Introduction

The need to effectively improve access to care for youth with mental disorders is a global challenge. The World Health Organization has reported that mental disorders comprise the largest single category of medical disability in young people [1, 2]. Adolescence has been identified as a critical period for the onset of the majority of psychiatric illnesses [3]. Left untreated, mental disorders increase risk for numerous negative short- and long-term outcomes such as suicide, unemployment, criminal activity and poor school performance [4]. While mental disorders in young people can be effectively treated, many who need care do not receive it [5]. There are numerous barriers facing youth trying to access mental health services including lack of identification of illness by self or others, stigma, lack of knowledge about mental health, lack of funding, wait times, parental involvement, the youth’s geographic location and fragmentation of services [6–10]. Some of these barriers can be addressed by promoting early identification of youth with mental disorders and enhancing access to care in the school setting [11–14]. With most youth spending extensive periods of time at school, educators and other school personnel are well positioned to identify those who may be showing signs of mental illness. While there is substantial evidence that providing educators with mental health literacy training impacts on student’s knowledge and stigma reduction [12, 15–21], there is little—if any—research that measures the impacts of these school mental health literacy-based interventions on referrals of students from schools to mental health services. This study addresses that gap employing third party blinded data source characterizing the clinical profile of referrals from schools and other sources to a regional publicly funded child and adolescent mental service.

As mental health literacy programs have the potential to improve knowledge and reduce stigma, this may help youth and trained educators distinguish between normal everyday negative experience and a potential mental health problem or mental disorder; and may enable them to seek help needed accurately and appropriately. As such, we hypothesized that these population health outcomes may be closely linked to positive clinical outcomes in mental health care, such as early access to care, accurate identification of cases to various levels of care systems (e.g., primary care, community services, specialized clinics, and emergency and inpatient tertiary care systems), and number of referrals from mental health staff in schools to mental health care. More specifically, the current study was designed to test the following hypothesis: School-based mental health literacy training will affect the frequency and quality of referrals positively, such as global function, strengths/concerns, and clinical urgency and severity (e.g., more urgent and severe cases); demographic variables (age, sex, and family composition) (e.g., earlier identification); system variables (frequency of referral, repeat admissions and emergent or scheduled service level, and referral reasons) (e.g., referrals with more severe cases), as determined by a number of validated tool to evaluate the mental health profile of patients admitted to services in the current study.

In September 2013, a mental health literacy intervention School Mental Health Literacy Improvements for Educators and Students (SMILES) was initiated by the Child and Adolescent Addictions Mental Health and Psychiatry Program (CAAMHPP) in the Calgary Zone of Alberta Health Services (AHS). The CAAMHPP represents regional community and hospital-based child and adolescent mental health services that were integrated in 2001. The three-year project, known by schools as the Mental Health Literacy Project, was funded by the Alberta Children’s Hospital foundation and included training educators in local school boards in the identification and referral of students at risk of mental disorders to the appropriate mental health professionals. School staff were trained through one-day workshops that discussed common mental disorders found in youth, their presentation in the school setting and steps to making effective referrals, as well as strategies to reduce stigma towards mental illness. This paper focuses on the impact of the one of the SMILES training components: the "Go-To Educator" training [16] (described below) with educational professionals and its subsequent impact on student access to mental health services and clinical profiles of students entering those services.

Employing standardized data from blind assessment of referrals to regional child and adolescent mental health services, the effect of “Go-To Educator” training on referral rates and the measured quality of referrals was examined under the hypothesis that school-based mental health literacy would may improve the frequency and quality of referrals.

Materials and methods

Participants

The sample was drawn from consecutive referrals (n = 54158; 54.86% female; 208 schools) (Table 1) with those in the before and after groups having a current school noted on their registration record in the CAAMHPP program of the AHS, Calgary Health Zone. There were significantly more female than male participants (p < .00001) in the sample. The Go-To Educator training had been designed for secondary school staff who interact with students between the ages of 12–18 years. However, schools in Calgary are broken into primary (Kindergarten to Grade 6: up to age 12 years), junior high (Grades 7–9: ages 13–14 years) and high school (Grades 10–12: ages 15–17 years). Hence, schools with staff from every grade level could be exposed to training.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Description of independent variables by dependent groups and bivariate multinomial logistic regression of independent variables by dependent groups.

All individual WCWL CMH-PCS Item data calculated based on the same number of participants in the total score calculation and therefore were not included in the individual cell.

https://doi.org/10.1371/journal.pone.0277695.t001

Results

Table 1 describes the sample for each of the dependent variable groups across all independent variables and provides both the 95% CIs for each group for the purpose of comparison. Table 1 further presents the results of bivariate multinomial logistic analysis noting the magnitude of the significant differences. Odds ratios greater than the value 1 indicate that the likelihood of group membership is greater in the indicated group as compared to the base group representing admissions after Go-To Educator intervention Exposure. Of the forty-eight independent variables, there were eighty-two significant differences at p < .05 across the three groups in comparison to exposure to the Go-To Educator training before admission group. Forty-six differences were significant at p < .0001 across the three groups in comparison to exposure to the Go-To Educator training before admission group.

