Saturday Night Bonus: Kid News

I hope that your weekend is going…well? Do we say that anymore?

I felt like a lot of scary sounding news came out in the last few days about kids and I thought you might like some answers. I can at least provide context, if not necessarily reassurance. I thought about waiting until next week but…well, sometimes, people just need to know.

In no particular order.

Florida, Kids

The Florida Sun-Sentinel published this story, headline that 30% of kids tested in Florida for COVID-19 were positive. Many people seem to have interpreted this to say that 30% of kids in Florida are positive, or they make up a disproportionate share of cases. Neither is true. The same article reports 17,000 of Florida’s 287,000 cases are in children under 18. This is about 6% of cases; by comparison this age group is about 25% of the population. There have been, tragically, 4 pediatric deaths. But as a share of all deaths, people under 18 make up just 0.08%.

It seems likely the high positive rate reflects selection in testing. Testing is annoying (and uncomfortable) and there isn’t much asymptomatic testing among kids. Which means most of the kids tested are symptomatic, so a high positive test rate isn’t that surprising. It is also possible this is just not right, since further reporting suggested many labs in Florida simply were not reporting negative results.

What we do learn here is that kids do get COVID-19; this shouldn’t be news, but I think sometimes it seems that it is given (incorrect) rhetoric that kids “can’t get it.” They can, they are just less likely.

South Korea, Schools

Just tonight, the NY Times and others reported on a very helpful new study out of South Korea. The NYT headline and subhead are terrifying: “Older Children as Likely as Adults to Spread Virus” and “…study suggests school reopening will trigger new outbreaks.”

Let’s dig in.

Here is what the study did. South Korea, unlike the US, has very good contact tracing (let’s not get into this here). The study starts with about 5,700 known COVID-19 patients and tracks about 59,000 of their contacts. They include both in-household and non-household contacts. Through this tracking, they look for how many of the contacts are infected. They divide the sample by the age of the patient, and calculate the share of contacts infected.

(One important caveat to the paper is that they do not actually know the direction of transmission. They identify the “index patient” as the first person who they knew to be infected, but it is possible that this wasn’t the first person infected.)

Before interpreting, let me just state the results, focusing on the kids. The authors divide the children in to two groups: kids 0 to 9 and 10 to 19.

Younger Kids Children 0 to 9 make up 0.5% of cases in the data. For comparison, this age group is about 6% of the total population so kids are much less likely to be infected, which is not new to these data. Among the 57 household contacts of these kids they traced, 3 were positive (5.3%). Among the 180 non-household contacts, 2 were positive (1.1%).

Older Kids Children 10 to 19 make up 2.2% of cases in the data (comparison: they are about 7% of population). Among the 231 traced household contacts in this group, 43 were positive (18.6%). Among the 226 non-household contacts, 2 were positive (0.9%).

We can compare these rates to adults. Among adults over 19, about 11.6% of their household contacts were infected, and 1.9% of their non-household contacts. Older adults are much more likely to infect non-household contacts. Young adults (20-29) are less likely to infect household contacts, probably due to the nature of their households (i.e. roommates rather than family).

Now, to interpretation. I think there are glass-half-full and glass-half-empty ways to take this (the Times clearly went with the latter). Here is what I think is the good news. Outside of household contacts, transmission from kids (and people in general) is really low. If your big worry is kids are out spreading around in the world that didn’t really happen here. Transmission is under 1%.

(NOTE THAT IN SOUTH KOREA THEY WEAR MASKS AND WERE SOCIALLY DISTANCING. SO WEAR A MASK AND THANK YOU FOR COMING TO MY TED TALK.).

Also good news is that among the youngest age group (basically, elementary school kids) the transmission even within household is really low. I actually find this pretty surprising. Think about all the contact you have with a little kid if you are, like, their mom. Still, only 3 of 57 contacts were infected. If your little kid has COVID-19, there seems to be only about a 5% chance you’ll get it.

And final good news is kids are less likely to be infected in the first place. Which we knew, but is always good to confirm.

On the glass half empty side, the older kids do seem to transmit in household as efficiently as adults. It would have been helpful to be more granular on the ages here (like, are 10 to 14 different from 15 to 19) but this is what we have. I think this does raise more concerns than we might have had before about older kids bringing the virus home. Again, their transmission to non household contacts is really limited and they are less likely to be infected than adults overall.

What does this say about schools in the US? I’m not sure how much I am updating here. The headline suggests this implies schools will be super-spreader locations, but given the non-household evidence I am not sure that is necessarily what we would conclude. But this isn’t a study of schools; at this time schools in South Korea were mostly closed, so for evidence on open schools we probably want to look more to Europe. I will also say that South Korea and America are not that similar in a lot of ways (mostly ways which imply schools in America will be higher risk, so that’s not a reassuring comment, just a note).

What I will say is that this may further build a case for focusing school reopen efforts on younger children. They appear to be at the lowest risk, and are least likely to spread. This data confirms that in a large sample. They also happen to be the group who learns least well at home. If we need more space, we may be best off holding high school kids out and using their space for the under-10 crowd. (I am not just saying this because my oldest child is 9.)

Babies in Texas

Various outlets have reported a fact, from a Medical Examiner in Texas, that 85 infants under 1 in a county have tested positive (out of about 8,000 cases). I wish there was more information on this. This is a very high share; it would mean that infants are more likely than adults to be infected as a share. No more information was provided here, about their condition or anything else. I am very much hoping more details will come out.

I do think this is a time when we want to step back to our priors for a minute, which I generally encourage people to do. At this point we have a lot of other data suggesting that kids, including infants, are less likely to be infected. When you see a “new” fact like this, there is always a temptation to whiplash to thinking, “Well, I guess we were wrong before.” But, in fact, the older evidence still stands. And in this case the older evidence is probably more compelling.

We should try very hard to find out more about this incident. But I think it is premature to revise our entire view on COVID in infants at this time.

Update: The county clarified that this is the total number of positive infant tests through the pandemic. In addition, they clarified they have been doing extensive testing of family members of high risk individuals, which may account for their higher numbers.

Bottom Line

Deep breath. News is scary. More information is good; it helps us make better choices. But we always need context.