(1/22) Here's a new paper from me about some of the biases affecting research on children and COVID-19, and more importantly, some solutions.

It hasn't been typeset yet (the PDF might be hard for non-English speakers to read?), so I'll do a quick summary.
(2/22) First, "bias" has a different meaning in epidemiology compared to everyday life.

It doesn't mean someone's done something wrong. It just means there's a methodological issue that affects the results.

This article explains some of the common types.
(3/22) Now, on to the paper. It's been argued that children are less susceptible to infection with SARS-CoV-2 than adults and play only a minor role in transmission.

This conclusion is likely premature, because it's often difficult to detect infections in children.
(4/22) What we do know is most children don't get seriously ill. The infection fatality ratio (mortality in everyone who's infected) is ~0.002% in children, compared to 0.1% for middle-aged adults, and more than 1% for older adults.

However, children can get long COVID & MIS-C.
(5/22) Data from the Office for National Statistics in the UK suggest that up to 13% of children still have symptoms at 5 weeks after infection (compared to about 22% for the population overall).

However, most children are asymptomatic - the prevalence may be up to 50%.
(6/22) One way to try to work out how susceptible children are to infection is to test everyone in a household after someone gets infected.

Some studies have found children are much less likely to be infected than adults, while others have found the risk is similar.
(7/22) But some studies only tested people with symptoms, and there's evidence children have been tested less than adults.

This is called differential bias.

There's also evidence to suggest that cases in children may be detectable for a shorter period than adults.
(8/22) For this reason, antibody studies might be a better way to detect infections in children.

When we compare the results of antibody and PCR tests, we see that infections are frequently missed for adults and children alike, but they are missed in children more often.
(9/22) In some antibody studies, young children and adolescents were similarly likely to test positive.

This suggests that young children are not less susceptible to infection as is widely believed.
(10/22) But antibody studies have some limitations, too.

Compared to adolescents and adults, young children are less likely to develop the kinds of antibodies that antibody tests look for.

This is due to the way that young children fight off the virus.
(11/22) Back to the household studies.

There's another problem, which arises because adults have often been index cases in these studies (because schools were closed, and children had less opportunity to get infected).
(12/22) Who had the closest contact with the adult index cases? Generally their partners.

For that reason, we see that partners are more likely to be infected than other adults in the same household.

This also means we're not comparing like with like when we look at children.
(13/22) Given adult partners of index cases are at high risk & typically make up most of the adults in these studies, we would expect to see more adults infected than children.

The adults had closer contact!

So higher adult attack rates don't mean children are less susceptible.
(14/22) It's also possible parents tried to shield their children from infection (healthcare workers have been over-represented in some studies).

We also see other evidence of circumstantial factors.
(15/22) In one household study, children were twice as likely to be infected when the index case was their mother, rather than father.

Another study found children were more likely to be infected if they were children of the index case.
(16/22) A big problem has been testing. In the early days of the pandemic, it was hard to for children to get tested because they didn't meet testing criteria. They still might not if they have atypical symptoms.

In some countries, they're still not tested unless seriously ill.
(17/22) I said that cases in children might be detectable for a shorter period. How can we get around that? One option is anal swab testing because faecal shedding is prolonged.

A study in Wuhan found about one-fifth of children had negative nasal swabs, but positive anal swabs.
(18/22) Another thing we need to do more of, is random testing of people without symptoms.

At the time I wrote the paper, random testing showed that children and teenagers in England were more likely to be infected than any other age group.
(19/22) Data from England also show that in the last part of 2020, children and teenagers were more likely to bring the virus home than adults, and were more likely to transmit it to their household members.
(20/22) What about these studies we see where children aren't transmitting the virus at school?

Well, we have to remember that most people (children AND adults) don't transmit.

About 5-20% of people account for 80% of secondary cases, & about 70% of people don't infect anyone.
(21/22) In Victoria, Australia, 66% of infections in childcare centres and schools were limited to a single case.

That sounds reassuring, until you remember about 70% of people don't transmit.

