Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jan 3.
Published in final edited form as: J Clin Psychol. 2021 Aug 10;77(12):2978–2993. doi: 10.1002/jclp.23236

Depression-related emotional problems mediate the relation between hopelessness and suicidal ideation severity

Roberto López Jr 1, Lia Follet 2, Annamarie B Defayette 1, Emma D Whitmyre 1, Jennifer Wolff 3, Anthony Spirito 3, Christianne Esposito-Smythers 1
PMCID: PMC8722668  NIHMSID: NIHMS1763315  PMID: 34378203

Abstract

Objective:

Cross-sectional research with adult samples suggests that hopelessness may indirectly affect suicidal ideation (SI) through overall depressive symptom severity. However, particular depressive symptom constellations, rather than overall symptoms, may underlie the association between hopelessness and SI. Yet, the cross-sectional nature of these studies precludes examination of the temporal associations among these constructs.

Methods:

Using path analysis, the present study examined whether depression-related emotional problems mediate the relation between hopelessness and SI in a clinical sample of 110 adolescents over a 6-month period. The specificity of depression-related emotional problems as a mediator was also evaluated.

Results:

After accounting for covariates, results supported the specificity of 3-month depression-related emotional problems as a mediator of the association between baseline levels of hopelessness and 6-month SI.

Conclusion:

Results suggest that treatment targeted specifically at hopelessness may help reduce depression-related emotional problems and lower SI, and ultimately, adolescent suicide risk.

Keywords: adolescent, depressive symptoms, hopelessness, suicide

1 ∣. INTRODUCTION

Within the last decade, death by suicide among adolescents ages 15–19 has increased substantially, with rates increasing by a third among males and doubling among females (Centers for Disease Control and Prevention, 2017). Suicidal ideation (SI) confers significant risk for suicidal behavior. For instance, over one-third of adolescents receiving mental health services who experience SI attempt suicide, with the transition occurring one to 2 years after the onset of SI (Glenn et al., 2017). Therefore, understanding mechanisms underlying the onset and maintenance of SI during adolescence is crucial for intervening in this at-risk population.

1.1 ∣. Hopelessness, depression, and SI

Hopelessness has been defined as a pessimistic view of the future as well as a view of the self as helpless in the face of stressors (Liu et al., 2015). The Hopelessness Theory of Suicide (Abramson et al., 1998) posits that hopelessness plays a central role in the onset and/or maintenance of SI. This theory suggests that hopelessness increases risk for suicide by eliciting and/or exacerbating psychiatric symptoms, from which individuals may desire to escape. Consistent with theory, many studies, particularly those which are cross sectional in nature, have linked hopelessness to severity of SI in community and clinical adolescent samples (for reviews, see Cha et al., 2018; Ribeiro et al., 2018). However, several longitudinal studies have failed to demonstrate hopelessness as a significant predictor of SI, particularly after controlling for depressive symptoms (Cha et al., 2018). While methodological differences between studies may account for nonsignificant results, it is also possible that a third variable, such as depressive symptoms, may account for the relation between hopelessness and severity of SI.

Both theory and research support the role of hopelessness in the onset and development of depressive symptoms among adolescents (for a review, see Liu et al., 2015). Existing literature suggests that hopelessness contributes to the onset of depressive symptoms, such as depressed mood, sleep disturbances, self-blame, and reduced self-esteem (Liu et al., 2015). There also exists a plethora of research to suggest that depressive symptoms are associated with SI using both cross-sectional and longitudinal research designs (Cha et al., 2018). In accordance with the Escape Theory of Suicide (Baumeister, 1990), aversive self-referential cognitions and emotions found to be associated with depressive symptoms (Hards et al., 2020) may be experienced as unbearable and consequently drive the motivation to end one's distress through suicide. Yet, to our knowledge, no studies have examined depressive symptoms as a mediator of the association between hopelessness and SI. Two separate studies have found that hopelessness was indirectly associated with SI through depressive symptoms in samples of adolescents and young adults (Spann et al., 2006; Wang et al., 2015). However, both studies were cross-sectional in nature and thus precluded examination of temporal associations between these variables. Ballard and colleagues' (2015) cross-sectional study with young adults also found support for the indirect association between hopelessness and SI through depressive symptoms, though these researchers tested specific depressive symptom constellations. Specifically, hopelessness indirectly affected SI through depression-related emotional problems (e.g., negative mood, guilt).

1.2 ∣. Depressive symptom constellations

The study of depressive symptoms as a unitary construct in relation to suicidality is commonplace, yet such practices may obscure unique associations between these constructs (Cha et al., 2018). For instance, cross-sectional research with adult samples has found that ratings of emotional problems associated with depression (e.g., negative mood, guilt) are more strongly related to SI than overall depressive symptom severity ratings (Ballard et al., 2015; Keilp et al., 2012). Longitudinal work suggests that prospective changes in such symptoms, in addition to neurovegetative symptoms (e.g., fatigue, sleep disturbances), are most strongly linked to changes in SI severity (Keilp et al., 2018). Similar research supports the differential association between particular depressive symptom clusters, as measured by the Children's Depression Inventory (CDI; Kovacs, 1985), and SI among youth (e.g., Esposito & Clum, 1999; Ivarsson et al., 2006).

