Mechanisms of Change in the Prevention of Depression: An Indicated School-Based Prevention Trial at the Transition to High School

Abstract

Objective:

Depression represents a major public health concern and prevalence increases significantly during adolescence. The high school transition may exacerbate risk of depression for youth with pre-existing vulnerability. The High School Transition Program (HSTP) is a brief skills-based intervention that has demonstrated efficacy in preventing depression in adolescents. The current study aimed to evaluate the theorized mechanisms of change of the HSTP intervention by testing a multiple mediation model including school attachment (SA) and self-esteem (SE) as two mediators of treatment outcomes.

Method:

Students (N= 497; 61.5% girls) with elevated depressive symptoms, identified for the intervention program via an 8^th grade screening were randomized to a brief intervention (n=247) or the HSTP (n=233) from 2003-2008. Participants completed measures at 5 time points. The first assessment occurred at the start of the second semester of 8^th grade and the last assessment occurred at the end of 9^th grade. A multiple mediation model tested whether SA and SE contributed to changes in depression for youth in the HSTP.

Results:

The mediation model, including contemporaneously assessed SE and SA, was not supported. There was evidence of sequential mediation, such that students who participated in the HSTP intervention reported higher SA, which in turn predicted improved SE, and in turn contributed to amelioration of depressive symptoms.

Conclusion:

The HSTP intervention ameliorated depressive symptoms by targeting factors specific to the school transition (i.e., SA). Results suggest youth at risk for depression may benefit from prevention efforts that enhance students’ capacity to effectively manage identified environmental stressors, such as school transitions.

Clinical trial registration information:

Middle School to High School Transition Project: Depression and Substance Abuse Prevention; https://clinicaltrials.gov/; NCT00071513

Introduction

Adolescent depression represents a major public health concern with as many as 12.8% of youths ages 12-17 experiencing a major depressive episode within the past year¹ and prevalence rates appear to be increasing for this age group.² Moreover, risk for depression continues to increase significantly from early through late adolescence.³ Negative sequalae for youth with subclinical depressive symptoms include decline in academic performance and subsequent unemployment or under-employment,⁴ decreased social functioning,⁵ and increased risk of suicide.⁶ Thus, there is a pressing need for developmentally appropriate prevention interventions for adolescents at-risk for depression. As explicated in the diathesis-stress model, risk of depression increases when an individual with pre-existing vulnerability is subjected to a significant stressor.⁷ Youth at-risk for depression may lack the coping skills or support necessary to effectively manage environmental stressors, which in turn contributes to depressive symptoms. During adolescence, the high school transition represents a developmentally normative process that may represent a significant stressor for some youth.⁸ In particular, the high school transition may exacerbate risk of depressive symptoms for predisposed youth; therefore, prevention efforts that facilitate youths’ transition into high school may be especially important for decreasing risk of future depressive symptoms.

Prior research demonstrated that indicated prevention interventions for adolescent depressive symptoms produce modest outcomes.^9–12 One explanation for the failure to find stronger effects is that previous preventions have utilized interventions that are either irrelevant, or too general, to adequately address the specific needs of the population.¹³ Given the precipitous incidence of depressive symptoms following the high school transition, targeted prevention interventions during the transition from middle school to high school, which emphasize skills and support related to the transition, may be particularly well-suited for preventing adolescent depressive symptoms. To that end, the High School Transition Program (HSTP) was designed to build social and academic problem-solving skills and engagement during this period of vulnerability for adolescents. Adapted from the Coping and Support Training for the Transition (CAST) intervention,^14,15 the HSTP is 12-session group skills-based intervention that has demonstrated effectiveness in preventing depressive symptoms in adolescents transitioning to high school in a single study.¹⁶ Specifically, youth randomized to the prevention intervention demonstrated a significantly sharper decrease in their rates of depressive symptoms from baseline to 18-month follow-up relative to youth in the control condition.¹⁶ Given the outcomes of the HSTP the current study sought to evaluate mediators of treatment outcome for depressive symptoms.

Considering proposed limitations in existing adolescent depression prevention programs (i.e., broad and/or lack of pertinent content), the HSTP intervention integrated evidence-based elements of both depression prevention^10,12,17 and school transition programs.¹⁸ The HSTP theoretical model (see Figure 1) proposes that adolescent depressive symptoms may be prevented through individual skill-building targeting students’ risk factors, such as low self-esteem, while also increasing students’ environmental support (i.e., school attachment) .

