Help-seeking attitudes and behaviours for mental health problems in adolescents before and during the first COVID-19 school closures in Germany

Aim: Comparing measures of psychological wellbeing and help-seeking in youths before and within the first school closures due to the coronavirus disease 2019 (COVID-19) pandemic enables a better understanding of the effects the pandemic has for those seeking professional help for mental health problems. Methods: Data were obtained from the Germany-based ProHEAD school study. Pre-lockdown and lockdown samples ( n = 648) were compared regarding pupils' psychological wellbeing, help-seeking attitudes and help-seeking behaviour. Results: Participants from the lockdown sample showed greater positive attitudes towards seeking professional help, whereas psychological wellbeing and help-seeking behaviour remained stable. Conclusions: Possible explanations may include an increased public discourse on mental health or self-selection bias for participation during lockdown.


| INTRODUCTION
The coronavirus disease 2019  pandemic with its associated school closures and physical distancing requirements starting in spring 2020 placed a particular challenge for the mental health of adolescents (Racine et al., 2020). At the same time, access to health care services such as psychotherapy or outpatient psychiatric treatment was restricted, posing additional barriers to professional care, which is already under-utilized by young people in need (Sanci et al., 2010).
Thus, one could assume that the lockdown increased the need for professional mental health care in youths, whereas utilization was lower than before. The ongoing ProHEAD (Promoting Help-seeking using E-Technology for Adolescents) project allows for a comparison of the mental health status as well as help-seeking behaviours and attitudes of adolescents in Germany before vs. within the nation-wide school closures that were implemented in Germany on 16th March 2020.

| METHODS
ProHEAD is a multi-centre consortium situated at six study sites across Germany and led by the managing site at the Department of Child and Adolescent Psychiatry at the University Hospital Heidelberg (Kaess & Bauer, 2019). Ethical approval had been obtained by the lead Ethics Committee of the Medical Faculty at the University of Heidelberg (Study ID: S-086/2018) and subsequently at all involved study sites. Participation in the ProHEAD project comprises a detailed, usually school-based, screening that includes measurements of mental health problems (Strengths and Difficulties Questionnaire, SDQ; Goodman, 1997), help-seeking behaviour (Actual Help-Seeking Questionnaire, AHSQ; Rickwood & Braithwaite, 1994) and help-seeking attitudes (Inventory of Attitudes towards Seeking Mental Health Services questionnaire, IASMHS; Mackenzie et al., 2004).
Due to the nation-wide school closures in Germany from 16th March, the school-based screenings were paused and pupils who had provided written informed consent from themselves and their legal guardians were invited for online participation, this means they completed the baseline screening at home after an Email invitation. A total of n = 5408 participants who were recruited in five regions of Germany and attend school grades 6-13 (aged 12-20 years) had completed the ProHEAD baseline screening at the time of analyses. Inclusion criteria for participation in the ProHEAD study were an age of at least 12 years and the provision of written informed consent of the participant and a parent or legal guardian. Of the total n = 5408 participants, n = 324 participants completed the ProHEAD baseline screening between 16th March and 31st August 2020 ('lockdown sample'). Within this time, most pupils were taught via home-schooling. Although, the exact circumstances under which teaching took place varied dynamically between schools and over time, all participants in the lockdown sample experienced extensive constraints to their usual daily lives by physical distancing regulations and other containment efforts. From the adolescents who were screened since the beginning of the ProHEAD surveys in November 2018 and before 16th March ('pre-lockdown sample'), n = 324 participants were individually selected as a comparison group by matching participants for age, sex and school type (based on the following German school types: Oberschule/Gymnasium, terminating with the general qualification for university entrance; Realschule, terminating with a secondary school level certificate; Hauptschule/Werkrealschule, basis of vocational education; Gemeinschaftsschule/Stadtteilschule, all qualifications are possible). In total, a sample of n = 648 adolescents (50% each prelockdown and lockdown, 68% female, M = 14.93 years, SD = 1.88) arose for this analysis. Group differences in help-seeking and wellbeing between the pre-lockdown and lockdown samples were analysed using two-sample t-tests. Furthermore, changepoint analyses (Eckley et al., 2011) implemented in R using the 'changepoint' package (Killick & Eckley, 2014) were used to explore patterns of changes in help-seeking over time. Changepoint analyses estimate points at which the statistical properties of an observation change and contribute meaningfully to differences in test statistics between adjacent segments. Thus, this method can be used to detect separate homogenous segments in time series data (Eckley et al., 2011;Killick & Eckley, 2014). Here we defined a minimum length of n = 50 consecutive observations to constitute one segment for analyses.

