The Break Free from Depression training module was successfully implemented among a large number of students and teachers spread over large areas in South-West Nigeria. There appears to be more female participation than males. Though data for the school enrolment rate in most part of southern part of Nigeria have shown a higher female enrolment than males [38], the difference is not enough to explain the female preponderance in the present project. It is possible that aside higher female enrolment rate, girls may be more easily recruitable into this kind of project than boys. It is also possible that the recruiters may have been inadvertently biased towards girls as part of the pervading sense of the deliberate need for inclusion of the girl-child in all health programs. Expectedly, there were more student participants from public schools than private schools because public schools are more likely to have a higher student density than private ones. However, there are more teacher-participants from private schools who agreed to participate in the program. This is also understandable as, despite a lower student density compared with public schools, an average private school in Nigeria has higher teacher: student ratio, with teachers that are less burdened with work and with have better incentives. All of these factors  may have enhanced their willingness to participate in an extra-curricular project like the present one.

Turning to the main objective of the project, the successful execution of the modified version of the Break Free from Depression training module in Nigeria is, however, a signal that such large-scale school-based mental-health programs is feasible and potentially impactful. It also confirms earlier needs-assessments in Nigerian schools which have shown that teachers and school administrators recognized the need for and will welcome school-based training programmes that would improve mental-health literacy within the school ecosystem [39, 40]. The overall finding is that the intervention resulted in significant improvement in the three target domains of depression literacy (knowledge, attitude, and confidence) among the students and their teachers. The highest effect size, which is of moderate size, was seen in the knowledge domain, followed by the confidence domain, among the two groups. Other small-scale controlled studies conducted in Nigeria and other parts of Africa have demonstrated improvement in aspects of mental-health literacy after a short training program among students and teachers alike [31, 32, 41].

However, for both groups, that is; students and teachers in the present study, attitude was the least impacted, in terms of both the significance of change and the magnitude of impact. In the present study, the change in attitude for teachers fell short of statistical significance by a few points. While one is inclined to conclude that our intervention improved attitude in students but not teachers, one must be mindful of the difference in sample size of the students compared with teachers in the present project and how this might have affected the statistical power of the analysis and the resulting determination of significance in both groups. The number of students that we were able to include in the study was at least 13 times larger than the number of teachers. It is possible that though the effect size of the change for attitude was the smallest, the sample size of the teachers may have just been below what would have been required to obtain a statistically significant result. The implication is that it is plausible that the intervention was successful in changing knowledge, confidence and attitude for both teachers and students. Suffice to say, however, that there have been mixed results on the impact of mental-health training programs on attitude among students and teachers. Some studies, including ones conducted in Africa, have found significantly improved attitude after mental-health literacy training programs [41, 42]. Other training programmes, however, have observed sustained poor attitude despite intervention [31, 32].

Recent systematic reviews showed that intervention program designs which involved extended contact-time spread over days or weeks and with multiple components (such as survivors lived-experiences and film/ documentaries) had better impact on attitudinal change compared with ultra-short and abridged training programs [43]. Future depression-literacy programs must be mindful of this design consideration. However, such designs must be mindful of the possibility that, in a resource-constrained setting such as the setting of the present project, a lengthier and more involved intervention may also not be as desirable given that it might not be as scalable as the present intervention. Furthermore, compared to the teachers’ group, the students’ group had an almost consistent significantly higher post intervention scores in more domains. This suggests that adolescence may be the golden period for achieving the best results when delivering mental health literacy interventions [44]. Adolescents are probably more likely than adults to find new knowledge interesting and be open minded enough for significant attitudinal change.

In terms of predictors, pre-intervention scores for the three domains of knowledge, attitude, and confidence were predictors of positive change in knowledge, attitude, and confidence, at least among the students. This finding is intuitive because it is expected that those with a pre-existing high-level of knowledge, attitude and confidence may also have better innate abilities to learn or are more likely to connect better with the training. The finding also suggests that mental health literacy interventions should be a continuing exercise, as those who are more likely to benefit from new information are those with better pre-existing information. Additionally, being a student in private school predicted improvements in attitude and confidence, while older adolescents and girls had better improvements on their post-intervention confidence in handling depression and suicidality issues compared with peers, all other factors kept constant. The sociodemographic nuances of impact of interventions are yet to be examined in the adolescent mental-health literacy literature and as such, we are unable to situate these observations. However, suffice to note that students of public and private schools in Nigeria often have wide differences in socio-economic background. Therefore, there might be certain socio-cognitive factors at play which are in favour of the demographics of the private-school students, but which were not measured in the present study. Likewise, there might be certain drivers of age and gender differences in the impact of the training program on attitude and confidence, which were not captured in our design.