Figure 1.

High School Transition Program (HSTP) Theoretical Model

Putative Mediators

School attachment.

According to life course theory,^19,20 school transitions represent a developmentally-typical social transition that is modulated by individual (e.g., coping skills), contextual (e.g., school environment), and social (e.g., social support) characteristics that determine future functioning. In this respect, intervening on any of these characteristics (e.g., individual coping, school variables) influences outcomes of the transition.¹⁹ By definition, school transitions change students’ social environments through introduction of new peers, teachers, and structure. Typically school transitions involve students from multiple schools merging and changes in students’ social networks, which may result in changes to perceived peer support.^8,21 Further, as students have to adjust to new academic expectations, environment, and teachers, they may feel less attached to their new school and less supported by their teachers.^22,23 For youth with preexisting vulnerabilities, these changes may deplete their available coping abilities and consequently increase adolescents’ risk of depressive symptoms. Likely as a consequence of these changes, school transitions are associated with declines in students’ academic functioning⁸ and decreases in students’ self-esteem.²²

Enhancing school attachment during the transition to high school may be especially important considering a significant proportion of youth report feeling disconnected from school by the time they reach high school.²⁴ Changes in perceived school support may ultimately confer risk for depressive symptoms, as evidenced by the increased incidence of adolescent depressive symptoms following the high school transition.²⁵ Conversely, increasing school attachment-students’ perception of and use of social support (e.g., peer and teacher support) and students’ sense of belonging -- may decrease risk of depressive symptoms as students may feel better equipped and supported in navigating challenges associated with the school transition. Indeed, students’ perceived school attachment (including sense of belonging and perceived school social support) longitudinally predicts decreased depressive symptoms across adolescence and into adulthood.²⁶ Interventions that enhance youths’ connection to school and increase their agency to seek social supports may decrease future depressive symptoms. Taken together, disruptions to students’ school attachment, as a result of school transition, likely confer risk for depressive symptoms, and therefore represents an important target for depression prevention interventions.

Self-esteem.

Self-esteem reflects a multifaceted construct that includes an individual’s perception of self-worth, self-respect, and general beliefs about oneself.²⁷ The perceptions of self that comprise self-esteem are often driven by assessment of one’s own competence across different domains (e.g., academics, social skills, problem solving abilities) as well as through social environmental factors, such as peer approval).²⁸ Youth with low self-esteem may have a poor view of their own resources and social supports, which may in turn affect their ability to adaptively cope or seek social support when stressed, thereby increasing risk of depressive symptoms. This conceptualization indicates that interventions may influence self-esteem through skills acquisition and enhancement of one’s sense of belonging and social support. Taken together, self-esteem represents a malleable construct that may represent an important target of treatment given its role in concurrent and future functioning.²⁹ A wealth of research suggests that low self-esteem both predicts, and co-occurs with depressive symptoms in adolescents.³⁰ Furthermore, self-esteem appears to be particularly relevant during the high school transition. As a protective factor, positive self-esteem facilitates students’ successful school transition;³¹ however, decreases in self-esteem may also be a casualty of the transition.²¹ Interventions that improve student self-esteem may be well-suited to prevent deleterious consequences of the transition.

The HSTP (see Figure 1) aimed to reduce risks of depressive symptoms among students transitioning to high school by increasing: 1) acquisition of coping skills competencies (self-esteem); 2) engagement in positive social activities (self-esteem and school attachment); 3) social support resources by building a supportive peer network (self-esteem and school attachment); and 4) parent support during the transition period. Theoretically, the HSTP enhanced youths’ self-esteem through adaptive skills training delivered within a supportive peer-group context and by improving their social support (both at school and with caregivers). The HSTP also aimed to increase students’ school attachment by increasing social supports and promoting students’ participation in positive, school-based activities with peers, thereby decreasing isolation and increasing students’ available resources.

