Emotional and behavioral outcomes among youths with mental disorders during the first Covid lockdown and school closures in England: a large clinical population study using health care record integrated surveys

The Covid-19 pandemic has changed many aspects of the lives of children and young people (CYP) globally. Since the outbreak, the government of the United Kingdom (UK) has implemented a series of national lockdowns during which workplaces and community spaces were closed (Supplementary material, Sect.  1.1, and Fig. S1 for timeframe). Schools were also closed to pupils, unless considered vulnerable or children of keyworkers, and education was provided remotely. Although schools gradually reopened from June 2020, disruption has continued due to ongoing restrictions and local outbreaks. Quarantine measures also meant a rapid re-configuration of Child and Adolescent Mental Health Services (CAMHS), which initially suspended most routine in-person appointments to focus on emergency work and/or moved to online appointments.

Concerns have grown regarding the impact of prolonged social distancing measures and school closure on the mental health of previously healthy CYP, as demonstrated by the world-wide increased rates of depression, anxiety, inattention, problematic eating and alcohol and cannabis use [1‐4]. This increase has been particularly evident during the first period of lockdown [5]. Disruption of routines, family and peer relationships, education, and support from services were identified as risk factors [6, 7], while a supportive network and adaptive coping strategies as protective factors [4, 6]. There have been fewer studies on CYP with pre-existing mental disorders but the existing ones mostly reported a negative impact, with worsening of pre-existing symptoms and/or emerging of new complaints, which they linked to their vulnerability and pandemic-related reduced support from services [8‐12]. However, other studies did not observe worsening but stable or improved symptoms [13‐15], which suggests that distinct clinical populations may be differentially affected according to diagnosis, socio-demographic characteristics, and pandemic-related factors. Most prior studies investigating the effects of lockdown included mixed clinical samples [9, 13, 16] or only CYP with a specific diagnosis [8, 12, 17]; and the majority of them focused on neurodevelopmental disorders. Thus, it is unclear whether CYP with distinct diagnoses may have been differentially affected. Further, prior studies including clinical samples reported that age, parental mental health, and financial challenges were associated with worsening of CYP’s symptoms [9, 12, 17]. However, other potential contributing factors, such as ethnicity, housing adequacy, and the type of restrictions, were less investigated. Finally, the relationship between remote education experience and mental health in CYP with pre-existing mental disorders has received limited attention [18, 19], thus the impact of changes in education provision in this vulnerable population is unclear. Addressing these questions is of importance as it may guide policy and clinical practice, and thus help mitigate the effects of this pandemic and inform the response to any future ones. For instance, child mental health services are still unclear about how to organize their limited resources, as there is considerable uncertainty over which diagnostic groups have been most affected; what pre-Covid socio-demographic and pandemic-related contextual factors conferred most risk; and what was the impact of changes in education provision.

This study aims to address these questions and is part of a larger clinical population-based prospective cohort study, which surveyed CYP and families under CAMHS at three time points during the pandemic. This specific report focuses on cross-sectional data based on parent responses to the first survey (June–September 2020) integrated with pre-pandemic socio-demographic and clinical information as routinely recorded on CAMHS electronic health records (EHR).

Our results showed that among CYP receiving support from mental health services pre-pandemic, those with ASD or ADHD were the most affected during social restrictions. However, irrespective of pre-pandemic diagnosis, the negative effects on family life, education and mental health have been extensive; and poor parental mental health and challenges with education were most strongly associated with worse emotional and behavioral outcomes.

Results that may be most informative for care pathway and resource planning are those on the differences among diagnostic groups, showing that CYP with ADHD or ASD were the most adversely affected by restrictions. Previous studies mainly focused on small samples with a single or mixed diagnosis and did not provide comparisons. These studies reported either a worsening of pre-pandemic symptomatology or development of new symptoms [12, 24, 25], especially in those with more severe illness pre-pandemic [17, 26]. A prior study in a small mixed sample showed that behavioral problems were more evident in CYP with ADHD and worsening prosocial behavior in those with ASD [10]. We observed that CYP with ADHD had similar behavioral but less emotional deterioration than those with ASD, and that both had more difficulties with social distancing and housing adequacy during lockdown as compared to those with EmD. Their coping difficulties may be partially related to different underlying mechanisms as, for instance, CYP with ADHD struggled the most with spending time at home, which may be related to impulsivity and hyperactivity; whilst those with ASD had more Covid-related worries and spent less time outdoor even for allowed activities. Finally, we observed that CYP with EmD were less adversely affected, and a greater proportion showed an improvement in emotional symptoms. There are different potential explanations for these findings, as CYP with EmD may be more resilient and/or may thrive more in the home environment, where they are less exposed to the school-associated stress that often occurs in EmD [18, 27].

