I wrote this for my Substack. Posting it here as I hope that I am wrong, and I'm looking for people to debunk my arguments. I found this post and this post, but both are old and contain numerous factual errors (I don't think they were wrong at the time of writing, but better data has come out since they were written. The point of this post isn't to debunk them directly, and I can obviously expand if requested, but for example, we now know that long COVID in children appears quite common).

Imagine a world where hundreds of millions of people are too sick to work. Could society continue to function?

It may sound hyperbolic, but barring some incredible fortune1, it looks like that’s where we are headed: A decade from now, almost one third of the world’s population could suffer from a severe chronic illness — the consequence of repeated COVID infections.

This condition, known as long COVID or post-COVID, isn’t rare. It affects many people who catch COVID. Even if they caught COVID before. Even if they are young and healthy. Even if they are vaccinated (or not vaccinated).

There are still a lot of unknowns, but the story that’s emerging goes like this:

  • Given that COVID is highly contagious, and immunity short-lasting, every person on the planet will probably catch COVID many times over their course of their life.
  • Every time a person catches COVID, there’s a chance they will develop long COVID. So with each reinfection, the cumulative probability of developing long COVID increases.
  • All people with long COVID suffer from a reduction in their quality of life, but some are worse than others. Many people with long COVID are too disabled to work.
  • Some people recover from long COVID, but for others it may be lifelong. And even when people do recover, a COVID reinfection often triggers a relapse back into long COVID. Reinfections can also make existing long COVID more severe.

Over time, as people rack up more and more COVID reinfections, the total number of people suffering from long COVID will increase. Some of these people will be too sick to hold a job. More and more people will be forced to turn to government disability programs. The labor force will shrink. Healthcare systems, and society in general, will be tremendously strained.

What is long COVID, and what causes it?

There are famous actors (source), athletes (source), politicians (source), and entertainers (source) who have publicly struggled with long COVID.

The CDC estimates that around 15 million Americans have long COVID as of June, 2023 (source). Loss of smell and taste, extreme exhaustion after physical activity, chronic cough, and brain fog are common symptoms (source) but there are at least 196 others (source).

Long COVID may be several distinct, occasionally overlapping illnesses. In some cases, long COVID is the result of organ damage from a COVID infection: lesions on the brain (source), scarring of the lungs (source), or heart damage (source).

In other instances, symptoms may be due to a persistent, long-lasting COVID infection (source) — COVID may be hiding out in parts of the body the immune system cannot reach. COVID has been found in stool samples of people months after they seemingly recovered (source), and it has been found in the brain and spinal cord (source).

Another possibility is that long COVID is an autoimmune disease (source) like multiple sclerosis, which is thought to be caused by the Epstein-Barr virus (source). Perhaps COVID is similar.

For some, long COVID could be the result of a vaccine injury (source). But long COVID predates the existence of COVID vaccines (source), and there are studies documenting people with long COVID who were not vaccinated (source).

1. Given that COVID is highly contagious, and immunity short-lasting, every person on the planet will probably catch COVID many times over their course of their life.

About 77% of American adults and teenagers were estimated to have caught COVID at least once (source) as of the end of 2022. At this point in the pandemic, most people know someone who has had COVID multiple times.

Immunity from vaccination or an infection wanes after only a few months (source), leaving people vulnerable to reinfection again and again. In fact, there are documented cases of people getting reinfected in less than 3 weeks (source), and reports of people having been infected with COVID at least 7 times (source).

Given that COVID is now known to be among the most contagious viruses in history (source) and immunity is short-lived, some experts believe the average person may contract COVID at least once per year for the foreseeable future (source).

2. Every time a person catches COVID, there’s a chance they will develop long COVID. So with each infection, the cumulative probability of developing long COVID increases.

What are the odds of developing long COVID?

  • The WHO says 1 in 10 COVID infections will result in long COVID (source).
  • Researchers at UCLA say 30% of COVID patients develop long COVID (source).
  • The CDC found that 20%-25% of COVID patients develop long COVID (source).
  • Chinese University said that 70% of people infected with COVID in Hong Kong developed some form of long COVID (source).
  • Researchers at the Imperial College of London said 1 in 3 people who had COVID have long COVID symptoms (source).