Current Study

Given the positive primary outcomes of the HSTP intervention,¹⁶ the current study aimed to directly evaluate the HSTP’s theorized mechanisms of change by testing a multiple mediation model including school attachment and self-esteem as two mediators of treatment outcomes. As the HSTP includes interventions directly targeting school attachment and student self-esteem, we hypothesized that prevention effects will be mediated through increases in student perceived school attachment and self-esteem. To explore the best fitting model, both multiple mediation and sequential mediation were evaluated as it is also possible that there is a sequential effect of the intervention. It is possible that HSTP enhanced students’ perceived school attachment, which in turn improved their self-esteem and subsequently ameliorated depressive symptoms. Conversely, it is possible that the HSTP improved students’ self-esteem which facilitated students’ perception of positive school attachment following the high school transition and in turn prevented future depressive symptoms.

Method

Six middle schools in a Pacific Northwest urban area collaborated in carrying out the indicated preventive intervention (HSTP) trial from 2003-2008. Seattle Public Schools agreed to participate in this project. This district has 12 middle schools and following district approval, individual schools were approached via contact with the principal. Seven middle school were invited to participate in the study. Schools were selected to represent the demographic diversity within the district. One school declined to participate. Administrative personnel and teachers in this school were open to participating. However, the school declined because their counseling staff indicated that they were already overburdened by several new school programs and they were not comfortable that some students would be randomly assigned to the Brief Intervention condition only. This school’s demographics during the study periods were as follows: 31.1% Asian American, 18.3% African American, 10.1% Latino/Hispanic, 0.5% American Indian, and 40% European American. The study was approved by the Institutional Review Board, active informed consent and assent were obtained from participants. Universal screening of the 8^th grade of participating schools was utilized to identify eligible participants. More information about universal screening and participant identification can be found in Makover and colleagues (2019). The ClinicalTrials.gov registration number is NCT00071513.

Participants

Four hundred and ninety seven 8^th grade students (61.5% girls) considered at-risk for depression were randomized to a brief intervention (n=247) or the prevention intervention program (n=233). Risk status was defined by elevated depression symptoms based on an adapted Moods and Feelings Questionnaire (MFQ).³² In consultation with Dr. Angold, modifications were made to the measure and the cut-point was based on 0.5 SD above the mean to reflect elevated depressive symptoms in the Seattle Public School middle school context. Dr. Angold noted several publications present information pertinent to the selection of MFQ cut points for use in various circumstances.^33,34 In the current study, two items were added (i.e., I felt uneasy or anxious, and I felt annoyed or irritated) and three suicide-related items (I thought about death and dying; I thought my family was better off without me; I thought about killing myself) were eliminated due to the impersonal setting in which the questionnaire was administered and the research team’s inability to adequately and practically follow up immediately with positive endorsements resulting in a 32-item scale with 30 MFQ items and 2 other items. The 30-item MFQ has been used frequently in school-based studies.³⁵ Thus, students were identified as eligible for the randomized controlled trial if their 32-item MFQ score was 15 or above. Eligibility also required that students score below the clinical cutoff for conduct problems on the Youth Self Report Aggressive Behaviors subscale,³⁶ understand English at a 3^rd grade level, and were not enrolled in a self-contained class for Serious Behavioral Disturbance Universal. The Clinical Evaluation interview allowed for additional clinical information regarding student problems and functioning and included the Screen for Youth Suicide Risk (SYSR); if indicated, parents and school counselors were notified and an appropriate care plan established. We did not screen for psychosis but did determine that 18 students who were receiving services for Severe Behavioral Disturbances/Special Education and 14 students receiving intensive outpatient mental health services were beyond the appropriate scope of this project.

A total of 2664 8th grade students were screened. This represented 60.3 % of those who were eligible, in all respects other than consent. Of non-participants, 476 parents declined (11% of the potential pool), 1222 parents (28%) did not respond, and 42 students (0.09%) declined. Screened students were 52.7% girls, and 47.3% boys. Ethnicity data of students eligible for screening at the school level was not collected; however, the final sample was generally representative of middle school students within the Seattle Public School system (see Table 1).

Table 1.

Demographic Information for Seattle Public School Students and Recruited Participants

Characteristic	Middle School Students	HSTP
	n (%)	n (%)
Female gender	4,885 (47.9)	314 (63.2)
Race/ethnicity
Euro-American	4,536 (44.5)	274 (55.1)
Black	2,446 (24.0)	68 (13.6)
Asian	2,403 (23.6)	78 (15.6)
American Indian	813 (8.0)	15 (3.0)
Hispanic	1152 (11.3)	62 (12.4)

Note: Direct comparison between groups was not completed as the school data represents all students within middle school (not just 8^th grade students). Demographic categories in the table reflect how Seattle Public Schools presented the data. HSTP = High School Transition Program.

