New findings from Swiss seroprevalence and an good opportunity to learn some more infectious disease epidemiology!

A similar prevalence in children aged 6 - 18 y (23%) than the adult population (~27%)

Lower in kids <6y (~15%) and over 65y (<15%)

What can we learn?

1/8
First, another lesson in age binning for kids (will we ever learn?)

6 - 18y is a weird age bin which I would not recommend

It hides big epidemiological differences

In fact, we have seen it in related data before...

2/8
The pre print of the first study from this team used age bin 5-19 y and reported no statistically different prevalence vs adults

medrxiv.org/content/10.110…

They changed for publication and found significantly lower prevalence in children <10y

thelancet.com/journals/lance…

3/8
Unsure why they would do the same thing again here, especially as a much lower prevalence in the under 6s hints it is very likely lower in the under 10's again

This is a much more important category policy wise as encompasses primary education

But we will move on...

4/8
Lets just assume for now the prevalence is the same for adults and children. Would this imply children are just as susceptible as adults?

Prevalence of infection is determined by 2 things:
- Susceptibility
- Exposure

5/8
Even if my susceptibility is low, if I lick door handles on a COVID ward I'll probably get infected

If my susceptibility is high and I lock myself in a cupboard for 6 months, I won't get infected

Seroprevalence alone cannot inform us about susceptibility, but can hint...

6/8
Children are back at school, with many times more contacts than adults, and therefore significantly more exposure

This is precisely the scenario modelled by Zhang et al in June

-Children less susceptible
-Increased exposure evens out prevalence

science.sciencemag.org/content/368/64…

7/8
In summary:

-Don't use silly age bins for children (please), split by policy implications
-Seroprevalence cannot in and of itself tell us about susceptibility
-These results are as expected if young children were less susceptible to infection (evened out by exposure)

8/8

• • •

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

Keep Current with Alasdair Munro

Alasdair Munro 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!

PDF

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 @apsmunro

15 Dec
Here is a good lesson in basic epidemiology

This data is difficult to interpret in a number of ways:
1. There is no denominator, either of institutions or populations
2. It counts the number of outbreaks, not cases
3. It doesn't tell you who had the infections

This matters

1/5
An extreme example of why this is important from earlier this year

After lots of excitement about outbreaks in schools it turned out the median number of children involved was...

Zero

2/5
For some denominator context there are:

- 24,000 schools (9 million pupils) not inc. university
- 11,000 care homes (410,000 residents)
- 117 prisons (79,000 prisoners)
- 26,000 restaurants (many not/partially open)

We are comparing apples and oranges

Epidemiology 101

3/5
Read 5 tweets
11 Dec
It's time to clarify some things about children, schools and #COVID19 🧵

Summary: Young children seem significantly less susceptible, probably less likely to transmit. Less clear for teens. Schools mainly follow community trends, but secondary much higher risk than primary

1
The best way to determine susceptibility is through household contact tracing, as it controls for *exposure* - everyone gets more or less the same

There are many of these. Results vary, which we expect because infection is complicated

That's why we need to combine results

2
Here's 4 meta analyses; all find young children are much less susceptible than adults. Some that teens are too

Zhu (RR 0.6)
academic.oup.com/cid/advance-ar…

Goldstein (RR~0.5)
dx.doi.org/10.1093/infdis…

Viner (OR 0.41)
jamanetwork.com/journals/jamap…

Maidwell (RR ~0.5)
medrxiv.org/content/10.110…

3
Read 24 tweets
27 Nov
Really interesting pre print looking at school transmission in Northern Italy during the second surge

Suggests huge age dependant differences in risk of transmission in children in these settings (who would have guessed?!)

Let's take a look...

doi.org/10.1101/2020.1…

1/8
Detailed contact tracing of 43 cases
- 5 teachers
- 38 children

Whole classes tested regardless of symptoms, and swabbed a second time 10-14d later if the first was within 6 days of exposure

How much onward transmission?

2/8
Pre school (6 kids, 2 teachers) - 0/156 (0%)
Primary school (14 kids) - 1/266 (0.44%)
Secondary school (23 kids, 5 teachers) - 38/572 (6.64%)

Notice a pattern here?

3/8
Read 8 tweets
26 Nov
Some people are *still* lumping "children" all together in talking about transmission, schools etc

That is a dumb thing to do

It doesn't take a paediatrician to tell you a 2yr old is different from a 17yr old

Case study using data from the @ONS

ons.gov.uk/peoplepopulati…

1/8
The @ONS performs large scale, *random* population testing via rt-PCR

People often complain test data in children is biased by lower symptom burden - that is not the case for this data

Hence, it is very informative (why aren't more countries doing this?!)

2/8
Here is a graph of the last few weeks modelled incidence by age group

Huge initial spike in 17 - 24yr olds
Up behind then followed 12 - 16yr olds
Then 25 - 34yr olds
Then 35 - 49yr olds
Then 50 - 69yr olds

Eventually the 2 - 11yr olds get dragged up behind

3/8
Read 8 tweets
23 Nov
IT WORKS 🥳🎉

Great news from the Oxford/AZ vaccine phase III interim analysis today!

ovg.ox.ac.uk/news/oxford-un…

But if you thought you had questions after previous vaccine announcements, boy oh boy does this leave us wanting more!

Quick thread...

1/6
What are the interim results?

-131 cases in Brazil and the UK
-Overall efficacy 70% for the Oxford vaccine compare to control (MenACYW vaccine)
-No safety concerns identified

BUT

Split by dosing regime:
-2x high dose 62% effective
-Low dose then high dose 90% effective

2/6
That is a truly intriguing finding, and we'll need to see the breakdown of the data to understand the uncertainty around those estimates

I'll leave it for clever immunologists to theorise why a lower dose priming shot might be more effective

But wait, there's more...

3/6
Read 6 tweets
13 Nov
An absolute smorgasbord of data around children, schools and #COVID19 this week

We've got @PHE_uk , the @ONS survey, REACT study, and a new SAGE report!

Lets take a birds eye view of the key findings 🧵

1/17
First in @PHE_uk

Cases flattening in secondary age children after a sharp rise, rates still much lower in primary age

School outbreaks flat/declining (half term will have played a role here)

assets.publishing.service.gov.uk/government/upl…

2/17
Of course these tests are symptom based, and may miss cases due to lower symptom burden in children

Let's look at the UK's 2 big, random sample studies which overcome this issue!

3/17
Read 17 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!