Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Jun 15.
Published in final edited form as: J Affect Disord. 2019 Apr 25;253:176–183. doi: 10.1016/j.jad.2019.04.089

Trajectories of Change in Maternal and Adolescent Depressive Symptoms in the Depression Prevention Initiative

Carolyn Spiro-Levitt 1,1, Robert Gallop 2, Jami F Young 3
PMCID: PMC6620130  NIHMSID: NIHMS1528178  PMID: 31051322

Abstract

Background

Given the prevalence and consequences of adolescent depression, depression prevention has become an important area of research. While prevention programs like Interpersonal Psychotherapy – Adolescent Skills Training (IPT-AST) have demonstrated effectiveness, little research to date has studied the relationship between maternal depression and adolescent outcomes in these programs.

Method

The current study investigated the relationship between maternal and adolescent depressive symptoms in 167 mother-adolescent dyads who were enrolled in the Depression Prevention Initiative (DPI), a randomized controlled trial that compared IPT-AST to group counseling (GC). First, the study examined the relationship between initial levels of adolescent and maternal depressive symptoms. The study then investigated whether maternal depressive symptoms improved over the two-year study period. Finally, the study assessed whether maternal and adolescent symptoms changed concurrently across time.

Results

Results indicated that initial levels of maternal and adolescent symptoms were positively associated. Additionally, maternal symptoms improved across the two-year period. Maternal and adolescent outcomes were related across time: as adolescents improved in our study, their mothers also improved.

Limitations

The study utilized self-report data only and did not allow for the testing of causality in the relationship between mother-youth depression.

Conclusions

These findings add to the literature demonstrating that as one part of the mother-child dyad improves, the other improves as well. These findings extend the current understanding of the relationship between maternal and adolescent depressive symptom outcomes, and have important implications for the prevention and treatment of depression.

Keywords: Prevention, Maternal Depression, Adolescent Depression, IPT-AST

Introduction

Prevention of Adolescent Depression.

Depression in adolescence is a common and debilitating disorder. By the end of adolescence, nearly one in five individuals experience a depressive episode (Thapar, et al., 2012). Given the high rates of depression in adolescence and its significant public health implications, the prevention of depression has increasingly become a focus of clinical research. Depression prevention programs have demonstrated small to moderate effects on depressive symptoms (Cuijpers, van Straten, Smit, Mihalopoulous, & Beekman, 2008; Horowitz & Garber, 2006; Merry et al., 2012; Stice et al., 2009). Given the importance of depression prevention, recent research has begun to investigate factors that may contribute to or impact adolescent outcomes. The current paper explores one such factor, maternal depressive symptoms, and explores the relationship between adolescent and maternal depressive symptoms across two depression prevention programs.

The magnitude of preventive intervention effects depends on the risk status of the targeted population. The Institute of Medicine (IOM) classifies three categories of preventive interventions: universal prevention programs include all individuals of a given population, selective prevention programs are administered to individuals who are classified as “above average risk” due to a known risk factor, and indicated prevention programs include individuals with subclinical symptoms of the targeted disorder (Horowitz et al., 2007). While Merry and colleagues (2012) found evidence that all levels of prevention are likely to be effective in reducing depressive symptoms when compared to no intervention, effects are greatest in selective and indicated programs (Horowitz & Garber, 2006; Stice et al., 2009).

One promising indicated prevention program is Interpersonal Psychotherapy - Adolescent Skills Training (IPT-AST) (Young & Mufson, 2003). IPT-AST focuses on the interpersonal context in which depression occurs. IPT-AST has demonstrated promising results in four efficacy studies (Horowitz, Garber, Ciesla, Young, & Mufson, 2007; Young, Mufson, & Davies, 2006; Young, Mufson, & Gallop, 2010; Young et al., 2016). The largest study of IPT-AST to date was the Depression Prevention Initiative (DPI) (Young et al., 2016), which compared the effects of IPT-AST to group counseling (GC) in schools through two-years of follow-up. The two interventions will be discussed in greater detail below, as the current study utilizes data from this large randomized controlled trial (RCT). From baseline to 6-month follow-up, adolescents in both interventions experienced improvements in depressive symptoms. However, youth in IPT-AST experienced significantly greater improvements in depressive symptoms than youth in GC (Young et al., 2016). Between 6-month and 24-month follow-ups, GC youth continued to show a decrease in depressive symptoms, while IPT-AST youth showed a non-significant increase in depressive symptoms. Across the entire study period, youth in both interventions experienced significant reductions in depressive symptoms, with no significant differences in overall rates of change between the two conditions (Young et al., 2018). The current paper builds on this larger study to investigate the relationship between adolescent outcomes and their mothers’ outcomes. While mothers were not directly targeted in our study, we believe that investigating this link is an important area for clinical research, as it can better inform future prevention efforts. The rationale for investigating maternal depression outcomes is outlined in more detail below.

Covariation of Adolescent and Maternal Depression Outcomes

The increased risk of depression in children of depressed mothers is well-documented (i.e., Weissman et al., 2016). A meta-analysis completed by Goodman and colleagues (2011) demonstrated that depression in mothers is associated with a broad range of psychopathology in children, including internalizing problems. Of relevance to the current study, research has begun to examine whether maternal and adolescent depressive symptoms are related and whether they change in tandem across depression intervention studies, within both the child and adult treatment literature. Across studies of evidence-based treatments including cognitive behavioral therapy (CBT), interpersonal psychotherapy (IPT), and pharmacotherapy, results generally show significant associations across mothers’ and their children’s treatment trajectories with few exceptions (i.e., see small pilot study by Verdeli and colleagues (2004)). For example, several studies examining the relationship between maternal and youth depression in the context of treatment for youth depression have found significant associations in mother-youth depressive symptom trajectories and outcomes (i.e., Kennard et al., 2008; Perloe, Esposito-Smythers, Curby, & Renshaw, 2014; Wilkinson, Harris, Kelvin, Dubicka, & Goodyer, 2013). Additionally, studies such as the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study investigated this association in the context of treatment for mothers, and also found associations between maternal-youth depression outcomes (Weissman et al., 2006).