The parents of eligible and interested students were contacted by phone. Of 697 families eligible to participate, 68 students declined, 2 students could not be reached, 61 parents declined, and 69 parents could not be reached within the study recruitment timeframes. Thus, 71.3% of the eligible sample or 497 students/families were randomly assigned, 241 to the HSTP condition and 256 to the Brief Intervention condition. For recruitment and enrollment details see Figure 2. Randomization was done via blind drawing of group assignment controlling only for equivalent distribution of gender across the two conditions. Demographic characteristics including gender, race/ethnicity, and SES are presented in Table 2.

Figure 2.

CONSORT Diagram

Note: HSTP = High School Transition Program.

Table 2.

Demographic and Clinical Characteristics of Recruited Sample

Characteristic	Intervention (n = 241)	Control (n=256)	p
	n (%)	n (%)
Female gender	149 (61.8)	165 (64.5)	0.54
Parent highest level of education			0.27
Less than high school	7 (3.1)	16 (6.6)
High school graduate/GED	37 (16.4)	37 (15.2)
Some college	71 (31.6)	84 (34.4)
College graduate or more	110 (48.9)	107 (43.8)
Race/ethnicity			0.70
Non-Hispanic White	130 (53.9)	144 (56.3)
Black	32 (13.3)	36 (14.4)
Asian	38 (15.8)	40 (15.6)
American Indian	6 (2.5)	9 (3.5)
Hispanic White	35 (14.5)	27 (10.5)
Hollingshead Score at screen	43.83 (17.5)	45.40 (15.1)	0.28
SMFQ score at screen	9.75 (5.0)	9.78 (4.9)	0.94
SMFQ score at Q1	8.67 (5.8)	8.23 (5.6)	0.35
Self-esteem at Q1	3.78 (1.3)	3.79 (1.2)	0.14
School support at Q1	3.85 (0.97)	3.73 (0.95)	0.94

Note: Table adapted from Makover et al.¹⁶ See original article for more detailed demographic information. SFMQ = Short Mood and Feelings Questionnaire; Q1 = Time 1

Assessment Procedures

Initial assessments were conducted in the home where the goals and purposes of the study were reviewed, consent/assent obtained, and assessments administered. The assessments were administered one month following screening (T₁), after subject selection and prior to random assignment. The first follow-up assessment occurred at the time of the final HSTP session (T₂ 3 months post-screening). Additional follow-up assessments were conducted at the beginning of the fall quarter of 9^th grade prior to the booster sessions (T₃ 9 months post-screening), at the end of the first semester of 9^th grade (T₄: 12 months post-screening), and at the end of the second semester of 9^th grade (T₅: 18 months post-screening).

Measures

Depressive symptoms.

Short Mood and Feelings Questionnaire (SMFQ)³⁷, which is a 13-item scale designed to measure depressive symptoms in children and adolescents age 8-17 years. SMFQ items are derived from the Diagnostic and Statistical Manual (DSM) criteria for major depression and dysthymia³⁷ and other features of childhood depression. Items are rated on a three-point scale (true, sometimes, not true) for the period of the past 2 weeks. The SMFQ has been shown to have high internal reliability (Cronbach’s alpha = 0.90), and results of exploratory factor analysis suggest that it is a unifactorial scale.³⁷ High correlations have been found between scores from the SMFQ, the Children’s Depression Inventory, and the Diagnostic Interview Schedule for Children depression scale.^37,38 Higher scores indicate more depressive symptoms. Internal consistency was strong across time points (αs=.89-.90).

Self-esteem.

Self-esteem was measured using a set of items from the High School Questionnaire (HSQ). Developed by the Reconnecting Youth research group¹⁴ the HSQ was administered at each of the at five time points over a period of 18 months from January of eighth grade through June of ninth grade. Most HSQ items are derivatives of established scales, reduced in length through empirical analyses with samples of both high risk and “typical” youth. Four items drawn from the Rosenberg Self-Esteem Scale²⁷ were used to assess self-esteem (i.e., I have many good qualities, I feel useless [reverse scored], I wish I could have more respect [reverse scored], I take a positive attitude). Items were scored on a Likert scale from 0 (never) to 6 (always). Higher scores reflect higher self-esteem. Internal consistency was adequate (αs=.69-.73).