Among teachers, improvement in knowledge and confidence was not dependent on age, gender, type of school where they taught, nor their pre-intervention assessment scores. Only their pre-intervention scores on attitude marginally predicted improvement in attitude. Aside from documenting the impact of the program on knowledge, attitude, and confidence, most of the mental-health literacy trainings that have been conducted hardly document the socio-demographic and other predictors of individual benefits. Therefore, there is hardly any previous study to compare these observations with. However, if these findings are replicated in future studies, it may have implications for design of similar interventions for maximal impact. These are areas for future exploration and replication.

There are a few other markers of success of the training program. Up to half of adolescents who participated in the training admitted to knowing someone (including themselves) who has sought help or advise from an adult for depression-related issues within the week of the training. While we did not determine the nature of the inquiry or the quality of the advice received, the finding indicates that the training stimulated conversation around depression and suicidality within the community. Such conversation may eventually lead to help-seeking. Furthermore, close to two-thirds of students who participated in the training program agreed that the program was a good way to learn about depression and suicidality. In the context of general mental-health literacy, school-going adolescents in Nigeria have found other school-based training programs acceptable [32]. This is another indicator of the acceptability of school-based mental-health literacy programs in Nigeria. Lastly, majority of the participating adolescents expressed willingness to be peer-trainers on the program if such opportunities were available. This observation has two different implications. One is that the finding seems to indicate a change in attitude that may not be adequately captured by the items on the survey which are explicitly designed to measure the construct of attitude change. Second is that the suggestion that respondents in this project may embrace peer-led mental-health literacy interventions. Though not yet implemented in the region of Africa, peer-led depression literacy programs has been shown to be acceptable and effective elsewhere [45]. Peer-led programs are viable alternatives to specialist-driven mental-health literacy programs in LMICs, such as Nigeria, where child mental-health specialist practitioners are in short supply and services are constrained [46]. For this reason, peer-led programs are more likely to be scalable in Nigeria than specialist-driven programs such as the present one and should be a future direction. The self-reported enthusiasm of the participating adolescents in the present study to serve as peer-educators is, therefore, encouraging.

The involvement of teachers (and to some extent, parents) is one of the major strengths of the present study. This is because the involvement of teachers and parents is a key addition to primary mental-health interventions, including psychoeducation programs for the adolescents. Though we had to abridge the “Break Free from Depression” depression-literacy program before implementation due to budgetary constraints, this was concurrently done with a cultural adaptation to be more culturally accessible, increase feasibility, and enhance ease of administration. The ability to modify the intervention to improve the fit to the community as well as to fit the resources and funding which are available is of paramount importance in doing effectiveness research. This is especially important to keep in mind with the idea of the intervention being scalable in order to benefit the maximum possible number of youths. Keeping these goals in mind, the fact that the intervention was found to be successful in improving knowledge, confidence and attitudes while being tailored to the needs of the treatment population increases the soundness of the study and its conclusions.

There are, however, a few design limitations of the present study that must be acknowledged. The first is that the schools which served as setting for the study were selected based on convenience. Therefore, the findings may not be easily generalizable. However, the schools were not selected based on any other bias that may have significantly influenced the results. Secondly, due to resource constraints, we were unable to expand the current project to include examination of other potential contributing barriers to service utilization such as stigma of mental-health. However, although we did not measure or intervene in stigma specifically, it is a fact that any effort geared towards improving mental-health literacy, as carried out in this project, will and have been shown to have some spill-over effect on reducing stigma [24]. For instance, an improved attitude, which we demonstrated in this study as one of the outcomes of our intervention, is a proxy for stigma reduction. Future studies may want to include a direct measure of and interventions specific for stigma in this kind of project. For the same reason, we report only the post-intervention outcomes within the week of intervention; we were unable to gather delayed outcome data. As such, we were not able to determine sustained impact. Other small-scale controlled trials of school-based mental-health literacy programs in Nigeria have documented sustained impact up to three months [31, 32]. Despite these limitations, the present study is still the largest school-based depression-literacy training program in Africa to date. Other future directions in subsequent studies will include child-parent co-training or interaction in discussion on depression-literacy program rather than the single-day parent briefing event as we did in the present project. This will enhance a continuous parent–child exchange long after the end of the project. Also, qualitative inquiry can be conducted among the students or teachers will be helpful, especially for the small but significant number who did not express helpfulness of the program as a way of learning about depression and suicidality. Such qualitative engagement will shed more light on better ways of organizing the project for improved acceptability.