Thread by @P_Bruijning

Our model-based analysis; Potential effects of school-based measures to reduce contacts between children, including school closures, markedly depends on the reduction in the effective reproduction number achieved
by other measures. medrxiv.org/content/10.110…

Let me try to explain what our new modelling study does and does not suggest. First, this is a model and it is currently being peer reviewed. Ganna Rozhnova did this work, and if you thinks it’s good all credits go to her (and the other co-authors). (1/10)

Results based on 2 aspects: (1) contact rates in different age groups. For children (<20yr) we distinguish school-based & non-school-based contacts; (2) what is the susceptibility of age groups to be infected with SARS-CoV2 after being contacted by an infectious person. (2/10)

Model is fitted to observed incidence and seroprevalence data in NL. Based on that we estimate that children (0.2) are less susceptible than adults (0.6), who are less susceptible than elderly (1). Naturally, that susceptibility will gradually increase within age groups. (3/10)

We investigate reducing either school-based or non-school-based contacts in 2 situations: when Re=1.3 & some measures in place (like end of Aug in NL); when Re=1 with extended measures in place (like end of Nov in NL). And then we estimate the effects of reducing contacts. (4/10)

With Re=1.3: only reducing school-based contacts will not bring Re<1. Reducing non-school-based contacts to levels March-April (& not reducing school-based contacts) brings Re to 0.8. Hence only extending summer leave for schools would not have prevented 2nd wave. (5/10)

Then: the pursued reduction to 0.8 not reached & in Nov Re=1, either bc measures failed (& more imposs/unwanted) or 0.8 not yet reached. Then adding reduction school-based contacts reduces Re (up to 0.8). Yet, further reducing non-school-based contacts will do the same. (6/10)

Because of differences in susceptibility and contact patterns reducing school-based contacts will have most effect in older children. Important assumption is that reduced school-based contacts are not replaced by non-school-based contacts. (7/10)

I think we learned that effects of school-based measures will be different in different stages of the pandemic. (8/10)

The study does not demonstrate that schools must be closed now. We should investigate if we can reduce contact without closure (e.g., with more testing, spreading presence) and realize that opportunities for reducing non-school-based measures are far from exhausted. (9/10)

By doing so, a 30-40% reduction in school-based contacts would already lower Re by 0.1 in a much more sustainable way than closing them.