The CDI, now in its second iteration (CDI-2; Kovacs, 2011), is among the most widely used measures of adolescent depressive symptoms (Stumper et al., 2019). Derived using factor analysis, the CDI-2 contains two scales, depression-related emotional problems, and functional problems. The depression-related emotional problems scale is comprised of two subscales: negative mood/physical symptoms and negative self-esteem. The functional problems scale is comprised of the interpersonal problems and ineffectiveness subscales. Depressive symptoms captured by the emotional problems scale include negative mood, decreased self-esteem, fatigue, and sleep disturbances. This study specifically examines depression-related emotional problems, rather than functional problems, because there is existing research on the unique relations between specific depression-related emotional problems and SI in cross-sectional work with adolescents and adults (e.g., Ballard et al., 2015; Esposito & Clum, 1999), as well as longitudinal research with adults (Keilp et al., 2018). Thus, it is possible that depression-related emotional problems mediate the association between hopelessness and SI severity among adolescents.

1.3 ∣. Alternative model

Though theoretical and empirical work support the role of depression-related emotional problems as a potential mediator in the association between hopelessness and SI, it is also possible that depression-related emotional problems lead to hopelessness, which in turn, predict SI severity. Existing research suggests a bidirectional relationship between general depressive symptoms and hopelessness (Gibb & Alloy, 2006). Furthermore, as described previously, hopelessness and SI severity are closely linked (Ribeiro et al., 2018). Thus, it is also possible that hopelessness mediates the association between depression-related emotional problems and severity of SI. However, to our knowledge, no studies to date have tested this model.

1.4 ∣. Current study

The purpose of the current study is to examine depression-related emotional problems as a mediator of the association between hopelessness and severity of SI in a clinical adolescent sample. As suggested above, the association between hopelessness, depressive symptoms, and SI is well-established (Ribeiro et al., 2018), with some evidence suggesting that hopelessness indirectly affects SI through depressive symptoms (Ballard et al., 2015; Wang et al., 2015). However, there exist several limitations in this research. First, prior studies rarely examine whether particular constellations of depressive symptoms account for the association between hopelessness and SI, a step that may improve our understanding of mechanisms that underlie SI (Cha et al., 2018). Second, existing research in this area with adolescents exclusively assessed this model cross-sectionally, precluding a true test of mediation (Selig & Preacher, 2009). Third, the average follow-up length among extant studies that examined hopelessness and depression as predictors of SI in general is approximately 9 years, limiting the clinical value of this research (Ribeiro et al., 2018). To address these limitations, the current study used a multiwave, longitudinal design over the span of 6 months to assess the potential mediating role of depression-related emotional problems between hopelessness and SI severity in a clinical adolescent sample. The current study utilized a cross-lagged panel model (CLPM) design to allow for a stringent study of longitudinal processes, including temporal direction of effects, while simultaneously accounting for concurrent levels of all variables at each wave of measurement (Selig & Preacher, 2009). Further, to provide a conservative test of the specificity of the hypothesized model, an alternative model with hopelessness as a mediator of the relation between depression-related emotional problems and SI severity was tested simultaneously. The following hypotheses are offered for the original and alternative models, respectively.

First, after accounting for covariates and baseline depression-related emotional problems, greater levels of hopelessness at baseline and the 3-month follow-up will predict greater depression-related emotional problems at 3- and 6-month follow-up, respectively. Greater depression-related emotional problems at baseline and 3 months will positively predict SI severity at 3 and 6 months, respectively. Finally, depression-related emotional problems at 3 months will mediate the relation between hopelessness at baseline and SI severity at 6 months.

Second, in an alternative model, after accounting for covariates and baseline levels of hopelessness, greater levels of depression-related emotional problems at baseline and the 3-month follow-up will predict greater hopelessness at 3- and 6-month follow-up, respectively. Greater hopelessness at baseline and 3 months will positively predict SI severity at 3 and 6 months, respectively. Finally, hopelessness at 3 months will mediate the relation between depression-related emotional problems at baseline and SI severity at 6 months.

2. ∣. MATERIALS AND METHODS

2.1 ∣. Participants and procedures

Before the start of data collection, study procedures were approved by both university and hospital Institutional Review boards. The sample consisted of adolescents and caregivers enrolled in a randomized controlled trial conducted in a community setting comparing a cognitive-behavioral intervention to enhanced standard care for dually diagnosed adolescents (Wolff et al., 2020). The current study is a secondary analysis of this data and not an examination of the main outcomes of the trial.

Recruitment was completed at a community mental health facility located in the Northeastern United States. All adolescents and their legal guardians seeking intensive, home-based mental health services were asked to participate. After consent/assent was obtained, research assistants administered self-report measures to adolescents and their guardians, including at screening, baseline, and follow-up assessments. After completion of the baseline assessment, participants were assigned into either an enhanced treatment as usual condition or the experimental cognitive-behavioral treatment using urn randomization (Wolff et al., 2020). Inclusion criteria for this parent study included: (1) English-speaking adolescent and legal guardian(s); (2) adolescent assent (consent for 18 years old); (3) consent of a legal guardian; and (4) enrolled in the intensive outpatient, home-based program for co-occurring substance use and psychiatric symptoms at the clinic. Exclusion criteria included presence of severe mental health symptoms, such as 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 ∣. Hopelessness

The Hopelessness Scale for Children (HSC; Kazdin et al., 1983) is a 17-item true/false self-report questionnaire with each item (e.g., “I don't think I will get what I really want”) scored 0 (“False”) or 1 (“True”). Total scores range from 0 to 17, with higher scores reflecting greater hopelessness. The HSC has been shown to have adequate reliability and validity in clinical adolescent samples (Spirito et al., 1998). The HSC was administered to adolescents at every assessment period, and internal consistency (α) was good (0.83–0.84) across all time points.

