Low confidence, slapdash job. Just putting this up in case other people want to compare notes.

(Update January 2022: Since writing this, I've adjusted my risk assumption down a bit based on newer studies suggesting lower rates of long-COVID than earlier studies (maybe a factor of 1.5 or something?). Then I adjusted down even more because it seems like omicron is less liable to cause serious illness than delta, holding everything else equal (maybe a factor of 2?). Granted there's more uncertainty about long-term effects of omicron, because it hasn't been around that long. But short-term effects have tended to correlate with long-term effects. Also, this is kinda unjustified based on my ignorance and lack of research, but in regards to permanent or long-term damage, I kinda have a good feeling about the fact that omicron is less likely to cause loss of taste and smell, it suggests to me that maybe it's more liable to leave nerve cells alone, and nerve cells are very important and (in some cases) bad at healing, cf. polio. (OTOH omicron seems to cause more brain fog, but that doesn't strike me as a nerve damage thing.) For the boosted adults and vaccinated kids in my family, I figure COVID is sufficiently low-risk that we should basically forget about it and live our lives. If I were the type of person to go bar-hopping, I would be bar-hopping. Except for one catch: our too-young-to-vaccinate kid. I'm figuring 99.5% chance that catching COVID wouldn't be a huge deal even for him. And avoiding catching COVID seems like it would require painful and damaging life disruption (e.g. pulling the kids out of school). So we're not doing that, and therefore we're figuring there's a very good chance we'll all catch it including him, and we're OK with that. We're still not indoor dining etc. during the current insane omicron peak (by contrast we were indoor dining just last month) because indoor dining is not something we care about very much, and because 60% chance of him catching it is still meaningfully better than 95% chance.)

(Update late February 2022: We're back to indoor dining!)


Now that all the adults in my family have been (at least partially) vaccinated, my kids will soon be the most COVID-vulnerable members of my family. Therefore it is newly decision-relevant to get a good sense for exactly how worried I should be about them catching COVID-19. This is April 2021, Boston area, two kids under age 7.

Target audience: Frequent lesswrong.com readers. Everyone else, go away. This is written for people who treat 1-in-10,000 risks as dramatically, wildly, viscerally scarier than 1-in-100,000 risks, people who understand that “zero risk” is a thing that does not exist in our universe, people who understand that life is full of tradeoffs, including between mental health and physical health, etc. etc. etc.

1. Death from COVID-19:

According to this paper written in August 2020 Fig. 2, the IFR is ~3/100,000 for age 0-4, <1/100,000 for age 5-9. I personally can probably adjust that downward from the known lack of risk factors. So that's very low—not worth sacrificing significant quality-of-life over. (That’s like a month or two out of a reasonable fatality risk budget, I figure—even less since we’re not making decisions that swing the risk of COVID-19 infection all the way from 0% to 100%.)

2. MIS-C:

“Multisystem Inflammatory Syndrome in Children” is a frightening syndrome that can produce severe problems including heart problems, neurological symptoms, strokes, and so on. CDC says (via NYT) that they know of 3185 cases (of which 1% were fatal, but death is already included in the previous section) as of this writing. I'll ignore the possibility that there are more MIS-C cases that the CDC doesn’t know about—this is a pretty serious and well-publicized condition, I presume that most kids with MIS-C are being hospitalized and diagnosed. CDC says most cases of MIS-C were ages 1-14, which is I guess a population of 60M in the USA. I dunno how many kids have been infected with COVID total, but if it’s similar to the prevailing rate (figure 28% including undetected cases), then we’re around 2-in-10,000 risk of getting MIS-C, conditional on catching COVID. (The number of detected cases in kids is disproportionately low compared to the rest of the population, I think, but I’ve always just been figuring that they’re less likely to be symptomatic than adults and therefore have an unusually low detection rate.)

Mayo clinic says “In rare cases, MIS-C could result in permanent damage or even death.”, which (accidentally) implies that almost all the time, kids who get MIS-C fully recover without permanent damage. That’s not a great source, but whatever. Also, this says that 7/7 MIS-C cases at a particular hospital were “fully recovered”.

