Two independent searches were carried out during June 2018 and then updated in March 2019. A total of 90 studies were included in this review, combining both barriers and facilitators (n=54) and the intervention (n=36) questions. PRISMA diagrams displaying the number of papers retrieved and the process of selection of the included studies is available in Figures 1 and 2. Regarding the inter-rater reliability for this review, the agreement between the researchers screening the papers was high, with a 85% accuracy and 95% precision (Kappa=0.954). Adolescents identified a range of formal and informal help-seeking options across studies, such as GPS, psychologists, psychiatrists, teachers, social workers (formal), and friends, family, sporting coaches, and online communities (informal). Regarding question 1, most of studies focused on identifying barriers and facilitators towards formal sources of help, whereas intervention studies had a wider variety of sources of help, depending on help-seeking behavior attempted to promote.
[Figure 1 and 2: Prisma flow chart for question 1 and 2]
Question 1: Help-seeking barriers and facilitators
Fifty-four studies that reported barriers and/or facilitators including a total of 56,821 participants were considered in the narrative synthesis (Table 1). Most of the studies (n=18) were conducted in Australia, followed by the United States (n=12) and the United Kingdom (n=5). The majority of the studies were cross sectional (n=36) [26-27-28-29-30-31-32-33-34-35-36-37-38-39-40-41-42-43-44-45-46-47-48-49-50-51-52-53-54-55-56-57-58-59-60-61], thirteen studies were qualitative [62-63-64-65-66-67-68-69-70-71-72-73-74] and six used a mixed-method design [75-76-77-78-79-80]. Three PhD dissertations and one conference abstract were included in the grey literature. The age ranged from 8 to 26 years old. Three articles included adolescents and their parents, while one article included just adolescents’ mothers.
The majority of studies were conducted in educational settings, such as schools (n=24) and tertiary education (n=11) focusing in non-clinical samples. Sixteen studies included participants from other community settings and two studies were conducted in mental health care facilities. Among the studies that include actual help-seekers (n=7), the most common reason for seeking help was suicidal ideation, self-harm, depressive symptoms, and general mental health concerns (e.g., anxiety/nervousness/fear). Therefore, the conclusion drawn by the majority of the articles were based on help-seeking intentions rather than actual behaviours, since the participants were not experiencing mental health problems and focused on hypothetical scenarios.
Stigma is defined as the fear of being socially sanctioned or disgraced leading to hiding or preventing certain actions or behaviours, including the misreporting of mental health problems . More than half of the included studies (n=30) made reference to this and other negative attitudes towards mental health problems as the main obstacle to help-seeking behaviours in adolescents. Of these, twenty-five studies referred to stigma as the primary obstacle, describing it through different concepts such as, “stigma”, “fear of stigmatisation”, “community stigma”, “perceived stigma” and “self-stigma. Other negative attitudes towards mental health problems included shame, fear, and embarrassment.
The second most mentioned barrier was associated to adolescents’ family beliefs toward mental health services and treatment (n=15). Barriers related to problem with communication and distrust towards health professionals, negative past experiences with mental health services, and believing that the treatment is not going to be helpful. This was especially true for studies including immigrant and refugee populations, which referred to cultural barriers including mistrust of mental health diagnosis and practitioners, and lack of cultural sensitivity in services as a significant barrier.
Mental Health Literacy
Mental health literacy refers to the ability to use mental health information to recognise, manage and prevent mental health disorders and make informed decisions about help-seeking and professional support . Almost one-third of the articles (n=14) referred to problems related to mental health literacy as a significant barrier including poor recognition of mental health conditions (self and others) and lack of awareness of available sources of help.
Adolescents’ attitudes towards help-seeking revealed a perceived need of self-sufficiency and autonomy were recognised as a relevant barrier in twelve studies, as well as fears of confidentiality breaches.
Other help-seeking barriers
To a lesser extent, problems regarding service and personnel availability and other structural factors (such as cost, transportation and waiting times) were mentioned as obstacles to help-seeking (n=8). This was a significant barrier for studies including rural and immigrant populations, and in studies that included parents in their sample.
