Talk:Long COVID

The lead section says "As of 2024, the prevalence of long COVID is estimated to be about 6–7% in adults, and about 1% in children", which is repeated further down the article. Are these percentages of all adults/children, or percentages of those adults/children who have been infected with COVID at some point? —Mahāgaja · talk 15:38, 1 October 2025 (UTC)

That's the overall prevalence —Femke 🐦 (talk) 16:06, 1 October 2025 (UTC)

Wow, that's astonishingly high. What is the prevalence among people who have been infected with COVID? —Mahāgaja · talk 09:38, 2 October 2025 (UTC)

Estimates are between one third and one half of the US population depending on whether you count those who might not have had symptoms. So in the Us that at least doubles the claimed 7% prevalence of long covid 9 (albeit needs adjusting for the claimed much lower prevalence amongst kids). Interesting test would be, although it's impossible to run, how many people have long covid who never had covid in the first place. 2A00:23C5:6429:601:50C:3D9C:8555:256E (talk) 22:40, 6 October 2025 (UTC)

I think there may be some incidence and prevalence confusion around this. 6–7% of adults can have this at any one point in time without 6–7% of adults having it for the rest of their lives.

For inclusive definitions ("if you're not feeling perfect after four weeks, then you have long COVID"), a whole lot of people who are just a bit slower than average to recover, whose asthma gives them a lingering cough for every respiratory illness, who happened to have a particularly nasty case, who got an unrelated virus but incorrectly blamed COVID for it, etc. have "long COVID" when it's really just a slow resolution of the acute phase.

Imagine if someone said that everyone with Infectious mononucleosis had "long mono" if they weren't over it in four weeks, or if Whooping cough used that cutoff. Basically anyone with those famously lengthy infectious diseases would have "long mono" or "long whooping cough". WhatamIdoing (talk) 01:49, 16 October 2025 (UTC)

The source now cited doesn't say explicitly what definition they use (4 weeks vs 3 months). The Lancet paper published in the same month does, but only gives UK numbers. That is a much lower prevalance of around 2% for 3+ months. I'll dig a bit more. —Femke 🐦 (talk) 07:24, 16 October 2025 (UTC)

I truly don't wish to be controversial in any way, but it does seem odd that this article does not include a reference to the plausibility of psychosomatic causes. I won't put a section in at this stage unless other editors agree. 2A00:23C5:6429:601:50C:3D9C:8555:256E (talk) 14:20, 6 October 2025 (UTC)

A section like that would require very strong sources, such as review articles from a reputable medical journal. If you find suitable sources, you could post them here. WhatamIdoing (talk) 01:07, 7 October 2025 (UTC)

Thanks for this. It's very helpful (I forgot to sign in above). I can cite a number of useful papers which I hope people won't find contentious and will improve the article. I'll think on the best way of putting them in. All the best, Emmentalist (talk) 14:46, 8 October 2025 (UTC)

We currently have a paragraph in this around misdiagnosis. The top sources describe it in that context, but always curious to know what other review articles say. I tend to rely most on overview sources for how to frame those contentious labels. I've not exploited the most recent reviews as much as I could, so might have another dig at that at some point. —Femke 🐦 (talk) 15:19, 8 October 2025 (UTC)

I don't think we cite the most recent Lancet review yet: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)01136-X/fulltext. It touches on psychological support, doesn't use the word psychosomatic, but does say that there's frequent misdiagnosis of various mental health conditions. —Femke 🐦 (talk) 15:25, 8 October 2025 (UTC)

• I'm proposing a paragraph like that below which recognises wider public discourse around the phenomenon of Long Covid. I've taken WP:NOTMEDJOURNAL, WP:UNDUE, WP:MEDRS and WP:RS in particular into account. I wonder if other editors have any thoughts?

'Some media outlets and commentators have speculated that psychosomatic or psychological factors may contribute to the experience of Long Covid symptoms. While these discussions are part of the broader public discourse, most medical research continues to lay stress on biological mechanisms as the primary drivers of the condition. Coverage of psychosomatic explanations should be understood as reflecting public discourse rather than as a statement of medical consensus'.