Our findings suggest that both pre-Covid socio-demographic and Covid-related contextual factors play a role when considering the impact of the pandemic on CYP. In agreement with previous studies, we found that behavioral problems were associated with younger age and emotional problems with older age and female sex [9, 10, 28, 29]. Unexpectedly, parents of Black and ethnic minorities reported lower levels of emotional and behavioral difficulties, which contrasts with the growing concern on the disproportionate impact of the pandemic on minority ethnic groups [30, 31]. This may be due to a selection bias and reflect the characteristics of families that responded to the survey. Alternatively, as these communities are highly represented in our catchment area, it is possible that the reliance on a large family/community network could buffer the effects of the pandemic [30]. Similarly, neighborhood deprivation did not appear to be a significant associated factor, in contrast to previous studies [17, 26]. This may be due to residual confounding when controlling for an area-based measure of socio-economic deprivation [32], or to the fact that our catchment area mainly included very deprived boroughs, thus we may have been less powered to detect differences between higher and lower income neighborhoods. Regarding contextual factors, poor parental mental health was among the most important. The association between children’s and parents’ mental health has long been recognized and linked to both the direct effects of problematic parenting and indirect factors, such as family conflict and socio-economic deprivation [33]. Prior studies also highlighted the reciprocal interaction between children’s and parents’ mental health during the pandemic, especially when the child had additional needs [26, 34]. For instance, a child’s externalizing behavior was associated with increased parental stress [35] and parental low mood was associated with worse child’s ADHD symptoms [24]. In sum, the worsening of CYP’s symptoms may increase strain on parents, at a time when they may struggle themselves due to their own mental health, financial concerns, and lack of social support, and may be less able to respond to their children’s increased needs [17].

Finally, results on the impact of changes in education may be important for policy around education provision and for teachers. We observed that CYP with neurodevelopmental disorders, especially those with ASD, struggled more with remote education than those with EmD. Similarly, a prior study reported that CYP with ADHD had more remote learning difficulties than those without ADHD [19]. These results may reflect their reliance on the routine and physical environment provided by school. Further, low education enjoyment/engagement were associated with greater emotional and behavioral deterioration. Although these associations are based on cross-sectional data, and may potentially reflect a reciprocal interaction, they suggest that CYP’s experience of education is strongly related to their mental health outcomes.

This study has several strengths. It is one of the largest surveys internationally of CYP with pre-existing mental disorders with available pre-Covid clinicodemographic information; it is based on a well-characterized and ethnically diverse clinical population; and considers a comprehensive range of pandemic-related variables. All these factors enhance the generalizability of findings. They also allowed us to obtain relatively precise estimates; provide clinicodemographic characteristics of both responders and non-responders; and include ethnic groups that may be otherwise hard to reach. However, limitations should also be considered. Considering the study design, this is the first (cross-sectional) report of our longitudinal study, and our results need to be complemented by longitudinal data to draw more robust conclusions on the relationship between exposures and outcomes. It also focuses on the initial period of the pandemic but, as its impact is evolving over time, our results need to be complemented by the data we collected at following time points, for which they represent a valuable baseline reference. Considering the sample, we surveyed a large clinical population, but we did not have a normative sample; thus, further studies should investigate whether children with pre-existing mental disorders fared worse than those in the general population. Only 32% of caregivers/CYP completed the survey and some submitted incomplete questionnaires. We noted that questions with lower response rates were those on parent circumstances and respect of social distancing measures (Table 2). Thus, findings may not be completely generalizable to the whole clinical population served by out Trust; however, results suggest that responders and non-responders had similar baseline clinicodemographic characteristics and the inclusion of all available data, instead of completed surveys only, reduced potential bias. We were unable to control for baseline severity at the start of the pandemic as many CYP did not have an up-to-date structured measure of symptoms or function recorded at that time. We considered the most common diagnoses, but those that received less attention in the Covid literature need also to be investigated (e.g., eating disorders and psychosis). Considering the outcomes, we measured symptom-based outcomes, but others can be considered in future studies (e.g., the increasing rates of A&E presentations or self-harm in CYP). The survey did not include validated measures of symptom severity; however, questions were based on the standard CRISIS questionnaire, which has been used in many pandemic-related studies, and we used factor analysis to confirm the underlying structure. We asked parents to rate symptom change as compared to before the pandemic, as the survey was developed at hoc to investigate the effects of the pandemic and, thus, we did not have pre-pandemic scores for comparison. Finally, we considered a broad range of factors potentially affecting mental health outcomes in a large sample, and tested the robustness of findings in sensitivity analyses, but did not correct for multiple comparisons. This was due to the exploratory nature of the analysis and the lack of well-known robust associates of worse mental health outcome in the literature of CYP with mental disorders. Nevertheless, our findings may provide a reference for future studies aiming at the identification of the most relevant risk factors in this vulnerable population.

In conclusion, our findings suggest that mental disorders, especially ASD and ADHD, may negatively affect CYP’s resilience and flexibility to adapt to rapid changes, especially when they experience additional challenges in their support network. Knowledge of the burden that social restrictions placed on CYP and families is important to guide clinical recognition, resourcing, and policy. Finally, increasing the availability of evidence-based treatments that reduce parental stress and support education may help mitigate the mental health impact of future pandemics.