Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Sep 1.
Published in final edited form as: J Affect Disord. 2020 May 23;274:662–670. doi: 10.1016/j.jad.2020.05.120

Depressive symptom severity mediates the association between avoidant problem-solving style and suicidal ideation

Roberto López Jr 1, Leslie A Brick 2, Annamarie B Defayette 1, Emma D Whitmyre 1, Jennifer Wolff 2, Elisabeth Frazier 2, Anthony Spirito 2, Christianne Esposito-Smythers 1
PMCID: PMC7401318  NIHMSID: NIHMS1600557  PMID: 32664000

Abstract

Background:

The contemporaneous association between avoidant style, a maladaptive social problem-solving strategy, and adolescent suicidal ideation has been well established. However, the mechanisms underlying this association are not well understood. Using cross-lagged panel modeling, the present study examined whether depressive symptom severity mediates the relation between avoidant style and severity of suicidal ideation. The specificity of depressive symptom severity as a mediator was also evaluated by simultaneously testing whether avoidant style mediates the association between depressive symptom and suicidal ideation severity.

Methods:

The sample included 110 adolescents enrolled in a randomized controlled clinical effectiveness trial. Avoidant style as well as depressive symptom and suicidal ideation severity were assessed via self-report with the Social Problem-Solving Inventory-Revised, Children’s Depression Scale-2, and Suicidal Ideation Questionnaire-Junior, respectively, at baseline, 3- and 6-months.

Results:

After accounting for participant age, sex, and treatment condition, path analyses supported the specificity of 3-month depressive symptom severity as a mediator of the association between baseline levels of avoidant style and 6-month suicidal ideation severity.

Limitations:

Results may not be generalizable to less severe samples. Causality cannot be inferred from study results. Data were exclusively collected via self-report.

Conclusions:

Findings suggest that avoidant style is indirectly related to suicidal ideation through depressive symptom severity. Thus, treatment targeted at improving social problem-solving skills, particularly avoidant style, may help reduce depressive symptoms and lower suicide risk.

Keywords: social problem-solving, depression, suicidal ideation, adolescence

1. Introduction

Suicide is the second leading cause of death among adolescents ages 15 to 19 in the United States, and rates continue to rise (Centers for Disease Control and Prevention [CDC], 2017). Suicidal ideation (SI) almost always precedes suicidal behavior (Turecki & Brent, 2016), and the transition from thoughts to behavior occurs quickly for many youth. Nock and colleagues (2013) found that up to 41% of adolescent females and 23% of adolescent males engage in suicidal behavior within the first year of onset of SI. Thus, the identification of factors that influence SI is of critical importance in adolescent suicide prevention.

Social problem-solving deficits have been documented as a relatively robust risk factor for adolescent SI (for a review, see Speckens & Hawton, 2005). While a large body of literature in this area exists, there are several limitations in this research. First, existing studies have not consistently assessed how different facets of social problem-solving, such avoidant style, may be uniquely related to SI (D’Zurilla & Nezu, 1982; Speckens & Hawton, 2005). Second, prior research typically assesses social problem-solving deficits and SI contemporaneously. This precludes study of temporal associations or mechanisms through which problem-solving deficits may lead to SI prospectively, such as depressive symptom severity (Speckens & Hawton, 2005). To improve upon existing research and provide a more rigorous test of the relationship between social problem-solving deficits and SI, the current study used a multi-wave, longitudinal design to examine the temporal association between one facet of social problem-solving deficits, avoidant style, and SI severity. Additionally, this study explored whether depressive symptom severity mediates this association.

1.1. Avoidant Style, Depressive Symptoms, and Suicidal Ideation

Avoidant style, defined as passivity or inaction in the face of problems (D’Zurilla & Nezu, 1982; Maydeu-Olivares & D’Zurilla, 2002), has been positively associated with severity of SI in community and psychiatric adolescent samples, with the relationship stronger among youth who reported higher (versus lower) levels of stress (Clum & Febbraro, 1994; Reinecke et al., 2001). Avoidant style has also been robustly linked with depressive symptom severity in adolescents (Becker-Weidman et al., 2010; Reinecke et al., 2001; Schäfer et al., 2017; Seiffge-Krenke & Klessinger, 2000). While avoidance of problems may lead to short-term relief, such as decreased negative affect in that moment, the problems causing the depressive symptoms often remain unresolved and may even escalate (Sheppes et al., 2015; Trew, 2011). Moreover, failure to solve problems may lead to and/or strengthen cognitive distortions (e.g., “I can’t do anything right” or “I’m worthless”), which have been shown to play a causal role in the onset and maintenance of depressive symptoms (Reinecke et al., 1998; Scott et al., 2019). Thus, avoidant style may confer risk for more severe depressive symptoms among adolescents.

Depressive symptom severity, in turn, is a strong and robust risk factor for adolescent SI (Gould et al., 2003; Wolff et al., 2018). Depressive symptoms, such as depressed mood, anhedonia, feelings of worthlessness, and negative self-esteem, may be experienced as intolerable and fuel the motivation to end one’s suffering via suicide (Hawton et al., 2012; Nock & Kazdin, 2002). Indeed, research suggests that suicide is a form of experiential avoidance (for a review, see Brereton & McGlinchey, 2019). Depressive symptoms can also lead to social isolation and loneliness (Matthews et al., 2016), which in turn, increase suicide risk (Calati et al., 2019). Notably, theory and research also suggest significant overlap in cognitive distortions that influence the onset and/or maintenance of depressive symptoms and suicidality (Spirito & Esposito-Smythers, 2008). Thus, depressive symptom severity may confer risk for more severe SI among adolescents.

Contemporary theory supports the possibility that depressive symptom severity may serve as a mechanism through which avoidant style and SI are related. According to The Three-Step Theory of Suicide (3ST; Klonsky & May, 2015), individuals tend to avoid situations or stimuli that provoke emotional (and/or physical) pain, which in turn, often results in long-term costs such as chronic stress. Chronic stress may then lead to depressive symptoms, such as intense sadness and worthlessness (Tennant, 2002; Trew, 2011). When these depressive symptoms intertwine with hopelessness and a lack of connectedness to others (see Klonsky & May, 2015), SI emerges and intensifies. One study by Reinecke and colleagues (2001) lends some initial support for this model. Using a clinical sample of adolescents, these researchers found that avoidant style indirectly affects SI severity through depressive symptom severity, though the design was cross-sectional so the direction of the association cannot be determined (Reinecke et al., 2001).