While further research is needed, these studies suggest that adolescent and maternal depressive symptoms may change concurrently while adolescents or mothers participate in depression interventions, at least in the treatment literature. Together, these findings suggest that treatment of either the adolescent or the parent may lead to a positive cycle; as either member of the pair improves, the other does as well, leading to further improvements in each member across time. To date, no depression prevention studies have examined whether maternal depressive symptoms improve as a function of their adolescents participating in an indicated prevention program.

The Current Study

There is a growing body of literature demonstrating the association between maternal and youth depression intervention outcomes, yet no study to date has examined the associations between maternal and adolescent depression outcomes in the depression prevention literature. Understanding the breadth and specificity of these associations could lead to important advances in the prevention and treatment of adolescent depression. In order to fill these gaps in the literature, the current study investigated three aims: (1) Were initial levels of maternal depressive symptoms correlated with initial levels of adolescent depressive symptoms as measured by a self-report depression scale in our prevention programs? We hypothesized that there would be a significant correlation between level of maternal depressive symptoms and adolescent depressive symptoms at baseline. (2) Did maternal depressive symptoms change over the course of the two adolescent depression prevention programs? Based on the treatment literature, we hypothesized that maternal depressive symptoms would improve over the course of the interventions. (3) Did changes in maternal depressive symptoms covary with changes in adolescent depressive symptoms across time in IPT-AST and GC? We hypothesized that there would be an association between the outcome trajectories of adolescents and their mothers; as adolescents improved over time, their mothers would also improve.

Method

Participants

The study utilized data collected from DPI (Young et al., 2016; Young et al., 2018). One hundred eighty-six adolescents who were enrolled in the 7th to 10th grades and had elevated depressive symptoms were randomized to either IPT-AST (N = 95) or enhanced group counseling (GC) (N = 91). The current study examined data from 167 mother-adolescent pairs (GC=76, IPT-AST=91); in the remaining 19 families a father or other parent completed the parent-report measures and therefore were not included in the present analyses. Among these 167 adolescents, 67.7% were female, and the average age was 13.49 years (SD=1.21). Racial minorities represented one third of the adolescent sample, with 21.6% of participants identifying as African-American, 4.8% as Asian-American, 0.6% as American Indian, and 6.6% as other or mixed race. Regarding ethnicity, 37.7% of adolescents identified as Hispanic. When examining the sample of mothers included in the study, racial minorities also represented one third of the group; 19.8% of mothers identified themselves as African American, 4.8% as Asian-American, and 2.4% as other or mixed-race. Regarding ethnicity, 38.3% of mothers identified as Hispanic.

Procedures

Youth who gave signed assent and whose parents gave signed consent participated in the study. Adolescents with elevated symptoms of depression were identified through a two-stage screening procedure. At the initial screening, adolescents completed the Center for Epidemiologic Studies-Depression Scale (CES-D; Radloff, 1977), a scale which measured depressive symptoms over the past week. Adolescents with a CES-D score of 16 or higher were eligible to be approached for the prevention project. As the second stage of the eligibility process, trained evaluators administered the Schedule for Affective Disorders and Schizophrenia for School-Age Children (K-SADS-PL; Kaufman, Birmaher, Brent, & Rao, 1997). The K-SADS-PL has been found to generate reliable and valid psychiatric diagnoses in ages 7 to 17 (Kaufman et al., 1997; Leffler, Riebel, & Hughes, 2015). Youth were eligible if they had at least two current threshold or subthreshold depression symptoms on the K-SADS-PL, at least one of which was a criterion A symptom (depressed mood, irritability, or anhedonia), but did not meet criteria for a depression diagnosis (see Young et al., 2016 for additional details on exclusion criteria and procedures).

Given that this study was conducted in a school setting, parents’ baseline evaluation occurred separately from their adolescents’ baseline evaluation. Parents completed the baseline CES-D at the time of consent, which occurred on average 2.44 weeks (SD = 1.37) weeks after the adolescents’ initial screening evaluation. Adolescents completed the baseline CES-D at the baseline evaluation which occurred on average 7.37 (SD = 1.66) weeks after the initial screening evaluation. Adolescents and parents also completed assessments at post-intervention, and at 6, 12, 18, and 24-months post-intervention. At each assessment, adolescents met with a trained clinical evaluator to complete a diagnostic interview and self-report forms. At these time points, parents completed measures about their adolescents, as well as self-report forms including the CES-D over the phone. Evaluators were naive to random assignment.

Interventions

IPT-AST.

IPT-AST is a manual based intervention (Young, Mufson, & Schueler, 2016) consisting of two individual pre-group sessions, eight group sessions, and an individual or dyadic (adolescent-parent) mid-group session. Four individual booster sessions were also conducted after conclusion of the group during the six-month follow-up period. In the pre-group sessions, the adolescent and group co-leader collaboratively identified the adolescent’s interpersonal goals for the group. The group focused on psychoeducation and interpersonal skill-building. Adolescents learned communication strategies such as using “I statements,” and practiced these strategies through group activities and role-plays in session. There were 18 IPT-AST groups, ranging in size from 3-7 youth. All groups were conducted by co-leaders, typically consisting of a clinical psychologist and a graduate student in clinical psychology.