Even using the most charitable estimate, about 10% of people will develop long COVID following a COVID infection. The rest may not. And if they’ve had one or two COVID infections without issue, they make think they’re safe — but they’d be mistaken.

With reinfection, there’s some evidence that a person’s odds of developing long COVID following that particular infection may decline (source) but they do not fall to 0%. Thus, their cumulative odds of developing long COVID increase following every reinfection.

In other words, people who have had COVID twice are more likely to have long COVID than those who have only had it once. And those who have had it three times are more likely to have long COVID than those who have only had it twice (source).

"You will have many patients come to us…with maybe the third, fourth, fifth infection, and now having finally developed post-Covid symptoms" said Dr. Marc Sala in an interview with Fox (source).

It’s not clear if some COVID variants are more likely to cause long COVID than others, but more recent omicron variants, like prior strains, are known to cause long COVID (source, source).

Unfortunately, there doesn’t seem to be much a person can do to avoid long COVID if they get infected. Existing COVID vaccines do not protect from long COVID. They reduce the odds, but only marginally — about 15%, according to a study of more than 13 million people (source). Those with pre-existing health conditions and older people are more likely to develop long COVID (source, source), but healthy people are also getting long COVID in large numbers: about one third of people living with long COVID had no pre-existing health conditions (source).

Taking Paxlovid during a COVID infection appears to reduce the odds of long COVID more than vaccination, but again, it’s only marginal — a 19%-26% reduction (source). Metformin was found to be better still, with a 41% reduction (source) but not foolproof.

At this point, the only way to reliably avoid long COVID is to avoid COVID.

3. All people with long COVID suffer from a reduction in their quality of life, but some are worse than others. Many people with long COVID are too disabled to work.

Long COVID varies in severity. Some people have only smell or taste loss, while others become completely bed-bound. But even a relatively minor case of long COVID results in tangible impacts to quality of life and a significant increase in depression (source). Some people with long COVID have a lower quality of life than people with stage 4 lung cancer, and it has a bigger impact on their day-to-day activities than a stroke (source).

But at what point does this suffering have a tangible impact on society? Perhaps when people can no longer work. The data on that varies, but it’s common for people suffering from long COVID to reduce their work hours or to stop working entirely:

  • A survey of patients seeking treatment at a long COVID clinic found that 20% of them couldn’t work (source).
  • A report from the New York State Insurance Fund found that 18% of people receiving workers’ compensation for long COVID could not return to work for over a year (source).
  • A study from the Minneapolis Fed found that 26% of people with long COVID had their work impacted — about 40% of them couldn’t work at all, and the rest worked fewer hours (source).
  • The Brookings Institute estimated that 4 million Americans were out of work in 2022 because of long COVID (source).

4. Some people recover from long COVID, but for others it may be lifelong. And even when people do recover, a COVID reinfection often triggers a relapse back into long COVID. Reinfections can also make existing long COVID more severe.

Some people with long COVID do recover eventually, but it often takes years. Others have still not recovered since the pandemic began (perhaps they will at some point, but there is reason to believe they may never recover2). A study of over 31,000 people infected with COVID found that 6% of them had not recovered when the study ended 18 months later (source). A more recent study looking specifically at people with long COVID found that over 92% of them did not recover after 2 years (source).

Reinfections appear to be particularly dangerous for people living with long COVID, or those who have recovered from it. A survey of people living with long COVID found that 45% of them who were reinfected had long COVID symptoms that had previously cleared up return with reinfection; worse, 41% said reinfection brought additional long COVID symptoms. For those who had recovered from long COVID entirely, about 60% of them developed long COVID again following reinfection (source).

A decade from now, perhaps one third of the world’s population could suffer from a severe chronic illness

The probabilities are highly uncertain, but let’s settle on a few assumptions and see what the math looks like.