1.2. Alternative Model

Though theory and research offer support for depressive symptoms as a mediator in the proposed model, it is also possible that depressive symptoms lead to social problem-solving deficits, such as avoidance, which in turn, predict SI. Negative emotions have been shown to exacerbate problem-solving deficits, including avoidant style (Chang, 2017; Frederickson, 2001). Relatedly, depressive symptom severity has been associated with deficits in executive functioning (for reviews, see McDermott & Ebmeier, 2009; Snyder, 2013), which is believed to underlie the implementation of emotion regulation strategies, such as social problem-solving, in the face of stressors (Zelazo & Cunningham, 2007). Thus, youth experiencing depressive symptoms may be more prone to implement maladaptive emotion regulation strategies, such as avoidant style. As depressive symptoms become more severe, adolescents may be disposed to use more extreme forms of avoidance to reduce their distress. Suicidality, considered to be one of the most severe forms of avoidance (Brereton & McGlinchey, 2019), may thus become an adolescent’s only perceived solution to ending their distress. Therefore, it is equally plausible that avoidant style mediates the relation between depressive symptom and SI severity. To the authors’ knowledge, however, no studies to date have examined this relationship.

1.3. Current Study

In summary, ample research supports an association between avoidant style and SI among adolescents. Theory and existing cross-sectional work also support the possibility that depressive symptom severity mediates this association. However, this model has not been tested using a longitudinal design. Using a cross-lagged panel model (CLPM), the present study tested this mediation model in a clinical sample of adolescents enrolled in a multi-wave, intervention trial. CLPMs allow for a more precise examination of temporal processes, such as directionality of effects, while accounting for contemporaneous levels of all study variables at each time point (Selig & Preacher, 2009). To provide a conservative test of this model, youth’s age and sex, which have been shown to influence social problem-solving deficits, as well as depressive symptom and SI severity (Adrian et al., 2016; D’Zurilla et al., 1998; Van Orden et al.; 2010), were covaried in the final, optimized model (see Figure 2). Treatment condition assignment in the parent study was also entered as a covariate. Study hypotheses were as follows:

Figure 2.

Figure 2.

Open in a new tab

Final cross-lagged panel model with variables of interest.

Note. BL = baseline; 3M = 3-month; 6M = 6-month. AS = avoidant problem-solving style; DP = depressive symptom severity; SI = suicidal ideation severity. Solid black paths are statistically significant. Grey paths are statistically non-significant. Not pictured for ease of interpretation are the contemporaneous paths between avoidant style, depressive symptom and suicidal ideation severity, as well as the direct paths from covariates to these variables at each respective time point. These direct and contemporaneous paths were included in this final statistical model and are discussed in the text. Parameter estimates presented here are standardized. For additional parameter estimates, see Tables 3 and 4.

*p < .05. *p < .001.

1) After accounting for covariates and baseline depressive symptom severity, greater levels of avoidant style at baseline and the 3-month follow-up would predict greater depressive symptom severity at 3- and 6-month follow-up, respectively. Greater depressive symptom severity at baseline and 3-months would positively predict SI severity at 3- and 6-months, respectively. Finally, depressive symptom severity at 3-months would mediate the relationship between avoidant style at baseline and SI severity at 6-months (see Figure 1).

Figure 1.

Figure 1.

Open in a new tab

Cross-lagged panel model (i.e., Model 1) with all hypothesized autoregressive and cross-lags between variables of interest.

Note. BL = baseline; 3M = 3-month; 6M = 6-month. AS = avoidant problem-solving style; DP = depressive symptom severity; SI = suicidal ideation severity. Black solid paths were retained in the final model (see Figure 2). Not pictured for ease of interpretation are the contemporaneous paths between variables of interest. These paths were included in all models and are discussed in the text. For parameter estimates, see Supplementary material.

2) After accounting for covariates and baseline levels of avoidant style, greater depressive symptom severity at baseline and 3-month follow-up would predict greater levels of avoidant style at 3- and 6-month follow-up, respectively. Greater levels of avoidant style at baseline and 3-months would predict SI severity at 3- and 6-months, respectively. Finally, levels of avoidant style at 3-months would mediate the relationship between depressive symptom severity at baseline and SI severity at 6-months (see Figure 1).

2. Method

2.1. Participants and Procedures

All study procedures were approved by both university and hospital Institutional Review Boards prior to the start of data collection. The sample consisted of adolescents enrolled in a randomized controlled pilot effectiveness trial comparing a cognitive-behavioral intervention to enhanced standard care for dually diagnosed adolescents (Wolff et al., under review). The current study is a secondary analysis of this data and not an examination of the main outcomes of the effectiveness trial.

All adolescents and legal guardians who presented for intensive, home-based mental health services at a community-based facility located in the Northeastern United States were approached for recruitment. After consent/assent was obtained, a trained interviewer administered semi-structured diagnostic interviews and research assistants (RAs) delivered the self-report assessments to adolescents and their guardians. This includes screening, baseline, and follow-up assessments.

The inclusion criteria for this parent study included: 1) English-speaking adolescent, 12 – 18 years of age and legal guardian(s), 2) adolescent assent (consent for 18-year-olds), 3) consent of a legal guardian, and 4) referred to the intensive home-based program for co-occurring substance use and psychiatric symptoms. Youth were ineligible if they exhibited serious psychiatric symptoms (e.g., active hallucinations, thought disorder, or a primary diagnosis of obsessive-compulsive disorder or disordered eating), acute homicidality, or violent behavior upon screening, indicating need for a specialized and/or higher level of care.

2.2. Constructs and Measures

2.2.1. Avoidant style.

The Social Problem-Solving Inventory-Revised (SPSI-R; Maydeu-Olivares & D’Zurilla, 2002) is a 52-item self-report measure of current, perceived problem-solving abilities that employs a 5-point Likert scale from 0 (“Not at all true of me.”) to 4 (“Extremely true of me.”). Only the 7- item avoidant style subscale was used in the present study. Raw subscale scores range from 0 – 28. Higher scores reflect a more avoidant problem-solving style (e.g., “When a problem occurs in my life, I usually put off trying to solve it for as long as possible”). Adequate internal consistency as well as factorial and convergent validity have been demonstrated with community and psychiatric adolescent samples (Sadowski et al., 1994). The SPSI-R was administered to youth at every assessment period (baseline, 3- and 6-months) and internal consistency (∝) for the avoidant style subscale was good to excellent (.84 – .93) across all time points.

2.2.2. Depressive symptom severity.

The Children’s Depression Inventory-2 (CDI-2; Kovacs, 2011) is a 28-item self-report measure of depressive symptoms during the past two weeks. Items on the CDI-2 are rated on a scale from 0 (e.g., “I am sad once in a while.”) to 2 (e.g., “I am sad all the time.”). Higher scores indicate greater depressive symptoms. For this study, the single item asking about suicidality was excluded to avoid overlap between measures of depressive symptoms and SI. Therefore, possible raw scores ranged from 0 – 54. The CDI-2 has been shown to have good reliability and construct validity in adolescents (Kovacs, 2011). The CDI-2 was administered to youth at every assessment period, and internal consistency (∝) was good to excellent (.85 – .91) across all time points.

