Three years of COVID-19-related school restrictions and mental health of children and adolescents in Japan

Our study explored the effects of persistent mitigation measures in school on children’s mental health and psychological well-being. Unlike in other countries, children in Japan lived with strict mitigation measures at school for a long time, even though the schools reopened in May or June 2020. 16.7 percent of the children experienced cancellation of school excursions—events that are generally regarded as special —in 2020 and 2021. As of March 2023, 71% of the respondents in our survey ate lunch silently in schools. The mask mandate in schools was lifted off for the first time on April 1, 2023.

While many have studied the effects of school closures, during the initial waves of the COVID-19 pandemic, on the mental health of children, there have been a limited number of studies on the impact of persistent mitigation measures in schools. Mental health problems in childhood and adolescence cannot only worsen their current well-being but also have long-term adverse health and social consequences^40,41,42. Accordingly, the mental health effects of these long-lasting restrictions must be a key consideration when setting mitigation measures in schools.

With this goal in mind, we found that the cancellation of school excursions was associated with a higher risk of developing depressive symptoms in children. A high cancellation rate of school events was positively associated with the tendency to report dissatisfaction with schools and friendship, and negatively associated with the tendency to report high satisfaction with school experiences. The results indicated that long-term restrictive school experiences disrupt children’s mental health and psychological well-being.

In addition to the overall negative effect of cancellation of school events, we uncovered heterogeneous effects according to the characteristics of children. First, we found that girls and children with no extracurricular activities tended to exhibit depressive symptoms due to the cancellation of school excursions. The result seemed to be justified, since psychological provisions, such as intimacy and support, are priorities in girls’ friendship⁴³, and children who do not participate in extracurricular activities tend to have fewer opportunities to develop friendship with others. Emotional care of these children would especially be important if a future pandemic forces the mass cancellation of school events again. Second, we found positive, significant association between the cancellation rate of school events and depressive symptom among boys and among those participated in extracurricular activities. If boys are more enthused by sports events at school than girls⁴⁴, the frequent cancellation of these events may affect the mental health of boys more strongly than that of girls. If school events serve as a venue for students to demonstrate their talents nurtured through extracurricular activities to friends, the cancellation of school events may affect the mental health of those who engage in extracurricular activities more strongly than those who do not.

We further found that children who were skeptical about the effectiveness of infection control measures in schools were more likely to develop depressive symptoms (OR 1.420 [95% CI 1.111–1.814]). This tendency was more pronounced in boys, children who did not participate in extracurricular activities, and junior high school students. The results suggested the importance of carefully explaining the rationale of strict mitigation measures to children and flexibly adjusting the measures as the knowledge of their effectiveness gets updated.

Finally, the effects of no school excursions and low evaluation of mitigation measures are different according to the outcome variables. First, no school excursions are statistically significantly associated with depressive symptoms, but only moderately with satisfaction measures (for example, OR 1.543 [95% CI 1.109–2.148] for WHO-5 index versus OR 0.659 [95% CI 0.432–1.005] for friendship satisfaction). This moderate association may arise if satisfaction is influenced by the comparison with peer groups^45,46. For example, students may feel more dissatisfied with the cancellation of a school excursion if neighboring schools go ahead with the excursion, but they may feel more satisfied if other schools also cancel, as they are more likely to think that the cancellation is inevitable. Second, low evaluation of mitigation measures is statistically significantly associated with depressive symptoms, but not with several satisfaction measures (for example, OR 1.423 [95% CI 1.113–1.820] for WHO-5 index versus OR 1.140 [95% CI 0.852–1.525] for dissatisfaction with overall school experience). This lack of association may reflect the possibility that students may see strict mitigation measures as inevitable when strict mitigation measures are also implemented in neighborhood schools and adults support them⁴⁷, even if these measures are detrimental to their mental health.

Since our study was based on a cross-sectional framework, several cautions are needed when we interpret the results. Among those, we explicitly discuss two major aspects. First, the estimated effects of mitigation measures may be confounded by the effects of other concurrent policies on COVID-19. For example, school events might have been cancelled more frequently in areas with high infection, and children in such areas might have experienced longer school closures than others. While this confounding effect is a concern, it is unlikely to be large after controlling for prefecture-fixed effects, since the prefectural governors in charge have the right to request schools to close, and the duration of school closures is mainly determined at the prefectural level. When we compared the results with and without prefecture-fixed effects in SI Appendix, Figure SI3, they were almost identical.