References (there are many options): https://www.thelancet.com/journals/eclinm/article/PIIS2589-5370%2824%2900335-3/fulltext?utm_ https://www.psychologytoday.com/us/blog/common-sense-science/202503/is-long-covid-being-mistaken-for-mental-illness?utm_source=chatgpt.com https://www.thetimes.com/uk/healthcare/article/why-is-britain-now-the-capital-of-long-covid-grjpvzfvw. Additionally, I would cite the Wikipedia article https://en.wikipedia.org/wiki/Psychosomatic_medicine.

— Preceding unsigned comment added by Emmentalist (talk • contribs) 08:29, 15 October 2025 (UTC)

The only one of those that's even potentially compliant with Wikipedia:Identifying reliable sources (medicine) rules is the first, which is PMID 39764180. What "some media outlets and commentators have speculated" is WP:UNDUE.

We should normally use the name somatic symptom disorder instead of psychosomatic, which is often perceived as derogatory and misunderstood as being voluntary. However, the Lancet article doesn't even apply that to long COVID; instead, it talks about "persistent somatic symptoms".

Looking through a few primary sources, it appears that the situation is approximately if you already have depression (which causes fatigue), then you're at higher risk of long COVID (the most common symptom of which is fatigue). WhatamIdoing (talk) 01:41, 16 October 2025 (UTC)

Hi User:Emmentalist. If you want to use AI for finding sources, I can recommend perplexity AI over ChatGPT. My experience is that chatGPT is accurate in around 20% of recommended sources, whereas perplexity AI reaches 40% or so. Both are atrocious, but perplexity AI might be faster than Google if you make sure to double and triple check the sources. One thing that trips some people up is the word 'somatic', which means physical. —Femke 🐦 (talk) 07:20, 16 October 2025 (UTC)

I think the goal of the proposed paragraph is to say something like "Some people say it's all just in your head, but real researchers say that's wrong, and it's a physical disease". WhatamIdoing (talk) 07:38, 16 October 2025 (UTC)

No, psychosomatic illnesses are as real as those with biological markers. Emmentalist (talk) 08:52, 16 October 2025 (UTC)

Thanks, @Femke As I said, there are many possible sources (I put a few quickly for illustrative purposes only, two from Chatgpt). The citations in the actual published paragraphs would not be these ones as there are plenty better ones. Emmentalist (talk) 08:52, 16 October 2025 (UTC)

I think we should be consistent with Wikipedia elsewhere as per MoS. Psychosomatic is used publicly and is well understood. See, for example, Psychosomatic medicine. Emmentalist (talk) 08:56, 16 October 2025 (UTC)

On the contrary: the word psychosomatic is so poorly understood that the psychiatrists removed the word from the Diagnostic and Statistical Manual of Mental Disorders decades ago. The last version that used that language was the DSM-III-R, which was published in 1987. WhatamIdoing (talk) 23:58, 16 October 2025 (UTC)

Psychosomatic is a well understood word and indeed in wide public usage, including here at Wikipedia. Etymologically, it's a simple construct; essentially a portmanteau referring to the connection between mind and body. Its deconstruction in DSMs IV and V, and its replacement by other medical terms, is noteworthy and that would be relevant for inclusion in the article. A key policy here, I think, is WP:NotMedJournal. best wishes Emmentalist (talk) 06:58, 17 October 2025 (UTC)

Could you please link that policy? That shortcut isn’t pulling anything up for me. Innisfree987 (talk) 07:10, 17 October 2025 (UTC)

There's WP:NOTJOURNAL. It's point is that we shouldn't be using technical jargon that average readers can be expected to misunderstand. WhatamIdoing (talk) 07:20, 17 October 2025 (UTC)

Etymologically, it's susceptible to the Etymological fallacy.