2.2.3. Suicidal ideation severity.

The Suicidal Ideation Questionnaire-Junior (SIQ-JR; Reynolds, 1987) is a 15-item self-report measure that assesses severity of SI over the past month. Items are rated on a 7-point Likert scale ranging from 0 (“I never had this thought.”) to 6 (“Almost every day.”). Raw scores range from 0 – 90. Higher scores indicate greater SI severity. The SIQ-JR has been shown to have excellent reliability and construct validity with adolescents (King et al., 1997; Reynolds & Mazza, 1999). The SIQ-JR was administered to youth at every assessment period, and internal consistency (∝) was excellent (.93 – .97) across all time points.

2.3. Data Analysis

This study utilized a CLPM design using Mplus version 8.1 (Muthén & Muthén, 1998-2017) to examine the contemporaneous and longitudinal relationships between avoidant style, as well as depressive symptom and SI severity. To inspect the data for bivariate relations associated with high multicollinearity (r’s > .60; Grewal et al., 2004), descriptive and correlational analyses were conducted using SPSS version 19.0. Multicollinearity diagnostics (i.e., variance inflation factor [VIF] and tolerance) were also calculated in SPSS by simultaneously regressing all variables of interest at baseline, 3-, and 6-month on 6-month SI. Additionally, missing data patterns, particularly in the criterion variable, can influence path estimates (Cohen et al., 2003). Therefore, adolescents with complete versus missing data for SI severity at 3- and 6-months were compared on SI severity at baseline using independent t-tests. Adolescents were also compared across a variety of key demographic variables (i.e., age, sex, treatment condition, and race/ethnicity) with independent t-tests and chi-square tests. Avoidant style and depressive symptom severity across all time points were examined with independent t-tests. The Benjamini-Hochberg procedure, with the false discovery rate set at 5%, was used to correct for multiple comparisons (see Benjamini & Hochberg, 1995).

Consistent with existing literature (see Little, 2013), a CLPM without covariates (i.e., Model 1) was created with all hypothesized cross-lagged as well as autoregressive and contemporaneous paths freely estimated (see Figure 1). To obtain the most parsimonious model, non-significant paths were then constrained in a series of models, allowing for nested comparisons per Satorra and Bentler’s (2010) guidelines. If two models fit the data equally well, the model with more constrained paths (i.e., the more parsimonious model) was selected.

Using the above procedure, a parsimonious model was identified by testing four additional nested models via the sequential constraining of non-significant autoregressive and cross-lagged paths. To create Model 2, the autoregressive path between baseline and 6-month avoidant style was constrained (a path in Figure 1), as it was not essential to testing study hypotheses. Relatedly, while traditional statistical methods of testing mediation require direct effects be calculated between predictor and criterion variables (see Baron & Kenny, 1986), more recent research argues that this relationship need not be estimated (see Hayes, 2009; Little, 2013). Therefore, building off Model 2, Model 3 was created by constraining the direct paths between hypothesized predictors (i.e., depressive symptom severity and avoidant style at baseline) and the criterion variable at subsequent time points (i.e., SI severity at 3- and 6-months; b paths in Figure 1). Finally, we began to constrain non-significant cross-lagged paths that were directly involved in our hypotheses. Model 4 was created by retaining constraints from Models 1 – 3 while also constraining the non-significant path from 3-month depressive symptom severity to 6-month avoidant style (c path in Figure 1). Building off Model 4, Model 5 was created by constraining the non-significant cross-lags directly involved in the hypothesized mediational pathway of the alternative model. Specifically, we constrained the path from baseline depressive symptom severity to 3-month avoidant style as well as the path from 3-month avoidant style to 6-month SI severity (d paths in Figure 1). Once the most parsimonious model was identified (see Figure 2), all covariates (i.e., age, sex, and treatment condition) were regressed on avoidant style, depressive symptoms, and SI across all time points to stringently test study hypotheses.2

To account for missing data as well as potential non-normality, all models used full information likelihood estimation (FIML) and a robust maximum likelihood estimator (MLR; Muthén & Muthén, 1998-2017). Individual model fit was assessed using the adjusted chi-square statistic, the comparative fit index (CFI), the Tucker-Lewis Index (TLI), and the root-mean-square error of approximation (RMSEA). A non-significant, adjusted chi-square suggests that the model fits the data well (Bentler & Bonnett, 1980). Additionally, good fit is assumed for values ≥ . 95 for CFI and TLI as well as ≤ .05 for RMSEA (Hu & Bentler, 1999). Acceptable fit is assumed by values ≥ . 90 for CFI and TLI and < .08 for RMSEA (Little, 2013).

Consistent with the mediation literature (Hayes, 2009; Little, 2013), indirect effects were only evaluated if the parameter estimates for both alpha and beta paths were statistically significant in the most parsimonious model. Therefore, only the indirect effect of 3-month depressive symptom severity on the relationship between baseline avoidant style and 6-month SI severity was estimated and tested using a bias-corrected bootstrapping approach with 5000 bootstrap re-samples; indirect paths are considered significant if the 95% confidence intervals do not contain zero (Preacher & Hayes, 2008). This approach makes no assumptions about the distribution of the indirect effect and has been considered superior to using traditional bootstrapping methods or the Sobel test (Hayes, 2009; MacKinnon et al., 2004).

3. Results

3.1. Descriptives, Bivariate Analyses, and Multicollinearity Diagnostics

A total of 111 adolescents were enrolled in the parent study. However, one youth did not complete any of the self-report measures for the variables of interest at any time point in the present study and was thus excluded from analyses. Thus, the present sample consisted of 110 adolescents (Mage = 15.71; SD = 1.18). The sample had slightly more males (n = 63; 57.3%) than females (n = 47; 42.7%) and self-identified as: 70.9% White (n = 78), 10.9% Black or African American (n = 12), 11.8% Multiracial (n = 13), and 2.7% Other (n = 3). Three participants (3.6%) declined to provide their racial identity. Thirty percent (n = 33) of participants identified as Latino. The annual family income ranged from < $5,000 (10.9%) to > $100,000 (10.9%), with median income between $26,000 and $49,000. Approximately half of participants (n = 61; 55.5%) were enrolled in the experimental condition. Table 1 includes means, standard deviations, and correlations for study variables.

Table 1.

Means, standard deviations, and bivariate correlations for all study variables.