School Attachment.

School attachment was also measured by a subset of items on the HSQ that have demonstrated reliability and validity.^14,15,39,40 Consistent with past use of this subscale, school attachment was conceptualized as the degree of attachment to teacher, peers, and conventional school goals (i.e., involvement, achievement, enjoyment and belonging) in their classes. Ten items reflecting satisfaction with classes, teacher, and peer support, as well as effort, sense of belonging and conflict in school was measured. Items were scored on a 7-point Likert scale (0 – never, 6-always). Higher score reflect higher attachment to school. Internal consistency was strong (αs=.85-.87).

Demographic information.

Students self-reported sex and race/ethnicity.

Randomization

Randomization was done via blind drawing of group assignment controlling only for equivalent distribution of gender across the two conditions.

Interventions

The High School Transition Program (HSTP).

The 8^th grade in-school intervention was adapted from the Coping and Support Training (CAST) small-group curriculum developed by the Reconnecting Youth research group at the University of Washington.⁴¹ Specific adaptations were made to address aspects of the school transition, for example, increasing school attachment by fostering communication between students and school personnel both before and after the transition. Content was delivered to six to eight adolescents per group in 12-sessions over the course of 6-weeks. Sessions were conducted during regular school hours with student attendance recorded and rotated through different school periods to minimize students missing the same regular class. The HSTP focused on reducing development of depressive disorders in at-risk youth as they completed the normative but challenging transition to high school. The overall objectives of the HSTP were: 1) to increase the acquisition of coping skills competencies; 2) to increase social support resources by building a supportive peer network; 3) to increase engagement in positive social activities; and 4) to motivate parents to increase their support during the transition period. Following the transition to high school, HSTP participants received 4 booster sessions in a 1:1 format to provide a bridge for the high school transition, which included support connecting with peers and teachers in the new school, improving communication skills (e.g., with new teachers), and goal-setting to support management of the new demands of high school. All student interventions occurred in the school setting and were led by a Master’s level clinician. A parent component was delivered by parent intervention specialists in four home visit sessions, two during 8^th grade, and two during 9^th grade, focused on psychoeducation about how school transition stress can affect their child and ways to encouraging students’ participation in social/school activities.

Brief Intervention.

Students randomized to the comparison group completed the one-on-one standardized interview and clinical follow-up with a trained clinician, with a phone call to parents to review concerns and make recommendations for additional services, as needed.

Attrition and Missing Data

Details of study enrollment and participant flow through the study can be found in the CONSORT diagram (Figure 2). Of the 497 participants initially randomized to study conditions, 480 completed time 5 measures (attrition = 3.42%). Across measures and time points missing data ranged from .004%-3.62%.

Data Analyses

Initially, we evaluated Pearson product-moment correlations to determine zero-order associations between variables. Mediation models were estimated using the SPSS PROCESS macro,⁴² which utilizes pairwise deletion in handling missing data; this was deemed appropriate given that rates of missing data were minimal (.004-3.62%). Further, the PROCESS macro evaluates missingness prior to estimating both main effect and indirect effect models which may result in more robust estimation when compared to mediation analyses involving individual regression models.⁴² The first set of mediation models included mediators (i.e., school support and self-esteem) that were measured contemporaneously. Subsequent models included serial mediation models that estimated school support (time 2 and time 3) and self-esteem (time 3 and time 4) at successive time points.

Results

Demographic and Clinical Statistics

Table 1 includes descriptive statistics and clinical characteristics of the study sample. Examination of between-group differences indicated that there were no significant differences across study conditions for baseline school attachment, self-esteem, or depressive symptoms (see Table 1). Additionally, there were no significant differences in group composition when considering sex or ethnicity. Sex was positively associated with depression at baseline (r=.23, p<.001) and follow-up (r=.22, p<.001), and negatively associated with self-esteem (r= −.19, p<.001) at baseline, such that girls reported greater depressive symptoms and lower self-esteem than boys. Ethnicity was not significantly associated with any study variable (|rs|=.009-.05, ps=.26-.85). Baseline depressive symptoms and depressive symptoms assessed at follow-up were negatively correlated with self-esteem (|rs|=.27-.62, ps<.001) and school support (|rs|=.15-.38, ps<.001) across time points. School attachment was positively correlated with self-esteem across time points (rs=.21-.41, ps<.001). Variables demonstrated positive associations within each construct across study time points (e.g., depressive symptoms at baseline were positively associated with depressive symptoms at follow-up).