According to sources such as PMID 15232041, psychosomatic frequently has "a pejorative meaning, such as “imaginary” or “made up”." This website says "Previously, terms like “psychosomatic” and “psychogenic” were more often used, but health care professionals recognized that many patients found those terms dismissive and insulting—and rightly so". The Academy of Psychosomatic Medicine changed its name in 2018 because "the name Psychosomatic Medicine was problematic. Many expressed the concerns that the name had a pejorative connotation". WhatamIdoing (talk) 07:19, 17 October 2025 (UTC)

I don't think that's an appropriate example of the etymological fallacy. Your first source (from 2004) describes the use of the term in newspapers, which we all know often misrepresent and misuse terminology. It notes that "In March 2003, the American Board of Medical Specialties unanimously approved the creation of a seventh psychiatric subspecialty, “psychosomatic medicine,” that defines all of those working in consultation-liaison psychiatry and general hospital psychiatry". Unfortunately it reports "the word “psychosomatic” is frequently used in newspapers to depict illness in a pejorative way. Specifically, it is often used to imply that an illness is not important or is imaginary, malingering, a sign of madness, or a character flaw." However they also note "We also found that when its use was not pejorative, the word “psychosomatic” was usually used to suggest a psychological problem or the effects of the mind on the body rather than a reciprocal interaction between body and mind. Some patients may find this meaning also stigmatizing." That, to patients who will only accept a physical or external cause for their illness, there really isn't going to be a terminology that they don't find stigmatising or pejorative. But on the other hand, the term doesn't adequately describe the consequences on mental health of having a long term physical illness, and the need to treat that also. They, possibly naively, concluded "If the new speciality of psychosomatic medicine is not to be dogged by stigma, it appears that we must work harder with the media both to rehabilitate the word “psychosomatic” to its proper meaning and to convey the many positive advances in psychosomatic research and treatments." The name change of the Academy of Psychosomatic Medicine to the Academy of Consultation-Liaison Psychiatry doesn't just reflect an admission that "rehabilitating" the "proper meaning" failed, but a widening of the scope of the name to explicitly include psychiatric support for illnesses with known physical and external causes.

I don't think there is anything useful to be learned from the second source, which is written in the style of an extended twitter rant, where enemy health professionals are described as stupid and fraudulent. From a scan of their blog, this author appears to be in the camp of those who wouldn't accept any terminology that suggests a psychological cause or effect for any ailment. There is a counter argument that those in this camp are themselves prejudiced against those with mental illness, associating it with a character flaw, madness, or imaginary diseases. And so even though the word "psychosomatic" (or any alternative term) is perceived by them as pejorative, it is questionable that Wikipedia should give such a prejudicial response any merit when deciding what word to use.

In summary, I don't think we should be swayed about word usage (a) because newspapers get it wrong or (b) because activists opposed to the entire concept in principle will find it pejorative. Some of our readers will find the term pejorative, which may suggest alternatives would be perceived more neutrally. But it might suggest we may never find a term that is accepted by all or isn't merely a euphemism.

I do agree that the article isn't strong on the problems with Long COVID as a diagnostic entity. The Diagnosis section mentions some issues but doesn't address the consequence of that, which is that some don't find it a useful entity for that very reason. The Epidemiology section has numbers that effectively look random and the source studies have a huge range of figures underlying them. For example, Prevalence of mental health conditions and brain fog in people with long COVID: A systematic review and meta-analysis states "It should be noted that the research summarised in this review is fundamentally descriptive, so causal attribution to acute COVID-19 disease is not possible." What is currently being described as long COVID might end up being a variety of conditions, some with physical cause attributable to the viral disease, and some not. -- Colin°^Talk 09:56, 17 October 2025 (UTC)

I agree psychosomatic is an overly technical term that doesn't help our readers much.

I don't have time to do another review of the literature, but have been toying with the idea of bringing this up to FA when the science is slowly starting to settle next year or the year after. We discuss in the terminology section that it may be a variety of conditions, but the current literature is slowly moving away from this idea, as subgroup analysis / clustering isn't showing a lot of differences between groups.

We discuss the misdiagnosis in one direction (people being told that it's anxiety instead), but not really discuss the misdiagnosis the other direction yet. From ME/CFS research, it seems plausible that 25% is misdiagnosed. Maybe slightly lower here, as people do have access to viral testing during acute infection, but still. —Femke 🐦 (talk) 16:14, 17 October 2025 (UTC)

On our article currently, we have a diagram of "over 50 long-term effects of COVID-19". The source here makes that claim in its title. That these are "effects of COVID-19". That's a causal effect claim, for which we usually demand more than just "X happened and at some point later, person reports Y ailment". We've just experienced the circus of a president claiming paracetamol is the cause of autism. So we know the danger of cause-effect claims that aren't really backed up with evidence. The UK health service reckon over 80% of the population caught covid at least once. And this list of 50 symptoms includes, well it includes pretty much most of the common things one might regard as an ailment and maybe visit the GP about. It isn't like nobody ever got fatigue, or weight loss, or chills until covid. The source says "We identified a total of 55 long-term effects associated with COVID-19 in the literature reviewed" which sounds like they are taking all the literature at face value.