Variables M (SD) 1 2 3 4 5 6 7 8 9 10 11 12
1. BL SI 17.16 (19.32) -
2. 3M SI 8.18 (10.65) .64** -
3. 6M SI 9.70 (14.51) .63** .48** -
4. BL DP 16.58 (9.82) .65** .37** .50** -
5. 3M DP 12.58 (7.57) .32** .50** .45** .61** -
6. 6M DP 11.37 (7.05) .39** .36** .70** .54** .59** -
7. BL AS 16.13 (10.04) .23* .25* .12 .30** .38** .17 -
8. 3M AS 14.21 (8.57) .05 .18 −.02 .36** .35** .05 .45** -
9. 6M AS 14.28 (7.94) −.02 .32** .06 .21 .32** .16 .26* .56** -
10. Age 15.71 (1.18) .30** .25* .14 .30** .28** .18 .22 .07 −.08 -
11. Sex - .22* .08 .16 .32** .15 .19 −.23 −.20 −.09 .14 -
12. Group - .05 .07 −.06 .07 .01 .02 −.07 .21 .14 .12 .07 -
Open in a new tab

Note. BL = baseline: 3M = 3-month: 6M = 6-month. AS = avoidant problem-solving style: DP = depressive symptom severity; SI = suicidal ideation severity; Age = participant’s age at baseline: Sex = participant’s sex (0 = male; 1 = female); Group = Treatment condition (0 = treatment as usual; 1 = experimental condition).

*

p < .05.

**

p < .01.

Multicollinearity can significantly bias CLPM path estimates when: 1) sample size is small (i.e., N < 100); 2) correlations between endogenous variables are > .60; and 3) reliability of measures is low (∝ < .80; Grewal et al., 2004; Mason & Perrault, 1992). Alternatively, VIF and tolerance values greater than > 10 and < .10, respectively, indicate high multicollinearity (Cohen et al., 2003). Although significant correlations between variables of interest ranged from .23 – .70, the present sample size (N = 110) was larger than established guidelines and all measures demonstrated good to excellent reliability across all time points. Furthermore, all VIF (range: 1.59 – 3.06) and tolerance (range: .33 – .70) values were below established thresholds for high multicollinearity. Taken together, these results suggest that multicollinearity did not significantly bias parameter estimates.

3.2. Missing Data

Of the total 110 youth, 15% had missing 3-month SI severity data and 16% at 6-months. Moreover, 20% had missing 3-month depressive symptom severity data and 15.5% at 6-months. Lastly, 24.5% had missing 3-month avoidant style data and 19.1% at 6-months. After correcting for multiple comparisons, adolescents with complete versus missing SI severity data at 3- or 6-months did not differ significantly on SI severity at baseline (p’s > .05). Additionally, youth with complete versus missing SI severity data at 3- or 6-months did not significantly differ across demographic variables (p’s > .05), avoidant style (p’s > .05) or depressive symptom severity (p’s > .05) across anytime points.

3.3. Model Comparisons

Using a nested comparison approach (see Table 2), fit for each of the five models was compared to select the most parsimonious. Fit for Model 1, which estimated all hypothesized cross-lagged paths, was acceptable: S-Bχ2(8) = 13.48, p = .09; CFI = .98; TLI = .92; RMSEA = .08. Model 2 fit the data equally well (Δχ2(1) = 0.15, p = .71) and was thus selected as the next point of comparison. Relative to Model 2, Model 3 also fit the data equally well (Δχ2(4) = 3.30, p = .51) as did Model 4 relative to Model 3 (Δχ2(1) = 3.42, p = .06). Relative to Model 4, Model 5 demonstrated significantly worse fit (Δχ2(2) = 10.09, p = .01), suggesting that Model 5, while more parsimonious than Model 4, represented the data poorly. Therefore, Model 4, which demonstrated good fit (S-Bχ2(14) = 13.48, p = .13; CFI = .98; TLI = .94; RMSEA = .06), was selected as the most parsimonious and accurate representation of the data. To provide a conservative test of study hypotheses using Model 4, age, sex, and treatment condition were regressed on all variables of interest across every time point. This final model (see Figure 2) represented the data well (S-Bχ2(14) = 21.36, p = .09; CFI = .98; TLI = .92; RMSEA = .07) and accounted for 50% of the variance in 6-month SI severity (R2 = .50, p < .001).

Table 2.

Nested model comparisons.

Model number & descriptions Model fit Model comparison
S-Bχ2 (df) p CFI TLI RMSEA AIC BIC S-BΔχ2 (Δdf) p
1. Reciprocal model with all paths included 13.48 (8) .09 .98 .92 .08 5929.03 5907.89 - -
2. Constrained path (a) that is not relevant to study hypotheses: 13.66 (9) .14 .98 .95 .07 5927.17 5906.49 0.15 (1) .71
  BL AS → 6M AS
3. Constrained paths (b) that estimate direct effects in original and alternative models: 17.17 (13) .19 .98 .96 .05 5922.14 5903.29 3.30 (4) .51
  BL AS → 3M SI
  BL AS → 6M SI
  BL DP → 3M SI
  BL DP → 6M SI
4. Constrained path (c) that is not relevant to mediation hypotheses in alternative model: 20.13 (14) .13 .98 .94 .06 5922.70 5904.32 3.42 (1) .06
  3M DP → 6M AS
5. Constrained alternative mediation paths (d): 30.14 (16) .02 .95 .90 .09 5928.23 5910.76 10.09 (2) .01
  BL DP → 3M AS
  3M AS → 6M SI
Open in a new tab

Note. a = path constrained to create Model 2 (see Figure 1); b = paths constrained to create Model 3; c = path constrained to create Model 4; d = paths constrained to create Model 5. BL = baseline: 3M = 3-month: 6M = 6-month. AS = avoidant problem-solving style; DP = depressive symptom severity, SI = suicidal ideation severity. S-Bχ2 Satorra-Bentler adjusted chi-square: df = degrees of freedom; CFI = comparative fit index; TLI = Tucker-Lewis index; RMSEA = root-mean-square error of approximation; AIC = Akaike information criteria; BIC = sample-size adjusted Bayesian information criteria; Δ = change in parameter.

3.4. Path Estimates

3.4.1. Covariates.

Table 3 provides an overview of all direct path estimates from covariates to the variables of interest within the final model (see Figure 2). At baseline, older (versus younger) adolescents reported higher levels of avoidant style (β = .21, p = .001) as well as significantly greater depressive symptom (β = .26, p = .005) and SI severity (β = .31, p < .001). At baseline, females (relative to males) reported lower levels of avoidant style (β = −.27, p = .002) and greater depressive symptom severity (β = .29, p = .001). Age and sex, however, were not significantly associated with any of the variables of interest at 3- or 6-months. Treatment condition was not significantly associated with any variables of interest across any time point.

Table 3.

Unstandardized and standardized parameter estimates of direct paths between covariates and variables of interest at each time point in the final model.