Let’s assume:

  • Every person catches COVID, on average, once per year going forward.
  • The odds of long COVID are 10% after the first infection, and 5% after every reinfection.
  • The average person suffering from long COVID recovers after 2 years.
  • When a person with long COVID recovers, then becomes reinfected, they have a 60% chance of developing long COVID again.
  • 25% of people with long COVID work fewer hours or don’t work at all.
  • There’s no cure for long COVID (currently the case).

Under these assumptions, after 10 years:

  • 40% of people will have had long COVID at some point.
  • 30% of people will have active long COVID.
  • 7.5% of people will be so disabled from long COVID it will impact their work.

The pandemic has been ongoing for over 3 years already, but mitigation efforts were in place to some degree for 2 of those years, which may have limited the number of infections and reinfections.

If we assume that at this point, 80% of people have had COVID, and the average person who has had COVID has had it once, then the assumptions above predict that 8% of people have active long COVID, and 2% of people are so disabled by long COVID it is impacting their ability to work.

The CDC pulse survey puts the percentage of American adults suffering from long COVID at 6% and the percentage of adults where long COVID is having a significant impact on their life at 1.6% (source), so the assumptions above may be slightly too pessimistic, but not wildly so.

The best data to predict where we are headed would be from a large sample of people who have had COVID many times (10+). If most or all of these people had long COVID, then my projections may be close to the truth. Unfortunately, as testing has been scaled back (source), it may be difficult to find such people.

The evidence that is out there seems to point in that direction: a study of over 400,000 US veterans who contracted COVID showed that the more times they had been infected with COVID, the more health problems they had afterwards (source).

Long COVID will cost trillions of dollars.

In a recent paper, Harvard economist David Cutler pegged the cost of long COVID at $3.7 trillion (source). Notably, he looked only at people who are currently suffering from long COVID today (source). He assumes about 9.6 million Americans are living with severe long COVID, or about 3% of the population.

Given that people are continuing to be infected, and continuing to develop long COVID, Cutler’s projection is a gross underestimate. If my assumptions are correct, and that number doubles or triples in the years ahead, the total cost of long COVID could be in excess of $10 trillion, or almost 50% of US GDP. Of course, that is only looking at the US — long COVID affects every country on Earth.

“The oncoming burden of long COVID…is so large as to be unfathomable,” wrote a group of immunology professors in a recent analysis in Nature (source).

I disagree — although uncomfortable, the consequences of long COVID can be understood. The data is out there, and the direction we are heading in is obvious. Hopefully, once more people understand this problem, we can begin to solve it.

1Some possibilities: (1) The virus evolves in such a way that future variants no longer cause long COVID. (2) Hygienic infrastructure (e.g. air filtration, devices that detect the presence of COVID in the air) or a sterilizing vaccine is deployed at scale and it sharply reduces the number of infections. (3) Effective treatments/cures are developed for long COVID.

2It’s a different coronavirus, but SARS-CoV-1 produced a similar phenomenon to long COVID. Far fewer people were infected with SARS than with COVID, but of those that were, some of them developed “long SARS” and they have not recovered even 20 years later (source).

New Comment
15 comments, sorted by Click to highlight new comments since:

I don't have time to check all of these links to see if this is the one time people didn't do this, but I've been pretty skeptical of the "Everyone is going to get Long COVID" narrative since the arguments for it seem to always make the same two mistakes:

  • Defining Long COVID inconsistently, including people who cough occasionally as having Long COVID when counting cases, but then focusing on people with severe symptoms when talking about how bad it is
  • Ignoring reverse causation -- getting COVID several times is correlated with bad outcomes, but how much of that is because people who have health problems are more likely to get sick?

Rather than collecting a bunch of different statistics and hoping that naively multiplying them works, I'd be more convinced by a simple graph of people who can't work due to Long COVID over time that shows a clear upward trend that doesn't seem to be leveling off (and ideally, comparing that to other sources of disability, and breaking it down by age so we can see if this is "suddenly 20 year-olds are getting freak disabilities" or "80 year olds are getting Long COVID along with all of the other health problems at that age").