Mediation Analyses

In order to understand the mechanisms of change in the HSTP intervention we conducted an iterative series of mediation analyses in line with the theoretical model of the intervention. Each model included sex, ethnicity, and baseline depressive symptoms, self-esteem, and school attachment as covariates. Initially, we estimated a multiple mediation model including self-esteem and school attachment measured at time 2 (i.e., immediately following the intervention) predicting depressive symptoms at final follow-up (time 5). This model was not supported as the indirect effects for school attachment (95%CI: −.03, .04) and self-esteem (95%CI: −.0002, .06) were not significant (see Figure 3).

Figure 3.

Initial Mediation Model

Note: Solid lines represent all estimated pathways. Indirect effects for both school support (95%CI: −.03, .04) and self-esteem (95%CI: −.0002, .06) were not significant.

Next, we estimated a sequential mediation model with school attachment at time 2 predicting self-esteem at time 3, which in turn predicted depressive symptoms at time 5 (see Figure 4). Results indicated that each pathway of interest was significant and in the anticipated directions. The indirect pathway from the intervention to school attachment to self-esteem to depressive symptoms was significant (95%CI: −.02, −.0005). Furthermore, after accounting for self-esteem, direct effects for group condition (B=−.11, p=.18) and time 2 school attachment (B=.02, p=.79) were not significantly associated with time 5 depressive symptoms.

Figure 4.

Serial Mediation Model

In order to determine the validity of this sequential model, we estimated a second mediation model evaluating school attachment at time 3 (i.e., after the school transition) predicting self-esteem at time 4 and finally depressive symptoms at time 5. Once again results supported a mediation model for the intervention influencing depressive symptoms through school attachment by way of self-esteem (95% CI: −.04, −.003).

Finally, in order to confirm the specific sequence of effects we estimated a reverse sequential model whereby the intervention influenced self-esteem, which in turn influenced school attachment to predict depressive symptoms at follow-up. This model was not supported for either time 2 to time 3 mediators (95% CI: −.0006, .004) or time 3 to time 4 mediators (95% CI: −.008, .002). Results of analyses using Time 4 depression were similar to the analyses described previously; therefore, we opted to retain results using Time 5 depression as the longterm follow-up period is a notable strength of the current study.

Discussion

Building on the evidence that HSTP, an adapted program designed to promote a positive transition from middle school to high school, for youth at-risk for depression demonstrated a positive prevention effect on depressive symptoms,¹⁶ the present study examined theorized mediators of intervention: self-esteem and school attachment. Identification of treatment mechanisms will allow further refinement of interventions and is a priority within the field.⁴³ Several mediational models were evaluated in accordance with the HSTP theoretical model. Results did not support HSTP’s direct effect on self-esteem, as the intervention was not directly related to self-esteem in the multiple, or serial, mediation models; however, results indicated that HSTP decreased depressive symptoms through improvement in school attachment, which in turn improved self-esteem. This trend was maintained when considering multiple time points (i.e., time 2 school attachment to time 3 self-esteem, and time 3 school attachment to time 4 self-esteem), and there was no support for a reverse-sequential model (i.e., self-esteem improving school attachment), thereby providing incremental support for the HSTP mechanisms of change. These results are important considering that interventions often indirectly target self-esteem through improvement in individuals’ skills (e.g., adaptive coping) prior to involving external resources (e.g., teacher support) to increase their school attachment, self-esteem, and depressive symptoms. The sequential mediation model is consistent with the broader literature considering the school ecological framework and strategies to promote successful transition to high school for more adolescents. Specifically, the intervention helped ameliorate depressive symptoms by providing students with positive peer-groups, supporting involvement in positive activities, and by building skills to adaptively cope with the school transition (e.g., communicating with new teachers). Though the current study included an indicated prevention intervention, findings support the inclusion of social-emotional curricula that strengthen teacher-student relationships, and encourage peer support, which may be applied from a universal prevention approach.

