Long-term Cross-over Effects of the Family Check-Up Prevention Program on Child and Adolescent Depression: Integrative Data Analysis of three Randomized Trials.

Abstract

Objective.

The present study examined prevention effects of the Family Check-Up (FCU) prevention program on longitudinal changes in youth depression, using harmonized data collected across three prevention trials, including one trial initiated in early childhood and two initiated in early adolescence (total N = 2322).

Method.

Data from parent and youth reports of youth depression were harmonized using Moderated Nonlinear Factor Analysis (MNLFA), which provides a robust means to examine differential item functioning (DIF) across subgroups of participants (e.g., age-groups, ethnic groups), and creates scale scores based on all available items while accounting for individual differences. Long-term intervention effects were tested using a multi-informant growth model examining changes in depression from baseline to up to 14 years post-baseline.

Results.

Across trials, significant long-term effects of the FCU on reductions in depression were observed, although effects were found to wane after approximately ten years.

Conclusion.

FCU effects on depression across trials were attained with a relatively brief parenting program designed to reduce behavior problems and improve relational functioning that emphasized parental motivation to change while supporting positive parenting strategies. Implications of these results are discussed, along with directions for future work in this area.

Nearly 20 to 25% of teens will experience a major depressive episode by the end of adolescence (Kessler, Avenevoli, & Merikangas, 2001). Levels of depression escalate in early adolescence, particularly in girls, who are nearly twice as likely as boys to experience significant depressive symptoms following the pubertal transition (Salk et al., 2017). Such experiences with depression are associated with a range of deleterious impacts on social and academic functioning, as well as heightened risk of suicidality over time (e.g., Lewinsohn et al., 2003). Moreover, adolescent depression is related to family risk processes including low levels of parental warmth and support, high levels of parental criticism and negative affect, as well as family conflict and stress (see Yap et al., 2014). Depression is also associated with substance use and conduct problems in adolescence (Spirito & Esposito-Smythers, 2006), and such co-occurrence may predict particularly serious depression-linked outcomes, including higher rates of suicidal ideation and attempts in adolescence (Vander Stoep et al., 2011).

Evidence Based Interventions for child and adolescent depression

The array of risk factors associated with depression provides potentially important targets for prevention and intervention programs. However, there are several limitations with existing evidence-based interventions (EBIs) for depression in children and adolescents. In particular, recent meta-analyses of both prevention and intervention studies for youth depression report small to moderate effect sizes (although prevention trials targeting high-risk samples produce larger effects versus universal prevention; Eckshtain et al., 2020; Werner-Seidler et al., 2017). Moreover, substantial relapse rates are often observed (e.g., 30% within 2 years; Birmaher et al., 2000), and few studies examine longer-term durability beyond the first year (Beardslee et al., 2013). Further, despite evidence of family-risk mechanisms for depression development in youth, and evidence that problematic family functioning predicts worse outcomes in child-focused interventions (e.g., Feeny et al., 2009), parents remain poorly integrated into EBIs for depression (Yap et al., 2016), although there are notable exceptions, such as Parent-Child Interaction Therapy-Emotion Development in early childhood (Luby et al., 2018) and Attachment Based Family Therapy in adolescence (Diamond et al., 2010).

In contrast to the EBIs for child and adolescent depression, family-focused prevention programs are well-established for conduct problems and substance use in childhood and adolescence (McMahon & Pasalich, 2020). Given high rates of co-occurrence and overlapping family risk factors (e.g. high levels of family stress and conflict, low parental warmth and support, and hostile and coercive parent-child interactions; Dishion and Patterson 2006), and developmental models highlighting early behavior problems as a potent risk for depression development (e.g the Dual-Failure model; Patterson & Stoolmiller, 1991), such programs may have substantial implications for preventing depression. Indeed, the identification of programs with broader effects beyond those directly targeted (i.e., “cross-over effects”), may highlight novel prevention pathways, and foster the further development of prevention programs to enhance such cross-over effects. It is worth noting that several family-focused prevention programs targeting conduct problems have demonstrated cross-over effects on internalizing symptoms across adolescence (e.g., Trudeau et al., 2016; Perrino et al., 2016).