Group Counseling (GC).

Group counseling was chosen as the comparison group because it reflected the type of groups run in schools. Although groups typically run in these schools had shorter and less frequent sessions, counselors agreed to hold eight weekly group sessions equal in length to the IPT-AST groups in that school. Counselors also agreed to meet with adolescents for a pre-group session, a mid-group session, and four booster sessions during the six-month follow-up period. No limits were given on the techniques to be used in GC groups in order to have GC reflect practices as normally delivered in schools. Some counselors ran manual-based, structured groups while others ran groups that were more flexible. There were 16 GC groups, ranging in size from 2-8 youth, which were typically run by a single group leader.

Measures

Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977).

The CES-D is a 20-item measure that assesses depressive symptoms over the past week. Scores range from 0-60, with higher scores indicating higher symptomatology. The CES-D has been shown to have high internal consistency, reliability, and validity in adolescent (Roberts, Andrew, Lewinsohn, & Hops, 1990) and adult (Weissman, Sholomskas, Pottenger, Prusoff, & Locke, 1977) samples. A score of 16 or above has been found to have both high sensitivity and high specificity for major depressive disorder in adults (Beekman et al., 1997) and has been considered the cutoff score indicating elevated depressive symptoms. While various cutoff scores have been identified for adolescents, we utilized the adult criterion as we had in prior studies in order to identify the greatest number of youth with elevated depressive symptoms (Young et al., 2016). The CES-D was administered to both adolescents and their mothers at each assessment. Cronbach’s alpha for the CES-D administered to mothers across time ranged from 0.90-0.92. Cronbach’s alpha for the adolescent CES-D ranged from 0.85-0.91.

Data analysis

The current study had three aims: (1) Assess whether initial levels of maternal depressive symptoms were correlated with initial levels of adolescent depressive symptoms, (2) Assess whether maternal depressive symptoms (CES-D) changed over the course of the intervention, and (3) Assess whether maternal outcomes correlated with simultaneous change in adolescent depressive symptoms across time in the two interventions.

Aim 1.

To achieve the first aim, we examined the relationship between maternal and adolescent depressive symptoms pre-prevention. When investigating the adolescents’ main outcomes, we discovered that there was a large decline in adolescents’ CES-D scores from the screening to the baseline evaluation before any intervention was delivered. These reductions suggested that symptom improvement began for some adolescents after completing the consent process and psychodiagnostic evaluations (Young et al., 2016). Therefore, we investigated the correlation between maternal and adolescent initial levels of depression in three ways: adolescents’ screening CES-D and mothers’ baseline CES-D, adolescents’ baseline CES-D and mothers’ baseline CES-D, and adolescents’ CES-D change scores (difference between screening and baseline CES-D) with mothers’ baseline CES-D. These correlations allowed us to investigate whether youth whose mothers had higher depressive symptom severity experienced less spontaneous improvement from the screening to the baseline evaluation prior to the start of the prevention groups. We investigated associations between depressive symptoms in mothers and their adolescents across interventions and between interventions. With a sample size of 167, we had over 80% power to detect a medium effect (r=0.25), which would be considered a clinically meaningful correlation (Cicchetti, 1994, 2008).

Aim 2.

To achieve the second aim, we first examined total change in maternal depressive symptoms over the course of the intervention. We paralleled the prior analyses that investigated the main outcomes of the interventions in the current RCT in two phases of change using hierarchical linear modeling (HLM) (Young et al., 2018), with the outcome being maternal CES-D score. The first phase of change investigated outcomes during the active intervention (baseline through the completion of booster sessions at 6-month follow-up) and the second phase of change examined 6-month follow-up through 24-month follow-up. As in Young et al. (2018), we produced total change estimates for the overall sample as well as both interventions. Within intervention effects assessed whether there was evidence of significant change for the respective intervention. Between intervention effects assessed whether there was significant differential change between IPT-AST and GC over the respective time period of interest. Maternal CES-D scores required a square-root transformation to normalize the residuals. While HLM was flexible in handling missing data, pattern-mixture models were used to assess whether important estimates were dependent on missing data patterns (Hedeker & Gibbons, 1997). With the pattern-mixture approach, separate intervention effects were estimated for specified missing data patterns. A differential intervention effect across the attrition patterns would have provided evidence of an informative missing data mechanism. Model-based estimates did not yield any evidence to suggest an informative missing data mechanism; therefore, missing data were treated as missing-at-random.

Aim 3.

To achieve the third aim, we then used multivariate multilevel modeling to explore the relationship between the simultaneous rates of change on the CES-D in adolescents and mothers, using square-root transformations for both adolescent and maternal CES-D scores. We paralleled the method used by Baldwin et al. (2014) for our analysis of simultaneous change. Because the data in our sample were longitudinal, the repeated observations within an individual were correlated. Additionally, because we had two simultaneous outcomes per time point (maternal depressive symptoms and adolescent depressive symptoms), the two measures were correlated. The random effects accommodated the individual change over time separately for the maternal depressive symptom scale and the adolescent depressive symptom scale as well as for the correlation within each dyad (i.e., mother and adolescent pair at each time point) (Singer & Willet, 2003). Statistical assessment of the respective correlation coefficients was based on the produced variance-covariance matrix of the random effects, which yielded pairwise Wald Chi-square statistics for the significance of each term in the variance-covariance matrix. The Wald Chi-square statistics were considered significant below the set alpha level of α = 0.05. We assessed the effect across all subjects regardless of intervention, and also assessed each intervention effect as well as differences between interventions through the entire 24-month follow-up period using the piecewise model. The models were fit using SAS 9.4.

