• Adolescent Health
• Epidemiology
• Psychology

Evidence to support the effectiveness of global school closures in controlling COVID-19 is sparse. There is continued uncertainty about the degree to which school children are susceptible to and transmit COVID-19. Balancing the potential benefits with harms involves explicit trade-offs for governments, but there has been little recognition that low-income and middle-income countries face a very different set of trade-offs around school reopening from those in wealthy countries.

Both reopening schools and keeping them closed carry risks that actively require mitigation.

Schools remain closed in many countries globally as part of efforts to control the COVID-19 pandemic.1 National governments face mounting dilemmas about when and how to reopen schools. We review the benefits and risks of school closures during the COVID-19 pandemic and outline key principles for reopening schools.

Data from previous outbreaks suggest that schoolchildren may play only a relatively small role in the transmission of coronaviruses.2 Data from COVID-19 are sparse. Those under 20 years appear to be around half as susceptible as adults to severe acute respiratory syndrome coronavirus 23 4 and much less likely to be symptomatic.4 Yet data on viral load suggest that children may have COVID-19 viral load similar to that of adults.5 Data on transmission in schools are sparse. Population-based contact-tracing data on transmission in schools in Australia have identified almost no transmission.6

Given this uncertainty, the impact of reopening schools on transmission and the potential for a second pandemic wave is unclear. However, there is no evidence that children are more likely to transmit than adults, unlike in some respiratory viruses. When children do get COVID-19, there is also clear evidence that they are very unlikely to have severe illness or die.4 Together these data suggest that children, particularly primary schoolchildren, are likely to be among the safer groups to begin relaxation of social distancing.

In contrast, the harms related to prolonged school closure are well documented.2 In addition to impacts on learning, these include reductions in physical activity and a range of impacts on mental health and well-being due to social isolation, reduced social support, increased exposure to violence at home, exclusion of children from school-delivered public health interventions (eg, vaccination, worming and feeding programmes) and exclusion of the most vulnerable students from social safety nets operating through schools. Indirect harms to broader society include short-term loss of healthcare and other key workers, as well as reductions to broader productivity due to parental absenteeism from work. These harms are greater in more deprived families, thus worsening health and educational inequalities. Balancing the potential benefits with harms involves explicit trade-offs for governments, decisions that are not without significant risk.

The European Union recommends that relaxing of social distancing measures such as reopening schools should only occur after there is clear evidence that spread has decreased for a significant period, there is sufficient health system capacity to cope with future peaks, and countries have sufficient monitoring and testing capacity to quickly detect and isolate infected people.7 Some countries have never closed primary schools (Iceland, Sweden and Taiwan), and some have commenced reopening schools, for example, Denmark, Finland, Norway, France and Germany.

However, there has been little recognition that low-income and middle-income countries (LMICs) face a very different set of trade-offs around school reopening from those in wealthy countries. Low-income countries have much higher proportions of children in the population and smaller proportions of the elderly. Impacts of school closures on COVID-19 transmission are predicated on the effectiveness of wider approaches to social distancing while children are at home. While young people’s adherence may be suboptimal in high-income countries, it is likely to be much less optimal in LMICs where broader distancing is challenged by population density and households being forced to leave their homes to access income, food and other necessities in the absence of social safety nets and distribution systems. Schools in LMICs also deliver a broader range of health interventions, with closures having significant impact on programmatic effectiveness. For example, mass drug administration for worm infection for one billion school-age children globally is currently stalled. School feeding programmes, a critical safety net in many societies, are also halted. The World Food Programme estimates that 370 million children are currently not receiving school meals,8 driving hunger, impacts on cognition and indirect effects on rural food chains.

Issues of inequality related to cessation of education have particular urgency in LMICs, especially for young women. Family poverty and hunger may require children and young people to gather food or seek work, and those who drop out are unlikely to return to school. School closures and disruption during the Ebola outbreaks in West Africa led to an increase in child marriages and failure to return to school postcrisis. Financial hardship and parent mortality may also result in girls leaving education to take on more domestic responsibilities. The Malala Foundation estimate that approximately 10 million more secondary-aged school girls could be out of school after the current crisis has passed.9

As policymakers debate when and how to reopen schools, efforts must be made to mitigate the effects of closure on children and young people, their families and broader society. High-income countries are providing education and mental health support via online routes and alternative provision for school meals, although these are less possible for LMICs. Mitigation policies in LMICs need to focus on those at most risk of not returning to education, particularly girls and young women.

The major challenge for countries is how as well as when to reopen schools. We propose five key principles to guide decisions.