The Family Check-Up Model

The current study examines cross-over effects on depression from the Family Check-Up (FCU) prevention program across three trials, two of which were initiated in early adolescence (age 11–12), and one of which was initiated in early childhood (age 2). The FCU is a family-centered model developed to reduce risk of conduct problems and substance use during periods of developmental transitions (i.e., toddlerhood, early adolescence) by improving parenting and family relationships (see Dishion, Stormshak, & Kavanagh, 2012). The FCU follows an adaptive framework, in which intervention strategies are tailored to the needs of individual families in an effort to enhance family engagement and investment (Collins, Murphy, & Bierman, 2004). At the core of the intervention model is the FCU assessment, a brief intervention based on motivational interviewing techniques, designed to enhance family engagement and trigger the behavior change process (Miller & Rollnick, 2012). The FCU assessment involves 3 sessions, including an initial interview, where the family coach explores parent concerns and stage of change readiness, and motivates involvement in the assessment, and a second session involving a formal assessment that includes observations of parent-child interaction. In the third feedback session, the family coach summarizes assessment results using motivational interviewing strategies and explores follow-up treatment sessions that support family management practices. Subsequent sessions involve family-based intervention tailored to the individual goals of the family based on results of the FCU assessment, including positive parenting practices (e.g., positive reinforcement, effective limit setting, and family-problem-solving), relational concerns (e.g., co-parenting), and contextual issues (e.g., school-communication, housing). In analyses from individual trials, the FCU has produced reductions in conduct problems and substance use (the originally targeted outcomes), decreased deviant peer affiliation, improvements in academic functioning and sexual-risk behavior, and increases in parent monitoring and communication in adolescence (e.g., Dishion, Nelson & Kavanagh, 2003; Stormshak, et al., 2011). Similarly, the early childhood version of the FCU has resulted in improvements in a range of child and parent(ing) outcomes, including emotional and behavior problems from early childhood through adolescence (e.g., Dishion et al., 2014), parenting, and maternal depression (Shaw et al., 2009).

Although the FCU was not developed to reduce depression, many family-risk mechanisms for conduct problems targeted by the FCU are also associated with depression, and so it is reasonable to predict cross-over effects of the FCU on depression. Indeed, we have also observed such effects on youth depression and suicidality in both adolescent and early-childhood FCU trials (e.g. Connell et al., 2008, 2016, 2018, 2019; Reuben et al., 2015; Shaw et al., 2009), particularly among youth with co-occurring emotional and behavior problems (e.g., Connell et al., 2008; Connell et al., 2018). However, individual trial analyses have been hampered by the limited measurement of depression, and changes in measures of depression that were employed over time within trials. For instance, in the first large-scale trial of the FCU program in early adolescence (the Project Alliance 1 study), depression measures were initially only administered to a subset of youth identified by teachers as being at high-risk for conduct problems, limiting statistical power (see Connell & Dishion, 2008). Similarly, changes in measurement approaches in the individual trials have limited the scope of longitudinal analyses. For instance, in the Early Steps Multisite study, initiated when children were 2 years old, youth reports of depression were not collected until early adolescence. In the Project Alliance 2 study, different measures were used in middle-school versus emerging adulthood assessments, limiting longitudinal analyses.

Current study

Such measurement challenges may be addressed by modern Integrative Data Analysis (IDA) techniques. The goal of the current work was to employ IDA techniques to harmonize depression data across three trials of the FCU, to enhance our ability to examine prevention effects on depression trajectories. We employed Moderated Nonlinear Factor Analysis (MNLFA; Bauer & Hussong, 2009) to facilitate item-level analyses across trials, despite the use of different measures and items across studies. MNLFA provides a robust means to adjust for differential item functioning across subgroups of participants (e.g., racial/ethnic groups), while creating scale scores based on all available items. These scale-score estimates are then used in hypothesistesting analyses. In the current analyses, we sought to examine the long-term effects of the FCU on youth depression across three randomized trials. These trials cover a broad developmental span with one trial beginning at age 2 and the other two beginning at age 11, and including follow-up assessments up to 14 years post-baseline. Given the advantages of MNLFA, the current analyses are able to estimate FCU effects on depression symptom trajectories across a longer time-span than would be possible in analyses of any individual trial, given the measurement limitations previously described. Therefore, the current analyses provide the most comprehensive, “birds-eye view” of FCU effects on depression to date.

We also examined possible moderation of FCU effects on depression by child race/ethnicity and gender. In the individual trials, very little evidence of differences in intervention outcomes have been observed across racial/ethnic groups, which may be attributed to the emphasis placed on multicultural competence in the training and supervision of interventionists in FCU trials (for discussion, see Smith, et al., 2014). More broadly, meta-analyses of depression-focused interventions have found that effect sizes are not significantly related to the proportion of minority youth in samples (Eckshtain et al., 2020), and so we did not have strong reason to expect to observe evidence of race/ethnicity moderating the effects of the FCU on youth depression. We examined gender moderation in light of significant disparities in levels of depression emerging in early adolescence (Salk et al., 2017). As for race/ethnicity, little evidence of gender differences in FCU effects have been observed on a range of outcomes in individual trials (e.g. Smith et al., 2014; Van Ryzin et al., 2012). Further, meta-analyses of the child and adolescent depression treatment literature suggest that effect sizes for males and females do not differ significantly (e.g. Eckshtain et al., 2020). Therefore, we did not have strong reason to expect gender to moderate FCU effects on youth depression in the current analyses.