Paths β b (se) 95% CI
Age →
 BL AS .27* 1.71 (0.56) .10 – .43
 3M AS −.13 −0.79 (0.65) −.35 – .07
 6M AS −.09 −0.42 (0.42) −.26 – .08
 BL DP .26* 2.13 (0.79) .07 – .43
 3M DP .02 0.13 (0.61) −. 17 – .21
 6M DP −.01 −0.01 (0.68) −.21 – .23
 BL SI .31* 5.13 (1.57) .13 – .47
 3M SI −.01 −0.02 (1.02) −.22 – .23
 6M SI −.11 −0.01 (0.68) −.33 – .08
Sex →
 BL AS −.27* −4.03 (1.33) −.43 – −.09
 3M AS −.17 1.22 (1.47) −.05 – .36
 6M AS .04 0.40 (1.15) −.16 – .24
 BL DP .29* 5.70 (1.77) .12 – .46
 3M DP .08 1.22 (1.47) −.25 – .07
 6M DP −.05 −0.64 (1.45) −.25 – .16
 BL SI .15 5.96 (3.78) −.04 – .35
 3M SI .01 0.12 (1.99) −.18 – .18
 6M SI −.02 −0.64 (1.45) −.20 – .17
Treatment Condition →
 BL AS −.08 −1.22 (1.35) −.27 – .09
 3M AS .17 2.34 (1.51) −.05 – .36
 6M AS .06 0.63 (1.07) −.14 – .24
 BL DP .02 0.43 (1.71) −.15 – .19
 3M DP −.09 −1.35 (1.24) −.25 – .07
 6M DP .01 0.08 (1.24) −.17 – .18
 BL SI .01 0.24 (3.62) −.18 – .19
 3M SI .01 0.21 (1.76) −.16 – .17
 6M SI −.06 −1.76 (2.55) −.22 – .12
Open in a new tab

Note. BL = baseline; 3M = 3-month; 6M = 6-month. AS = avoidant problem-solving style; DP = depressive symptom severity; SI = suicidal ideation severity. CI = confidence interval for standardized parameter estimates.

*

p < .05.

3.4.2. Contemporaneous paths.

Contemporaneous associations between variables of interest varied prospectively (see Table 4). At baseline, levels of avoidant style were positively associated with depressive symptom (r = .37, p < .001) and SI severity (r = .22, p = .02). Depressive symptom and SI severity were similarly related (r = .60, p < .001). At 3-months, levels of avoidant style were not significantly related to depressive symptom (r = .18, p = .157) or SI severity (r = .17, p = .157). However, depressive symptom and SI severity were positively associated (r = .43, p < .001). At 6-months, levels of avoidant style were not significantly related to depressive symptom (r = .06, p = .658) or SI severity (r = .02, p = .893). Depressive symptoms and SI severity were also positively associated (r = .60, p < .001).

Table 4.

Unstandardized and standardized parameter estimates of autoregressive, cross-lagged, contemporaneous, and indirect paths between levels of avoidant style as well as depressive symptom and suicidal ideation severity in the final model.

Paths β b (se) 95% CI
Autoregressive
 BL AS → 3M AS .32* 0.30 (0.12) .07 – .54
 3M AS→ 6M AS .56* 0.44 (0.09) .36 – .72
 BL DP → 3M DP .54* 0.42 (0.08) .36 – .70
 BL DP → 6M DP .27* 0.20 (0.08) .05 – .48
 3M DP → 6M DP .51* 0.49 (0.12) .27 – .72
 BL SI → 3M SI .62* 0.34 (0.06) .44 – .75
 BL SI → 6M SI .52* 0.13 (0.19) .28 – .74
 3M SI → 6M SI .09 0.38 (0.11) −.22 – .34
Cross-lagged
 BL AS → 3M DP .17* 0.17 (0.08) .01 – .32
 3M DP → 6M SI .33* 0.64 (0.28) .07 – .55
 BL DP → 3M AS .29 0.21 (0.11) −.01 – .55
 3M AS → 6M SI −.20 −0.42 (0.25) −.39 – .01
 3M AS → 6M DP −.20 −0.21 (0.11) −.42 – .01
Indirect
 BL AS → 3M DP → 6M SI .06* 0.11 (0.07) .01 – .14
Contemporaneous r 95% CI
 BL AS ↔ BL DP .37* .14 – .50
 BL AS ↔ BL SI .22* .03 – .40
 BL DP ↔ BL SI .60* .41 – .73
 3M AS ↔ 3M DP .18 −.20 – .32
 3M AS ↔ 3M SI .17 −.06 – .41
 3M DP ↔ 3M SI .43* .25 – .59
 6M AS ↔ 6M DP .06 −.20 – .32
 6M AS ↔ 6M SI .02 −.21 – .25
 6M DP ↔ 6M SI .60* .36 – .73
Open in a new tab

Note. BL = baseline; 3M = 3-month; 6M = 6-month. AS = avoidant problem-solving style; DP = depressive symptom severity; SI = suicidal ideation severity. CI = confidence interval for standardized parameter estimates. Bolded paths are part of the hypothesized indirect effect.

*

p < .05.

3.4.3. Autoregressive paths.

While simultaneously accounting for effects attributable to covariates as well as contemporaneous and cross-lagged associations between avoidant style, depressive symptoms, and SI, temporal stabilities for the variables of interest were estimated and tested with autoregressive paths (see Table 4). All autoregressive paths between these variables were significant, with the exception of the path between SI severity at 3- and 6-months (β = .09, p = .522).

3.4.4. Cross-lagged paths.

Cross-lagged paths suggested only one direction for effects (see Table 4). Specifically, baseline levels of avoidant style positively predicted 3-month depressive symptom severity (β = .17, p = .024). However, 3-month levels of avoidant style did not significantly predict 6-month depressive symptom severity (β = −.20, p = .056). Three-month depressive symptom severity significantly predicted 6-month SI severity (β = .33, p = .008). In contrast, paths between baseline depressive symptom severity to 3-month levels of avoidant style (β = .29, p = .067) as well as 3-month levels of avoidant style to 6-month SI severity (β = −.20, p = .096) were not significant.

3.4.5. Indirect path.

The indirect path was examined to test whether 3-month depressive symptom severity indirectly affected the relationship between baseline levels of avoidant style and 6-month SI severity. This indirect path was significant: β = .06, 95% CI [.01, .14].

4. Discussion

A robust cross-sectional relationship exists between social problem-solving deficits and adolescent SI. However, there is a paucity of research examining whether different aspects of social problem-solving deficits, such as avoidant style, are uniquely related to suicide risk over time, as well as potential mechanisms underlying this relationship. Drawing on existing literature in this area and suicide theory, the primary purpose of this study was to evaluate the potential role of depressive symptom severity as a mediator between avoidant style and SI severity using a longitudinal, multi-wave design with a clinical adolescent sample. The secondary aim of this study was to evaluate the specificity of depressive symptom severity as a potential mediator.