So what happened in schools was actually consistent with what happens generally.
(22/22) So, to wrap up:

➡️ Children are more susceptible to infection than commonly thought.

➡️ Children probably play an important role in community transmission.

➡️ We need to design better studies

➡️ We need to interpret the data carefully

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Dr Zoë Hyde

Dr Zoë Hyde Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @DrZoeHyde

24 Oct
The AY.33 sub-lineage of the delta variant might not be readily detectable with current PCR or rapid antigen tests. These can be updated, but in the interim, cases may go undetected.

This is one reason why Australia’s plan to end hotel or home quarantine is premature.
This isn’t the first time variants have caused testing issues. One of the reasons the alpha variant was noticed was because tests for the spike gene returned false negatives.

Fortunately, the tests also looked for other parts of the virus, which had not changed significantly.
A variant also emerged in France that was difficult to detect. Viral loads in the upper (but not lower) respiratory tract were often below the detectable limit in people infected with the B.1.616 variant.
Read 4 tweets
23 Oct
Do you know what’s really funny? I’ve published more on geriatrics than @DrKGregorevic.

And while *I* have published papers on cognitive impairment and dementia, I can’t find a single paper on that subject by Dr Gregorevic in PubMed.
For example, not long ago, I published:

Incidence and predictors of cognitive impairment and dementia in Aboriginal Australians: A follow-up study of 5 years

And one of the projects I’m currently working on is a cluster RCT to improve dementia detection.
It’s interesting that when people don’t like me talking about the science of COVID-19, they resort to using ad hominem attacks.

And not particularly good ones, either.
Read 4 tweets
18 Oct
Long COVID is real, and we need to protect children from it. In Israel, long COVID clinics for children are busy. The clinic at the Schneider Children’s Medical Center in Petah Tikva has about 150 patients and hundreds more are waiting for treatment. 🧵
Liat Ashkenazi-Hoffnung, a clinic doctor says that in some children long COVID "appears as a direct continuation of severe illness but in very many of the children, there is a severe illness, followed by a lull of several months and only then do the symptoms of long COVID begin."
Many children fully recover, but it can take time.

Ashkenazi-Hoffnung says "we had a boy here who was a competitive swimmer and came down with long COVID and was very anxious and in pain. After half a year he went back to swimming and even broke a personal record."
Read 14 tweets
14 Oct
A new study of weekly testing of children and staff at a Belgian primary school shows what we’ve always suspected: if mitigation measures aren’t in place, transmission is common between children and adults at school, and it spills over into households.
In this study, the researchers found:

✅ adult-to-adult transmission
✅ adult-to-child transmission
✅ child-to-adult transmission
✅ child-to-child transmission
The researchers found that the virus readily spread from the school into households.

Teachers infected their partners, and children infected their parents.
Read 4 tweets
29 Sep
Although there are issues with waning immunity, current COVID-19 vaccines offer excellent protection. But this might not always be the case. Future variant-specific boosters may preferentially boost responses to the original strain and be less effective.🧵
The theory works like this: a person exposed to strain A of the virus (either by vaccination or infection) may prime their immune system such that the ability to make future antibodies specific to a future strain (strain B) is reduced.

This is known as immune imprinting.
In that scenario, a vaccine booster for strain B will give some protection against the new strain B, but the immune system will preferentially produce antibodies against the original strain A.

In certain situations, this has the potential to be harmful.
Read 11 tweets
21 Sep
Study of people with mild or moderate (but not hospitalised) COVID-19 from the first wave in Geneva. 7-9 months later, at least 25% had >=1 persisting symptom. Most common: fatigue (14%); loss of smell/taste (11%); headache (7%); shortness of breath (8%).
Note: These proportions were calculated using the entire study sample as the denominator. However, one-third of people were lost to follow-up and their health status was unknown. It’s therefore possible that the proportion of people experiencing persistent symptoms was higher.
Of those with fatigue, 27% said they were limited in strenuous activity. 60% of those with shortness of breath experienced this when walking up a slight hill or when hurrying. Most people with headache or loss of smell or taste reported at least moderate symptoms.
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!