Table 2 presents the results of three logistic analyses comparing each group separately to the base group: Go-To Educator invention exposure before admission. Each was analyzed separately to overcome the effects of potential multicollinearity, due to the fact that all variables are measuring some aspect of clinical psychopathology.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Estimates comparing admissions after smiles exposure to all other groups in three logistic models.

https://doi.org/10.1371/journal.pone.0277695.t002

Compared to smiles exposure before admission, smiles exposure after admission was increased in the following variables: ace item: physical neglect; WCWL items: 15-family functioning or factors affecting child, 7-externalized symptoms/disruptive behavior, 11-significant biological family history of mental illness, 1-danger to self, 6-internalized symptoms, age, and length of stay from index to last admission.

Compared to smiles exposure before admission, smiles exposure after admission was reduced in the following variables: CGAS on index admission, total WCWL score on index admission, high risk sex (female/self-defined) ace items incarcerated household relative and exposure to domestic violence.

Compared to smiles exposure before admission, no exposure admission before 2014 was increased in the following variables: referral reason: adjustment problems, WCWL items 13-social/friendships/community functioning, 7-externalized symptoms/disruptive behavior, 1-danger to self, 6- internalized symptoms, as well as problem severity on index admission referral reason internalizing/emotional issues and age.

Compared to smiles exposure before admission, no exposure admission before 2014 was reduced in the following variables: length of stay from index to last admission, number of admissions, total WCWL score on index admission, emergency/inpatient admission and referral reasons eating issues and other.

Compared to smiles exposure before admission, no exposure admission after 2013 was increased in the following variables: ace item: physical neglect, parental separation or divorce, household substance abuse, emotional neglect, exposure to domestic violence, sexual abuse, household mental illness, emotional abuse, and incarcerated household member and in WCWL items 12- school and/or work difficulties, 13-social/friendships/community functioning, 7-externalized symptoms/disruptive behavior, as well as problem severity on index admission, total WCWL score on index admission and age.

Compared to smiles exposure before admission, no exposure admission after 2013 was reduced in the following variables: length of stay from index to last admission, WCWL items 6-internalized symptoms, 16-prognosis without further intervention, 5-children’s global assessment of function scale (CGAS), 17-degree of likely benefit with further intervention, 8-comorbid medical conditions, psychotic symptoms, 9-comorbid psychiatric conditions, danger to self, harmful substance use/misuse as well as high risk sex (female/self-defined), total ace score on index admission, emergency/inpatient compared to scheduled admission and referral reasons school/learning/attention problems, eating issues and other.

Discussion

The results of this study show evidence that access to care may be influenced by specialized training of educational staff. These employed methods similar to another study focusing on mental health literacy in primary care [32]. In general, the Go-To Educator training significantly changed the demographics, clinical profile, and treatment effects of students presenting to tertiary mental health services and substantially increased the referral rate of associated cases compared to untrained schools and schools pre-training. Changes were observed in respect to referral reasons, demographic characteristics (e.g., age, sex, and family compositions), clinical profiles (e.g., comorbid psychiatric diagnoses, danger to self/others, development progress, externalized symptoms, substance use, problems at home), most of the system variables (e.g., length of stay, repeat admissions), and some treatment effect variables (e.g., admission/discharge strength concern). Therefore, mental health literacy training in schools had a contextual effect on a substantial number of the independent variables compared to the non-exposed groups and pre-group. Taken together, the results are consistent with the mandate of mental health literacy training in schools, indicating that mental health literacy training in schools appears to improve identification of students generally with greater clinical severity and therefore more in need of care.

Findings in this study are responses to the call for school-based mental health programs by Mental Health Commission of Canada [33] to effectively address early identification of mental health problems or mental disorders among youth. While research suggests that the system-level conditions are essential for school mental health programs to flourish [34], there are system-level gaps such as protocol and agreements defining the pathway to services in the school community and insufficient professional development for school staff [35]. This study built a mechanism for education and health systems to initiate the dialogue and presented an example how between-system collaborations may be established by sharing data and therefore to facilitate the pathway through care.

Changes in referral reasons are in good alignment with how Go-To Educators were trained to identify at-risk students. For example, students referred to services from after-training schools were more likely to demonstrate adjustment problems, thought disturbances, and harmful behaviours/thoughts to others, and this may be because trained teachers had realized these challenges were related to potential mental disorders, such as anxiety or psychosis, rather than behavioural challenges alone. Meanwhile, we observed fewer students reported as having danger to self at intake, possibly because students were being sent for services before becoming suicidal or engaging in problematic self-harm behaviours. We found fewer students presenting to tertiary services for social/family issues perhaps because the Go-To Educator training taught that most of these challenges can be dealt with in the community without need to access tertiary care mental health services.