A closer look at the regional variation in the cancellation of school events further supported the idea that mitigation measures in schools are subject to local idiosyncratic factors and that the confounding bias is unlikely to be large. In SI Appendix, Figure SI4, the implementation rate of school excursions in public high schools in 2020 and 2021 is shown based on the official statistics on school events⁴⁸. The figure shows that prefectures with similar epidemic statuses, such as Tokyo and Osaka, can have significantly different implementation rates. Public high schools in Tokyo cancelled nearly all school excursions in 2020 and 2021, whereas the implementation rates in Osaka were 52% in 2020 and 100% in 2021. Some rural prefectures, such as Shimane, exhibited very low implementation rates, whereas others, such as Oita and Kagoshima, exhibited high implementation rates. Anecdotal evidence suggested that the strictness of mitigation measures significantly depends on the characteristics of the governor-in-charge. The large disparities in the cancelation or implementation status within relatively homogenous prefectures suggested that the cancelation or implementation of school events was subject to idiosyncratic local factors, indicating that the simple regression analysis in this study could possibly be sufficient to uncover the causal effects of the cancellation of school events.

Second, in this study, we relied on the binary decision regarding whether each school event was held or not, ignoring the manner in which it was held, if at all. For example, even if a school excursion was held, it might have been scaled down to a large extent (such as, location of the school trip may have been changed from overseas to a nearby city). Sports days in many schools were often held without family audience, even if they were held. Therefore, we are, in part, comparing children who experienced the cancellation of a school event with those who experienced the event even if in a restricted manner, as opposed to those who experienced the event without any restriction. Since the occurrence of restricted execution of school events was prevalent, our study underestimated the negative impact of school event cancellation during the COVID-19 pandemic.

Third, as discussed early, we asked the parents to complete the survey with their children. It is possible that children hide their true feelings and thoughts from their parents. Because we cannot determine the direction of bias associated with this feature of our survey, it would be useful to supplement this study with another qualitative, face-to-face survey in future research.

Despite the limitations, our study has several strengths. To the best of our knowledge, our study was the first to shed light on the cumulative effects of long-term school restrictions on children’s mental health and psychological well-being. Since previous studies generally focused on the effects of school closures^{3,5,19,20,21,22} rather than the potential negative effects associated with long-lasting restrictions at re-opened schools, the current study fills an important gap in the literature. In addition, we implemented the survey in a timely manner, in March 2023, when students in Japan graduate from their respective schools (sessions in Japanese schools start in April and end in March), and hence, the respondents were children in the final year of their schools. Had the survey been conducted later, it might have suffered from a more severe recall bias and memory fading, since the respondents would have already moved into a different stage of life. Had the survey been conducted earlier, it might have underestimated the cumulative effects of the three-year-long school restrictions.

From the perspective of policy implications, other mitigation measures could have implemented better balance of infection control and children’s well-being in school. For example, a greater promotion of hand sanitization^49,50 and more frequent PCR testing at schools⁵¹ might have been able to mitigate the spread of infection without generating the mass cancellation of school events and disrupting friendship among children, which some psychological studies claim is a concern in terms of emotional recognition^15,52,53,54 and cognitive performance⁵⁵.

Finally, we would like to note that children’s well-being should be a key factor in determining the degree of mitigation measures in school. The WHO Constitution stipulates that mitigation measures should be implemented in a manner that respects individual liberties and always considers the least burdensome methods for children⁴⁶. The UN Convention on the Rights of the Child requires that in all actions concerning children, whether undertaken by public or private social welfare institutions, courts of law, administrative authorities, or legislative bodies, the best interests of the child should be a primary consideration (Article 3). It would be important to strike a balance between protecting children and their parents from COVID-19, and protecting children from the potential harm associated with strict mitigation measures. The potential harm we highlighted in this paper—deterioration in mental health— has not received much attention till date in the policy debate on the pandemic response in Japan. Lack of attention to this issue was particularly concerning in 2022 and 2023, after vaccination became widespread and the case fatality rate dropped dramatically even among the adult population; moreover, several systematic reviews till date have shown that children and schools might play only a limited role in COVID-19 transmission^49,50. Looking forward, it is desirable for the Japanese society to discuss ways to smoothly guide children to return to their pre-pandemic lifestyle. In future pandemics, it will be important to carefully examine the ethical issues and effectiveness of schools’ mitigation measures from an interdisciplinary perspective.

For future studies, it would be interesting to explore how mental health disruption in children during the pandemic will have persistent effects on their later life. Although several studies have already shown that a negative experience and deteriorated mental health during childhood causes long-term adverse health and social consequences^40,41,42 in the normal times, researchers have only just begun to explore the medium- and long-term impacts of mitigation measures during pandemic. The accumulation of such studies will contribute significantly to the development of research on how childhood experiences are linked to outcomes in adulthood.