Our diagram of "Long-term effects of COVID-19" is based on a paper submitted in January 2021. It claims "80% of the infected patients with SARS-CoV-2 developed one or more long-term symptoms". By "long-term" they mean "14 to 110 days post-viral infection". The upper limit of 110 days appears to be because, well, covid hadn't been around very long for those cited studies. There's quite a difference between someone still being fatigued three weeks after being hospitalised with covid, and someone still being fatigued six months after getting covid and feeling rotten for a week. These are all lumped together. We'd frankly be surprised if someone wasn't fatigued for some time after a spell in ICU.

We claim "hearing loss or tinnitus" affects 15% of people who got COVID. But the two sources that looked at this examined 425 cases, not the millions of us who caught covid. One letter to the editor from July 2020 examined 138 hospitalised patients and followed them up at 8 weeks with a detailed questionnaire that included a question about hearing loss and tinnitus. Eight reported hearing loss and four already had pre-existing loss. Eight reported tinnitus and three had pre-existing tinnitus. That's 11% if we combine the two, but 6% each condition (nobody developed both).

The second study first published September 2020 looked at 287 "survivors", 80% of whom had a mild infection, 15% required oxygen at home and 5% ICU. The follow up duration isn't clear but was a minimum of 20 days after negative PCR. They report 16% ticked tinnitus in their questionnaire, but don't report on hearing loss at all. They suggested "All subjects recovered from COVID‐19 should undergo long‐term monitoring" which would be a remarkable burden on the health service.

Today is five years later. Have we no better sources on hearing loss and tinnitus as a symptom of long COVID? This is a letter to the editor in the early days of the pandemic reporting 8% of hospitalised patients developed tinnitus at 8 weeks, and a study two months later showing mostly mild cases reported 16% developed tinnitus at some unspecific point (can't be more than weeks) after. No long term data. Not really sure why both of these hearing conditions were even lumped together and reported as 15%.

If over 80% of the UK population caught covid at least once, it clearly can't be true that 12% (80% x 15%) of the UK population, or 8 million people, now have hearing loss and tinnitus as a post-viral consequence.

That's just looking at one claim. -- Colin°^Talk 17:21, 17 October 2025 (UTC)

There are better sources out now, yes. The article is roughly stuck at what was available in 2023. I might start from scratch when I work towards an FA next year, as the typical 5 years of meddate doesn't work well. Will try to use mostly sources from the last 2 or 3 years. —Femke 🐦 (talk) 18:06, 17 October 2025 (UTC)

Found one source for potential updated figures at least (saving this for later): https://www.cell.com/action/showPdf?pii=S0092-8674%2824%2900886-9. —Femke 🐦 (talk) 19:21, 18 October 2025 (UTC)

There is a vast amount of material which discusses the possible psychosomatic origin or role in some or many Long Covid cases. There's really no question of WP:Undue. We must also note WP:NotMedJournal. MoS also seeks consistency across articles and there are already articles, as I note below, which use the term psychosomatic. I understand that it is often seen as a pejorative and that is part of the discursive phenomenon: that's worthy of inclusion in itself. The Wikipedia article on Somatic disorders frames the disorder as a noun, whereas psychosomatic is generally employed as an adjective and remains the most commonly used term, so I think it's best to use the latter here. Emmentalist (talk) 09:09, 16 October 2025 (UTC)

• Thanks for those comments. No-one commented on the actual paragraph, though. Here it is again, shortened. Should it go in, with appropriate citations, in this form, in an amended form or not at all?