So I figure, conditional on a kid catching COVID, there’s a 2-in-10,000 risk of getting MIS-C, going through a somewhat terrifying ordeal, but eventually fully recovering. And, there's a, I dunno, 1-in-100,000 risk of permanent problems. Again, combine that with the fact that I’m not making decisions that swing the risk of COVID-19 infection all the way from 0% to 100%, and I find this a pretty much acceptable price in cases where I’m spending it on real benefits in my kids’ mental health and quality-of-life. Unless my numbers are wrong of course. So I'm pretty much ignoring MIS-C too. The next two categories seem much worse than that.

3. Long COVID:

Children with long covid” (New Scientist, Feb 2021) says “Evidence from the first study of long covid in children suggests that more than half of children aged between 6 and 16 years old who contract the virus have at least one symptom lasting more than 120 days, with 42.6 per cent impaired by these symptoms during daily activities.” What??? 43%? No way. That’s way too high. This article calls it rare. 43% is not rare.

The 43% statistic comes from Preliminary Evidence on Long COVID in children. It seems like a helpful article but I don’t know what to make of the selection bias. Where exactly did they get these kids? “This cross-sectional study included all children ≤18 year-old diagnosed with microbiologically-confirmed COVID-19 in Fondazione Policlinico Universitario A. Gemelli IRCCS (Rome, Italy).” That should disproportionately sample sick kids, and especially severely sick kids, right? So I’m going to ignore that.

The New Scientist article also says “The UK Office for National Statistics's latest report estimates that 12.9 per cent of UK children aged 2 to 11, and 14.5 per cent of children aged 12 to 16, still have symptoms five weeks after their first infection.” That’s this link. I’m guessing that the population here is "initially-symptomatic kids" as opposed to "all infected kids". So divide by 2 or 3? And not all of those 12.9% are catastrophic. Some may be kids who are easily-fatigued for 5 weeks then recover, which kinda sucks but isn’t that big a deal. I dunno, figure, conditional on a kid catching COVID-19, 2% chance that it’s, like, a really really long and miserable slog that everyone will deeply regret. The rest of the time it's at worst in the ballpark of adding up 1 unusually severe flu + 1 broken leg—lots of pain, hassle, doctors visits, medical bills, missed activities, and so on, but not worse than that. Life goes on.

4. Long-term complications:

Maybe you catch COVID as a kid and then there’s a 1% higher risk of heart disease decades later. Or something else. How would we even know?

My general impression is that kids’ bodies are generally good at recovering and rebuilding themselves over time. But that’s not really based on anything. An example in the opposite direction is polio: I guess polio kills nerve cells in a way that’s unrepairable, and which gradually gets worse and worse over decades after apparent recovery? Is the nervous system unusually hard to repair?? Because, um, COVID often impacts the nervous system too, right?! Yeesh.

I have no idea, I’m out of my depth here.

I guess I'll say 1% chance of a big-deal long-term latent problem, conditional on catching COVID-19. That's not really based on anything, but I need a number because I have to make decisions and weigh tradeoffs. Happy for any input here.


So, conditional on a kid catching COVID, I guess I'm currently thinking that I should mainly be weighing a ~2% chance of a miserable months-long ordeal until they recover, plus (overlapping) ~1% chance of a big-deal long-term latent problem that will show up later in life.

OK, I guess when I multiply everything out right now…

(UPDATE 2 MONTHS LATER TO ADD: When I wrote this list below, COVID prevalence in my region was 16× higher than it is today. So I'm obviously being more relaxed now. Different regions are different; I suggest using the microcovid calculator.)

  • Any masked outdoor activity
    • Risk rounds to zero, do it without thinking twice.
  • Fully-vaccinated adults (6+ weeks past 1st dose) spending time indoors unmasked with my kids
    • If the adults have not recently been spending extensive (or unmasked) time indoors with unvaccinated people, then no problem, don’t even think twice about it. If they have, then try to avoid it, but maybe it’s OK from time to time if there’s sufficient social benefit.
  • Somewhat-vaccinated adults (2-6 weeks past 1st dose) spending time indoors unmasked with my kids
    • If necessary. Depends on what the adult has been up to and how much social benefit we’re getting out of it.
  • Kids go to school (masked)
    • We're already doing that. Our local school is pretty good about ventilation and masking, and has mandatory universal weekly PCR testing. Any remaining risk is more than compensated by the very large benefits for both kids & parents.
  • Kids spending time inside with other kids
    • I guess on rare occasions if there’s a sufficient social benefit. But definitely try to keep such activities outside, until community spread goes down from its current high levels. Oh, I guess there should be an exception for school classmates, since they're already spending time together inside masked, every day at school. But go for masks and open windows.
  • Kids tag along shopping indoors, masked
    • Check the microcovid calculator, but probably not if it’s avoidable, at least not until community spread goes down from its current high levels.
New Comment
22 comments, sorted by Click to highlight new comments since:

Thank you for looking into this and posting about it. This question has been on my mind a lot.