You can see that the number of people in the US with a disability (according to the Fed) has surged since COVID. https://fred.stlouisfed.org/series/LNU00074597

That's not perfect, but it's a compelling proxy.

that graph:

  • uses raw total, not adjusted for population size
  • is not adjusted for age of population either
  • doesn't define disability on that page
  • shows a marked decrease in disability levels in 2020, suggesting. There are reasonable explanations for that, but it only makes me more skeptical the methodology is a match for the question you're trying to answer. 

I haven't done the math, there could still be an uptick in age-adjusted risk of disability. But this graph doesn't show it.

Sure. Here's the thing: We don't have that data.

What we do know is that from 2010 - 2021, the number of disabled people in the US rose by about 11%, while the population rose by about 6%. Then in 2021, the number of disabled people begins to spike, and is now up 13% alone in the last 2.5 years, while the US population is only up about 1% in the same timespan.

That is a huge increase. Something is obviously causing a sharp increase in disabled people. What could that be? Could it be the novel virus that has a proven mechanism to make people disabled that began spreading through the population as restrictions were lifted in 2021?

Downvoted because I don't think the evidence presented here is strong enough to justify the claims made by, e.g. the title. After reading this I'm left with a lot of questions like:

  • are we currently suffering a disaster caused by long influenza?
  • are some countries locally suffering a disaster of long SARS?
  • long AIDS?
  • our systems seem to route around a large chunk of people not working, are we sure this would be a disaster?
  • what makes us so sure long COVID will be uniquely bad?

At least personally I want to see posts on Less Wrong that, when they make bold, confident assertions like this post does, put in substantial effort to back them up OR make it clear that they are bold, speculative assertion. I'm judging this post based on its tone, and as a post in the category of bold, confident assertions, and the lack of evidence presented to back the assertions makes me dislike it.

I thought about giving the long flu example, but flu is much less contagious than covid and does not infect everyone yearly. That holds even more for SARS or MERS. 

Nope.

According to the CDC pulse survey you linked (https://www.cdc.gov/nchs/covid19/pulse/long-covid.htm) the metrics for long covid are trending down. This includes: currently experiencing, any limitations, and significant limitations categories.

The big issue with your essay is actual definition of "long COVID." As you touched on, a very wide range of symptoms with different levels of severity and potential mechanisms are all being lumped together in this category. Your argument assumes generic "long COVID" falls towards the more severe end and extrapolates from there. 

Rather than dealing with long COVID as a monolith, it's probably necessary to break it down by specific symptoms, magnitude, and duration. If you want to talk about large economic consequences, you need to filter out the large proportion of symptoms that are legitimate problems but do not rise to the level of macroscopically affecting employment.

For example, a lot of COVID cases seem to retain a persistent cough for weeks or months after the acute period. Are these people suffering from long COVID? Technically, I'd say yes, but that won't affect their position in the work force. Similarly, many people develop "brain fog" that doesn't get in the way of their job pouring concrete or shuffling papers. These are conditions that might more severely disrupt an athlete or scientist respectively, but would be lost in the noise for the average person.

Most people in the US aren't really healthy to begin with. They often lack exercise and consume a lot of low-quality food. The noise floor for poor health is pretty high. A guy who eats at McDonald's all the time and hasn't run since high school is already in bad enough shape that these kinds of nebulous long COVID symptoms will be lost in the noise. 

It remains unclear to me what the real likelihood of various post-COVID ailments really is. There have been so many poor studies that relied on self-reporting through surveys or lacked proper uninfected controls (not even nucleocapsid antibody tests), and of course the totally inconsistent definition of what "long COVID" actually is. We also need to account for vaccination status including the particular vaccine generation and the specific strain of SARS-CoV2 at hand if possible. The real question now is what are the chances of Symptoms A, B, C... with magnitudes X, Y, Z... right now, with XBB-strain virus and BA.5 or XBB vaccination, not what was happening to naive populations in 2020 or with mismatched vaccines. 

