How much could closing schools slow a close-contact epidemic?
People had fewer and differently mixed contacts at weekends and during school holidays, which would lower an infection's reproduction number enough to suggest school closure could meaningfully slow spread in most countries studied.
Source
Estimating the impact of school closure on social mixing behaviour and the transmission of close contact infections in eight European countries
Study at a glance
- Design
- Cross-sectional — Secondary analysis of a one-day social contact diary survey; weighted negative binomial regression of contact counts and age-structured contact matrices used to compute relative changes in the basic reproduction number.
- N
- N=7290 · 7290 participants recording 97904 contacts on one day each; holiday comparisons were possible only for Belgium, Great Britain, Luxembourg and the Netherlands.
- Population
- Broadly representative samples of all ages (children and adolescents deliberately oversampled) in Belgium, Great Britain, Finland, Germany, Italy, Luxembourg, Poland and the Netherlands, 2005-2006.
- Outcome
- Number of daily conversational and physical contacts, age-by-age contact matrices, and the relative change in R0 for weekends versus the week and holidays versus regular periods.
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The authors reanalysed a large diary survey in which people in eight European countries listed everyone they talked to or touched on one assigned day. They modelled how many contacts people made depending on age, household size, day of the week and holiday period, and built age-by-age contact matrices. Treating weekends and school holidays as natural stand-ins for school closure, they estimated how much the basic reproduction number (R0) of a close-contact infection would change under those mixing patterns.
What they found
People reported roughly 30% fewer contacts at weekends than on the average weekday and about 9% fewer during holidays than in regular periods. For close contacts, estimated R0 was 12% to 26% lower at weekends in six countries, with no significant change in Germany and Finland. Holidays lowered R0 by 10% in Belgium, 17% in Great Britain and 45% in the Netherlands, with no significant change in Luxembourg. Weekends and holidays also shifted mixing away from same-age contact toward more contact between generations.
The limits
What it doesn't show
Weekends and holidays are imperfect proxies for pandemic school closure: parents may still work, and in a real pandemic other activities would also stop, so the true effect could differ. The survey was not designed to compare these periods, holiday data existed for only a few countries, and some samples (especially the Dutch holiday estimate and older age groups) were small. R0 changes depend on the assumption that reported conversations and touches stand in for transmission-relevant contacts, and no actual infections were measured.
Key terms
- Basic reproduction number (R0)
- The average number of people one infectious person would infect in a fully susceptible population; an epidemic grows when it is above one.
- Social contact matrix
- A table giving the average number of daily contacts people in each age group have with people in every other age group.
- Social contact hypothesis
- The assumption that transmission rates between age groups are proportional to their recorded contact rates.
- Assortative mixing
- The tendency to have most contacts with people similar to oneself, here of similar age.
- Negative binomial regression
- A model for count data that allows the variance to exceed the mean, as with highly variable numbers of daily contacts.
- Next-generation matrix
- A matrix giving how many secondary infections in each age group arise from one case in each other age group; its largest eigenvalue is R0.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
What data source did this study analyse?
Common questions
Why use weekends and holidays instead of real school closures?
Real closures during an epidemic were not available to study, but schools are shut at weekends and in holidays, so contact patterns then give a directly observed hint of how people mix when children are out of school, including compensating contacts such as grandparents caring for children.
Why focus on close contacts?
Physical or close contacts are thought to be better proxies for the contacts that transmit airborne and droplet infections than brief conversations.
Why might Germany show no benefit?
The German survey collected data differently from the other countries, and the authors think its results are more biased, so the null finding there should be treated cautiously.
More on Infectious disease