COVID-19 Pandemic School Disruptions and Acute Mental Health in Children and Adolescents

Key Points Question Were COVID-19 pandemic–related school disruptions associated with youth mental health emergencies? Findings In this cross-sectional study of 13 014 pediatric emergency department visits at 9 university hospitals in Italy, school opening following the disruption was associated with an increase in acute psychiatric emergencies. In particular, visits for suicide ideation/suicide attempt increased during the school calendar and especially among females. Meaning The findings of this study suggest that school may be a major source of stress for youths; factors mediating school-associated mental health disturbances in youth should be investigated.


Introduction
Concern has been raised about a possible role of school pressure in mental health disturbances in youth, but few studies have investigated whether there is an association. 1 Examining psychiatric emergency department (ED) visits, which represent the most severe mental disorders, may help evaluate a possible association with school opening.An association between school days and higher rates of ED visits has been found in local samples from the US and UK, 2,3 especially regarding visits for self-harm and aggression. 4,5[11] The COVID-19 pandemic disrupted school for millions of students worldwide at different times and with variable intensity.This unique situation can offer an opportunity to study the association between school and psychiatric emergencies.The initial wave of the pandemic in the spring of 2020 was accompanied by a marked overall decrease in all pediatric ED visits but a less pronounced decrease for psychiatric visits.[16] A decrease in the rate of adolescent suicides was observed in the US at the onset of the COVID-19 pandemic when schools were closed, with a subsequent increase in the fall of 2020, when many schools returned to in-person instruction. 5,17Transitioning from online to in-person schooling was accompanied by a 12% to 18% increase in adolescent suicide. 5,18Conversely, there are also suggestions that the prolonged duration of COVID-19 school closure was associated with worse adolescent mental health. 19 Italy, the strictest COVID-19 preventive restrictions (social lockdown with limitation of in-person schooling) started in the spring of 2020 and were in force for most of that year, with gradual attenuation, based on the fluctuating incidence of the infection, in 2021.School was online in many areas, at least during some weeks, until the summer of 2021.In-person school reopened in September 2021. 20e role of school is confounded by other factors, especially the restrictions to social and recreational activities, so it can be difficult to separate school from social lockdown effects. 21The seasonal pattern of ED psychiatric visits should also be considered given higher rates of psychiatric hospitalizations and suicide in spring and summer. 22,23Another factor to consider is a general increase in youth psychiatric ED visits worldwide. 6Over the past 10 years, the proportion of pediatric ED visits for mental health reasons has approximately doubled, with a 5-fold increase in suicide-related visits. 7,15is study used the diverse levels of strictness in the lockdown measures mandated by the health authorities in Italy during the COVID-19 pandemic, which entailed varying degrees of restriction in social contacts, recreational activities, and school opening, including complete school closure, online education, and regular in-person schooling.The level of school restrictions and that of the other social and recreational activities did not always change in parallel, and this discordance allows the association with school to be estimated in the context of the other pandemic measures.
Thus, the study aimed to examine whether and how child and adolescent psychiatric ED visits changed in volume, demographic factors, and psychiatric presentations during and soon after the COVID-19 pandemic, with respect to school opening (complete closure, online schooling, and in-person attendance).These analyses accounted for the social containment measures enacted during the pandemic, time trends, and patient's age, sex, and socioeconomic status.

Design
This was an observational cross-sectional study of the hospital clinical records of all the ED visits of children and adolescents (age, 0-17 years) between January 1, 2018, and December 31, 2021, at 9 university hospitals in Italy.The hospital EDs were located at Brescia, Cagliari, Florence, Genoa, Rome (2 hospitals), Sassari, Trieste, and Turin, and served a cumulative area of about 7 million people.The study consisted of the collection and analysis of anonymized data of naturalistically treated patients, without active recruitment of participants.The study received institutional ethical approval by the ethics committee at Regina Margherita Children's Hospital, Turin.The report followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. 24e hospital medical records were systematically examined by trained clinicians under the supervision of senior child and adolescent psychiatrists (C.D., G.A., T.A., R.A., D.C., A.C., S.C., T.P., A.T., and S.U.) and psychiatric reasons for the ED visit were extracted.Rater reliability was assessed by randomly selecting 5 visits from each participating site and asking 2 raters to categorize them independently.The interrater reliability was 0.90 (Fleiss κ).A complete description of data extraction methods can be found in the eMethods in Supplement 1.