More generally, I'm surprised the potential long-term consequences of COVID on intelligence (i.e. brain fog) don't receive more discussion. This is a community where some are so concerned with eking out a few more IQ points that they're willing to consider experimental genetic selection IVF techniques on their offspring or self-medicating with ostensibly nootropic substances. Perhaps people feel that avoiding it is hopeless, so it's easier to dismiss it as a concern. 

Isn't there also evidence that long covid is partly psychosomatic? (random paper that lists some studies)

The number of people prone to psychosomatic symptoms is probably not going to go up, so your growth rate should be overestimated. 

There are also other risk factors involved, same argument applies to those.

In the extreme scenario those people prone to develop long covid already have it and very few other people will get it. 

The assumptions in your simulation also seem consistent with that possibility:
Maybe the reinfection long covid probability of 5% is mostly the 60% of the 10% ... ;-)

For what it's worth I know zero people with long covid and I have also never heard anybody mention an acquaintance with long covid. 

Isn't there also evidence that long covid is partly psychosomatic

I think an equally compelling explanation is that people with symptoms are more likely to attribute them to covid than other causes because it leads to better treatment. The implications for long covid in particular are identical though. 

It's the opposite, actually. I have family members with obvious long COVID symptoms they don't acknowledge because they're committed to a political ideology which asserts that COVID is "over". Most people with long COVID will have to visit half a dozen doctors before they are properly diagnosed. And long COVID treatment is literally non-existent. There are no approved treatments or therapies at this point. None. So, rather than get "better" treatment, they won't get any. There are long COVID clinics at various universities. They have waiting lists that are several months long, and their treatments involve things like coaching on CFS pacing strategies (e.g. listen to your body, don't exert yourself).

Apologies- I didn't mean to suggest they got better medical treatment. You're correct that the state of the art is terrible. But doctors, employers, and family members will be nicer to people labeled with long covid than they will be to people with a different label of the same constellation of symptoms.

TBC I'm not disputing your description of "takes 6 doctors to get diagnosed, and that last doctor still can't help".  If anything I think it glosses over how awful the first 5 doctors are on average. But 6 doctors, followed by a socially legitimized diagnosis, is still better than the average fatigue-inflammation-brainfog-metabolic-neuro-??? cluster because that cluster is treated miserably.

Social legitimacy varies by social group, of course. Some people are assholes about long covid and that's devastating for their friends, family, and employees who suffer from it. But I have trouble imagining any of them would be nicer about a CFS or chronic Lyme diagnosis, while I think there are a decent chunk of people who mock those diagnoses while taking long covid seriously. 

Well, we know that COVID damages the brain, and some people have had psychotic episodes following a COVID infection.

But long COVID is not psychosomatic, at least probably not in most cases. As it is novel, there's no straightforward test at this point, but there has been a decent amount of research in the last couple of years, so physical markers have been found. Some I listed in the essay, but others include things like inflamed vagus nerve, different enzyme levels in the blood, etc. And of course, a lot of long COVID is actual organ damage which can be viewed on an MRI or other tests.

I personally know many people with long COVID, but they often didn't identify it as such. I'm shocked that you don't at least know one person with long term smell/taste loss.

[-][anonymous]00

I feel like more effective, robust vaccines ought to have been invented at this point. Perhaps the standards were lower for the first ones, but it seems they were all approved quite quickly and none of the new ones since have ever made my headway.

If what this article claims is true, advancement due to AI and work automation will likely be more accepted among a population partially incapacitated by illness. There will be less need to keep the pretense that ones right to exist or value as a person is dependent on how good of a worker they are when AI can do any human job better. AI of that caliber could also likely develop therapies to effectively combat long COVID anyway.

What, exactly, is a clinical definition of “long COVID”? Are the people experiencing symptoms long after COVID still testing positive? If not, how can we be confident to call it “long COVID” as opposed to some other cause?

I don’t hang out here a lot. I was “invited” but I’m very different from most people in here. I see “bio security” as an overt attempt at “absolute, centralized genetic control” and consider most international organizations (WHO, UN, etc..,) worthless. It’s better to risk death by hurricane than let anyone control the weather.