In addition, developmental concerns were no longer major referral reasons compared to no-training schools. This may have resulted because developmental concerns were not a key focus of the Go-To Educator training, which addresses mental disorders commonly start during adolescence such as anxiety, Attention Deficit Hyperactivity Disorder, Depression, and Schizophrenia. Eating disorders were less identified as well because the training effectively taught how to distinguish between normal dieting patterns and eating disorders so that Go-To Educators obtained the capacity for appropriate referral. Similarly, the number of students referred to services for addiction or legal issues dropped as a function of the training given; Go-To Educators were directed to refer students with a substance use disorder to Youth Addiction Services who do not currently share the same registration system with the rest of Child and Adolescent Services. This may have impacted the specificity of the referral.

Further, based on the fact that students referred to services were younger we hypothesized that the Go-To Educator training had improved early identification capacity among educators in the after-training schools. Identified students were also more from single parent family, indicating that educators may have become more sensitized to the needs of single parent families for mental health support.

We also found that although students from the after-training schools waited longer to be admitted to services, once in care they had more repeat admissions and they stayed longer in the service. These factors speak to the complexity and severity of these cases and the level of required services these students required. The more complex the cases, the more admissions and the longer duration of care provided. These proxy measures thus may suggest that the students referred as a result of the Go-To Educator training had more complex mental health care needs than those referred from control sites. While their symptoms were more severe from after-training schools at admission, the symptom severity was reduced significantly at discharge, meaning treatment was effective helping students to services. These are very encouraging results, implying not only the quality of referral was improved, but also students in need of care received appropriate care.

It should be highlighted that the Go-To trainings were open to any professional affiliated with a school. This included support staff, school-based nurses, occupational therapists, family school liaison workers and school-based police resource officers, in addition to the teachers. Creating a larger shared literacy regarding mental health may have had important impacts on the referral number and profile and is worth further investigation. As such, the educational system may consider to further stabilize the Go-To Educator training in the school setting, including deciding the number of Go-To Educators and number of training sessions in each school, establishing a coordinator role to facilitate the training in each school board, securing consistent funding and linking with the local mental health services. Therefore, the Go-To training intervention may facilitate the links between health and education to effectively address adolescent mental health needs in a collaborative manner in order to sustain the impacts of Go-To training in the long term. This goal may be achieved through the collaborative work by policy makers in both education and mental health in the future.

While an economic analysis was not conducted as part of this study, given the findings presented herein, there may be significant frugalities realized with this approach compared to usual school—clinic referral patterns. The cost of project implementation was minimal—the master trainer had his costs covered by the mental health service. The project formed training teams out of volunteers from health and education to lead the workshops, no additional funds to cover participant time were used for this purpose. Infrastructure costs of the program were negligible, comprising mostly part-time administrative support. The vast majority of trainings occurred on time designated to support existing professional development creating minimal need for substitute teachers to “back-fill” for educator participants. ’School hosted’ trainings eliminated the need for travel for participants, and site training costs such as space rental were covered by the schools in which the training occurred. However, these preliminary cost considerations now need to be appropriately calculated using appropriate economic analysis and will be addressed in future studies [36].

Limitations included the possibility that trained Go-To Educators might move between schools. While such migrations happened, this was not the majority of trained teachers, who were equally likely to migrate to other exposed schools. Students with identified mental problems could also move between schools. Both forms of migration would bias the results in favor of the null hypothesis. Some students with mental problems may have sought assistance from private services (not CAAMHPP), which would similarly bias the results in favor of the null hypothesis.

Conclusion

To summarize, the Go-To Educator training was shown to be effective to potentially help educators early identify students at risk of mental disorders, accurately distinguish the difference between a mental disorder and everyday challenges or problems, understand different functions of health agencies for appropriate referral of at-risk students, and to support the health system to effectively treat students in need of care. Previous research on mental health training of Canadian educators has been focusing on knowledge improvement, stigma reduction or enhanced help-seeking [14, 17, 37], or the training of in-school mental health professionals (e.g. guidance counsellors, social workers, and health nurses) to address mental health needs of young people in the setting. Our approach to blend teachers and in-school mental health professionals is, to our knowledge, the first of its kind to closely effectively link education and health together and has the potential to reduce the burden of youth mental illness through early identification and improved quality of referral and treatment. This is also the first study using relatively objective independent measures, rather than self-report measures, to evaluate the impact of community based mental health literacy interventions on the mental health system, which may provide information on how to evaluate such interventions in the future.

Staff migration between schools during the training period may have impacted their knowledge of mental health and referrals. Such migration would have dampened the impacts in trained schools and enhanced in untrained schools. Other mental health initiatives that were not coordinated with the Go-To Educator training and were running concurrently may have created impacts in the referral profiles (YAS outreach and Headstrong events). As these were also being applied in comparison schools the training showed impact in spite of this potential confound. Future studies will need to better control for concurrent mental health related activities that may influence referral rate and quality. The study is not a randomized controlled trial design which may lead to biased results due to the potential different demographics of students referred to the services between the exposed and unexposed schools.

Acknowledgments

The authors would like to thank Mina Hashish and Rebecca Alaffe from the Teenmentalhealth.org Team for their work on preparation and submission of the manuscript.