Variables
School opening was categorized as an ordinal variable from 1 to 3, with in-person schooling assigned the highest value (3), online schooling a medium value (2), and school closure the lowest value (1).
For each week since March 2020, lockdown intensity was recorded according to the official administrative categorization for each geographic area.The lockdown severity was quantified using scores from 1 to 5 (the strictest lockdown measures) (eMethods in Supplement 1).The socioeconomic status of each patient's neighborhood was estimated through the computation of a deprivation index (eMethods in Supplement 1). 25

Statistical Analysis
Data analysis was conducted from July 1 to August 31, 2023.Descriptive statistics were applied to the sociodemographic and clinical data.The weekly psychiatric ED visits total count and the weekly psychiatric ED visits count separately for suicide attempts, suicidal ideation, eating disorders, and psychomotor agitation were modeled using generalized mixed models, with visit count following a Poisson distribution in a bayesian framework.Each model had fixed effects for year, school opening, lockdown severity, and median deprivation index.All regression models included study sites as a random intercept, and we allowed the regression slopes for time (both yearly and weekly variations) to differ across study sites.For models assessing SI with SA, we considered the weekly sum of the visits for SI with SA as the dependent variable.eMethods in Supplement 1 provides more detail.

Results
A total of 1 017 885 pediatric ED visits occurred during the study period: 13 014 (1.3%) were psychiatric visits.Mean (SD) age was 13.8 (3.8) years, 63.2% of the individuals were females, and 36.8% were males.
These last 3 factors significantly increased during the study period, with increments of 294.8% for eating disorders, 297.8% for SI, and 249.1% for SA.
Tables 2, 3, and 4 present the incidence rate ratios (IRRs) for the variables under study for all psychiatric visits, psychomotor agitation, eating disorders, and SI with SA. eTables 2-7 in Supplement 1 detail the IRRs for psychiatric visits by sex, age (<14 years vs Ն14 years), and psychiatric visits due to suicide attempt.2).
To examine whether the observed findings were specific to psychiatric visits or mirroring other pediatric conditions, we incorporated an offset for the weekly number of all the pediatric visits into the same model.The analysis revealed that, when accounting for the total number of pediatric ED visits, the lockdown severity was associated with an increase of the proportion of psychiatric visits over the total (IRR, 1.18; 95% CI, 1.09-1.28)(Table 2), while the associations observed for the other analyzed variables were similar to the model without the offset (Table 2).
Given this difference between the 2 models (with and without the offset) and the need for exploring the specificity of mental health conditions while accounting for the changing trends of ED visits, we decided to incorporate, for all the subsequent models (Table 3 and Table 4; eTables 2-7 in [4.2] years) was significantly lower (analysis of variance with post hoc Tukey-corrected comparisons, F 3,1310 = 9.70; P < .001)than in the subsequent years (13.91 [3.5] years for 2019; 13.75 [4.0] years for 2020; and 13.94 [3.8] years for 2021) (

Table 1 .
Demographic and Clinical Characteristics of Patients With ED Psychiatric Visits

Table 2 .
Longitudinal Mixed Model With Poisson Distribution of All Psychiatric ED Visits

Table 3 .
Longitudinal Mixed Model With Poisson Distribution of ED Visits for Eating Disorders Monthly trends of psychiatric ED visits are shown in the Figure.To explore the main factors associated with the weekly trends of psychiatric visits, we used a longitudinal mixed model applied to a Poisson distribution.School opening was associated with the largest increment in the number of visits (IRR, 1.29; 95% CI, 1.23-1.34)(Table2, Figure).A 19% increase in psychiatric visits was also found for every year since the beginning of the study (IRR, 1.19; 95% CI, 1.16-1.22)(Table 2, Figure).