Introduction

In January 2020, WHO first identified the novel coronavirus (COVID-19), later declaring the spread of COVID-19 as a global pandemic in March 2020 [29]. Subsequently, many countries imposed national lockdowns, closing schools and workplaces, leaving people to learn virtually, enforcing social distancing measures, and implementing restrictive measures that prevented individuals from going to public places or from meeting people from other households [45].

Quarantines and lockdowns are states of isolation that are psychologically distressing and unpleasant for anyone who experiences them [14, 50]. Young people, who are at higher risk of developing mental health problems than adults [32], may be particularly vulnerable to the adverse effects of isolation, including school closures, due to the disruption lockdown causes on their physical activity and social interaction [101].

Previous systematic reviews and meta-analyses have looked at the impact of COVID-19 on the mental health of the general population [108] and healthcare workers [31]. One previous systematic review looked at the psychological burden of quarantine associated with exposure to contagious diseases on children and adolescents but included only three articles on COVID-19 [49]. To our knowledge, this is the first comprehensive systematic review focusing exclusively on the impact of the COVID-19 pandemic response lockdown on child and adolescent mental health.

This systematic review aims to summarise the literature exploring the effects of COVID-19 lockdown on a wide range of mental health outcomes in children and adolescents. We further explore the risk factors and protective factors for developing mental health outcomes in the context of COVID-19 lockdown.

Methods

The format of the methods and results was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [70] (eTable 1) (study protocol registered on PROSPERO: CRD42021225604).

Search strategy and selection criteria

A systematic search was conducted by two independent researchers (UP, MF) on Embase, Ovid MEDLINE (R), Global Health, Web of Science, and APA PsycINFO from inception until the 1st of April 2021. The search terms used can be found in eMethods 1. As this is an emerging topic, we looked at medRxiv, psyArXiv, and bioRxiv pre-print databases to identify further relevant studies. A manual search of the references of the included studies and reviews related to this topic was conducted using Google Scholar. Articles identified were screened as abstracts. After excluding those that did not meet our inclusion criteria, the full texts of the remaining articles were assessed for eligibility and decisions were made regarding their final inclusion in the review.

The inclusion criteria were as follows: (1) individual studies with original data, including grey literature, (2) conducted in children and adolescents aged ≤ 19 years, (3) exposed to COVID-19 lockdown, as operationalised in each study (see eTable 2), (4) evaluating mental health outcomes (see eTable 3 for the full list of outcomes), (5) in English. The exclusion criteria were as follows: (1) conference proceeding, abstracts, case reports or reviews, (2) studies including adults > 19 years, (3) studies in which children and adolescents were not exposed to COVID-19 lockdown, (4) studies focusing on physical health outcomes only.

Data extraction

Independent researchers (UP, MF) carried out data extraction. Any discrepancies arising were resolved through consensus, consulting another researcher (GSP) if an agreement was not attained. The variables extracted included: lead author/year, country, study design (cross-sectional, cohort, qualitative, mixed methods), sample size, sex (% females), age (mean ± SD, range), exposure data (lockdown definition, length of lockdown), instruments, outcomes (see eTable 3), report (parent, children), quality appraisal (see below) and key findings.

Strategy for data synthesis

The results of the systematic review were summarised in tables and narratively synthesised. Results were stratified by poor mental health outcomes and risk factors, followed by good mental health outcomes and protective factors.

Quality appraisal

For study appraisal, this review used the Newcastle–Ottawa Scale (NOS) adapted for cross-sectional studies [69], which has been attached as a supplementary file (eMethods 2). This scale has three domains: selection, comparability, and outcome. The domain of selection has four categories assessing the representativeness of the sample, the sample size, the number of non-respondents, and the ascertainment of the exposure, with a maximum of five stars to be awarded. The domain of comparability has one category assessing if confounding factors are controlled for, with the maximum award of two stars. The final domain of outcome has two categories assessing the outcome and the appropriate usage of statistical tests, with the maximum award of three stars. All categories can score one star, apart from the ascertainment of the exposure and assessment of the outcome, both of which can score two stars. A total of 10 stars can be awarded if a study meets all the criteria specified.