Six studies focused on the relationship between symptomatology and help-seeking. These found that higher levels of psychological distress, suicidal ideation and depressive symptoms were linked to lower help-seeking behaviours.
Of the 56 included studies, 19 also referred to facilitators of help-seeking behaviours. Mental health literacy and prior mental health care were the most cited facilitators for help-seeking for mental health problems (n=10). Specifically, timely access to mental health was facilitated by having a previous positive experience with mental health services or help-seeking, being familiar with the sources of help, and good symptom and problem recognition. Higher engagement with the community and having a trusting and committed relationship with relevant adults such as parents, schoolteachers and counsellors also facilitated seeking help among adolescents. Further details of the included articles are available in Table 1.
Few studies identified a significant difference when comparing younger and older adolescents in relation to barriers and facilitators to help-seeking, with no conclusive findings being reached. Some findings suggested that older adolescents tended to establish to feel more comfortable with people with mental health issues  and had less help-seeking fears . In contrast, younger adolescents had greater knowledge about professional sources of help . Only one study found a significant difference between ages regarding help-seeking, with younger adolescents reporting higher intentions of seeking help .
Twenty-four studies examined possible gender differences in help-seeking barriers and facilitators. Seven studies did not find significant differences between genders [28-39-40-42-46-51-69]. One study reported higher help-seeking intentions in males experiencing suicidal intentions  and two studies found that females perceived more overall barriers [26-58]. However, this may be related to higher rates of females seeking help for mental health problems compared to males [31-33-37-42-48-53-58-61-76]. Studies reviewed did not evidence convincing differences between gender in relation to help-seeking.
Question 2: Help-seeking interventions
Thirty-six studies on interventions targeting help-seeking behaviour, including a total of 28,608 participants, were summarised in the review (Table 2). Most of intervention studies were conducted in Australia (14) and the United States (14), followed by Canada (4) and United Kingdom (3). All studies were conducted in educational setting including high school (n=35) and college (n=1). The majority of studies developed interventions for non-clinical samples, and their focus was the prevention of mental health problems and the promotion of healthy coping strategies via help-seeking behaviours. Outcomes varied between help-seeking intentions, attitudes and behaviours. Almost half of the studies focused on the effectiveness of the interventions, while sixteen were feasibility or pilot trials and study protocols. Most of the studies used a quasi-experimental design (n=21) followed by randomised controlled trials (n=15). The age of participants ranged from 11 to 19 years old, although one study that included participants under 29 years old was incorporated as more than half of the sample were adolescents. Interventions were delivered using four main methods: psychoeducation, outreach interventions, multimedia tools and peer leader training.
Types of intervention
Most of studies (n=23) used psychoeducation and classroom-based interventions. Although all the interventions focused on encouraging help-seeking behaviours, the emphasis and content differed among them, including general mental health topics, suicide and depression awareness and stigma.
Five studies developed programmes based on the notion that promoting mental health awareness could enhance mental health literacy and promote help-seeking [84-85-86-87-88]. Four interventions targeting help-seeking for suicide were identified within five studies [89-90-91-92-93]. Five interventions explicitly targeted help-seeking for depression in school-based settings their focus being to educate the school population about adolescent depression and thereby encourage help-seeking [94-95-96-97-98]. Two studies evaluated the effectiveness of an intervention combining depression awareness and a suicide prevention programme promoting early identification and self-referral [99-100]. Six classroom-based interventions addressing stigma were identified, two of which used psychoeducation to overcome myths regarding mental illness [101-102] and four focused on providing interpersonal contact with people with mental health conditions in order to improve acceptance and increase help-seeking intentions [103-104-105-106].
Three studies used outreach interventions to target mental health help-seeking [107-108-109]. These aim to establish contact with adolescents who may be experiencing psychological and emotional distress in order to help them get the attention they need and increase their access to health services. They were based on the Building the Bridges to General Practice (BBGP) programme, developed by Wilson et al. (2005), a programme that aims to target help-seeking obstacles for physical and psychological problems by promoting contact between high school students and general practitioners .