You're welcome but I hope you're not taking my word for anything. Note the warning at the top :-)

Interesting.  We are in somewhat the same boat. Fully vaccinated adults with a two year old.  I think where we come out is as follows.  

(1)  The risk to kids of COVID over the short term are clearly lower than for adults.  Over the long term, it is presently unknown.  

(2) It is highly likely (>90%) that we will be able to vaccinate young children by next year, so any risk reducing measures we take will be temporary. (Also, see (5).)

(3)   The risk from outdoor activities and from vaccinated people are very low.  Therefore, we are fine with outdoor activities masked or not and with socializing with fully vaccinated people.

(4) There are limited gains from indoor activities with unvaccinated people, so we will not bring our daughter indoors with unmasked unvaccinated people or unnecessarily indoors with people whose vaccine status is unknown.

(5) COVID prevalence here is dropping, whether for reasons of increased vaccination or otherwise.  If, due to increased vaccination, those rates stay down, we can relax these restrictions.

This seems reasonable, but I wonder whether "long-term complications" might be a bit underrated. It seems like there are a lot of viruses that have long-term effects or other non-obvious consequences. (I should add that I'm not a biologist, so this is not an informed opinion.)

The example I'm most familiar with is chicken pox causing shingles, decades later from the initial sickness. In that case, shingles is (I think) typically more severe than the original sickness, and is quite common: 1 out of 3 people develop it in their lifetime, according to the CDC.

Other examples that come to mind are measles erasing immune memory (which IIRC wasn't known until recently) and, though not a childhood illness, HPV causing cervical cancer.

Each of these examples has some big differences from SARS-CoV-2, but there isn't much experience with severe coronaviruses, so I don't know how to do better. Maybe the ideal would be to go through a list of reasonably well-understood viruses and check what proportion have known long-term effects or non-obvious consequences (and the rate).

We can get a lower bound from chicken pox and measles. If there are 10-20 common childhood illnesses (based on a quick search), then, using 2 as the numerator, at least 10%-20% of them have consequences that are not immediately obvious. If we go with the 1/3rd rate for shingles (since I don't know for measles), that would translate into a 3%-7% lower bound for covid.

Would I go with a >3% estimate of serious long-term effect or non-obvious consequence from covid to a kid that catches it? A persuasive counterargument that comes to mind is that the immediate experience of covid is less severe to a kid than chicken pox or measles, which would suggest that non-immediate effects are also less severe.

All-in-all, my confidence is extremely low, but hopefully this gives some food for thought.

Yeah thanks!

I guess was thinking that kids who don't get bad cases at the time are unlikely to have long-term effects. I think polio is like that. In particular, I assume that only the bad COVID cases get into the nervous system, where I'm especially concerned. So that's how I got a lower number. But I dunno either :-)

I've also been looking for posts like this, slapdash as you may feel it is (but well-organized!) I have been considering going to an outdoor cookout including our <10yo kids where maybe 25% of the adults are vaccine-hesitant and won't have masks. microCOVID didn't seem to have enough options for me to model this type of group, but I just realized I could do a calculation by only including the number of people likely to be unvaccinated. We decided to go with our kids masked (either surgical or filtered) and distanced, but it's a big step as the first event we will have attended outside our bubble.

Overall, this sounds a lot like flu side-effects:


I'm not saying "it's the flu". I'm saying that if Covid becomes another flu-like childhood disease, as Scott Alexander predicts, then we should expect comparable long-term effects. The question might be how much that changes the baseline of such conditions overall. 


What are the opinions now,with omicron being out there in some countries, and coming to others ? Our kids are 6 and 4 years old. The 6 year old had her second dose only yesterday, because our country was slow in allowing vaccination. The younger one is unvaccinated because there is no aproved vaccine for her.

I added an update at the top.

Keep an inventory of the new BinaxNow home tests and have guests pay for them.