Results

Aim 1

Table 1 displays mean scores for adolescent and maternal depressive symptom (CES-D) scores. We investigated the correlation between initial levels of maternal and adolescent depressive symptoms in three ways (maternal baseline CES-D and adolescent screening CES-D, maternal baseline CES-D and adolescent baseline CES-D, and maternal baseline CES-D and adolescent CES-D change scores), as displayed in Table 2. We first investigated these correlations in the overall sample, and subsequently investigated the correlations within each intervention. We failed to find a correlation between adolescents’ screening CES-D and maternal baseline CES-D scores (r = .02, p > .10). Conversely, we found a significant positive correlation between maternal baseline CES-D and adolescent baseline CES-D scores (r = .17, p < .05), signifying a small to medium effect size. For maternal baseline CES-D and adolescent CES-D change scores, the magnitude of the adolescent CES-D change score was inversely related to maternal baseline CES-D scores (r = −.17, p < .05); higher depression in mothers was associated with smaller change scores in adolescents. This effect was in the small to medium range. The strength of the correlations of maternal baseline CES-D scores with adolescent CES-D scores (at screening, baseline, and CES-D change scores) was not moderated by intervention (p > .40), which is reflective of the relatively similar correlations between the two interventions as illustrated in Table 2.

Table 1.

Descriptive Statistics for CES-D Scores

Adolescent CES-D Scores
Timepoint N Minimum Maximum Mean Standard
Deviation
Screening 167 16 44 23.91 6.68
Baseline 167 0 40 15.19 8.54
6-month 155 0 37 10.54 7.71
12-month 155 0 43 10.86 9.21
18-month 142 0 47 10.54 9.42
24-month 144 0 41 10.12 9.34
Change Screening to Baseline 167 −42 17 8.72 9.24
Maternal CES-D Scores
Timepoint N Minimum Maximum Mean Standard
Deviation
Baseline 162 0 48 11.34 10.08
6-month 142 0 44 9.87 10.28
12-month 141 0 44 8.78 9.97
18-month 127 0 56 8.61 9.90
24-month 132 0 45 7.27 8.73
Open in a new tab

Table 2.

Pearson Correlation for Initial Levels of Depressive Symptoms

Overall Sample
N = 162		 GC
N = 73		 IPT-AST
N = 89
Adolescent Screening CES-D and Maternal Baseline CES-D .02 −.02 .09
Adolescent Baseline CES-D and Maternal Baseline CES-D .17* .21# .12
Adolescent CES-D change score and Maternal Baseline CES-D −.17* −.22# −.10
Open in a new tab
*

Significant at p < .05,

#

trend at p < .10

Aim 2

Our focus in Aim 2 was to determine whether maternal depressive symptoms changed over the course of the study period. Using a piecewise model, we examined change from baseline through 6-month follow-up and change from the 6-month follow-up through the rest of the longitudinal period. Across both interventions, there was a significant decrease in maternal depressive symptoms from baseline to 6-month follow-up (t(150) = 2.70, p < .01) corresponding to an on-average 1.35 (SE=0.50) point decrease. Mothers of youth in GC reduced on-average 0.93 (SE=1.03) points which was not significantly different from 0 (t(150) = −1.47, p = . 14). Mothers of youth in IPT-AST experienced a significant decrease in depressive symptoms from baseline through 6-month follow-up (t(150) = 2.62, p < .01), reducing on-average 1.78 (SE=0.96) points on the CES-D. The difference in the amount of change between GC and IPT-AST was 0.85 (SE=1.33) points, which was a non-significant difference (t(150) = 0.74, p = .46).

During the follow-up phase, there were continued significant reductions in maternal depression scores across both interventions (t(150) = 1.98, p < .05) corresponding to an on-average 1.67 (SE=0.84) point decrease. For GC we found a non-significant reduction in CES-D scores over time (GC: t(150) = 1.04, p = .30), corresponding to a reduction of 1.07 (SE = 0.99) points. In IPT-AST, there was a significant reduction in maternal CES-D scores of 2.27 (SE = 0.90) points during the follow-up phase (t(150) = 2.02, p < .05). The difference in the amount of change between GC and IPT-AST was 1.20 (SE = 1.23) points, which was a non-significant difference in rates of change during the follow-up period (t(150) = 0.63, p = .53).

Over the entire study period, mothers in both conditions experienced significant reductions in CES-D scores (t(150)=4.85, p < .01). Mothers of youth on-average had a 3.02 (SE=0.62) point reduction over the entire study. There were also significant reductions in CES-D scores across the entire study period for both intervention groups (GC: t(150) = 2.58, p = .01; IPT-AST: t(150) = 4.89, p < .01). Mothers of youth in GC had an on-average reduction of 2.00 (SE = 0.99) points over the entire study period, whereas mothers of youth in IPT-AST experienced an on-average reduction of 4.04 (SE = 0.89) points. The difference in total change for IPT-AST compared to GC was 2.04 (SE = 1.23) points, which was not significant (t(150) = 1.43, p = .15). Change in maternal depressive symptoms is summarized in Table 3.

Table 3.

