak of giardia (and subsequent cases of ME/CFS) in Bergen, however that occurred in 2004.

That said, there might have been a lag between initial infection and the development of ME/CFS (and diagnosis). My health started to go downhill after a mystery stomach bug. Two years later, I got EBV and developed ME/CFS.

https://me-pedia.org/wiki/2004_Bergen,_Norway_outbreak

 

I think the general shape of the graph will be a sampling artefact. The peaks in 2000 and 2009 might be more real but still might be artefacts of the way things are remembered.

The likelihood of sampling artefact was my main concern in relation to the GWAS study and still is. I wonder if the Norwegian study here might give clues as to how one might be able to measure and adjust for such artefacts.

It would be interesting to know what the date of onset was for all current S4ME members. In fact one might study date of onset for all S4ME members who respond to a questionnaire about it and then a sample of those who did not respond initially but responded to further specific requests.

 

I'm pretty sure that the illness had already been present for a while when I first noticed the "onset" event.

I'm curious what the NIH intramural study will find. They have very strict entry criteria requiring a documented infection. That means their findings might not extend to patients where the onset was less definite.

 

I too find it hard. The only definite symptom I had near the beginning (I think!) was the leg weakness - difficulty walking up slopes. So I can guesstimate the year, I think.

 

It would be easy enough to do a members' poll on best-guess-at-year-of-onset, wouldn't it? Even if the software wouldn't allow a list of 50+ options, we could start by sorting people into decades.

A simple sudden vs gradual onset option might also be useful, as well as an indication of whether onset may have taken place over more than one year.

Or should we work towards a slightly more detailed questionnaire? No real idea what's useful and what isn't, but happy to help if I can.

 

There is potential confusion about onset, even with those whose ME onset is associated with an acute condition.

In one sense I could say my onset was between 2pm and 2.15pm on a specific date twenty eight years ago, this was the onset of a bad headache that seamlessly became glandular fever (later tested positive for EBV) that seamlessly became ME.

However a number of years later I identified I had non-celiac gluten intolerance, which I retrospectively suspect appeared several months before the onset of my glandular fever, raising the possibility that some change(s) had begun prior to the glandular fever.

Technically the glandular fever was not ME, and ME can only be diagnosed a period of time after the onset of any associated infection; when in the eight months or so (I was lucky to have tuned in GP) between my glandular fever onset and the formal diagnosis of ME did my glandular fever become ME? I suspect this is impossible to answer retrospectively because there was a continuity of symptoms.

Over a number of years of much reduced activity, I came to believe I had recovered from ME, though in retrospect I may have still been experiencing some PEM. However on another identifiable date (31st Dec 1999) I developed symptoms of influenza (my millennium bug). I was not formally tested to identify the specific virus. Associated with this influenza was a major ME relapse, which though since has fluctuated it has not achieved anything like the previous remission and is now not far off its most severe. Retrospectively I don’t think it is possible to say whether this was a relapse of my original ME or a new re occurrence of the condition.

So I have potentially five possible points of onset for my current ME spread over some seven years
- possible physiological changes predating the glandular fever
- the onset of the glandular fever
- the transition from glandular fever to ME (impossible to date retrospectively)
- the onset of influenza
- the transition from influenza to the reoccurrence or relapse of ME

Also though I still have the symptoms associated with the original glandular fever/ME diagnosis with each subsequent relapse new symptoms have developed, such that my ME over the last six years or so feels qualitatively different to the condition when it was first diagnosed in the early 1990s. Do I have the same condition now that I had then?

Personally I regard the onset of my ME as the same as the onset of my glandular fever and would base any answer on onset on that belief, but until there is better understanding of the underlying disease process there must be a degree of uncertainty.

 

Mine was similarly complicated, covering a period of several years. My family noticed that my stamina wasn't the same after measles, but I'm fairly sure I didn't have diagnosable ME then; even the onset of a clear ME pattern after exposure to EBV took several months.

But if we take it from the point [ETA: I mean the approximate year] at which we developed a marked impairment in our ability to live life as normal, it's probably good enough to plot it out?

 

Thoughts

• A life expectancy for ME in the mid 50's (3 studies?) many pwME onsets 1980's and earlier have died
• Time lag for diagnosis and engagement with right areas of support can be several years hence 'tail off' post 2013 (IOW many are still awaiting diagnosis and would therefore increase figures for recent years)
• Rightly or wrongly the work of Psychs brought ME to more prominence around 2007-2011
• Increase in access to social media at time of onset helped pwME access support groups post 2000
• If you know the name of the viral trigger, especially a high profile deadly pandemic, especially when others are long hauling, is it easier to insist on referrals to clinics? Is it easier to not doubt your sanity? If you pick up some random virus that no one else does and there is no media interest are you more likely to brush it off if mild, given a misdiagnosis of depression, psych or FND? Hence peaks for 2000/2009.
• Until there's a biomarker and the stigma is gone for ME, epidemiological studies are unreliable.
• Could/is the DecodeME / YouandME registry be set up to do something similar and gather some extra data with a 20000+ cohort being asked when they became ill, severity etc?

 

Thanks to everyone who contributed to the discussion. This is my attempt to summarise the main points about possible explanations for the surprising graph shape.

Possible explanations related to the illness itself.

1. A slow steady rate of recovery, so that over time, fewer and fewer people are left within any particular year of onset (see point 3 here). (This seems to contradict what we know about the illness.)
2. A genuine increase in cases of ME/CFS over time in Norway. @Ravn
However:

Explanations relating to an unrepresentative sample

1. Underrepresentation of older people (who will generally have onset earlier in time).
See point 2 in my post above.

2. More cases are diagnosed, as doctors get better at recognising the illness.
Either all the cases that were missed, are never diagnosed and so people don't show up in the survey. Or this "backlog" of missed cases are diagnosed later, contributing to the apparent growth (assumes that people are confused by the question about onset and instead report year of diagnosis). @Ravn

3. Uncertainty about the date of diagnosis.
Identifying the year of onset is is clearly an issue for people who have gradual onset @MeSci @strategist @Midnattsol. However, I'm not clear how some "fuzziness" about the onset date would lead to the particular shape of the onset graph. I suspect the same applies to sudden onset - unless it was common for people to have experience like @Peter Trewhitt, though even then it's not obvious how that would create the pattern seen in the graph. Please do point out out if I have got this wrong.

I'm not sure if I've quite understood you. Am I right to think that you believe there is an artefact, but you don't know what it is? And that data on onset year for all forum members would help shed light on the situation? And if so, how? I'm probably just being a little bit slower here and would like to know more.

@Fizzlou are you suggesting using existing registries together more onset data that would throw more light on the situation, as per Jonathan Edwards?

So that's just yet more questions from me. Also, anyone please point out if my summary is misguided. I do love stuff like this.