| RESULTS
Analyses revealed that participants completing the ProHEAD screening after 16th March (lockdown sample) reported statistically significantly more positive attitudes towards seeking help for mental health F I G U R E 1 Changepoint analysis of the IASMHS scores from first to last included participant (N = 648). The red horizontal line marks the separate segments identified, the red vertical line marks the cutoff between pre-lockdown and lockdown sample.

| DISCUSSION
Compared to the pre-lockdown sample, participants from the lockdown sample reported more positive attitudes towards seeking professional help for a mental health problem. There were no differences in the level of psychological distress between the two samples. scale, which describes the willingness and ability to seek help for a psychological problem (Mackenzie et al., 2004). Figure 1 shows that the increased attitude values declined after some time and reached the pre-lockdown level when the last participants were included in August 2020. By this time, most schools had established new routines and students were able to become accustomed to home schooling and, in some cases, return to face-to-face instruction. This return to a 'new normal' may have reduced the immediate effects of the initial closures. Another possible explanation for the temporary improvement in attitudes towards professional help could be that the adolescents who participated in the survey during the pandemic came from a particularly involved parental home. The barrier to participating in ProHEAD during the school closures was higher than before because the survey had to be completed independently from home instead of guided in class during lessons, which was also reflected in declining participation rates in the lockdown sample. It is possible that during this period, students who were already more open to the topic of mental health or whose parents had motivated them to do so were more likely to participate. However, at this point, possible explanations for the changes in adolescent help-seeking attitudes can only be of speculative nature, and further scientific evidence is needed to confirm underlying causal relationships. The temporary increase in positive attitudes towards help-seeking did not result in higher rates of actual help-seeking though. This might be explained by the impeded access to health care services within the pandemic, hampering adolescents from seeking actual face-to-face help. Instead, those in need turned to online services, as demonstrated by an increased utilization of the ProHEAD-online intervention (Kaess et al., 2020). Online interventions for adolescents with severe mental health problems, like implemented in line with the ProHEAD project, presumably could compensate for some of the ceased face-to-face contact points of professional help.
The extent of mental health problems was not influenced by the lockdown in the present matched samples; neither did these variables moderate any effects of the lockdown on help-seeking. Thus, regardless of their mental health status, participants in the lockdown sample did not have an increased need for treatment compared to those from the pre-lockdown sample. Although other studies reported increased loneliness and mental health deterioration in young people during the lockdown (Racine et al., 2020), pupils in the present sample did not show increased mental-health problems in response to the pandemic.
While it is well plausible that adolescents with increased mental-health problems refrained from participation in the study, thus introducing sampling bias, experiences with ProHEAD are different. Importantly, ProHEAD provides online support for adolescents with mental health problems and accordingly previous utilization rates within the group of adolescents in need were high in the past and increased during the pandemic (Kaess et al., 2020). This indicates that in fact adolescents in need participated in the present study, which is also supported by the stable values of symptom severity in the two samples investigated. However, it is still possible that this pattern changed during the pandemic and those wo experienced declines in their mental health during this time did decide to not participate in the study. Nevertheless, the present finding should encourage to question premature conclusions regarding the mental health status of adolescents during COVID-19, especially as most of the previous studies, unlike the presented data from the ProHEAD project, lacked comparative pre-lockdown samples (Racine et al., 2020). It must be noted that the effect size for the difference in help-seeking attitudes between pre-lockdown and lockdown sample was small (d = 0.224). However, as the observed effect was not due to a specific intervention targeting at help-seeking outcomes but became evident as a part of general political measures, even a small effect should be recognized and discussed.
Overall, our results indicate a temporary improvement in attitudes towards seeking professional support for mental health problems during the first COVID-19 associated school closures in Germany. Possible explanations for this could be, for example, an increased public discourse on mental health, or a self-selection bias for participation in the ProHEAD study during the lockdown. The underlying relationships and ways to improve attitudes in the long term need to be the subject of further research to improve the uptake of professional support among youth experiencing mental distress. LUSTIG ET AL. 333