Estimated Reduction of Maternal CES-D Scores Across Time

Intervention Estimate Std Error T-value P
Baseline Through 6 Month Follow-up
Overall Sample (GC + IPT-AST) 1.35 0.75 2.70 0.01
GC 0.93 1.03 1.47 0.14
IPT-AST 1.78 0.96 2.62 0.01
CONTRAST GC vs IPT-AST 0.85 1.33 0.74 0.46
6 Month Through 24 Month Follow-up
Overall Sample (GC + IPT-AST) 1.67 0.72 1.98 < 0.05
GC 1.07 0.99 1.04 0.30
IPT-AST 2.27 0.90 2.02 0.04
CONTRAST GC vs IPT-AST 1.20 1.23 0.63 0.53
Total Change Baseline Through 24 Month Follow-up
Overall Sample (GC + IPT-AST) 3.02 0.71 4.85 < 0.01
GC 2.00 0.99 2.58 0.01
IPT-AST 4.04 0.89 4.89 < 0.01
CONTRAST GC vs IPT-AST 2.04 1.23 1.43 0.15
Open in a new tab

Aim 3.

Our focus in Aim 3 was understanding whether there was a relationship between change in maternal and adolescent depressive symptoms over time. The mean profiles for both maternal and adolescent CES-D scores can be seen in Figure 1. Figure 2 illustrates the relationship between the average slopes over the entire longitudinal period for both the maternal and adolescent depressive symptom scales for each individual. As seen in Figure 2, there was considerable variability within each intervention arm in the respective on-average relationship between maternal CES-D and adolescent CES-D slopes. Within both intervention arms, there appeared to be a small but positive relationship between rates of change in maternal CES-D scores and rates of change in adolescent CES-D scores.

Figure 1.

Figure 1.

Open in a new tab

Mean profiles for maternal and adolescent CES-D scores over time.

Figure 2.

Figure 2.

Open in a new tab

Relationship between average slopes for maternal and adolescent CES-D scores over 24-month longitudinal period.

Across interventions, using a piecewise multivariate HLM, we found a significant correlation (r = 0.37 (SE = 0.15), z = 2.52, p = .01) between an adolescent’s change in depressive symptoms and the respective mother’s change in depressive symptoms over time across the first phase of change (baseline through 6-month follow-up). Additionally, we saw a significant correlation (r = 0.31 (SE = 0.12), z = 2.55, p = .01) between an adolescent’s change in depressive symptoms with the respective mother’s change in depressive symptoms over the second phase of change (6-month through 24-month follow-up). Correlations per intervention arm during the first phase of change were 0.44 (SE = 0.21, z = 2.11, p = .04) for GC and 0.29 (SE = 0.21, z = 1.40, p = . 16) for IPT-AST which were not statistically significantly different (χ2(3) = 2.60, p = .46). Correlations per intervention arm during the second phase of change were 0.39 (SE = 0.19, z = 2.02, p = .04) for GC and 0.26 (SE = 0.14, z = 1.65, p < .10) for IPT-AST which were not statistically significantly different (χ2(3) = 1.41, p = .70).

Discussion

The current study examined the relationship between maternal and adolescent depressive symptoms across two adolescent depression prevention programs (IPT-AST and GC) in three ways. The findings for each of these aims are discussed below.

Aim 1. Relationship between Baseline Maternal and Adolescent Depressive Symptoms

Due to the nature of the screening process, we measured initial parental depressive symptoms once (at the consent visit, which occurred temporally between the adolescent’s screening and baseline visit) and initial adolescent depressive symptoms twice (at both the screening and baseline visits). We found that baseline maternal symptoms were not correlated with adolescent symptoms at the screening assessment. While we had hypothesized that we would find a correlation, these measures were taken several weeks apart which may have contributed to the lack of an association. Unfortunately, we were unable to collect maternal CES-D scores at the screening assessment, and cannot know whether screening scores for adolescents and their mothers would have been associated at this time point.

As hypothesized, we found a small but significant positive correlation between adolescent and maternal depressive symptoms at baseline, demonstrating that higher levels of symptoms in adolescents were associated with higher levels in mothers. This finding is in line with much of the previous literature in intervention research which has found similar associations (i.e., Kennard et al., 2008; Wilkinson et al., 2013), with the exception of few studies in which such associations were not found (i.e., Perloe et al. 2014). The connection between depressive symptoms in mothers and their children has been well documented in non-intervention studies as well (i.e., Goodman et al., 2011). The current study provided further evidence for the link between adolescent and maternal depression, suggesting that adolescents of mothers with higher levels of depressive symptoms were more likely to have higher symptoms themselves (and vice versa).

As discussed earlier, we found a large decline in adolescent CES-D scores from the screening to baseline evaluations before any intervention was delivered. While we are not sure the exact mechanism of this early decline in scores, other prevention studies have found similar effects (i.e., McCarty et al., 2013; Wijnhoven, Creemers, Vermulst, Scholte, & Engels, 2014). We hypothesize that the consent process and diagnostic evaluations were therapeutic interventions (Young et al., 2016), as they provided psychoeducation and support from a trained clinical evaluator. As a part of this first aim, we chose to investigate how this change may have been related to initial maternal depressive symptoms. We found that the magnitude of the early change in adolescent CES-D scores from the screening to baseline evaluation was inversely related to initial maternal CES-D scores. Adolescents whose mothers had higher levels of depressive symptoms at baseline had smaller improvements in CES-D scores from their screening to baseline evaluations. Given that the mothers’ baseline evaluation occurred temporally between the adolescents’ screening and baseline evaluations, it is difficult to draw any firm conclusions about the mechanisms underlying this early drop in scores; however, this finding suggests that adolescents of mothers with higher depressive symptoms at the start of our study experienced less spontaneous improvement.