Results

Search results

A systematic electronic search identified a total of 2856 publications and 41 additional articles were found via backward searching of key papers. Of those, 324 publications underwent full-text screening. A total of 263 publications were excluded at the full-text screening stage and 61 articles finally met the criteria for inclusion, 3 of which were identified in the pre-print databases. Results of the search follow in the PRISMA 2009 flow diagram (Fig. 1).

Fig. 1
figure 1

PRISMA flowchart

Full size image

Study characteristics

The sample sizes within the included studies ranged 15–7,772 participants, (n = 54,999). The mean age of participants was 11.3 years (range: 1–19 years). 49.7% of participants were female. Most studies were cross-sectional studies (n = 45, 73.8%) while the rest were longitudinal studies (n = 16, 26.2%). Included studies took place across five continents including Europe (n = 35, 57.4%), Asia (n = 22, 36.1%), Australia (n = 1, 1.6%), North America (n = 1, 1.6%), South America (n = 1, 1.6%), and across more than one continent (n = 1, 1.6%). Most studies involved only parent self-reports (n = 21, 18,655) or solely child self-reports (n = 20, 25,327), while other studies involved both parent and child self-reports (n = 10, 7,931). A proportion of studies (n = 10, 2,321) utilised interviews, of which some were parent interviews (n = 2, 535), some were child interviews (n = 5, 2,629), and some interviewed both parents and children (n = 3, 261). Duration of lockdown was 52.3 ± 21.3 days in the included studies (range 30 [1, 4, 12, 17, 25, 27, 34, 39, 41, 57, 82]—100 days [80, 100]). The characteristics of the included studies and their findings are summarised in Table 1.

Table 1 Characteristics of the included studies
Full size table

Poor mental health outcomes and risk factors

Most commonly evaluated outcomes in the included studies (see eTables 3–4) were anxiety (n = 35, 57.4%) and depression (n = 24, 39.3%). Symptoms of anxiety exacerbation were reported during or related/associated to lockdown in 57.4% of studies [1, 3, 5, 6, 11, 12, 16, 18, 20, 27, 39, 41, 47, 48, 51, 53, 56, 62, 72, 75, 79, 80, 82, 86,87,88,89,90, 94, 103, 107, 109, 110, 112]. The prevalence of symptoms of anxiety ranged from 1.8% [110] to 49.5% [41] between studies. 59.6% of young people reported increased ruminations [11]. 13.4% of children were found to be experiencing severe anxiety [109]. Meanwhile, 3.2% of children and adolescents met DSM-5 criteria for PTSD [110]. Risk factors to anxiety included lack of routine (p < 0.001) [6], female sex (p < 0.001) [20] (p = 0.041) [62], adolescence (p = 0.005) [44], excessive COVID-19 information (p < 0.05) defined by repeated exposure to COVID-19 related information [110], media exposure (OR = 2.4) [51], and being previously referred for psychiatric treatment (OR = 4.4) [51] (Fig. 2). An increased social media usage was associated with a higher risk of developing anxiety symptoms or and depression symptoms (OR = 1.83, p = 0.001) [89]. Children with Autism Spectrum Disorders (ASD) showed more anxiety than children without ASD during the lockdown [6].

Fig. 2
figure 2

Risk and protective factors for anxiety symptoms/ affective symptoms in children and adolescents. This diagram refers to the risk and protective factors that are mentioned more than once within included studies

Full size image

Symptoms of depression were the second most commonly reported outcomes (n = 24, 39.3%) [2, 7, 12, 18, 20, 25, 41, 47, 48, 56, 62, 63, 72, 80, 83, 86, 89, 94, 106, 107, 109,110,111,112]. The prevalence of symptoms of depression ranged between 2.2% [110] and 63.8% [41] amongst studies. 7% of young people reported anhedonia [78]. The prevalence of children and adolescents with severe depression increased from 10 to 27% [41]. The prevalence of non-suicidal self-injury (OR = 1.35, p < 0.001), suicide ideation (OR = 1.32, p = 0.008), suicide planning (OR = 1.71, p < 0.001), and suicide attempts (OR = 1.74, p < 0.001) increased from November 2019 to May 2020 after lockdown [111]. Common risk factors for depression included female sex (p < 0.001) [18, 25, 62], being an adolescent (p < 0.01) [18, 83], a high amount of COVID-19 cases in the area (OR = 2.3, p < 0.001) [83, 107], and being exposed to a relative doing first-line job responsibilities related to COVID-19 (p < 0.05) [20]. Anger and irritability were common outcomes within children and adolescents ranging from 30.0% [78] to 51.3% [86] and from 16.7% [6] to 73.2% [86], respectively.