Methods

Samples

Early Steps.

This prevention trial includes 731 mother–child dyads, originally recruited between 2002 and 2003 from Pittsburgh, Pennsylvania; Eugene, Oregon; and Charlottesville, Virginia (for complete details see Dishion et al., 2008). In the sample, 49.5% of youth were female, with diverse racial/ethnic representation (54.4% White, 28.2% African American, 10.7% Latinx, and 6.7% multiracial or other). Families with children between the ages of 2 years 0 months and 2 years 11 months were recruited from Women, Infants, and Children Nutritional Supplement (WIC) centers following a screen to ensure that they met the study criteria. High-risk status on at least two of the following three risk domains was required for study inclusion: (a) child behavior (conduct problems, high-conflict relationships with adults), (b) family problems (maternal depression, daily parenting challenges, substance use problems, teen parent status), and (c) socio-demographic risk (low educational achievement and low family income using WIC criteria). Families completed assessments at 10 study waves, from ages 2 through age 16 years. Retention was above 80% for most assessments, including age 16, with 73% retention at the lowest point (age 7.5). Families were individually randomized to intervention (50.2%) or control conditions (49.8%) at baseline. Control conditions for all three trials are described in the “Trial Details” section. Intervention families were offered the FCU and follow-up services as warranted, nearly annually on eight occasions from age 2 to 10.5 years. Of the 367 families randomized to the intervention condition, 343 (93.5%) took part in the FCU (an initial interview, assessment and feedback sequence) at least once between age 2 and 10.5, with such participation predicted by elevated parental depression, female child gender, being above the poverty line and parental age at childbirth (see Dishion et al., 2014). The percentage of families receiving the FCU at each wave from 66% (at age 5) to 76% (age 2). Most families receiving the FCU completed follow-up sessions focused on parenting, child development, and behavior management. The percentage of FCU-completing families opting for further sessions ranging from 65% (age 7.5) to 74% (age 4). The average number of sessions at each wave ranged from 2.3 (age 10.5) to 3.5 (age 5), with an average of 1 hour and 20 minutes (SD = 51) per session.

Project Alliance 1 (PAL1).

This trial includes 998 adolescents and their families, recruited from three middle schools within a predominantly lower-income metropolitan community in the northwestern United States, when the children were in sixth grade (for more complete details see Dishion et al., 2002). In the sample, 50.6%% of youth were female, with diverse racial/ethnic representation (49.2% White, 29.2% African American, 6.8% Latinx, and 14.8% multiracial or other). Parents of all sixth-grade students in two cohorts were approached for participation, and 90% consented to participate. Biological fathers were present in 585 families (58.6%). Youth were randomly assigned at the individual level to either control (498 youth) or intervention classrooms (500 youth) in the spring of sixth grade. Families completed assessments at 9 study waves, from age 11 years to age 28 – 30 years (although age 28 – 30 data were not included in longitudinal analyses, as no other study provided such long-term data and only youth-report data were collected at this final assessment). Retention was above 80% for most assessment points, with 75.6% retention at the final assessment point. In the intervention condition, 115 families (23%) received the FCU in grades 7 – 9, and 88 families (17.6%) received further intervention services after the FCU. FCUs also were offered in high school (in Grades 10–11), and 170 families (34%) received the FCU, 109 of whom had not received it during middle school. Therefore, 224 families (45%) received the FCU across the study, with an average of 8.9 hours of direct contact with intervention staff over the study (SD = 9.42 hours). Family receipt of the FCU was predicted by biological father absence from the home, youth reports of family conflict and peer deviance, and teacher reports of behavior problems in 6^th grade (Connell et al., 2007).

Project Alliance 2 (PAL2).