Aim 2. Change in Maternal Depressive Symptoms Across Interventions

Across interventions, maternal depressive symptoms decreased significantly over the course of the 24-month longitudinal period. Through both phases of change and across the entire 24-month period, we found significant reductions in maternal depressive symptoms. We further investigated the changes in maternal depressive symptoms by intervention, and found that during both the active intervention phase and the follow-up, mothers of youth in IPT-AST experienced significant reductions in depressive symptoms, while mothers of youth in GC experienced non-significant reductions in depressive symptoms. Across both phases of change and across the overall 24-month period, the magnitude of change was greater for IPT-AST than GC, but the differences in rates of change between the two interventions were not significant. The finding that mothers in IPT-AST experienced significant reductions in depressive symptoms across both phases of change is notable. In the larger RCT we found that while adolescents in IPT-AST experienced significant improvements in depressive symptoms through 6-month follow-up, the benefits dissipated across 24-month follow-up (Young et al., 2018). However, mothers in IPT-AST continued to experience improvements over time in depressive symptoms even during the follow-up period. It is possible that the communication skills taught in IPT-AST led to a beneficial impact on the mother-adolescent relationship, which may have contributed to the lasting improvements in mothers; however, we recognize that this is speculation. Future research investigating the mechanisms underlying improvement in maternal symptoms would add clarity to this picture.

Our finding that maternal depressive symptoms improved across time adds to the body of literature which has found that when one part of the mother-child dyad participates in an intervention, the other experiences improvements in symptoms (i.e., Kennard et al., 2008; Pilowsky et al., 2008). While we hypothesized that we would see a decrease in maternal symptoms, our findings were still striking, as maternal depression was not specifically targeted in our study and we saw sizable change in mothers’ symptoms across interventions. Notably, these improvements persisted across a two-year follow-up period, even when adolescents did not experience consistent reductions. Our findings that mothers improved across both interventions is also interesting. One might anticipate that mothers’ and adolescents’ CES-D scores would be more linked in the IPT-AST group given the intervention’s focus on interpersonal issues. Conversely, others might argue that the intervention, which teaches adolescents how to communicate effectively with others, may help break the association between maternal and adolescent symptoms. However, the general takeaway from these findings suggest that depression prevention programs, regardless of the specific interventions used, may have “trickle up” effects and beneficially impact mothers’ depressive symptoms. Future studies would benefit from including a no-intervention control group to assess intervention effects on mothers’ symptoms versus spontaneous improvement across time. Additionally, the inclusion of fathers in future studies would be interesting to see whether these same effects may be true for both parents.

Aim 3. Relationship between Change in Maternal and Adolescent Depressive Symptoms

Regarding concurrent change trajectories, we found a significant medium-sized correlation between rates of change in adolescent and maternal depressive symptoms during both the first and second phases of change across intervention condition. There were no differences across the two interventions in these associations at either phase of change. This finding adds to a growing body of literature demonstrating that youth depression outcomes were associated with mothers’ depression outcomes (i.e., Kennard et al., 2008; Perloe et al., 2014; Wilkinson et al., 2013). In these treatment studies, as adolescents improved across time, parents also experienced improvement in depressive symptoms. This study is the first to our knowledge to show this association in adolescents with elevated depressive symptoms enrolled in depression prevention programs. Our findings, in conjunction with the previous studies, suggest that intervening with one part of a depressed (or sub-clinically depressed) mother-child dyad might lead to improvements in the other. As such, preventive interventions for youth may have benefits for their parents as well.

Limitations

The current study has several limitations. First, maternal and adolescent depressive symptoms were measured in our study using only self-report data (CES-D). While the larger RCT conducted structured clinical interviews on youth, we only have self-report data for mothers. While the CES-D had been shown to have high reliability in assessing depressive symptoms, self-report data might not reflect the most objective picture of an individual’s mental health. Future studies of IPT-AST would benefit from the inclusion of a structured clinical interview to assess maternal depression diagnoses in addition to symptoms.

While our study provides further evidence for the association between maternal and youth depression outcomes, we do not have evidence to demonstrate that improvement in youth symptoms led to improvement in mothers. The lack of demonstrated causality has been a limitation in most research on mother-youth depression (i.e., Perloe et al., 2014; Weissman et al., 2006), as the myriad of possible confounding variables makes testing of causality difficult. For instance, it is possible that improvements in mothers’ symptoms had an effect on their children’s symptoms (reverse causation). Several studies in the literature, including the STAR*D trial (Pilowsky et al., 2008) have attempted to understand the directionality of effects but have had inconclusive results. Therefore, future studies should attempt to address this challenge and design methodologically rigorous studies to investigate causality.

Additionally, GC was a powerful control group, and our study did not include a no-intervention control group. As such, we were not able to determine how maternal symptoms would have changed across time if their adolescents had not participated in a depression prevention program. As stated earlier, without this type of comparison condition, it is unclear whether changes in maternal symptoms are attributable to the prevention programs or simply reflect regression to the mean. Future studies of IPT-AST would benefit from including a no-intervention control group. Additionally, we did not collect data on treatment utilization in mothers. As such, we do not know if improvements in maternal depressive symptoms were related to mothers seeking treatment rather than improvements in their adolescents’ depression. Future research is also warranted to better understand the mechanisms underlying the association between maternal and adolescent depressive symptoms.