Symptoms of ADHD were frequently reported (n = 12, 19.7%) [17, 25, 34, 42, 59, 64, 74, 75, 85, 88, 91, 100]. Particularly, difficulties concentrating ranged from 55.9% [25] to 76.6% (p < 0.001) [75] in children and adolescents exposed to lockdown. Hyperactivity/inattention difficulties increased during lockdown (p < 0.001) [34, 42]. Exacerbation in symptoms of ADHD were related to increases in activity levels (50.1%), irritability (45.8%), disruptive behaviour (47.7%) [88], and conduct problems [74]. Risk factors for symptoms of ADHD worsening included sleep problems [17], being male (p < 0.001) [64], being a child compared to being an adolescent (p < 0.05) [85, 100], and parental stress (p < 0.001) [91].

Sleep disturbances were reported in a portion of the included studies (n = 11, 18.0%). 20% of children [79] and 55.6% of adolescents reported difficulty sleeping [25]. The proportion of children with sleep disorders increased from 40 to 62% during lockdown [54]. Young people showed difficulties initiating and maintaining sleep, the frequency of parasomnia increased [54]. Most studies reporting on sleep disturbances found that young people slept for longer during lockdown (p < 0.001) [39, 54]. Children went to bed ~ 53 min later (p < 0.0001) and woke up ~ 66 min later (p < 0.0001) than before the lockdown [34].

Longitudinal research findings showed a rise in children’s depressive symptoms [12] and anxiety symptoms compared to before the lockdown (p < 0.001) [19, 62]. Their risk increased when spending more time on COVID-19 media reports (p < 0.05) [110]. Furthermore, 41% of children and adolescents experienced a reactivation in eating disorder symptoms post lockdown, with a more pronounced reactivation of disordered eating seen in adolescents [44].

In lockdowns that lasted one month, previous service contact helped to alleviate anxiety [1]. However, longitudinal research findings showed that in a lockdown that lasted three months, children exhibited a deterioration in mental health symptoms, as reported by their parents, with a 10% increase in emotional symptoms, a 20% increase in hyperactivity/inattention, and a 35% increase in conduct problems [100]. Children and adolescents with special educational needs and neurodevelopmental disorders (NDD) showed more emotional symptoms, conduct problems, and hyperactivity/inattention scores than those without special educational needs and neurodevelopmental disorders [100]. Young people with NDD (28%), specifically ADHD, showed more conduct problems through lockdown, in comparison to neurotypicals controls (9%, p < 0.01) [74]. A decrease in therapy and rehabilitation support predicted externalising behaviours in children with NDDs [11].

Good mental health and protective factors

31.4% of children, especially 9-year-olds (16.8%), were seen to be calmer during the pandemic than before it, and most children were able to cope and adapt to the lockdown measures (92.6%) [79]. Family relationships improved in 41.6% of households during lockdown [44]. Some children felt safe, relaxed, and happy when with their families [47, 48]. Healthy parent–child relationships were associated with positive parent–child communication [94]. Parents praised their children 67.6% more and spent 72.9% more time with them during the lockdown [88]; 58% of children were happy to spend more time with their families [89].

Some studies identified protective factors for mental health difficulties during the COVID-19 lockdown. Routines were associated with fewer symptoms of depression and improved mental health conditions in adolescents (p < 0.01) [38, 83]. Parent–child discussion was seen to mediate some anxiety (OR = − 1.6, p < 0.001) and depression (OR = − 1.9, p < 0.001) symptoms [94]. Parent–child discussion frequency was positively correlated to current life satisfaction (p < 0.05) [94]. A further protective factor for the mental health of children was play [47] (Fig. 2). Physical activity in children was associated with a lower hyperactivity-inattention risk (OR = 0.44, for 1–2 days activity a week; OR = 0.56, for < 2 days of activity a week) [59].