This prevention trial includes 593 families recruited when their child was in sixth grade, from three predominantly lower-income urban middle schools in the northwestern United States (for more complete details, see Stormshak et al., 2011). In the sample, 48.5% of youth were female, with diverse racial/ethnic representation (36% White, 15.2% African American, 18% Latinx, and 30.8% multiracial or other). Parents of all sixth grade students in two cohorts were approached for participation, and 76% consented to participate. Families were assigned to control (n=207; 35%) or intervention conditions (n = 386; 65%), using an unbalanced randomization approach to increase the power to detect heterogeneous patterns of intervention effects. Families completed comprehensive assessments at 7 study waves, from age 11 years to age 23 years. Retention was above 80% for most assessment points, with 78% of participants completing at least one of the early-adult assessments (Stormshak et al., 2018). Within the intervention condition, 42% (n = 163) of families received the FCU between grades 7–9, 80% of whom received follow-up intervention services. Average duration of intervention services was 337 minutes (approximately 6 hours). The FCU was offered to families again in early adulthood (age 20), at which 34.7% (n = 134) of those in the intervention condition families received the FCU, with an average of 1.92 hours of intervention time.

Trial Details

All trials focused on the FCU, which follows an adaptive intervention framework integrating motivational interviewing principles, with doses and intervention targets adapted to family needs, on the basis of the 3-session FCU-Assessment. The format of the FCU assessment was common across the three trials, and involves a three-session sequence in which multimethod data is collected regarding parent (e.g. depression, stress, positive/supportive and harsh/coercive parenting), youth (e.g. emotional and behavior problems, peer relationships, academic functioning), and family functioning (e.g. family conflict, family communication and problem-solving, interparental relationship quality, parent-child relationship quality). Assessment results are reviewed with family members during the feedback session, including a discussion of family strengths and areas of concern. The feedback sessions lead to a collaborative discussion of the assessment results, including identification of follow-up services consistent with the family goals. These follow-up services were based upon the Everyday Parenting Curriculum (EPC; Dishion et al., 2012), adapted to be age appropriate.

Early Steps.

Follow-up intervention services were adapted from the EPC to be appropriate for early childhood, including topics on parenting practices, family management concerns (e.g., co-parenting), and contextual issues (e.g., child care, partner relationship, housing, self-care).

PAL1 and PAL2.

PAL1 and PAL2 were both school-based prevention trials and employed a multilevel prevention framework. The universal intervention, designed to support positive parenting practices and engage parents of high-risk youth for the selected intervention, included the establishment of a Family Resource Center (FRC) in each middle school, offering brief parent consultations, feedback to parents on their student’s behavior at school, and access to videotapes and books. The selected intervention was the 3-session FCU assessment. The FCU feedback leads to a collaborative decision with parents on indicated services appropriate for their family, including a parent group intervention and individually based behavior family therapy, grounded in the EPC (Dishion et al., 2012), and focused on building positive parenting skills (e.g., positive reinforcement, limit-setting, problem-solving, and communication skills).

Control Conditions.

Comparable control conditions were used across trials. In the school-based PAL1 and PAL2 trials, the control condition was “school as usual,” in which control-condition families were not provided access to FCU prevention components but were free to access other school or community resources as normal. In the Early Steps study, control-condition families were not provided access to services associated with the FCU program but were allowed to access all normal WIC program or other community services as normal.

Measures.

Table 1 lists the measures administered in each sample at each assessment point.

Table 1.

Sample sizes and assessments by study at each assessment wave.

		Years Post-Baseline
Study	Reporter	0	1	2	3	5	7	9	10	11	12	14
PAL1	Parent:	140	144	81			648
	Youth:	389	407	239		796	808			818	856
	Measures:	2, 3, 4, 5, 6	2, 3, 4, 5, 6	2, 3, 4, 5, 6		1, 2, 3	3, 5, 6			1, 3	1, 3
	Youth Age:	11	12	13	14	16	18	20	21	22	23	25
PAL2	Parent:							323	273	222
	Youth:	585	524	508	482			415	388	360
	Measures:	4, 7	4, 7	4, 7	4, 7			5, 6	5, 6	5, 6
	Youth Age:	11	12	13	14	16	18	20	21	22	23	25
ES	Parent:	730	657	627	614	567	587	567			570	583
	Youth:							450			546	569
	Measures:	5	5	5	5	5	5	5			4, 5	4, 5
	Youth Age:	2	3	4	5	7.5	9.5	10.5	12	13	14	16

Notes: Abbreviations for studies are: Project Alliance 1 (PAL1); Project Alliance 2 (PAL2); Early Steps (ES).

Measures are identified as follows: (1) Brief Symptom Inventory; (2) Self-report Health measure; (3) Life Events Coping Inventory; (4) Child Depression Inventory; (5) CBCL-Parent; (6) CBCL-Youth; (7) Depression Symptom Checklist.

