The impact of adding school closure to other social distance measures, 17 March 2020
Updated 13 May 2022
© Crown copyright 2022
This publication is licensed under the terms of the Open Government Licence v3.0 except where otherwise stated. To view this licence, visit nationalarchives.gov.uk/doc/open-government-licence/version/3 or write to the Information Policy Team, The National Archives, Kew, London TW9 4DU, or email: psi@nationalarchives.gov.uk.
Where we have identified any third party copyright information you will need to obtain permission from the copyright holders concerned.
This publication is available at https://www.gov.uk/government/publications/the-impact-of-adding-school-closure-to-other-social-distance-measures-17-march-2020/the-impact-of-adding-school-closure-to-other-social-distance-measures-17-march-2020
Nick Davies, Adam Kucharski, Roz Eggo and John Edmunds on behalf of the CMMID COVID-19 Modelling Team. London School of Hygiene and Tropical Medicine.
17 March 2020.
Summary
- Adding school closure to the existing package of interventions is likely to further reduce deaths by around 9%
- Closing schools now or after Easter makes little difference to the overall impact as case numbers are low at present
- Compensating behaviour, whereby some children are looked after by elderly individuals reduces the impact of the policy, but is highly unlikely to negate the overall beneficial impact of school closure
Aim
To assess the potential additional impact of adding school closure either immediately or after the Easter break to the mixed package of interventions announced on 16 March 2020.
Methods
We use the LSHTM age-structured stochastic transmission dynamic model. We used a county-level model (London boroughs are treated separately), and aggregated the data to the national level. Model results are for England. We assume that 50% of cases of any age are asymptomatic. Counties and boroughs were seeded as before so that London boroughs were more likely to be seeded first and there was a roughly 30-to-35-day delay in peaks in an unmitigated epidemic. The seeding produced around 400 to 600 cases per day in London and similar numbers outside London on 16 March, which is roughly in line with the estimates from current nowcasting. Social distancing was put in place to capture those measures put in place on 16 March. This package of interventions are labelled “Intervention” in the figures and tables and they include:
1. case-isolation, which reduces the transmission of clinical cases by 35%;
2. cocooning of high risk individuals and those of the age of 70. We assume that 10% of adults under the age of 70 are high risk. Cocooning is assumed to result in a reduction in ‘Other’ and ‘Work’ contacts (as measured by POLYMOD) to 25% of their normal values.
3. home working: we assume reduces ‘Work’ contacts by 30%, and ‘Travel’ contacts by a similar amount;
4. a reduction in ‘Leisure’ contacts by 75%.
NOTE: the model is compartmental in nature and not able to simulate household quarantine, so may underestimate the impact of the package of measures revealed on the 16 March.
To this package of measures, school closure is added, either immediately (17 March) or after the Easter holidays. School closure is assumed to continue until September.
5. In the base case, school closure is assumed to reduce “School” contacts to zero with no compensating changes in mixing. Note that “School” contacts as measured in POLYMOD also include college and university contacts.
6. As a sensitivity analysis, we look at a number of compensating strategies where children increase their contact with the elderly as a result of school closure. We assume that either 10%, 20%, or 50% of school-age children make one contact with one additional elderly person per weekday. These are labelled E10, E20 and E50 respectively.
Results
The package of measures announced on 16 March 16 is expected to reduce the peak height and size of the epidemic (see figure and table). However, substantial numbers of cases and deaths remain, and peak demand is likely to far exceed the capacity of the health service. Adding school closure is expected to reduce this peak demand further (figure) and the total number of cases. Under base-case assumptions adding school closure to the package of interventions reduces deaths by a further 24,000 over the course of the epidemic (an additional reduction in deaths of about 9%). Closing schools now or after Easter makes little difference to the total size of the epidemic.
Increasing contact between children and the elderly as a result of school closure can reduce the impact of the policy on preventing additional deaths and hospitalised cases. However, the compensatory behaviour has to be very large for this to mitigate the effect of school closure entirely. For instance, even if 50% of children have an additional daily contact with an elderly individual, then the overall impact of school closure is still positive – reducing the median estimate of deaths from around 290,000 without school closure to 282,000 with school closure (table).
Even with school closure, the UK is likely to experience a large epidemic which will result in overwhelming demand for health services.
Table: Estimated number of cases, deaths and ICU/non-ICU hospital beds over the course of the epidemic under different mitigation strategies.
| Scenario | Cases (thousands), Median (IQR) |
|---|---|
| Base | 24100 (21900 to 25500) |
| Intervention | 19900 (16100 to 22000) |
| Measures (17 Mar school closures) | 18000 (16400 to 20000) |
| Measures (Easter school closures) | 17800 (16200 to 19900) |
| Measures (17 March) and E10 | 17900 (16300 to 20000) |
| Measures (17 March) and E20 | 17900 (16300 to 20100) |
| Measures (17 March) and E50 | 18100 (16500 to 20300) |
| Scenario | Deaths (thousands), Median (IQR) |
|---|---|
| Base | 447 (383 to 496) |
| Intervention | 290 (217 to 344) |
| Measures (17 March school closures) | 266 (222 to 317) |
| Measures (Easter school closures) | 264 (220 to 317) |
| Measures (17 March) and E10 | 268 (224 to 320) |
| Measures (17 March) and E20 | 271 (228 to 324) |
| Measures (17 March) and E50 | 282 (239 to 334) |
| Scenario | Non-ICU beds occupied (thousands), Median (IQR) |
|---|---|
| Base | 26100 (22700 to 28700) |
| Intervention | 18100 (13900 to 21200) |
| Measures (17 March school closures) | 16700 (14100 to 19600) |
| Measures (Easter school closures) | 16600 (14000 to 19600) |
| Measures (17 March) and E10 | 16700 (14200 to 19700) |
| Measures (17 March) and E20 | 16900 (14400 to 19900) |
| Measures (17 March) and E50 | 17400 (14900 to 20400) |
| Scenario | ICU beds occupied (thousands), Median (IQR) |
|---|---|
| Base | 3520 (3020 to 3910) |
| Intervention | 2290 (1710 to 2710) |
| Measures (17 March school closures) | 2090 (1750 to 2500) |
| Measures (Easter school closures) | 2080 (1730 to 2500) |
| Measures (17 March) and E10 | 2110 (1760 to 2520) |
| Measures (17 March) and E20 | 2140 (1790 to 2550) |
| Measures (17 March) and E50 | 2220 (1880 to 2630) |
Figure: Estimated national epidemic curves for no intervention (Base) the currently announced set of interventions (not including HH quarantine, “Intervention”) and school closure (the right hand 5 columns). Columns 3 and 4 are for immediate school closure, or school closure after Easter, respectively. The right hand 3 columns are for immediate school closure with compensating behaviour in which increasing proportions of children have an additional weekday contact with an elderly individual.