Consistent with study hypotheses based on the original model, 3-month depression severity mediated the association between baseline avoidant style and 6-month SI severity. Specifically, youth with greater baseline levels of avoidant style experienced greater depressive symptoms at month 3, which then resulted in greater SI severity by month 6. These results are generally consistent with previous cross-sectional research which suggests that avoidant style may be indirectly associated with greater adolescent SI severity through depressive symptom severity (Reinecke et al., 2001).

Contrary to study hypotheses for the alternative model, 3-month avoidant style did not mediate the relationship between baseline depressive symptoms and 6-month SI. Specifically, youth with greater baseline depressive symptom severity did not demonstrate more avoidant style at 3 months, which did not result in greater SI severity by 6 months. Taken together, these findings suggest that this alternative model was not an accurate representation of the longitudinal relationship among these three variables. Furthermore, the non-significant associations of all hypothesized paths in this alternative model provide evidence for the specificity of depressive symptom severity as a mediator between avoidant style and SI severity.

This pattern of results is consistent with the 3ST (Klonsky & May, 2015), which highlights the impact of avoidance on depressive symptoms, such as depressed mood, anhedonia, and worthlessness. When combined with a lack of connectedness and hopelessness, these symptoms contribute to the onset and escalation of SI (Klonsky & May, 2015). However, lack of connectedness and hopelessness was not explicitly tested in our current model. This is particularly important in light of the fact that not all adolescents who experience depressive symptoms develop SI (Hawton et al., 2012). Future multi-wave, longitudinal studies should incorporate constructs such as connectedness and hopelessness in examinations of the relation among avoidant style, depressive symptoms, and SI.

Overall, this study extends previous research in several ways. Whereas most previous research examines the relationship between problem-solving deficits generally and suicidality using one time point, the current study assessed the relationship between avoidant style, one facet of problem-solving deficits, and SI severity across three points. Importantly, the longitudinal relationship between these variables has not been previously examined, precluding an understanding of potential mechanisms, such as depressive symptoms, underlying the relationship between avoidant style and SI severity (Speckens & Hawton, 2005). Relatedly, the specificity of depressive symptom severity had not been evaluated. To address these gaps in the literature, the current study assessed these constructs using well-validated measures at three points across 6 months, allowing for a test of the direction of effects over time while accounting for the contemporaneous relationships between variables of interest. Age and sex were also controlled for in the final optimized model, two factors that influence the variables of interest. The final model accounted for half of the variance in SI severity at 6-months. Taken together, these findings highlight the important role that depressive symptoms play in the relation between avoidant style and SI severity.

4.1. Clinical Implications

The findings of the present study hold important implications for work with adolescents in clinical contexts. First, incorporating an assessment of problem-solving abilities at the start of treatment with adolescents with depressive symptoms may help inform treatment planning. Those with an avoidant style may benefit from the receipt of evidence-based treatments that help build effective problem-solving strategies, such as cognitive-behavioral therapy, which commonly includes instruction in problem-solving skills, to help reduce depressive symptoms and subsequent SI (see Zhou et al., 2015). Relatedly, use of newer therapies that help reduce the use of avoidant behavior through the clarification of values and contact with the present moment, such as acceptance and commitment therapy (ACT; Hayes et al., 2006), could also be beneficial. Indeed, preliminary work with adolescents suggests that ACT is a potentially efficacious treatment for depressive symptoms in youth (Hacker et al., 2016).

4.2. Limitations and Future Directions

Although this study has several strengths, including a multi-wave design, the use of repeated, well-validated, self-report measures, and stringent analytic strategy, results should be interpreted within the context of some limitations. First, study results may be limited in generalizability. Though the use of a clinical sample of adolescents at heightened risk for SI is another clear strength, it is unclear whether study results generalize to less severe samples and/or among youth not currently receiving mental health services. Relatedly, the current sample size may have precluded detection of all significant associations among study variables. Additionally, the sample was predominantly Caucasian and of non-Latino ethnicity. Thus, future research should include larger and more diverse samples in similar studies. Second, this study focused exclusively on SI severity. It is unclear whether or not avoidant style may confer risk for suicide attempts via depressive symptom severity. The small number of participants who endorsed suicide attempts barred examination of this hypothesis. Third, while using path analysis to evaluate the temporal direction of effects is a significant strength, causality cannot be inferred from such models (Cohen et al., 2003). Last, the present study included only self-report measures. Future research is warranted using multi-method approaches, including behavioral paradigms to assess problem-solving abilities.

Limitations notwithstanding, this study is among the first to examine the prospective relationship between avoidant style as well as depressive symptom and SI severity using a multi-wave design with a clinical sample of adolescents. Findings indicate that depressive symptom severity may account for the relationship between avoidant style and SI severity. Clinically, results suggest that treatments aimed at reducing use of avoidant problem-solving style could help reduce depressive symptoms and thus decrease suicide risk among adolescents.

Supplementary Material

1

Highlights.

  • Avoidant problem-solving style predicted depressive symptom severity prospectively

  • Depressive symptom severity predicted suicidal ideation severity prospectively

  • The indirect path through depressive symptoms was significant

  • The specificity of depressive symptoms as a mediator was supported

  • Results held after accounting for age, sex, and baseline levels of study variables

Acknowledgements:

The authors thank NIAAA for the financial support but recognize that the findings and conclusions are those of the authors and do not necessarily reflect the opinions of NIAAA.

This research was funded by a grant from the National Institute on Alcohol Abuse and Alcoholism (NIAAA; R01AA020705). The NIAAA was not involved in the preparation of this manuscript or in the decision to submit it for publication.

Funding: This work was supported by the National Institute on Alcohol Abuse and Alcoholism (NIAAA; grant number R01AA020705).

Footnotes

Declaration of interests: None.

2

Although the SIQ-JR assesses past month SI, one of the anchors assesses lifetime experiences (i.e., “I had this thought before but not in the past month). To determine if this anchor confounded study results, all models were re-run after altering the value of this anchor (i.e., from “1” to “0”). An equivalent pattern of results was found.

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 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.