2.2.2 ∣. Depression-related emotional problems

The Children's Depression Inventory-2 (CDI-2; Kovacs, 2011) is a 28-item self-report measure of depressive symptoms during the past 2 weeks. Items on the CDI-2 are rated on a scale from 0 to 2. Only the 13-item depression-related emotional problems subscale was used in the present study (e.g., Over the past 2 weeks, I am sad once in a while [0], many times [1], or all the time [2]). To avoid overlap between measures of depression-related emotional problems and SI, the single item assessing suicidality was excluded. Therefore, possible raw subscale scores ranged from 0 to 24. Higher scores indicate greater depression-related emotional problems. 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 (α) for the depression-related emotional problems subscale was acceptable to good (0.75–0.85) across all time points.

2.2.3 ∣. SI 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, which begin with “I thought…”, “I wondered…”, or “I wished…” followed by a statement indicative of SI (e.g., “… about death”), 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 to 90. Higher scores indicate greater SI severity. The SIQ-JR has been shown to have excellent reliability and construct validity with adolescents (Reynolds & Mazza, 1999). The SIQ-JR was administered to youth at every assessment period, and internal consistency (α) was excellent (0.93–97) across all time points.

2.3 ∣. Data analysis

This study utilized a CLPM design using Mplus version 8.4 (Muthén & Muthén, 1998–2017) to examine the contemporaneous and longitudinal relationships between hopelessness, depression-related emotional problems, and SI severity. Multicollinearity diagnostics (i.e., variance inflation factor [VIF] and tolerance) were calculated using SPSS version 19 by simultaneously regressing 6-month SI on all other variables of interest at baseline, 3 and 6 months. Since missing data patterns, particularly in the criterion variable, can influence path estimates (Cohen et al., 2003), adolescents with complete versus missing data for SI severity at 3 and 6 months were compared on SI severity, hopelessness, and depression-related emotional problems at baseline as well as age, sex, treatment condition, and race/ethnicity. The false discovery rate was set to 5% to correct for multiple comparisons (see Benjamini & Hochberg, 1995).

In line 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. Nonsignificant paths from the competing model were then constrained to equal zero per Satorra and Bentler's (2010) guidelines to obtain the most parsimonious model. Therefore, if two models fit the data equally well, the model with more constrained paths (i.e., the more parsimonious model) was selected.

A parsimonious model was found using the procedure outlined previously by testing a nested model where all nonsignificant cross-lagged paths from the competing model were constrained. Specifically, to create Model 2, the cross-lagged paths between baseline and 3-month depression-related emotional problems to 3- and 6-month hopelessness were constrained to equal zero. Additionally, cross-lags between baseline and 3-month hopelessness to 3- and 6-month SI severity were constrained to equal zero. Finally, the direct, cross-lagged path between baseline depression-related emotional problems and 6-month SI severity was also constrained to equal zero. Once the most parsimonious model was identified (see Figure 1), hopelessness, depression-related emotional problems, and SI across all time points were regressed on all covariates (i.e., age, sex, and treatment condition) to stringently test study hypotheses.1 With RMSEA as the measure of effect, an extension of the MacCallum-Browne-Sugaware framework (MacCallum et al., 2006) was used to ensure sufficient power to detect suitable fit for Model 2 before inclusion of covariates. Power estimates were conducted using a web utility (Preacher & Coffman, 2006) and Rweb. RMSEA for the baseline model (df = 8) and the most parsimonious (df = 13) were set to 0.01 and 0.09, respectively (see MacCallum et al., 2006). With α = .05 and N = 110, power approached .80 (1-β = .74).

FIGURE 1.

FIGURE 1

Open in a new tab

Final cross-lagged panel model with variables of interest. Note: BL = baseline; 3M = 3 months; 6M = 6 months. Solid black paths are statistically significant. Grey paths are statistically nonsignificant. Not pictured for ease of interpretation are the concurrent relationships between hopelessness, depression-related emotional problems, and suicidal ideation severity, as well as the direct paths from covariates to these variables at each respective time point. These direct and concurrent paths were included in the 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

All models utilized full information likelihood estimation and a robust likelihood estimator (MLR; Muthén & Muthén, 1998–2017) to account for missing data as well as potential nonnormality. Individual model fit was assessed using an adjusted chi-square statistic (see Satorra & Bentler, 2010), the comparative fit index (CFI), the Tucker-Lewis Index (TLI), and the root-mean-square error of approximation (RMSEA). A nonsignificant, adjusted chi-square suggests that the model fits the data well (Satorra & Bentler, 2010). Additionally, good fit is assumed for values ≥0.95 for CFI and TLI as well as ≤0.05 for RMSEA (Little, 2013). CFI and TLI values of 1 indicate good fit and can occur if the chi-square value is less than or equal to the degrees of freedom (Little, 2013).

Consistent with mediation literature (Little, 2013), indirect effects were evaluated if the parameter estimates for both alpha and beta paths were statistically significant. Therefore, only the indirect effect of 3-month depression-related emotional problems on the relation between baseline hopelessness and 6-month SI severity was estimated using a bias-corrected bootstrapping approach with 5000 bootstrap resamples; indirect paths are considered significant if the 95% confidence intervals do not contain zero (Little, 2013).

3 ∣. RESULTS

3.1 ∣. Descriptive, bivariate, and multicollinearity analyses

One hundred and eleven 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 as such, was excluded from analyses. Thus, the present sample consisted of 110 adolescents (Mage = 15.71; SD = 1.18; range: 13–18). The sample had slightly more males (n = 63; 57.3%) than females (n = 47; 42.7%). Participants 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 <$5000 (11.3%) to >$100,000 (11.3%), 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. All VIF (range: 1.42–2.73) and tolerance (range: 0.37–0.70) values did not cross established thresholds for high multicollinearity (VIF >10 and tolerance <0.10; Cohen et al., 2003), suggesting that multicollinearity did not significantly bias parameter estimates.