Early Steps:

Parents completed the 99-item Child Behavior Checklist for Ages 1.5–5 (CBCL; Achenbach, 1991), at ages 2, 3, 4, and 5, and the CBCL for Ages 6–18 (Achenbach, 1991) during the age 7.5, 8.5, 9.5, 10.5, 14, and 16-year assessments. The CBCL includes depression-symptoms rated on a 3-point scale (0 = not true, 1 = somewhat/sometimes true, 2 = very true/often true), with a 6-month time-frame. Youth completed the Child Depression Inventory (CDI; Kovacs, 1992), a 27-item self-report measure of symptoms over the past 2 weeks.

PAL1.

A stratified assessment approach was used at the age 11, 12, and 13 assessments, with depression data only collected for an elevated-risk subset of youth, based upon teacher-reports of risk for conduct problems (see Connell & Dishion, 2008). At ages 11, 12, 13, and 18, parents and youth completed the CBCL-ages 6–18 (Achenbach, 1991). Youth also completed the CDI (Kovacs, 1992) at ages 11, 12, and 3, and the Brief Symptom Inventory (BSI; Derogatis & Spencer, 1982) at ages 16, 22, and 23, which assesses past-week depression symptom severity. Youth completed a self-report health measure at age 11, 12, 13, and 16 years, that included an item regarding past-year occurrence of suicide attempts. Youth completed the Life Events Coping Inventory (LECI; Dise-Lewis, 1988) at ages 11, 12, 13, 16, 18, 22, and 23, which includes two items reflecting suicidal ideation and self-harm in response to stress.

PAL2.

Youth completed a depression symptom checklist at ages 11 – 14, that includes 14 items reflecting past-month depressive symptom severity (Klostermann, Stormshak, & Connell, 2016). Youth also completed the CDI (Kovacs, 1992) at ages 11, 12, and 13, as part of the FCU assessment. At ages 20, 21, and 22, parents and youth completed the CBCL (Achenbach, 1991).

Analytic plan

Overview of Data Harmonization.

Data harmonization analyses employed moderated non-linear factor analysis (MNLFA; Hussong, Curran, & Bauer, 2013). Compared to traditional meta-analysis, which aggregates effect size indices from previously conducted analyses in primary studies, MNLFA facilitates item-level analyses across datasets when items from different measures have been used across trials. MNLFA permits the estimation of a final latent-variable reflecting the construct of interest, adjusting for Differential Item Functioning. Parent-and youth-reports of depression were analyzed separately, using the R-based aMNLFA package (Gottfredson et al., 2019) to draw calibration samples and iteratively format Mplus command files. MNLFA analyses were conducted in Mplus 8.4 (Muthen & Muthen, 2020).

Following Gottfredson and colleagues (2019), a single time-point of data for each participant was randomly selected to generate a calibration sample. An iterative series of analyses were conducted with the calibration sample to examine invariance across covariates for factor means, variances, and item intercepts/factor loadings to obtain valid item parameter estimates adjusting for DIF. Ignoring DIF during scoring may lead to inferential errors because these group differences in measurement properties may appear to reflect true group differences rather than artifacts of measurement (see Gottfredson et al., 2019). The following Covariates were included: youth gender (0 = male, 1 = female), ethnic minority status (0 = European American, 1 = racial/ethnic minority), age at assessment, intervention-group (0 = Control group, 1 = Intervention), and study (with two orthogonal contrasts, comparing PAL1 with PAL2, and PAL1 with Early steps). First, in a series of analyses, we examined potential covariate differences in overall mean and variance in depression severity, in item discrimination (i.e., factor loadings), and in item difficulty (i.e., item intercepts). Second, results from these analyses were then included in a full model, simultaneously testing invariance across these model parameters for all covariates with significant effects in the first-step analyses. Third, to protect against type-I errors, a Benjamini-Hochberg (1995) family-wise error correction was applied to results from the second-stage analysis, to generate a final model including only significant effects that survived correction. Finally, model parameters were fixed to the estimates from this final model, to generate a scoring model. This scoring model was then used to generate depression-estimates using the full longitudinal data set, which were used in subsequent analyses.

Longitudinal models.

