Unequal access to diagnosis of myalgic encephalomyelitis in England, 2025, Ponting and Samms

Of course, "the ICD-10 code that best reflects ME/CFS symptoms" may not be, and probably isn't, the actual code used for every patient.

 

Can the Hospital Episode Statistics data be re-analysed to estimate the prevalence of different autoimmune diseases (or other conditions, Scheibenbogen claims Raynauds is more prevalent in ME/CFS patients) amongst those patients compared to the general population (a priori, to me, it might not seem like the best dataset to do that)?

 

Good question. A quick Google found this paper that looked at a cohort in the US, which says,

"Of the 19 constituent codes, we find 5 that meet our Bonferroni corrected significance threshold (see Supplementary Table 1). The only significant sub-code to D89 was D89.9 (immunodeficiency, unspecified). Four sub-codes were significant from G93. Of these, the most significant was G93.3 (postviral fatigue syndrome), which was 4.4 times more common in the post-COVID period than the pre-COVID period."

Identifying long-term effects of SARS-CoV-2 and their association with social determinants of health in a cohort of over one million COVID-19 survivors, 2022,
Mukherjee et al
Open access, https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-022-14806-1

 

https://www.meaction.net/2024/02/01...-update-prevalence-figures-based-on-new-data/

MEAction will now use 700000 as the prevalence for UK and Scotland although they note issues.

 

Maybe some. But I've seen a number of people reporting that according to their GP, they don't have Long Covid, they have depression, also they don't have any patient with Long Covid. So I guess they are all either not coded, or coded as depression, or who knows what else. And that probably explains why many GPs don't think that LC is a problem. And much of the growing "mental health crisis".

The only group of illness where what is happening to the patient has minimal significance compared to what the physician in the room happens to believe about it.

 

#MEA choosing to rely on the CDC study because it is the 'latest data' doesn't sound like the right call to me: we need credible estimates, and at 1.3% this is not: I think we would notice if 1 in 76 people had MEcfs.

The CDC data is suspect:
1. It asked people if a medic had told them they had ME or chronic fatigue syndrome. Yet we often hear that most in the US do not have a diagnosis so presumably, the 'real' prevalence rate is much higher.
2. The F:M sex ratio is low at about 1.9.
3. Crucially, the question asks people if they have been told they have ME or "chronic fatigue syndrome" - easily confused with plain chronic fatigue - which is far more common (up to 4%) and has a sex ratio <2.

The prevalence rate suggested by this new study (up to 0.92% female, 0.25% male, roughly 0.6% combined, seems more realistic.

 

This paper hasn't been peer-reviewed yet, but I'd like to make a few comments.

• I think this is the largest ME dataset in the world, with 100,000 people given a diagnostic code of G 93.3 (postal fatigue syndrome, the ICD code that most closely resembles ME, CFS).
• The F:M ratio of 3.9 isn't too far off the roughly 5:1 figure seen in many research studies (which, unlike this one, apply specific diagnostic criteria), and is much higher than is seen for chronic fatigue. That gives me some confidence that many of the subjects have MEcfs, rather than a generic diagnosis.
• A roughly 5x lower rate of diagnosis in other-than-white people is shockingly low even compared with underdiagnosis in other illnesses (fig 3a). There is evidence that, if anything, CFS is even more common in non–white groups.
• The 10–fold range in diagnostic rates across England is much bigger than can be explained by differences in the level of other-than-white populations. This strongly points to big differences in service provision/willingness to diagnose. I think this information should help advocacy for improving services by region, as well as by ethnicity.

Finally, it makes sense to me to assume that the highest rate (amongst white people) – for Cornwall – gives us the best estimate of true prevalence rates in England. Does this make sense to others as well?

 

Does anyone know why the areas with the highest prevalence are:
Norfolk & Suffolk,
Cornwall and the South West?

See Fig 1B below, prevalence by ICB (integrated care board - which doesn't align that well with NHS trusts), sex and white/other-than-white

Cornwall had Anthony Pinching, consultant immunologist, AfME medical advisor and NHS clinical lead for the South West. I think he retired in 2011. He was also deputy dean of the erstwhile Peninsula Medical School. I gather they currently work closely with patients.

 

In my province (unsure about the other ones) we are still at ICD-9 when it comes to coding. We are simply not counted by the health care system, and even less tracked through the years.

 

The South West might well be the influence of Crawley and crew, especially for the Bristol and Bath area.

 

Yes, especially for Bath and Bristol. I’m sure the local presence of Action for. ME is fact as well.

 

MEA research update summarises this study:
https://meassociation.org.uk/2024/02/me-cfs-research-published-30-january-5-february-2024/

 

Now published:

Link | PDF (BMC Public Health, April 2025, open access)

 

Age - thinking about the demographics for eg Devon. Unless that has changed it used to be pretty heavy on older folk year-round.

More time to get it and be diagnosed. But also I guess as people get older they might need care when it combines either age /other debility so ends up being recorded where it might slip off records at other ages thanks to various reasons.

the london numbers are consistently almost bottom of heap . Again there could be a few factors including age and how on earth you can continue to stay living there if you get it before you’ve a string hold on property market and if you’ve moved there for work. More might have had to move home.

otherwise the picture seems consistent with having had biomedical provision apart from somerset being a gap in my knowledge? Although of course the Bristol clinic is ambitious and isn’t that where eg action for ME were based

 

Age would be an interesting thing to compare the data with. Shouldn’t be too difficult for someone with access to the data and some R knowledge…

I also wonder about correlations with other things like poverty. There’s a breakdown here, different areas though
https://commonslibrary.parliament.uk/research-briefings/sn07096/

Edit: just noticed they included deprivation in the paper

The HoC library often has good data, reports and sometimes dashboards like this on constituency estimates for the prevalence of 20 health conditions, which again uses different area breakdowns but may be of interest
https://commonslibrary.parliament.uk/constituency-data-how-healthy-is-your-area/
Or this on UC Claimants
https://commonslibrary.parliament.uk/constituency-data-universal-credit-claimants/

 

I am a bit sceptical about all this focus on diagnosis rates. We don't even know if ME/CFS is a good diagnostic category. Doctors may handle patients well or badly irrespective of what name they give. For rheumatoid I had no interest in this sort of question because I knew that there were huge numbers of people for whom a diagnosis of RA would be equivocal - possible or probable. Medicine doesn't really work with numbers like this. The important thing is to find out what is going on and treat it.