SCIENCE science.orgsuggested that children were less likely to
become infected or transmit SARS-CoV-2
to others. Many of these studies, how-
ever, involved polymerase chain reaction
(PCR) testing of symptomatic contacts and
therefore missed asymptomatic and mild,
transient infections, which are more com-
mon in children than adults. Nonetheless,
educational settings were closed in many
countries as part of national lockdown miti-
gation strategies, and the net effect was a
rapid decline in cases across all age groups,
although the specific contribution of school
closures compared to all the other lock-
down measures implemented at the same
time was difficult to estimate.
As community cases fell during national
lockdowns, discussions about reopening
educational settings were highly divisive
among parents, educationalists, clinicians,
politicians, and policy-makers. S ome coun-
tries such as Iceland and Sweden had suc-
cessfully kept their preschools (age 1 to 6
years) and schools (age 7 to 16 years) open
throughout the pandemic. The UK, like
many other countries, opted to partially
reopen preschools and some primary and
secondary school (age 5 to 16 years) for in-
person teaching from 1 June 2020. Strict
physical distancing and infection control
measures were implemented, including
smaller class sizes and clustering staff and
students into bubbles, with face coverings
recommended for staff but not for stu-
dents. National surveillance conducted by
Public Health England (PHE) identified
very few infections and outbreaks during
the 6 weeks until mid-July 2020, with most
cases affecting staff and limited transmis-
sion among students ( 2 ). Notably, the risk
of SARS-CoV-2 infection and outbreaks in
educational settings correlates strongly
with local community infection rates ( 2 ,
3 ), which is expected because higher com-
munity rates allow more opportunities for
virus introductions into schools.
In June 2020, PHE also initiated active
surveillance in primary schools where staff
and students provided blood samples to
test for SARS-CoV-2 antibodies as a robust
marker of prior infection, irrespective of
symptom status or duration of infection
( 4 ). SARS-CoV-2 antibody positivity rates
were similar in students and staff and
comparable to local community serop-
revalence, which is consistent with similar
school serosurveillance studies performed
in other high-income countries. This is im-
portant because it indicated that children
were as likely to be infected and develop
antibody responses against SARS-CoV-2
as adults. Additionally, only 14% and 31%
of seropositive primary and secondary
school students, respectively, reported
COVID-19 symptoms compared to 53%
and 77% of staff, respectively, confirming
the higher proportion of asymptomatic
infections with younger age ( 4 ). After 6
weeks at school, only 4 out of 603 (0.7%)
4- to 12-year-old students and 1 out of
1015 (0.1%) staff sampled in schools across
England seroconverted, further confirming
very low risk of infection following partial
reopening of primary schools, albeit when
community infection rates were also low.
As occurred in many countries, the deci-
sion to fully reopen schools in late 2020 led
to further c oncerns about the difficulties in
implementing mitigations and physical dis-
tancing with full attendance in classrooms
with poor ventilation and lack of air quality
monitoring, using carbon dioxide detectors,
for example. Although the low risk of severe
COVID-19 in children was well-recognized,
there were increasing reports of pediat-
ric multisystem inflammatory syndrome
(MIS-C), a rare immune complication in
~0.03% of ≤20-year-olds that arises 2 to 4
weeks after SARS-CoV-2 exposure. There
was also increasing recognition of Long
Covid which, although lacking a formal case
definition, comprises persistent, intermit-
tent, relapsing, and new-onset symptoms
lasting several weeks to months after SARS-
CoV-2 infection. Additionally, the personal
risk, due to age and comorbidities, of educa-
tion staff being infected was a concern.
The UK experience of full reopening of
in-person school teaching was similar to
those of other countries, whereby cases in
students increased but lagged behind and
did not exceed local community infection
rates in adults ( 3 ) (supplementary fig. S1).
In November 2020, however, the UK im-
posed a second national lockdown because
of increasing cases, but schools remained
open. Cases declined first in adults and
then in children who continued to attend
full-time education. From late November,
there was a rapid increase in cases due to
the more transmissible Alpha (B.1.1.7) vari-
ant. This was first observed among second-
ary school–aged children because they were
still attending school, whereas adults re-
mained in lockdown.
One of the main concerns with the full re-
opening of schools was the risk of in-school
transmissions leading to large outbreaks,with onward transmission to household
members and potentially into the wider
community. In England, PHE surveillance
found that most schools did not experience
a COVID-19 outbreak. When they did oc-
cur, more than half involved only two cases
before they were controlled, most likely
because of mitigation measures: frequent
sanitization, physical distancing, class bub-
bles, staff masking, and rapid testing with
lateral flow devices and isolation of cases
and their bubble contacts. PHE surveillance
during September–October 2020 identi-
fied more frequent outbreaks in secondary
schools (10%) compared to primary schools
(2%) or nurseries (0.3%) ( 5 ). Although data
were collected over different periods, infec-
tion rates in nursery staff (24.3%) during
November 2020 were significantly higher
than in primary (9.8%) or secondary school
(4.0%) teaching staff during September–
October 2020. By contrast, infection rates
in children across the three educational set-
tings remained low (2.3, 0.8, 1.2%, respec-
tively) , especially when the index case was
also a child ( 5 ). The higher infection rates
in nurseries may be explained by difficulties
in maintaining physical distancing owing to
smaller premises and constant adult care
needed by infants and toddlers.
These investigations, however, did not
provide sufficient granularity to distinguish
new introductions of the virus from the
community compared with in-school trans-
mission. But recent US studies with active
contact tracing found that very low rates of
secondary infections occurred inside edu-
cational premises in K–12 (5 to 18 years)
schools. In North Carolina, surveillance of
11 school districts with >90,000 students
and staff attending in-person schooling for 9
weeks found 773 community-acquired SARS-
CoV-2 infections, whereas contact tracing
found only 32 additional infections acquired
in school ( 6 ). Among 17 rural Wisconsin
schools, COVID-19 incidence among 4876
students and 654 staff during 31 August–29
November 2020 was lower (3453 out of
100,000) than in the county overall (5466 out
of 100,000) ( 7 ). Of the 191 cases in students
and staff, only seven (3.7%) were linked
to in-school spread ( 7 ). In New York City,
COVID-19 prevalence in public schools was
similar to or less than community prevalence
during 9 October–18 December 2020 ( 8 ).
Additionally, of 36,423 school-based close
contacts, only 191 (0.5%) subsequently tested
positive for SARS-CoV-2, and the likely index
case was an adult for 78.0% of secondary
cases ( 8 ). Taken together, despite the inevi-
table biases and limitations of such studies,
the risk of acquiring SARS-CoV-2 in school
remains very low when appropriate mitiga-
tion measures are in place, although recom-(^1) Immunisation and Countermeasures Division, Public
Health England, London, UK.^2 Paediatric Infectious
Diseases Research Group, St. George’s University
of London, London, UK. The members of the sKIDs
Investigation Team are provided in the supplementary
materials. Email: [email protected]
A teacher prepares COVID-19 rapid tests for students at
a secondary school. These are one of several mitigation
measures to ensure in-person schooling is safe.
5 NOVEMBER 2021 • VOL 374 ISSUE 6568 681