Estimates of depression generated from aMNLFA were analyzed using a Multi-informant Autoregressive Latent Trajectory Model (ALT; Bollen & Curran, 2004). Latent depression variables were estimated at each study wave, based on youth- and parent-reports of depression. Time-specific item intercepts and factor loadings were held invariant over time, and latent intercept and linear slope parameters were estimated from multi-informant depression estimates. Time was modeled as “Years Post-Baseline,” with a time-score of “0” representing the pre-treatment baseline. The baseline assessment was not included in the estimation of the latent intercept or slope parameters but was allowed to predict subsequent depression levels (Bollen & Curran, 2004). The autoregressive paths captured time-specific variability around the underlying latent trajectory. Once acceptable model-fit was achieved, covariates (gender, ethnic minority status, intervention-assignment, and orthogonal contrasts comparing PAL1 with PAL2, and with Early Steps) were added to the model. Intervention assignment was not allowed to predict pre-treatment baseline depression, but was allowed to predict latent intercept (reflecting depression at 1-year post-baseline) and slope parameters. All other covariates were allowed to predict both the latent intercept and slope parameters, as well as baseline depression. Follow-up tests of gender and race/ethnicity moderation employed multiplicative interaction terms between intervention assignment and gender or ethnicity. Analyses employed Full Information Maximum Likelihood (FIML; Rubin & Little, 2002) estimation to accommodate missing data. Monte-carlo simulation analyses indicated this modeling approach has 80% power to detect at least small effect sizes (Cohen’s d > .20). Acceptable model-fit is indicated by non-significant Chi-square value, CFI and TLI values above .90, and RMSEA and SRMR values less than .08, although SRMR values < .10 are acceptable for large sample sizes (N > 1000) when RMSEA < .06 (Hu & Bentler, 1999).

Results

Descriptive Statistics.

The aggregated sample included data from 2322 families. The number of parents and youth providing depression-related data at each study wave is shown in Table 2. The sample included 49.7% female youth, and was racially/ethnically diverse (47.5% White, 25.4% African American, 10.9% Latinx, and 16.3% multiracial/other).

Data Harmonization.

We conducted separate MNLFA analyses for youth-reports and parent-reports of depression. Both analyses followed the same data-map. Following a comprehensive examination of available items and guided by diagnostic criteria as well as the example of Curran and colleagues (2014), we included items from all available measures that assessed 17 depression symptoms, including: 1) sad/blue, 2) irritable/moody, 3) anhedonia, 4) appetite disturbance, 5) sleep disturbance 6) restless/agitated, 7) low energy, 8) overtired, 9) feeling worthless/inferior, 10) feelings of guilt, 11) concentration problems, 12) suicidal ideation, 13) self-harm/suicide attempt, 14) thoughts of death/dying, 15) hopelessness, 16) loneliness, 17) cries a lot. A complete listing of items by measure mapped onto these symptom domains is show in Supplemental Table 1. MNLFA analyses included all individual items across measures. Youth reported on all symptoms across samples (for a total of 53 items across measures), while parental report was not collected for two symptoms -- thoughts of death/dying, and hopelessness (for a total of 15 items). Item responses were dichotomized to represent the endorsement of a given symptom across measures (0 = no, 1 = yes), an approach also used by Curran and colleagues (2014) to facilitate model convergence in the face of sparse responses to the upper range of response scales for some items and measures. Results for the final step in testing for DIF in MNLFA analyses are shown in Supplemental Table 2 for parent-reports of youth symptoms, and in Supplemental Table 3 for youth-reports. Youth- and parent-reported depression estimates were correlated within each time-point (r = .25 to r = .40 across waves, p < .05).

Multi-informant ALT model.

We first conducted a series of analyses to generate an acceptable-fitting unconditional model. The final model at this stage allowed several correlated residuals among youth-reports and among parent-reports across ages, reflecting informant effects over time. To avoid issues with negative residual variances, residual variances were fixed to zero for the linear slope and time-specific depression estimates at the following time-points: 1, 2, 3, 5, and 11 years post-baseline. The final unconditional model provided acceptable fit to the data by most criteria (χ2 = 271.96, df = 88, p < .05, CFI = .96, TLI = .94, RMSEA = .03, SRMR = .08). Next, covariates were added to the model. Model fit remained acceptable (χ2 = 530.31, df = 184, p < .05, CFI = .93, TLI = .91, RMSEA = .03, SRMR = .09). The latent intercept was scaled to zero, while the linear slope was significant and negative (estimate = −.004, SE = .002, p = .01). The autoregressive parameters were positive, ranging from .67 to 1.08 (all p’s < .001).

Random assignment to intervention was significantly related to both the intercept (estimate = −.014, SE = .005, p = .004) and slope parameters (estimate = .002, SE = .001, p = .01). These intervention effects represent medium effects (Cohen’s d = −.77 and Cohen’s d = .54, for intercept and slope effects). The intervention effect, depicted in Figure 1, shows that random assignment to the FCU condition predicted stronger declines in depression over time. However, these intervention effects also waned over time, such that intervention and control conditions showed similar levels of depression by the last follow-up, 14 years after the baseline assessment.

Figure 1.

Intervention effect on estimated depression trajectories.