References

  1. Adrian M, Miller AB, McCauley E, & Stoep AV (2016). Suicidal ideation in early to middle adolescence: Sex-specific trajectories and predictors. Journal of Child Psychology and Psychiatry, 57(5), 645–653. 10.1111/jcpp.12484 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Baron RM, & Kenny DA (1986). The moderator-mediator variable distinction in social psychological research: Conceptualization, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173–1182. https://doi:10.1037/0022-3514.51.6.1173 [DOI] [PubMed] [Google Scholar]
  3. Becker-Weidman EG, Jacobs RH, Reinecke MA, Silva SG, & March JS (2010). Social problem-solving among adolescents treated for depression. Behaviour Research and Therapy, 48(1), 11–18. 10.1016/j.brat.2009.08.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Benjamini Y, & Hochberg Y (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B (Methodological), 57, 289–300. 10.1111/j.2517-6161.1995.tb02031.x [DOI] [Google Scholar]
  5. Bentler PM, & Bonett DG (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88(3), 588–606. 10.1037/0033-2909.88.3.588 [DOI] [Google Scholar]
  6. Brereton A, & McGlinchey E (2019). Self-harm, emotion regulation, and experiential avoidance: A systematic review. Archives of Suicide Research, 0, 1–23. 10.1080/13811118.2018.1563575 [DOI] [PubMed] [Google Scholar]
  7. Calati R, Ferrari C, Brittner M, Oasi O, Olié E, Carvalho AF, & Courtet P (2019). Suicidal thoughts and behaviors and social isolation: A narrative review of the literature. Journal of Affective Disorders, 245, 653–667. 10.1016/j.jad.2018.11.022 [DOI] [PubMed] [Google Scholar]
  8. Centers for Disease Control and Prevention (2017). Preventing Suicide. Retrieved from: https://www.cdc.gov/violenceprevention/suicide/fastfact.html
  9. Chang EC (2017). Applying the broaden-and-build model of positive emotions to social problem-solving: Does feeling good (vs. feeling bad) influence problem orientation, problem-solving skills, or both? Journal of Social and Clinical Psychology, 36(5), 380–395. 10.1521/jscp.2017.36.5.380 [DOI] [Google Scholar]
  10. Clum GA, & Febbraro GAR (1994). Stress, social support, and problem-solving appraisal/skills: Prediction of suicide severity within a college sample. Journal of Psychopathology and Behavioral Assessment, 16(1), 69–83. 10.1007/BF02229066 [DOI] [Google Scholar]
  11. Cohen J, Cohen P, West SG, & Aiken LS (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum. [Google Scholar]
  12. D’Zurilla TJ, Maydeu-Olivares A, & Kant GL (1998). Age and gender differences in social problem-solving ability. Personality and Individual Differences, 25, 241–252. 10.1016/S0191-8869(98)00029-4 [DOI] [Google Scholar]
  13. D’Zurilla TJ, & Nezu A (1982). Social problem-solving in adults In Kendal PC (Ed.), Advances in Cognitive-Behavioral Research and Therapy (Vol. 1). San Diego, CA: Academic Press. [Google Scholar]
  14. Fredrickson BL (2001). The role of positive emotions in positive psychology: The broaden- and-build theory of positive emotions. American Psychologist, 56(3), 218–226. 10.1037/0003-066X.56.3.218 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gould MS, Greenberg T, Velting DM, & Shaffer D (2003). Youth suicide risk and preventive interventions: A review of the past 10 years. Journal of the American Academy of Child and Adolescent Psychiatry, 42(4), 386–405. 10.1097/01.CHI.0000046821.95464.CF [DOI] [PubMed] [Google Scholar]
  16. Grewal R, Cote JA, & Baumgartner H (2004). Multicollinearity and measurement error in structural equation models: Implications for theory testing. Marketing Science, 23(4), 519–529. 10.1287/mksc.1040.0070 [DOI] [Google Scholar]
  17. Hacker T, Stone P, & MacBeth A (2016). Acceptance and commitment therapy – do we know enough? Cumulative and sequential meta-analyses of randomized controlled trials. Journal of Affective Disorders, 190, 551–565. 10.1016/j.jad.2015.10.053 [DOI] [PubMed] [Google Scholar]
  18. Hayes AF (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 78(4), 408–420. https://doi:10.1080/03637750903310360 [Google Scholar]
  19. Hayes SC, Luoma JB, Bond FW, Masuda A, & Lillis J (2006). Acceptance and commitment therapy: Model, processes, and outcomes. Behaviour Research and Therapy, 44,1–25. 10.1016/j.brat.2005.06.006 [DOI] [PubMed] [Google Scholar]
  20. Hawton K, Saunders KEA, & O’Connor RC (2012). Self-harm and suicide in adolescents. The Lancet, 379, 2373–82. 10.1016/S0140-6736(12)60322-5 [DOI] [PubMed] [Google Scholar]
  21. Hu L, & Bentler PM, (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1–55. 10.1080/10705519909540118 [DOI] [Google Scholar]
  22. Kaufman AS (1990). Kaufman Brief Intelligence Test: K-BIT. Circle Pines, MN: American Guidance Service. [Google Scholar]
  23. King CA, Katz SH, Ghaziuddin N, Brand E, Hill E, & McGovern L (1997). Diagnosis and assessment of depression and suicidality using the NIMH Diagnostic Interview Schedule for Children (DISC-2.3). Journal of Abnormal Child Psychology, 25(3), 25(3), 173–181. 10.1023/A:1025739730823 [DOI] [PubMed] [Google Scholar]
  24. Klonsky ED, & May AM (2015). The three-step theory (3ST): A new theory of suicide rooted in the “ideation-to-action” framework. International Journal of Cognitive Therapy, 8(2), 114–129. 10.1521/ijct.2015.8.2.114 [DOI] [Google Scholar]
  25. Kovacs M (2011). Children’s Depression Inventory 2 (CDI 2) (2nd Ed.). North Tonawanda, NY: Multi-Health Systems, Inc. [Google Scholar]
  26. Little TD (2013). Longitudinal Structural Equation Modeling. New York, NY: Guilford Press. [Google Scholar]
  27. MacKinnon DP, Lockwood CM, & Williams J (2004). Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39(1), 99–128. 10.1207/s15327906mbr3901_4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Mason CH, & Perreault WD Jr. (1992). Collinearity, power, and interpretation of multiple regression analyses. Journal of Marketing Research, 28(3), 268–280. https://doi.org/10.1177%2F002224379102800302 [Google Scholar]
  29. Matthews T, Danese A, Wertz J, Odgers CL, Ambler A, Moffit TE, & Arseneault L (2016). Social isolation, loneliness, and depression in young adulthood: A behavioral genetic analysis. Social Psychiatry and Psychiatric Epidemiology, 51(3), 339–348. 10.1007/s00127-016-1178-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Maydeu-Olivares A, & D’Zurilla TJ (2002). Manual for the Social Problem-Solving Inventory-Revised. North Tonawanda, NY: Multi-Health Systems. [Google Scholar]