Quality appraisal

The quality appraisal of the 61 studies is summarised in eTable 5. The overall average stars achieved through the 61 included studies was 7.0 stars (range = 4–9), which is considered as moderate quality. The domain of selection scored an average of 4.2/5.0 stars. The domain of comparability scored an average of 0.5/2.0 stars. The domain of outcome scored an average of 2.3/3.0 stars.

Discussion

To our knowledge, this is the first systematic review to evaluate the effect of the COVID-19 lockdown on the mental health of children and adolescents. We found anxiety and depression to be the most common outcomes. A significant, substantial increase in depression and anxiety symptoms was seen in children during the lockdown compared to rates observed before the lockdown [12, 19, 62]. Other outcomes that seem to be associated with the COVID-19 lockdown are loneliness, psychological distress, anger, irritability, boredom, fear, and stress. Our results expand previous knowledge by identifying groups that may be at risk for mental health deterioration [6, 18, 20, 44, 48, 51, 62, 74, 83, 90, 100, 112]. During the lockdown, new psychiatric conditions may appear, while children and adolescents with previous mental health conditions, such as eating disorders, may experience a reactivation [44, 51].

The prevalence of PTSD seen in children exposed to COVID-19 was 3.2% [110]. This prevalence is lower than the one previously found in children quarantined or isolated due to the influenza A (H1N1) pandemic [92]. However, PTSD symptoms usually appear months after the traumatic experience, so it may be too early to estimate its scope at the moment. Furthermore, mental health in epidemics was more impaired in the phase following the acute outbreak, than in the initial phase [22]. Future research should evaluate a potential increase in PTSD symptoms and establish appropriate measures accordingly. Specifically, preventive measures in individuals at risk are recommended to avoid reaching these dramatically high rates observed in other health-related disasters. Teacher-based, resilience-focused interventions post-trauma have shown promising results [105]. Furthermore, meta-analytical evidence suggests trauma-focused psychotherapy might be effective for the prevention of PTSD in patients with acute stress symptoms [93].

Individuals with previous eating disorders have been among the most intensively affected. 41% of young people under clinical care experienced a reactivation in eating disorder symptoms post lockdown [44], particularly those with low self-directedness and less adaptive coping strategies [9]. Lack of weight monitoring during confinement may have played a role here [9]. Individuals suffering from eating disorders struggled to maintain feeding routines and research shows COVID-19 lockdown to significantly correlate with symptoms of disordered eating [61]. Considering eating disorders have the highest mortality rate [102], there should be an increased utilisation of digital tools to support those with eating disorders in the context of COVID-19 [28].

This review found sociodemographic characteristics influencing the development of poor mental health outcomes associated with COVID-19 lockdown to include older age (13–15 vs. 6–12, p < 0.03 [18, 83]) and female sex [18, 20, 62, 90]. Adolescents have been previously identified as a vulnerable group, going through an important period in their development [15] where peer relationships are of the most importance. Older adolescents displayed more depressive symptoms than younger adolescents during the lockdown [18]. This may be because the onset of depression increases as children transition into adolescence [68]. A further explanation is that adolescents are in particular need of social contact and interpersonal relationships. The period of adolescence is a motivator for peer connection [36] and the desire for peer and social support [37], which aids the development of identity [67]. However, during the lockdown, they need to attend online learning, cope with school closures and adapt to a mandatory decrease in social relations [60].