Additionally, female gender was significantly associated with higher baseline depression (estimate = .04, SE = .02, p = .01), a less negative linear slope (estimate = .002, SE = .001, p =.02) and more positive time-specific youth-reports of depression (estimates ranged from .18 to .56 over time, all p’s < .001). Racial/ethnic minority status was not significantly related to baseline depression, or the intercept and slope parameters. Relative to PAL1, Early Steps youth exhibited more positive baseline values of depression (estimate = .15, SE = .02, p = .001), a less positive latent intercept (estimate = −.02, SE = .01, p = .01), and a less negative slope (estimate =.003, SE = .001, p = .01). No significant differences between PAL1 and PAL2 were observed.

Post-hoc tests.

To probe the time-course of intervention effects, post-hoc tests were conducted by saving the time-specific latent variable estimates, and conducting a series of follow-up regression analyses predicting latent depression scores at each time-point, controlling for baseline scores as well as all covariates in the original model (gender, ethnicity, and the two study-level contrasts). The intervention effect was significant at all follow-up waves, from 1 to 11 years post-baseline. However, at 12- and 14-years post-baseline, the intervention effect was no longer statistically significant (12 years post-baseline, intervention effect = −.01, SE = .006, n.s.; 14 years post-baseline, intervention effect = .004, SE = .006, n.s.).

Moderation analyses.

Moderation by youth gender and race/ethnicity was examined in separate follow-up analyses. However, there were no significant interaction effects for either gender or ethnicity in relation to linear or quadratic slope parameters. Finally, in response to a reviewer, we examined differences in intervention effects across Early Steps and the two adolescent samples, given differences in ages and school versus community settings. Analyses adapted the multi-informant ALT model but included a single contrast comparing Early Steps and the combination of PAL1/PAL2 samples, along with a treatment by study-contrast interaction. The interaction was allowed to predict the intercept and slope parameters, although neither effect was significant (intercept effect = −.005, SE = .006; slope effect = .001, SE = .001).

Discussion

The current study examined the long-term effects of the FCU prevention program on depression in childhood and adolescence across three randomized trials. Using IDA analyses, these results represent a comprehensive test of cross-over effects of the FCU prevention approach on the development of depression, spanning a period of 14 years across childhood and adolescence. Although we have documented FCU effects on depression in each of the individual trials (Connell et al., 2008, 2018, 2019), changes in measures over time and idiosyncratic features of the assessment approach (e.g. administering depression measures only to a high-risk subset of youth in the initial years of the PAL1 trial), have limited our examination of long-term depression trajectories. By employing a rigorous data harmonization approach, we were able to conduct a powerful test of long-term effects of the FCU on depression across time and trials.

Using this harmonized data, we observed significant cross-over effects of the FCU on depression trajectories. Youth randomly assigned to the FCU condition exhibited significantly steeper declines in depression over time. These effects were small to moderate in magnitude and persisted over a decade following initiation of the FCU program. These effects were attained with a relatively brief parenting program that emphasized parental motivation to change while supporting positive parenting strategies designed to reduce behavior problems and improve relational functioning. However, the effects of the FCU on depression appeared to wane over time, as indexed by the small but significant positive effect on the linear slope. Indeed, post-hoc analyses showed that by 12 to 14 years following the initiation of the prevention program, youth in the FCU versus control condition were largely indistinguishable with respect to depressive symptoms. It is worth noting that, in each of the three trials, families were offered the FCU on multiple occasions over time. In Early Steps, families were offered the FCU through age 10.5 years (approximately 8 years post-baseline). In PAL1, families were offered the FCU through high school (grades 10 – 11, 5 to 6 years post-baseline). In PAL2, families were offered the FCU throughout middle-school, with a subsequent administration in early adulthood, 9 years post-baseline. It may be that offering FCU “booster-sessions” over time is needed to sustain effects on depression over longer time-spans. However, based on the relatively brief but targeted nature of the FCU intervention, and the serious long-term consequences associated with youth depression, such boosters are likely to represent a cost-effective approach to prevention.

Of note, we did not find evidence that youth race/ethnicity or gender moderated the effects of the FCU on trajectories of youth depression. The generalizability of prevention effects on depression across subgroups of youth are consistent with analyses in the primary trials that have found consistent FCU effects across racial/ethnic and gender groups on a range of outcomes, including conduct problems, substance use, and aspects of family functioning (e.g. Smith et al., 2014; Van Ryzin et al., 2012). These results are also consistent with the broader depression treatment and prevention literatures, which have generally found comparable benefits of prevention and intervention programs across subgroups (e.g. Eckshtain et al., 2020).