  31. McDermott LM, & Ebmeier KP (2009). A meta-analysis of depression severity and cognitive function. Journal of Affective Disorders, 119, 1–8. 10.1016/j.jad.2009.04.022 [DOI] [PubMed] [Google Scholar]
  32. Muthén LK, & Muthén BO (1998-2017). Mplus User’s Guide (8th Ed). Los Angeles, CA: Muthén & Muthén. [Google Scholar]
  33. Nock MK, Green JG, Hwang I, McLaughlin KA, Sampson NA, Zaslavsky AM, & Kessler RC (2013). Prevalence, correlates, and treatment of lifetime suicidal behavior among adolescents: Results from the national Comorbidity Survey Replication Adolescent Supplement. JAMA Psychiatry, 70, 300–310. https://doi:10.1001/2013.jamapsychiatry.55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nock MK, & Kazdin AE (2002). Examination of affective, cognitive, and behavioral factors and suicide-related outcomes in children and young adolescents. Journal of Clinical Child and Adolescent Psychology, 31(1), 48–58. 10.1207/S15374424JCCP3101_07 [DOI] [PubMed] [Google Scholar]
  35. Preacher KJ, & Hayes AF (2008). Asymptomatic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879–891. 10.3758/BRM.40.3.879 [DOI] [PubMed] [Google Scholar]
  36. Reinecke MA, DuBois DL, & Schultz TM (2001). Social problem solving, mood, and suicidality among inpatient adolescents. Cognitive Therapy and Research, 25(6), 743–756. 10.1023/A:1012971423547 [DOI] [Google Scholar]
  37. Reinecke MA, Ryan NE, & DuBois DL (1998). Cognitive-behavioral therapy of depression and depressive symptoms during adolescence: A review and meta-analysis. Journal of the American Academy of Child & Adolescent Psychiatry, 37(1), 26–34. 10.1097/00004583-199801000-00013 [DOI] [PubMed] [Google Scholar]
  38. Reynolds WM (1987). Suicidal Ideation Questionnaire. Odessa, FL: Psychological Assessment Resources. [Google Scholar]
  39. Reynolds WM, & Mazza JJ (1999). Assessment of suicidal ideation in inner-city children and young adolescents: Reliability and validity of the Suicidal Ideation Questionnaire-JR. School Psychology Review, 28(1), 17–30. [Google Scholar]
  40. Sadowski C, Moore LA, Kelley ML (1994). Psychometric properties of the Social Problem Solving Inventory (SPSI) with normal and emotionally disturbed adolescents. Journal of Abnormal Child Psychology, 22(4), 487–500. 10.1007/BF02168087 [DOI] [PubMed] [Google Scholar]
  41. Satorra A, & Bentler PM (2010). Ensuring positiveness of the scaled difference chi-square test statistic. Psychometrika, 75(2), 243–248. 10.1007/s11336-009-9135-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Schäfer JO, Naumann E, Holmes EA, Tuschen-Caffier B, & Samson AC (2017). Emotion regulation strategies in depressive and anxiety symptoms in youth: A meta-analytic review. Journal of Youth and Adolescence, 46, 261–276. 10.1007/s10964-016-0585-0 [DOI] [PubMed] [Google Scholar]
  43. Scott K, Lewis CC, & Marti N (2019). Trajectories of symptom change in the treatment for adolescents with depression study. Journal of the American Academy of Child & Adolescent Psychiatry, 58(3), 319–328. 10.1016/j.jaac.2018.07.908 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Seiffge-Krenke I, & Klessinger N (2000). Long-term effects of avoidant coping on adolescents’ depressive symptoms. Journal of Youth and Adolescence, 29(6), 617–630. 10.1023/A:1026440304695 [DOI] [Google Scholar]
  45. Selig JP, & Preacher KJ (2009). Mediation models for longitudinal data in developmental research. Research in Human Development, 6(2-3), 144–164. 10.1080/15427600902911247 [DOI] [Google Scholar]
  46. Sheppes G, Suri G, & Gross JJ (2015). Emotion regulation and psychopathology. Annual Review of Clinical Psychology, 11, 379–405. 10.1146/annurev-clinpsy-032814-112739 [DOI] [PubMed] [Google Scholar]
  47. Snyder HR (2013). Major depressive disorder is associated with broad impairments on neuropsychological measures of executive function: A meta-analysis and review. Psychological Bulletin, 139(1), 81–132. 10.1037/a0028727 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Speckens AEM, & Hawton K (2005). Social problem solving in adolescents with suicidal behavior: A systematic review. Suicide and Life-Threatening Behaviors, 35(4), 365–387. 10.1521/suli.2005.35.4.365 [DOI] [PubMed] [Google Scholar]
  49. Spirito A & Esposito-Smythers C (2008). Evidence based therapies for adolescent suicidal behavior In Steele R, Elkin D, & Roberts M (Eds.), Handbook of evidence-based therapies for children and adolescents (pp. 177–194). New York, NY: Springer. [Google Scholar]
  50. Tennant C (2002). Life events, stress, and depression: A review of recent findings. Australian and New Zealand Journal of Psychiatry, 36, 173–182. 10.1046/j.1440-1614.2002.01007.x [DOI] [PubMed] [Google Scholar]
  51. Trew JL (2011). Exploring the roles of approach and avoidance in depression: An integrative model. Clinical Psychology Review, 31, 1156–1168. 10.1016/j.cpr.2011.07.007 [DOI] [PubMed] [Google Scholar]
  52. Turecki G, & Brent DA (2016). Suicide and suicidal behavior. Lancet, 387, 1227–1239. 10.1016/S0140-6736(15)00234-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Van Orden KA, Witte TK, Cukrowicz KC, Braithwaite SR, Selby EA, & Joiner TE (2010). The interpersonal theory of suicide. Psychological Review, 117, 575–600. 10.1037/a0018697 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Wolff JC, Davis S, Liu RT, Christine CB, Cheek SM, Nestor BA, Frazier EA, Schaffer MM, & Spirito A (2018). Trajectories of suicidal ideation among adolescents following psychiatric hospitalization. Journal of Abnormal Child Psychology, 46, 355–363. 10.1007/s10802-017-0293-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Zelazo PD, & Cunningham WA (2007). Executive function: Mechanisms underlying emotion regulation In Gross JJ (Ed.), Handbook of emotion regulation (pp. 135–158). New York, NY, US: The Guildford Press. [Google Scholar]
  56. Zhou X, Hetrick SE, Cuijpers P, Qin B, Barth J, Whittington CJ, Cohen D, Del Giovane C, Liu Y, Michael KD, Zhang Y, Weisz JR, & Xie P (2015). Comparative efficacy and acceptability of psychotherapies for depression in children and adolescents: A systematic review and network meta-analysis. World Psychiatry, 14, 207–222. 10.1002/wps.20217 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1600557-supplement-1.docx (16.9KB, docx)

RESOURCES