Six types of multimedia interventions have been developed to address some of the difficulties of reaching an adolescent population, such as fear of confidentiality breaches, stigma and self-reliance [111-112-113-114-115]. The interventions included interactive films to engage students with mental health related topic and online platforms providing personalised information regarding the decision-aids process.
Peer training interventions
Peer training interventions are focused on the training of peers who act as active agents of change and social interactions incorporated into the daily activities within the school environment . All three programmes followed similar principles concerning improving the climate around mental health problems, promoting social connectedness, and challenging norms and behaviours associated with help-seeking [117-118-119-120]. “Peer leaders” acted as a link between the student population and mental health literacy, promoting the acceptability of seeking for help for mental health problems.
Further details of the included articles are available in Table 2.
No studies referred to significant differences concerning the effectiveness of help-seeking interventions when comparing ages. No significant gender differences were identified regarding the effectiveness of the help-seeking interventions [89-101-103-111]. However, before the intervention females tended to have higher mental health literacy and more adaptive attitudes regarding mental health problems [90-111], including greater help-seeking knowledge and intentions [107-112-113].
The main goal of this review was to describe the interventions targeting help-seeking in adolescents and therefore did not include an analysis of their effectiveness. Almost half of the included studies were study protocols and feasibility studies, so effect sizes were not reported. However, some findings are worth mentioning.
Four studies which looked at effectiveness of the interventions focused on psychoeducation about depression found a significant effect in increasing help-seeking. King et al.,  identified that there was an increase in future help-seeking behaviours after the interventions and that this was maintained at three months’ follow-up (t= 4.634/ p<.001). Strunk et al.,  found a significant increase of help-seeking (p<0.0005); however, this was not sustained at follow-up (p=0.014). Robinson et al., found that the intervention group was more likely to seek help at post-test (Odds ratio (95% C.I) =3.48 (1.93, 6.29), p<0.0001) and Ruble et al.,  found increased intention of help-seeking from others after the intervention (t= 13.658/ p<0.0001.).
The three studies that looked at the effectiveness of stigma reduction identified positive effects of the intervention on help-seeking. Two studies [101-104] found a significant reduction in self-stigma surrounding seeking help after the intervention (p<0.05) and one study  found a significant effect of the intervention in help-seeking intentions (Wilks’ L=.942, F (4,417)=6.428, p<0.001).
Finally, all the studies that focused on outreach found a significant effect of the intervention in help-seeking intentions. One detected an increase in intentions at three months follow-up (F(2,217)= 3.04/ p<0.05) , Rughani  found short terms improvements in help-seeking intentions (F (14,225) =1.87 p < .03) and Wilson  found a significant effect in the intention of seeking help for psychological problems after the intervention (F(2,598)=4.31 p<0.01).
The majority of the studies were low to medium quality with moderate to high risk of bias. Most of the cross-sectional studies did not state a clear inclusion and exclusion criteria and did not consider possible confounders affecting the interpretation of the outcome. Regarding qualitative research, the most common problem was linked to sample size and the difficulty of providing a clear strategy to address the subjectivity of the authors in the interpretations of the data. Mixed method studies presented some inconsistencies in addressing specific components of both quantitative and qualitative traditions, and in the process of integrating both approaches. Regarding intervention studies, it was difficult to identify to what extent the groups were similar at baseline. Although some studies included baseline measures of demographic information, most of them did not consider confounders or other factors influencing effectiveness, and some studies did not have any baseline measures. Also, few studies included follow-up and the ones that did, had high attrition rates and short follow-up periods (up to 6 months); therefore, it is not possible to attribute a long-lasting effect to the interventions. Quasi-experimental studies acknowledge possible selection and sample bias. Randomised controlled trials presented difficulties in terms of the blinding of the research team and participants at different stages of the process.
Overall there was inconsistency regarding the measurements of help-seeking, with most of the studies focusing on help-seeking intentions, which is not necessarily related to future behaviours. Moreover, many studies did not use valid and reliable instruments for measuring help-seeking. This is especially true for the experimental studies since most of them developed tools focused on their intervention rather than standardised help-seeking measures. Finally, most of the studies only used self-report measures, increasing the risk of bias of the findings. We did not assess the quality of study protocol, feasibility studies and pilot studies.