TABLE 1.

Means, SDs, and bivariate correlations for all study variables

Variables M (SD) 1 2 3 4 5 6 7 8 9 10 11 12
1. BL Suicidal ideation severity 17.16 (19.32) –
2. 3M Suicidal ideation severity 8.18 (10.65) 0.64** –
3. 6M Suicidal ideation severity 9.70 (14.51) 0.63** 0.48** –
4. BL Depression-related emotional problems 6.86 (5.22) 0.62** 0.44** 0.51** –
5. 3M Depression-related emotional problems 4.99 (3.83) 0.35** 0.49** 0.48** 0.64** –
6. 6M Depression-related emotional problems 4.54 (3.69) 0.37** 0.40** 0.73** 0.49** 0.54** –
7. BL Hopelessness 5.18 (4.09) 0.47** 0.21* 0.33** 0.51** 0.44** 0.40** –
8. 3M Hopelessness 4.46 (3.87) 0.15 0.36** 0.15 0.28* 0.45** 0.30** 0.52** –
9. 6M Hopelessness 4.03 (3.75) 0.07 −0.03 0.26* 0.23* 0.21 0.36** 0.35** 0.33** –
10. Participant's age at baseline 15.71 (1.18) 0.30** 0.25* 0.14 0.25** 0.25* 0.16 0.24* 0.19 −0.03 –
11. Participant's sex (0 = male; 1 = female) – 0.22* 0.08 0.16 0.32** 0.15 0.20 0.03 −0.13 −0.03 0.14 –
12. Treatment condition (0 = treatment as usual; 1 = experimental condition) – 0.05 0.07 −0.06 0.06 0.01 0.05 −0.02 0.16 0.20 0.12 0.07 –
Open in a new tab

Note: BL = baseline; 3M = 3 months; 6M = 6 months.

*

p < .05

**

p < .01.

3.2 ∣. Missing data

Of the total sample, 15.3% had missing 3-month SI severity data and 16.2% at 6 months. Additionally, 18.0% had missing 3-month depression-related emotional problem data and 14.4% at 6 months. Last, 27.9% had missing 3-month hopelessness data and 22.5% at 6 months. After accounting for multiple comparisons, adolescents with complete versus missing SI severity at 3 or 6 months did not differ significantly on SI severity at baseline (p's > .05). Moreover, youth with complete versus missing SI severity data at 3 or 6 months did not significantly differ across treatment condition (p's > .05), demographic variables (p's > .05), depression-related emotional problems (p's > .05) or hopelessness (p's > .05) across any time points.

3.3 ∣. Model comparisons

To select the most parsimonious model, fit for the two models was compared utilizing a nested comparison approach (see Table 2). Model 1, which estimated all hypothesized cross-lagged paths, demonstrated good fit: S-B χ2(8) = 7.61, p = .47; CFI = 1.00; TLI = 1.00; RMSEA = 0.01. Model 2, in which all nonsignificant cross-lagged paths from the competing model were constrained, fit the data equally well (Δχ2(5) = 5.23, p = .39). Thus, Model 2, which also demonstrated good fit (S-B χ2(13) = 12.78, p = .47; CFI = 1.00; TLI = 1.00; RMSEA = 0.01) was selected as the most precise and parsimonious representation of the data. To provide a rigorous test of study hypotheses using Model 2, all variables of interest across every time point were regressed on treatment condition, sex, and age. This final model (see Figure 1) represented the data well (S-B χ2(13) = 13.32, p = .42; CFI = 1.00; TLI = 1.00; RMSEA = 0.02) and accounted for 49% of the variance in 6-month SI severity (R2 = .49, p < .001).

TABLE 2.

Nested model comparisons

Model fit
 Model comparison
Model number and descriptions S-B χ2 (df) p CFI TLI RMSEA AIC BIC S-B
Δχ2 (Δ df)		 p
1. Reciprocal model with all paths included 7.61 (8) .47 1.00 1.00 0.01 5226.36 5205.22 – –
2. Constrained paths in competing model:
BL Depression-related emotional problems
→ 6M Suicidal ideation severity
BL Depression-related emotional problems
→ 3M Hopelessness
BL Hopelessness
→ 3M Suicidal ideation severity
3M Depression-related emotional problems
→ 6M Hopelessness
3M Hopelessness
→ 6M Suicidal ideation severity		 12.78 (13) .47 1.00 1.00 0.01 5220.81 5201.97 5.23 (5) .39
Open in a new tab

Note: BL = baseline; 3M = 3 months; 6M = 6 months. S-B χ2 = Satorra-Bentler adjusted chi-square; df = degrees of freedom; Δ = change in parameter.

Abbreviations: AIC, Akaike information criteria; BIC, sample-size adjusted Bayesian information criteria; CFI, comparative fit index; RMSEA, root mean square error of approximation; TLI, Tucker-Lewis index.

3.4 ∣. Path estimates

3.4.1 ∣. Covariates

At baseline, older (vs. younger) adolescents reported greater hopelessness (β = .24, p = .019), depression-related emotional problems (β = .27, p = .017), and SI severity (β = .31, p < .001). At baseline, females (relative to males) reported greater depression-related emotional problems (β = .27, p = .003). Age and sex, however, were not significantly related to any of the variables of interest at 3 or 6 months. Treatment condition was not significantly related to any variables of interest at baseline, 3 or 6 months (see Table 3).