Results did not support the initial theoretical model with simultaneous mediators, as the HSTP did not directly influence students’ self-esteem. Despite evidence from previous research that school-based behavioral interventions can positively impact students’ self-esteem,⁴⁴ there may be a number of reasons why the HSTP did not directly influence self-esteem. Firstly, it is possible that a longer intervention time frame is necessary for interventions to directly impact self-esteem. For example, in Guo and colleagues work teachers implemented the Positive Action program across multiple years, whereas the current study included a relatively brief (i.e., 12 week) intervention.⁴⁴ Furthermore, self-esteem reflects a broad construct that includes an individual’s own perception of self, including beliefs about self-worth, competence, and efficacy,²⁷ and generally interventions that demonstrate positive outcomes for self-esteem typically do so through targeting intermediate characteristics (e.g., self-regulation strategies, social supports).⁴⁵ Specifically, as guided by tenets of self-esteem enhancement theory,⁴⁶ Guo and colleagues described an intervention that included specific strategies designed to directly alter students’ views of self (e.g. through participation in prosocial activities);⁴⁴ whereas in the current study the intervention aimed to increase students’ adaptive coping, social support, and involvement in school activities which in turn may have benefited students’ self-esteem.

Results of the current study are especially informative given the need for prevention interventions that are tailored for developmentally crucial periods (i.e., the high school transition) in hopes of altering at-risk youths’ trajectories and reducing risk of long lasting consequences (e.g., decreased frontal cortical volume)⁴⁶ or adult psychiatric problems.⁴⁸ In accordance with the diathesis-stress model of depression, youth with vulnerability for depression may demonstrate depleted individual resources (e.g., low self-esteem) and antecedent risk factors (e.g., academic decline) that contribute to increased risk for depression over time. School transitions are normative, and similarly it is developmentally typical for youth to experience changes to their relationships with adult authority figures (e.g., teachers) during adolescence;⁴⁹ however, not all youth report disruptions to teacher relationships during this period.⁵⁰ Youth with pre-existing vulnerability for depression may be especially at risk for future relationship discord, which contributes to poorer psychological functioning, such as low self-esteem, and subsequently depression. Taken together with results of the current study, positive school attachment, including positive student-teacher relationships, may act as a buffer against depressive symptoms. Importantly, results of the current study suggest that adolescent prevention interventions may be improved through consideration of the developmental and implementation context in which they occur. Namely, prevention interventions that leverage school resources (e.g., school counselors and teachers) to support students’ perceptions of school attachment as they transition to high school may ultimately benefit students’ perceptions of their ability to effectively manage the transition and reduce risk of depression. Such strategies may be easily implemented within existing universal prevention programs (i.e., Tier 1) designed to support students through the high school transition.

While findings from the current study inform the field by providing important insights into developmentally tailored approaches to depression prevention interventions, certain limitations must be considered. For the current study, case finding procedures included students completing a screening measure to determine study eligibility, which may have inadvertently captured youth who met clinical criteria for a depressive disorder and therefore necessitated more intensive intervention. Future research may consider including diagnostic assessments in order to ensure inclusion of only youth with elevated, but not clinical, depressive symptoms. School-level demographic data were not available for students eligible for screening, and therefore the representativeness of the sample in the current study is limited. Additionally, though the current study included an extended follow-up period (i.e., time 5 occurred 18 months after initial screening), addressing a relative weaknesses in the existing literature,¹⁰ the last assessment occurred during the spring semester of students’ 9^th grade year. Given that the incidence of depression increases significantly (i.e., approximately doubling) through late adolescence,⁵¹ it is possible that the current study did not adequately capture this progression, and future research should include longer follow-up periods to better understand the long-term effects of prevention interventions provided during the high school transition.

Despite these limitations, results of the current study point towards a promising approach for prevention of adolescent depression. Interventions that directly target factors associated with school transitions, such as school attachment, can have downstream positive effects that may reduce onset of depression in young adolescents. Further, the current study demonstrates that prevention interventions can be successfully integrated into school settings and effectively engage students’ key social supports (i.e., teachers, peers, caregivers) to reduce the impact of the transitions on depressive symptoms.