Summary and Conclusions

Despite these limitations, the findings from the current study contribute to the growing literature investigating the relationship between maternal and adolescent depressive symptom outcomes across depression interventions. Our study found that maternal and adolescent depressive symptoms were related, and that mothers with higher baseline depressive symptoms had adolescents with higher baseline symptoms. Maternal depressive symptoms improved across our interventions, even though maternal depression was not directly targeted. We found these effects across both prevention programs, suggesting that mothers of adolescents enrolled in depression prevention programs may have experienced some indirect benefit. However, we recognize that without a no-intervention control group, we cannot know whether mothers of adolescents who did not receive a preventive intervention would also improve across time. Future studies of depression prevention programs would benefit greatly from including a no-intervention control group in their design. Such a design would allow direct comparison to see whether these improvements in mothers were specific to those whose adolescents received a preventive intervention. Lastly, mothers improved in tandem with their youth who participated in these prevention programs. These findings add further evidence to the literature that shows that as one part of the mother-child dyad improves, the other improves as well. However, we do not yet know the specifics of this relationship, and further research is needed to understand if targeting either mothers or adolescents will have the greatest impact on the mother-adolescent dyad. Nonetheless, these findings add to the current understanding of the relationship between maternal and adolescent depressive symptom outcomes, and have important implications for the prevention and treatment of depression.

Highlights.

  • Initial levels of maternal/adolescent depressive symptoms positively associated

  • Maternal depressive symptoms improved across time in adolescent prevention programs

  • Maternal and adolescent depressive symptom outcomes related across time in study

Acknowledgments

Role of the Funding Source

This research was supported by the NIMH (R01MH087481). NIMH had no involvement in the collection, analysis and interpretation of data, in the writing of the report, nor in the decision to submit the article for publication.

Funding: This research was supported by the NIMH (R01MH087481) grant awarded to Jami F. Young, Ph.D.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflicts of Interest

We wish to draw the attention of the Editor to the following facts which may be considered as potential conflicts of interest and to significant financial contributions to this work. Dr. Jami Young receives royalties from Oxford University Press. All other authors declare that they have no conflicts of interest.

We confirm that the manuscript has been read and approved by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the manuscript has been approved by all of us.

We confirm that we have given due consideration to the protection of intellectual property associated with this work and that there are no impediments to publication, including the timing of publication, with respect to intellectual property. In so doing we confirm that we have followed the regulations of our institutions concerning intellectual property.

We further confirm that any aspect of the work covered in this manuscript that has involved human patients has been conducted with the ethical approval of all relevant bodies and that such approvals are acknowledged within the manuscript.

Contributor Information

Carolyn Spiro-Levitt, Department of Child and Adolescent Psychiatry, Hassenfeld Children’s Hospital at New York University Langone.

Robert Gallop, Department of Mathematics, West Chester University

Jami F. Young, Department of Child and Adolescent Psychiatry and Behavioral Sciences, Children’s Hospital of Philadelphia, Department of Psychiatry, University of Pennsylvania Perelman School of Medicine