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 Hopelessness .24* 0.83 (0.36) 0.02–0.42
 3M Hopelessness .02 0.06 (0.33) −0.18–0.21
 6M Hopelessness −.14 −0.46 (0.37) −0.36–0.09
 BL Depression-related emotional problems .23* 1.13 (0.49) 0.04–0.41
 3M Depression-related emotional problems .04 0.15 (0.35) −0.16–0.22
 6M Depression-related emotional problems −.01 −0.05 (0.41) −0.24–0.22
 BL Suicidal ideation severity .31* 5.02 (1.57) 0.13–0.47
 3M Suicidal ideation severity −.01 −0.07 (1.08) −0.24–0.23
 6M Suicidal ideation severity −.09 −1.14 (1.33) −0.31–0.10
Sex →
 BL Hopelessness −.01 −0.01 (0.77) −0.18–0.18
 3M Hopelessness −.13 −1.04 (0.77) −0.32–0.06
 6M Hopelessness .01 0.11 (.80) −0.19–0.23
 BL Depression-related emotional problems .27* 3.22 (1.12) 0.09–0.45
 3M Depression-related emotional problems .01 0.08 (0.77) −0.17–0.19
 6M Depression-related emotional problems .13 1.12 (0.80) −0.06–0.30
 BL Suicidal ideation severity .16 6.20 (3.75) −0.03–0.36
 3M Suicidal ideation severity −.01 −0.25 (1.97) −0.20–0.16
 6M Suicidal ideation severity .05 1.30 (2.75) −0.13–0.24
Treatment condition →
 BL Hopelessness −.05 −0.42 (0.75) −0.23–0.13
 3M Hopelessness .11 0.91 (0.77) −0.08–0.30
 6M Hopelessness .17 1.29 (0.81) −0.03–0.37
 BL Depression-related emotional problems .04 0.48 (1.03) −0.13–0.21
 3M Depression-related emotional problems −.05 −0.39 (0.68) −0.20–0.11
 6M Depression-related emotional problems −.06 −0.47 (0.77) −0.23–0.13
 BL Suicidal ideation severity .01 0.06 (3.63) −0.18–0.19
 3M Suicidal ideation severity .02 0.31 (1.77) −0.15–0.17
 6M Suicidal ideation severity −.10 −3.02 (2.53) −0.25–0.08
Open in a new tab

Note: BL = baseline; 3M = 3 months; 6M = 6 months. CI = bias-corrected confidence interval for standardized parameter estimates. Sex: 0 = male; 1 = female. Treatment condition: 0 = treatment as usual; 1 = experimental condition.

*

p < .05.

3.4.2 ∣. Contemporaneous and autoregressive paths

Though the magnitude of contemporaneous relations among variables of interest varied over time, directionality remained consistent (Table 4). Greater hopelessness levels were concurrently associated with greater depression-related emotional problems at baseline (r = .53, p < .001), 3 months (r = .42, p < .001), and 6 months (r = .32, p = .013). Similarly, hopelessness was associated with greater SI severity at baseline (r = .43, p < .001), 3 months (r = .40, p < .001), and 6 months (r = .35, p = .005). Depression-related emotional problems were related to greater SI severity at baseline (r = .62, p < .001), 3 months (r = .45, p < .001), and 6 months (r = .71, p < .001).

TABLE 4.

Unstandardized and standardized parameter estimates of autoregressive, cross-lagged, contemporaneous, and indirect paths between variables of interest in the final model

Paths β b (SE) 95% CI
Autoregressive
 BL Hopelessness → 3M Hopelessness .56** 0.54 (0.10) 0.36–0.71
 BL Hopelessness → 6M Hopelessness .28 0.26 (0.13) −0.01–0.52
 3 M Hopelessness → 6M Hopelessness .19 0.18 (0.18) −0.15–0.53
 BL Depression-related emotional problems →3M Depression-related emotional problems .48** 0.35 (0.06) 0.31–0.64
 BL Depression-related emotional problems → 6M Depression-related emotional problems .19 0.14 (0.07) −0.01–0.38
 3M Depression-related emotional problems → 6M Depression-related emotional problems .30* 0.30 (0.12) 0.07–0.51
 BL Suicidal ideation severity → 3M Suicidal ideation severity .56** 0.31 (0.06) 0.30–0.76
 BL Suicidal ideation severity → 6M Suicidal ideation severity .49** 0.37 (0.11) 0.23–0.73
 3M Suicidal ideation severity → 6M Suicidal ideation severity .09 0.12 (0.17) −0.20–0.32
Cross-lagged
 BL Hopelessness → 3M Depression-related emotional problems .21* 0.22 (0.10) 0.04 −0.37
 3M Depression-related emotional problems → 6M Suicidal ideation severity .27* 0.92 (0.39) 0.06 −0.47
 BL Hopelessness → 6M Suicidal ideation severity −.02 −0.06 (0.28) −0.18–0.14
 BL Depression-related emotional problems → 3M Suicidal ideation severity .08 0.15 (0.06) −0.16–0.29
 3M Hopelessness → 6M Depression-related emotional problems .22* 0.24 (0.10) 0.03–0.42
Indirect
 BL Hopelessness → 3M Depression-related emotional problems→ 6M Suicidal ideation severity .06* 0.20 (0.15) 0.01–0.15
Contemporaneous r 95% CI
 BL Hopelessness ↔ BL Depression-related emotional problems .53** 0.36–0.66
 BL Hopelessness ↔ BL Suicidal ideation severity .43** 0.25–0.58
 BL Depression-related emotional problems ↔ BL Suicidal ideation severity .62** 0.43–0.76
 3M Hopelessness ↔ 3M Depression-related emotional problems .42** 0.22–0.59
 3M Hopelessness ↔ 3M Suicidal ideation severity .40** 0.18–0.59
 3M Depression-related emotional problems ↔ 3M Suicidal ideation severity .45** 0.27–0.59
 6M Hopelessness ↔ 6M Depression-related emotional problems .32* 0.03–0.54
 6M Hopelessness ↔ 6M Suicidal ideation severity .35* 0.07–0.57
 6M Depression-related emotional problems ↔ 6M Suicidal ideation severity .71** 0.52–0.82
Open in a new tab

Note: BL = baseline; 3M = 3 months; 6M = 6 months. CI = bias-corrected confidence interval for standardized parameter estimates. Bolded paths are part of the hypothesized mediation pathway.