More broadly, the current results add to the growing literature demonstrating that parent/family focused prevention programming may yield long-term benefits for the reduction of depression in youth (Trudeau et al, 2016; Perrino et al. 2016). Although there has been a recent growth in recognition of the importance of incorporating parent/family components into depression-focused prevention programming, such incorporation remains understudied. This study also adds to the emerging literature demonstrating long-term cross-over benefits of prevention programs designed to reduce conduct problems and substance use on other areas of youth functioning (e.g. Reider & Sims, 2016). The benefits of the FCU on depression may be due to the relatively broad developmental influence of parenting and family relational functioning across childhood and adolescence. As such, prevention programs emphasizing the cultivation of positive, effective parenting strategies, fostering supportive parent-child relationships, and the establishment of healthy family routines are likely to have broad influences across a range of outcomes, including depression, that were not directly targeted. As EBIs for depression in childhood have focused primarily on working directly with youth, to the relative neglect of meaningfully incorporating parents, the current results suggest that parent- or family-focused efforts represent a promising avenue for further prevention development.

Summary and Future directions.

Two strengths of the current work are worth emphasizing. First, our examination of long-term effects across trials was facilitated by the use of MNLFA, which provides a powerful tool to synthesize data across trials despite measurement differences, while accounting for differential item functioning across trials and covariates. In turn, the use of synthesized data provides the opportunity to test prevention effects beyond those that are possible in the individual trials. Second, the multi-informant growth model represents an additional strength, as the time-specific latent variables aggregate across youth and parent reports of depression across a broad time-frame and developmental range, mitigating bias associated with mono-informant methods. There are also important practical benefits to the multi-informant growth model, as it allowed for the estimation of individual trajectories of depressive symptoms, while accommodating missing data in either child or youth reports at individual assessment points. For instance, in the PAL2 study, parental reports of youth depression were not collected in the first four waves of the study, (only youth-reports were collected in these waves). Building a multi-informant model facilitated the estimation of cross-reporter latent trajectories despite such planned missingness.

As with any study, however, there are limitations with implications for future work. First, by harmonizing data across a range of measures, IDA involves a tradeoff between specificity (e.g., detailed examination of treatment effects by measurement approach), and generalizability. Specifically, the current analyses incorporated data across a range of measures, and across parent- and youth-report. Future work examining possible measurement or informant effects may be warranted. Second, all studies followed a randomized encouragement trial design, with families assigned to the intervention condition allowed to select how much to engage in intervention services. It is possible that greater engagement with services may be associated with stronger depression-related reductions. Although the current analyses were not appropriate to address this possibility, future work employing advanced analytic strategies for this question (e.g. Complier Average Causal Effect analysis) may be warranted. Third, there was substantial variability in in recruitment-setting (school-based vs. WIC-recruited), and in youth ages at baseline across trials. Age effects were controlled for in harmonization analyses, and study membership was included as a covariate in all analyses, such that the main effects of intervention reflect the benefits of the FCU approach across trials, adjusting for such differences. Moreover, in a follow-up analysis, intervention effects on growth parameters did not differ across the two trials initiated in early adolescence, compared to the trial initiated in early childhood. Our results highlight that the parent/family focus of the FCU, and the emphasis on parental motivation to change present common benefits for depression in youth across ages and study settings.

The present results suggest several avenues for future work. First, future work should examine potential mechanisms of the FCU’s cross-over effects on depression. Given the focus of the FCU, we suspect these effects may be driven by changes in parental use of positive, proactive parenting strategies, reductions in family conflict, and improvements in parental depression, as has been documented in individual FCU trials for reductions in behavior problems (e.g. Fosco et al., 2016; Shaw et al., 2009). Similarly, future work examining a broader range of moderators is needed and may identify factors associated with diminished treatment response that may drive future modifications of the FCU approach. Although limited evidence of moderation of FCU effects on depression has been observed in individual trials, the enhanced longitudinal scope afforded by data synthesis may yield novel insights. We are currently harmonizing data on child-level (e.g., behavior problems), and family-level factors (e.g., parental depression) that may serve as mediators or moderators, to facilitate future work on these issues. Finally, future work adapting the FCU to more specifically target depression development may be warranted. Although we observed long-term benefits of the FCU for depression, the program was not designed to target family (e.g., emotional overinvolvement and psychological control; Yap et al., 2014) or child factors (e.g., depressogenic cognitions; Hankin, 2006) specific to depression. Integrating such factors into the FCU program may further enhance effects on depression.

Public Health Significance:

This study synthesizes data across three randomized trials, showing that the FCU prevention program leads to significant reductions in youth depression that are observed over approximately 10 years. This family focused program, designed to prevent conduct problems and substance use, appears to offer long-term benefits for depression, as well.