Another vulnerable group identified by this systematic review are the children and adolescents and with previous mental health difficulties or with “special educational needs and disabilities” (SEND) and/or neurodevelopmental disorders [6, 44, 51, 100]. One of the reasons children and adolescents with neurodevelopmental disorders are highly vulnerable to suffering psychological distress is that while they usually prefer routine and predictable environments, the COVID-19 pandemic is a situation of fast-paced changes [24]. Children and adolescents with SEND, ASD and/or disabilities have had their carefully constructed routines suddenly disrupted [98] alongside affected support networks resulting in a higher risk of experiencing poor mental health and increased stress during the unprecedented lockdown [8]. With special education closed, these young people may struggle more with adapting to virtual schooling. Social factors in these children and adolescents are also important. 24% of teachers claimed families of those with SEND and/or disabilities don’t have access to sufficiently powerful devices or software required to download or access digital materials required [77], which further complicates their situation. As a result of lockdown, symptoms of ADHD were seen to worsen [25, 34, 42, 59, 75, 85, 88, 100]. Certain home environments (e.g., having a garden or adequate space at home) had a positive impact on ADHD symptoms. However, limited academic adjustments for children with ADHD were reported by parents, resulting in difficulties to carry out school-related tasks [13].

Identifying risk and protective factors is crucial for clinical practice to identify individuals who are more vulnerable to poor mental health outcomes and to develop clinical practices and public health strategies to reduce the negative impact of lockdown on children and adolescents. Risk factors include lack of routine [83], the form of internet usage [19], COVID-19 media exposure, and a relative doing first-line job responsibilities related to COVID-19 [20]. Quarantine affects the structure of children and adolescents days’ [46]. Therefore, schools play an important role over lockdown as they’re able to provide structure into young peoples’ days which is seen to be protective, as long as they don’t overburden young people [101].

In addition, school closure has been identified as a key stressor for some young people [7]. Significant associations have been found between emotional reactions and home-school experiences [53]. 56% of those experiencing psychosocial problems as a result of lockdown reported that this was related to not being able to attend school [89]. Furthermore, during school closure, child protection referrals from schools have decreased compared to previous years [99]. A decrease in help-seeking behaviour and access to care may have contributed negatively to the mental health of children and adolescents. The impact of school closure has not been equal for all. Children in the primary school reported fewer depressive symptoms compared to children in middle school [106].

Internet usage reduces the time being spent doing other beneficial activities and may adversely affect children’s emotional health and psychological wellbeing [66]. Problematic internet usage was seen to result in psychological distress characterised by excessive time spent gaming, on one’s smartphone, and on social media [19]. This has been supported by research finding that problematic internet use is associated with depression, anxiety, and other health problems [35]. Excessive time spent on the internet may occur as children are bored at home, isolated from peers, and cannot attend regular extracurricular activities. Research has found those in social isolation to have a higher level of media contact, with more exposure to COVID-19 related information [58]. During pandemics and epidemics, media exposure is reported to worsen severe mental health outcomes [21]. For instance, excessive COVID-19 media exposure has been associated with an increase in anxiety levels and stress [40]. It would be recommendable for parents to limit the time children and adolescents spend using the internet and to model positive coping behaviours [97] to reduce stress, encouraging children to carry out other activities, for example, listening to music [43], reading together, and playing board games together [55]. Physical activity also reduced hyperactivity-inattention risk in children [59], which could be encouraged or recommended by caregivers. Parents and health professionals should also make sure children and adolescents get only truthful and balanced information. These aspects are a real challenge for parents that need to work remotely and simultaneously take care of their children.

Previous research has shown that family environment, parental practices, and methods of coping affect children’s post-disaster mental health [23]. However, the lockdown has not negatively impacted everyone and may have been beneficial for some relationships to develop. Parent–child discussion was seen to be protective against child mental health, specifically anxiety, depression and stress, and is related to life satisfaction [94]. Perceived family relationship improvements may be a consequence of families being able to spend more quality time with one another due to remote working [104]; however, this can result in mental strain on some parents, especially parents of children with SEND [33]. Family relationships may serve to support child adjustment when faced with adversity [26]. In addition to this, experiencing collective family major life challenges may promote positive family transformations [65].