*

p < .05

**

p < .001.

Autoregressive paths assessed the temporal stability for variables of interest (see Table 4). Baseline hopelessness levels significantly predicted levels at 3 months (β = .56, p < .001) but not at 6 months. Relatedly, 3-month hopelessness levels did not significantly predict 6-month levels. Baseline levels of depression-related emotional problems significantly predicted levels at 3 months (β = .48, p < .001) but not 6 months. However, depression-related emotional problems at 3 months significantly predicted 6-month levels (β = .30, p = .008). Baseline SI severity significantly predicted 3- (β = .56, p < .001) and 6-month levels (β = .49, p < .001). However, 3-month SI severity did not significantly predict 6-month levels.

3.4.3 ∣. Cross-lagged and indirect paths

After constraining nonsignificant paths and accounting for effects attributable to covariates as well as autoregressive and contemporaneous relations between variables of interest, cross-lagged paths suggested only one direction for effects (see Table 4). Baseline levels of hopelessness positively predicted 3-month levels of depression-related emotional problems (β = .21, p = .012). Three-month levels of hopelessness levels also significantly predicted 6-month levels of depression-related emotional problems (β = .22, p = .021). Baseline levels of hopelessness did not predict 6-month SI severity. Though baseline levels of depression-related emotional problems did not significantly predict 3-month SI severity, 3-month levels of depression-related emotional problems positively predicted 6-month SI severity (β = .27, p = .013). The indirect path from baseline levels of hopelessness to 6-month SI severity through 3-month depression-related emotional problems was significant: β = .06, 95% CI (0.01, 0.14).

4 ∣. DISCUSSION

There exists cross-sectional research supporting the indirect effect of hopelessness on SI severity through depressive symptoms, particularly depression-related emotional problems. Yet, there are no longitudinal studies examining the temporal associations between these constructs among adolescents, precluding true tests of mediation. Furthermore, while the predictive utility of hopelessness and depressive symptoms for SI has been well-documented, the average follow-up is approximately 9 years, limiting the clinical utility of this research (Ribeiro et al., 2018). Based on extant theoretical and empirical research, the primary aim of this study was to build on existing research by examining the potential role of depression-related emotional problems as a mediator between hopelessness and SI severity using a short-term, prospective, multiwave design with an adolescent clinical sample. The secondary purpose was to evaluate the specificity of depression-related emotional problems as a possible mediator.

Consistent with the main study hypothesis, depression-related emotional problems mediated the relation between hopelessness and SI severity. Specifically, adolescents with greater baseline levels of hopelessness experienced greater depression-related emotional problems at Month 3, which in turn, predicted greater SI severity by Month 6. The final model accounted for almost half of the variance in SI severity at Month 6. This pattern of results parallels the cross-sectional relations found in prior cross-sectional research with adolescents and adults (Ballard et al., 2015; Spann et al., 2006). They are also entirely consistent with the Escape Theory of Suicide (Baumeister, 1990), which underscores the role of maladaptive cognitions and aversive emotional states in the development and maintenance of SI. Specifically, hopelessness-related cognitions about oneself and/or the future may precipitate distressing, depression-related emotional problems. These depression-related emotional problems may eventually be experienced as intolerable, which in turn, drives SI.

In contrast, hypotheses associated with the alternative model were not supported. Hopelessness did not mediate the relation between baseline depression-related emotional problems and severity. Specifically, adolescents with greater depression-related emotional problems at baseline did not experience greater hopelessness by Month 3, and hopelessness at Month 3 did not predict greater SI severity by Month 6. These results are consistent with prior research which suggests that temporal effects of hopelessness on depression-related emotional problems are unidirectional (Liu et al., 2015). Relatedly, the nonsignificant relations of the proposed paths in this alternative model offer support for the specificity of depression-related emotional problems as a mediator between hopelessness and SI severity.

4.1 ∣. Clinical implications

Current study results hold significant implications for therapeutic work with adolescents in clinical settings. Assessing for hopelessness at the start of treatment with adolescents may help inform treatment planning. For youth with high levels of hopelessness, skills designed to address hopelessness-related thoughts at the start of care may be indicated. This may include the use of traditional cognitive-behavioral therapy (CBT) techniques, such as cognitive restructuring and problem-solving, to bolster positive expectations of the future (Daniel & Goldston, 2012). To further augment this work, the use of acceptance and commitment therapy (ACT) techniques that promote a willingness to experience distressing emotions while simultaneously engaging in actions that move one closer to one's goals (Zhenggang et al., 2020) may also help decrease and/or protect against depression-related emotional symptoms and, subsequently, SI. As prior research also suggests that greater numbers of negative life events during adolescence are associated with slower declines in hopelessness over time (Giollabhui et al., 2018), CBT and ACT skills can also be used to help adolescents effectively tolerate and cope with, and in some cases prevent, negative life events (e.g., arrests, suspensions, etc.). This, in turn, may lead to quicker reductions in hopelessness as well as depressive symptoms and suicide risk. Finally, given that youth exist within multiple systems (i.e., family, school, etc.), and significant and/or chronic problems within these systems may fuel negative beliefs about the future, addressing problems within these systems as part of standard clinical care (e.g., via parent training, family work, parental referrals for care, work with school personnel, psychoeducational testing, etc.) may bolster support and hopefulness about the future.