'Some clinicians and commentators suggest that psychosomatic or psychological factors may contribute to the experience of Long Covid symptoms. While these discussions are part of the broader public discourse, most medical research stresses biological mechanisms as the primary drivers of the condition'. Emmentalist (talk) 10:47, 18 October 2025 (UTC)

It's difficult to comment without sources. Writing text and then finding sources to support it isn't usually the best way to deal with more contentious topics. Instead, find the highest-quality review papers, and summarise what they say on the topic. —Femke 🐦 (talk) 10:51, 18 October 2025 (UTC)

Perhaps this will be clear:

• Should these words be added? No, because we do not have appropriate citations.
• Should these words be added? No, because this "he said/she said" style is not a desirable and encyclopedic style for this article. We don't want to provide facts about what some people say ("Some clinicians and commentators suggest..."). We want to provide facts about long COVID ("This is/isn't a factor").
• Should something else be added about the relationship between pre-existing mental health status and the risk of distress post-COVID? Probably. But we need the sources before we can figure out what to say. And whatever we add shouldn't contain the word psychosomatic.

As a side note, Emmentalist, I find myself in the last few days assuming that you're not from the US. I don't have any basis for this belief (no obligation whatsoever to comment on this or to say whether I'm right or wrong!) except that mind–body dualism is hugely influential in Western culture, and you see happily free of it. WhatamIdoing (talk) 19:46, 18 October 2025 (UTC)

Happily the third item is already covered under risk factors, so I’m not sure anything needs to be added. Innisfree987 (talk) 20:17, 18 October 2025 (UTC)

I agree we shouldn't start with a POV and go finding sources for that. Looking at

• Mechanisms of long COVID and the path toward therapeutics paper Femke linked above
• Research on post-COVID syndrome: current gaps and future perspectives - International Journal of Infectious Diseases a survey of the current state of play

I see a lot of uncertainty. While it is desirable to provide facts vs arguing a case in front of the reader, some topics just remain problematic. We can talk with certainty about the diagnosis, the definition, the treatment for, the prognosis of, something like epilepsy. It's been stable for years and has solid foundations. Whereas Long COVID strikes me more like an ongoing journey of discovery, more appropriately covered by a narrative perhaps than fixed MEDMOS sections. Both sources make it clear there is deviance among what to call this (do the differences matter or not?), what the criteria are (how long, how severe, is a formal COVID diagnosis required, are self reported symptoms sufficient, etc). There are various theories as to cause. Unlikely to be one treatment. And we have the complication that probably most of what was published in the first year or two after COVID arrived is now junk. These sources also note the complication that widespread vaccination brings, meaning the cohorts who end up very sick or in ICU are vanishing, and more patients are ones who have caught it more than once. There isn't widespread testing for COVID any longer, so, unless you end up in hospital, it is unlikely you'd know if you had COVID or a cold.

If this article has mostly developed organically since the concept was first proposed, then ruthlessly rewriting it is likely a good approach. I don't want to discourage you, Femke, from improving this, but stability is an FA requirement. It might be reviewers think this is an evolving topic, rather than one where what we write today is still likely to be true in six months or a year. -- Colin°^Talk 18:16, 19 October 2025 (UTC)

The stability criterion is "it is not subject to ongoing edit wars and its content does not change significantly from day to day", which to be implies to me okay that it will be somewhat outdated within a year. Still, the reviews I read from 2024 already seem somewhat out of date to me.. Let's see if 2026 brings more stability.

I always think there is value in writing about topics while they're 'hot', as that's the moment putting down what is known and what isn't, is most useful.

The review papers I'm reading use the standard sections we're used to in mosmed, so I don't think it's necessary to change those. —Femke 🐦 (talk) 18:32, 19 October 2025 (UTC)

It's certainly worth writing about, and something "under development" is interesting in its own right. I'm just saying this isn't a mature topic, and I'm not sure how the FA reviewers will react to that.

Wrt WAID, "he said/she said" comment, of the few reviews I looked at, they all attribute a definition (e.g. three months post infection) and a name (lots of variants) to an official body and offer several choices before then adding their own. I don't think at this point it is possible to state as a fact what name it has (other than that we've chosen Long COVID) or the criteria. One review noted that every single study had its own criteria. The early studies seemed to be explicitly of an "information gathering" style of reporting what doctors saw in the field, rather than I feel any rigorous attempt to isolate ailments as likely to have a post-viral cause. -- Colin°^Talk 08:24, 20 October 2025 (UTC)