Numerous changes in mental health provision have occurred since the start of the COVID-19 pandemic lockdown to minimise the infection rate, such as a rise in community support services and implementing inpatient infection-control measures. To ensure continuity of mental healthcare for service users, mental health services have had to adapt mainly via adopting more telehealth methods [71, 84]. The COVID-19 lockdown has resulted in a rise in virtual, remote therapy, which may have future implications for service provision after COVID-19. For example, telehealth will allow those who live in remote areas to access mental healthcare more easily. Since COVID-19, telemedicine has been expanded, so more people are eligible for it and rules have been relaxed for health insurance providers and doctors. Telehealth reduces barriers to access, is more cost-effective, and has a wide availability of services within paediatric care [81]. However, telemedicine is limited by one’s technology literacy, psychological resistance to new methods, and cultural background [52].

The findings highlighted in the present work have further clinical implications. Governments should ensure that lockdowns be as short as possible to limit the psychological effects of lockdown on children and adolescents, while protecting their safety. Governments should also aim to release COVID-19 information, information about prevention measures, and lockdown updates while ensuring that the information provided is accurate [95]. The public should have access to age-appropriate resources such as improving sleep hygiene, maintaining a balanced diet, routine keeping, and mental healthcare [101] to educate young people on keeping healthy to prevent negative psychological effects. Sleep quality [17, 25, 34, 39, 72, 79, 86] and sleep disorders [10, 54, 109] were seen to worsen during the pandemic. Therefore, easily accessible sleep hygiene resources for children may be protective against adverse sleep effects. Moreover, an increase in funding allocation to mental health services needs to be provided along with trained staff to facilitate care and ensure continuity of care for vulnerable populations and cope with the long-term mental health effects the COVID-19 lockdown may have.

Another fundamental clinical implication highlighted by this review is that the break of care due to the lockdown [45] may have delayed access to treatment, pushing the course of recovery back [84]. Clinicians should follow up on those who have experienced a break of care as they may be more vulnerable to reactivation of symptoms post lockdown. This can be done by having regular mental health check-ups for vulnerable groups to assess their mental state. Unfortunately, a reduction in self-help-seeking behaviours has been observed. For instance, hospital presentations for self-harm decreased in 2020, compared to 2019 [76] regardless of an increase in these behaviours seen during the lockdown [111]. Attention should be paid to the more vulnerable groups post-lockdown when it comes to accessing mental health care and parents should also be provided psychoeducational resources to help identify psychological distress in their children [101].

We need to balance health and safety on one side and mental health and normal psychological development on the other. Short lockdowns seem to be better tolerated, especially with previous service contact [1]. However, the longer the lockdown lasts the more support children may need. For instance, children under lockdown for three months exhibited far more conduct problems [100]. Lockdowns should be made as short as possible and should assess the benefit/risk balance when deciding how long lockdowns should last to limit mental health consequences.

This study has some limitations that must be considered. First, this review does not meta-analytically evaluate the magnitude and consistency of the mental health outcomes described due to the heterogeneity of the outcomes and measurement methods. Second, evidence on the effect of lockdown on low-income households and low-income countries was limited. Further research is needed to draw conclusions on whether the impact of COVID-19 lockdown on the mental health of children and adolescents is different or not between low- and high-income countries. A third limitation is that some studies (62.3%) did not provide details about the duration of the lockdown established, precluding drawing further conclusions from our end.

Fourth, most studies included (73.8%) were cross-sectional, limiting causal inference. The inclusion of some cohort studies evaluating children and adolescents before and after the lockdown, allowed to evaluate more precisely the effect of lockdown while helped control some situational confounders. A fifth limitation would be that most included studies focussed on psychological reactions and symptomatology rather than the appearance of mental disorders, which has implications for practice. Future longitudinal studies should follow children and adolescents who experienced poor mental health during lockdown to see if they recover, mental health difficulties persist, or they crystalise into full-blown mental disorders. Lastly, most studies were conducted online, where it would be difficult for children to ask for clarification around the questions they did not understand. Due to the lockdown, children and adolescents could not attend research centres for their safety.

Conclusions

The COVID-19 lockdown has resulted in psychological distress and highlighted vulnerable groups such as those with mental health difficulties, and risk factors such as lack of routine and excessive COVID-19 media exposure. However, for some families being able to spend more quality time together has been positive. Supporting the mental health needs of children and adolescents at risk is key. Clinical guidelines to alleviate the negative effects of COVID-19 lockdown and public health strategies to support this population need to be developed.