Relatedly, among youth reporting depressive symptoms, study results suggest that depression-related emotional problems, in particular, may confer risk for more severe SI. These include symptoms of negative mood, decreased self-esteem, fatigue, and sleep disturbances. Thus, when administering the CDI-2 in clinical care, it will be important to attend to elevated scores on the depression-related emotional problems scale, even if the total score does not suggest substantial clinical concern. When the CDI-2 is not used, close attention to the constellation of symptoms that comprise the depression-related emotional problems scale is clinically warranted. For example, a fair amount of research suggests that sleep disturbance is associated with SI (see Liu et al., 2019). Thus, integration of sleep hygiene work and/or referral for a sleep specialist and related treatment when warranted may help decrease SI.

4.2 ∣. Limitations and future directions

The current study has several strengths, including a multiwave design, the use of repeated, well-validated, self-report measures with a clinical adolescent sample, and a rigorous analytic strategy. However, results should be considered in light of some limitations. First, though these results offer support for the EscapeTheory of Suicide and build incrementally upon existing research, not all adolescents who experience hopelessness and/or depressive symptoms develop SI (Cha et al., 2018). Therefore, future multiwave, prospective studies may aim to assess other constructs that have been implicated in the onset and/or maintenance of SI severity among adolescents, such as thwarted belongingness, perceived burdensomeness, and pain (see Joiner, 2005; Miller et al., 2015). For instance, in the presence of high (vs. low) levels of hopelessness, thwarted belongingness and perceived burdensomeness predict greater SI (Hagan et al., 2015). Relatedly, the Three-Step Theory (Klonsky & May, 2015) suggests that a combination of pain (emotional and/or physical) and hopelessness lead to the onset of SI. Second, results may be limited in generalizability. While the inclusion of a high-risk, clinical sample with co-occurring substance use and psychiatric problems is a significant strength, it is unclear whether study results generalize to adolescents with only internalizing symptoms or youth who were not receiving a skills-based treatment designed to address internalizing symptoms, such as CBT. Indeed, youth across both arms of the present clinical trial appeared to receive some CBT skills in the context of their sessions (Wolff et al., 2020). Under such conditions, it is possible that the constructs under study may have been more likely to change in tandem or mediation more likely. In a similar vein, participants predominately identified as White and as non-Latino, again, limiting generalizability. Moreover, the current sample size may have hindered detection of all significant relations among study variables. Therefore, future studies may aim to include larger and more diverse samples with a similar design. Third, this study focused solely on SI as an outcome. Thus, it is unclear whether hopelessness may confer risk for suicide attempts via depression-related emotional problems, which is another important avenue for future research. The small number of participants who endorsed suicide attempts precluded an examination of this question in the present study. Fourth, while using path analysis to examine the direction of effects between study variables is an important strength, causality cannot be deduced from such models (Cohen et al., 2003). Last, the current study exclusively used self-report measures. Relatedly, while all measures assessed current levels of each construct, time frames did not perfectly coincide (e.g., SIQ-JR assesses past month SI while CDI-2 assesses depressive symptoms during past 2 weeks). Future research that utilizes multimethod approaches is needed, such as the use of parental reports of adolescent's symptoms.

5 ∣. CONCLUSION

Nevertheless, the current study is among the first to evaluate the longitudinal relations among hopelessness, depression-related emotional problems, and SI severity in an adolescent clinical sample. Results suggest depression-related emotional problems may help explain the association between hopelessness and SI severity. Clinically, findings support treatments that bolster positive expectations of the future to reduce depression-related emotional problems and subsequently lower suicide risk among adolescents.

Supplementary Material

Supplementary Table

ACKNOWLEDGEMENTS

The authors thank NIAAA (Grant Number: R01AA020705) for the financial support but recognize that the findings and conclusions are those of the authors and do not necessarily reflect the opinions of the NIAAA.

Footnotes

CONFLICT OF INTERESTS

The authors declare that there are no conflict of interests.

SUPPORTING INFORMATION

Additional Supporting Information may be found online in the supporting information tab for this article.

1

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

DATA AVAILABILITY STATEMENT

Research data are not shared.