References

  1. Baldwin SA, Imel ZE, Braithwaite SR, & Atkins DC (2014). Analyzing multiple outcomes in clinical research using multivariate multilevel models. Journal of Consulting and Clinical Psychology, 82(5), 920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Beekman AT, Deeg DJH, Van Limbeek J, Braam AW, De Vries MZ, & Van Tilburg W (1997). Brief communication: Criterion validity of the Center for Epidemiologic Studies Depression scale (CES-D): Results from a community-based sample of older subjects in the Netherlands. Psychological Medicine, 27(1), 231–235. [DOI] [PubMed] [Google Scholar]
  3. Cicchetti DV (1994). Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological Assessment, 6(4), 284. [Google Scholar]
  4. Cicchetti D, & Gunnar MR (2008). Integrating biological measures into the design and evaluation of preventive interventions. Development and Psychopathology, 20(3), 737–743. [DOI] [PubMed] [Google Scholar]
  5. Cuijpers P, van Straten A, Smit F, Mihalopoulos C, & Beekman A (2008). Preventing the onset of depressive disorders: a meta-analytic review of psychological interventions. American Journal of Psychiatry, 165(10), 1272–1280. [DOI] [PubMed] [Google Scholar]
  6. Goodman SH, Rouse MH, Connell AM, Broth MR, Hall CM, & Heyward D (2011). Maternal depression and child psychopathology: A meta-analytic review. Clinical Child and Family Psychology Review, 14(1), 1–27. [DOI] [PubMed] [Google Scholar]
  7. Hedeker D, & Gibbons RD (1997). Application of random-effects pattern-mixture models for missing data in longitudinal studies. Psychological Methods, 2, 64–78. [Google Scholar]
  8. Horowitz JL, & Garber J (2006). The prevention of depressive symptoms in children and adolescents: A meta-analytic review. Journal of Consulting and Clinical Psychology, 74(3), 401–415. [DOI] [PubMed] [Google Scholar]
  9. Horowitz JL, Garber J, Ciesla JA, Young JF, & Mufson L (2007). Prevention of depressive symptoms in adolescents: a randomized trial of cognitive-behavioral and interpersonal prevention programs. Journal of Consulting and Clinical Psychology, 75(5), 693–706. [DOI] [PubMed] [Google Scholar]
  10. Kaufman J, Birmaher B, Brent D, & Rao U (1997). Schedule for affective disorders and schizophrenia for school-age children-present and lifetime version (K-SADS-PL): Initial reliability and validity data. Journal of the American Academy of Child & Adolescent Psychiatry, 36, 980–988. [DOI] [PubMed] [Google Scholar]
  11. Kennard BD, Hughes JL, Stewart SM, Mayes T, Nightingale-Teresi J, Tao R, … & Emslie GJ (2008). Maternal depressive symptoms in pediatric major depressive disorder: Relationship to acute treatment outcome. Journal of the American Academy of Youth & Adolescent Psychiatry, 47(6), 694–699. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. McCarty CA, Violette HD, Duong MT, Cruz RA, & McCauley E (2013). A randomized trial of the positive thoughts and action program for depression among early adolescents. Journal of Clinical Child & Adolescent Psychology, 42, 554–563. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Merry SN, Hetrick SE, Cox GR, Brudevold.Iversen T, Bir JJ, & McDowell H (2012). Cochrane Review: Psychological and educational interventions for preventing depression in children and adolescents. Evidence-Based Child Health: A Cochrane Review Journal, 7(5), 1409–1685. [DOI] [PubMed] [Google Scholar]
  14. Perloe A, Esposito-Smythers C, Curby TW, & Renshaw KD (2014). Concurrent trajectories of change in adolescent and maternal depressive symptoms in the TORDIA study. Journal of Youth and Adolescence, 43(4), 612–628. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Pilowsky DJ, Wickramaratne P, Talati A, Tang M, Hughes CW, Garber J, … & Alpert JE (2008). Children of depressed mothers 1 year after the initiation of maternal treatment: findings from the STAR* D-Child Study. American Journal of Psychiatry, 165(9), 1136–1147. [DOI] [PubMed] [Google Scholar]
  16. Radloff LS (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401. [Google Scholar]
  17. Roberts RE, Andrews JA, Lewinsohn PM, & Hops H (1990). Assessment of depression in adolescents using the Center for Epidemiologic Studies Depression Scale. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 2(2), 122–128. [Google Scholar]
  18. Singer JD, & Willet JB (2003). A framework for investigating change over time. Applied longitudinal data analysis: Modeling change and event occurrence, 315. [Google Scholar]
  19. Stice E, Shaw H, Bohon C, Marti CN, & Rohde P (2009). A meta-analytic review of depression prevention programs for children and adolescents: factors that predict magnitude of intervention effects. Journal of Consulting and Clinical Psychology, 77(3), 486–503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Thapar A, Collishaw S, Pine DS, & Thapar AK (2012). Depression in adolescence. The Lancet, 379(9820), 1056–1067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Verdeli H, Ferro T, Wickramaratne P, Greenwald S, Blanco C, & Weissman MM (2004). Treatment of depressed mothers of depressed children: pilot study of feasibility. Depression and Anxiety, 79(1), 51–58. [DOI] [PubMed] [Google Scholar]
  22. Weissman MM, Pilowsky DJ, Wickramaratne PJ, Talati A, Wisniewski SR, Fava M,… & Cerda G (2006). Remissions in maternal depression and youth psychopathology: a STAR* D-youth report. JAMA Psychiatry, 295(12), 1389–1398. [DOI] [PubMed] [Google Scholar]
  23. Weissman MM, Sholomskas D, Pottenger M, Prusoff BA, & Locke BA (1977). Assessing depressive symptoms in five psychiatric populations: A validation study. American Journal of Epidemiology, 106, 203–214. [DOI] [PubMed] [Google Scholar]
  24. Weissman MM, Wickramaratne P, Gameroff MJ, Warner V, Pilowsky D, Kohad RG,… & Talati A (2016). Offspring of Depressed Parents: 30 Years Later. American Journal of Psychiatry. 10.1176/appi.ajp.2016.15101327 [DOI] [PubMed] [Google Scholar]
  25. Wilkinson PO, Harris C, Kelvin R, Dubicka B, & Goodyer IM (2013). Associations between adolescent depression and parental mental health, before and after treatment of adolescent depression. European Youth & Adolescent Psychiatry, 22(1), 3–11. [DOI] [PubMed] [Google Scholar]
  26. Wijnhoven LA, Creemers DH, Vermulst AA, Scholte RH, & Engels RC (2014). Randomized controlled trial testing the effectiveness of a depression prevention program (‘Op Voile Kracht’) among adolescent girls with elevated depressive symptoms. Journal of Abnormal Child Psychology, 42, 217–228. [DOI] [PubMed] [Google Scholar]
  27. Young JF, Benas JS, Schueler CM, Gallop R, Gillham JE, & Mufson L (2016). A randomized depression prevention trial comparing Interpersonal Psychotherapy-Adolescent Skills Training to group counseling in schools. Prevention Science, 17, 314–324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Young JF, Jones JD, Sbrilli M, Benas JS, Spiro CN, Haimm C, Gallop R, Mufson L, Gillham JE (2018). Long-term effects from a school based trial comparing Interpersonal Psychotherapy-Adolescent Skills Training to group counseling. Journal of Clinical Child & Adolescent Psychology, 1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Young JF, & Mufson L (2003). Manual for interpersonal psychotherapy-adolescent skills training (IPT-AST). New York: Columbia University. [Google Scholar]
  30. Young JF, Mufson L, & Davies M (2006). Efficacy of Interpersonal Psychotherapy. Adolescent Skills Training: an indicated preventive intervention for depression. Journal of Child Psychology and Psychiatry, 47(12), 1254–1262. [DOI] [PubMed] [Google Scholar]
  31. Young JF, Mufson L, & Gallop R (2010). Preventing depression: a randomized trial of interpersonal psychotherapy-adolescent skills training. Depression and Anxiety, 27(5), 426–433. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Young JF, Mufson L, & Schueler CM (2016). Preventing Adolescent Depression: Interpersonal Psychotherapy-Adolescent Skills Training. New York: Oxford University Press. [Google Scholar]

RESOURCES