REFERENCES

  1. Abramson LY, Alloy LB, Hogan ME, Whitehouse WG, Cornette M, Akhavan S, & Chiara A (1998). Suicidality and cognitive vulnerability to depression among college students: A prospective study. Journal of Adolescence, 21, 473–487. 10.1006/jado.1998.0167 [DOI] [PubMed] [Google Scholar]
  2. Ballard ED, Patel AB, Ward M, & Lamis DA (2015). Future disposition and suicidal ideation: Mediation by depressive symptom clusters. Journal of Affective Disorders, 170, 1–6. 10.1016/j.jad.2014.08.029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baumeister RF (1990). Suicide as escape from self. Psychological Review, 97(1), 90–113. 10.1037/0033-295X.97.1.90 [DOI] [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. Centers for Disease Control and Prevention. (2017). QuickStats: Suicide rates for teens aged 15-19 years by sex—United States, 1975-2015. MMWR. Morbidity and Mortality Weekly Report 10.15585/mmwr.mm6630a6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cha CB, Franz PJ, Guzmán EM, Glenn CR, Kleiman EM, & Nock MK (2018). Annual research review: Suicide among youth – epidemiology, (potential) etiology, and treatment. Journal of Child Psychology and Psychiatry, 59(4), 460–482. 10.1111/jcpp.12831 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cohen J, Cohen P, West SG, & Aiken LS (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Erlbaum. [Google Scholar]
  8. Daniel SS, & Goldston DB (2012). Hopelessness and lack of connectedness to others as risk for suicidal behavior across the lifespan: Implications for cognitive-behavioral treatment. Cognitive and Behavioral Practice, 19(2), 288–300. 10.1016/j.cbpra.2011.05.003 [DOI] [Google Scholar]
  9. Gibb BE, & Alloy LB (2006). A prospective test of the Hopelessness Theory of Depression in children. Journal of Clinical Child & Adolescent Psychology, 35(2), 264–274. 10.1207/s15374424jccp3502_10 [DOI] [PubMed] [Google Scholar]
  10. Glenn CR, Lanzillo EC, Esposito EC, Santee AC, Nock MK, & Auerback RP (2017). Examining the course of suicidal and non-suicidal self-injurious thoughts and behaviors in outpatient and inpatient adolescents. Journal of Abnormal Psychology, 45, 971–983. 10.1007/s10802-016-0214-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Hagan CR, Podlogar MC, Chu C, & Joiner TE (2015). Testing the interpersonal theory of suicide: The moderating role of hopelessness. International Journal of Cognitive Therapy, 8(2), 99–113. 10.1521/ijct.2015.8.2.99 [DOI] [Google Scholar]
  12. Hards E, Ellis J, Fisk J, & Reynolds S (2020). Negative view of the self and symptoms of depression in adolescents. Journal of Affective Disorders, 262(1), 143–148. 10.1016/j.jad.2019.11.012 [DOI] [PubMed] [Google Scholar]
  13. Ivarsson T, Svalander P, & Litlere O (2006). The Children's Depression Inventory (CDI) as measure of depression in Swedish adolescents: A normative study. Nordic Journal of Psychiatry, 60, 220–226. 10.1080/08039480600636395 [DOI] [PubMed] [Google Scholar]
  14. Joiner TE (2005). Why people die by suicide. Harvard University Press. [Google Scholar]
  15. Kazdin AE, French NH, Unis AS, Esveldt-Dawson K, & Sherick RB (1983). Hopelessness, depression, and suicidal intent among psychiatrically disturbed inpatient children. Journal of Consulting and Clinical Psychology, 51, 504–510. 10.1037/022-006x.51.4.504 [DOI] [PubMed] [Google Scholar]
  16. Keilp JG, Ellis SP, Gorlyn M, Burke AK, Oquendo MA, Mann JJ, & Grunebaum MF (2018). Suicidal ideation declines with improvement in the subjective symptoms of major depression. Journal of Affective Disorders, 227, 65–70. 10.1016/j.jad.2017.09.018 [DOI] [PubMed] [Google Scholar]
  17. Keilp JG, Grunebaum MF, Gorlyn M, LeBlanc S, Burke AK, Galfalvy H, Oquendo MA, & Mann JJ (2012). Suicidal ideation and the subjective aspects of depression. Journal of Affective Disorders, 140(1), 75–81. 10.1016/j.jad.2012.01.045 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. 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]
  19. Kovacs M (1985). The Children's Depression Inventory (CDI). Psychopharmacology Bulletin, 21(4), 995–998. [PubMed] [Google Scholar]
  20. Kovacs M (2011). Children's Depression Inventory-2 (CDI-2) (2nd Ed.). Multi-Health Systems, Inc. [Google Scholar]
  21. Little TD (2013). Longitudinal Structural Equation Modeling. Guilford Press. [Google Scholar]
  22. Liu JW, Tu YK, Lai YF, Lee HC, Tsai PS, Chen TJ, Huang HC, Chen YT, & Chiu HY (2019). Associations between sleep disturbances and suicidal ideation, plans, and attempts in adolescents: A systematic review and meta-analysis. Sleep, 42(6), zsz054. 10.1093/sleep/zsz054 [DOI] [PubMed] [Google Scholar]
  23. Liu RT, Kleiman EM, Nestor BA, & Cheek SM (2015). The hopelessness theory of depression: A quarter-century in review. Clinical Psychology: Science and Practice, 22, 345–365. 10.1111/cpsp.12125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. MacCallum RC, Browne MW, & Cai L (2006). Testing differences between nested covariance structure models: Power analysis and null hypotheses. Psychological Methods, 11(1), 19–35. 10.1037/1082-989X.11.1.19 [DOI] [PubMed] [Google Scholar]
  25. Miller AB, Esposito-Smythers C, & Leichtweis RN (2015). A short-term, prospective test of the Interpersonal-Psychological Theory of suicidal ideation in an adolescent clinical sample. Suicide and Life-Threatening Behavior, 46, 337–351. [DOI] [PubMed] [Google Scholar]
  26. Muthén LK, & Muthén BO (1998-2017) Mplus User's Guide (8th Ed.). Muthén & Muthén. [Google Scholar]
  27. Preacher KJ, & Coffman DL (2006, May). Computing power and minimum sample size for RMSEA [Computer software]. Available from http://quantpsy.org/ [Google Scholar]
  28. Reynolds WM (1987). Suicidal Ideation Questionnaire. Psychological Assessment Resources. [Google Scholar]
  29. 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. 10.1080/02796015.1999.12085945 [DOI] [Google Scholar]
  30. Ribeiro JD, Huang X, Fox KR, & Franklin JC (2018). Depression and hopelessness as risk factors for suicide ideation, attempts and death: Meta-analysis of longitudinal studies. The British Journal of Psychiatry, 212, 279–286. 10.1192/bjp.2018.27 [DOI] [PubMed] [Google Scholar]
  31. 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]
  32. Spann M, Molock SD, Barksdale C, Matlin S, & Puri R (2006). Suicide and African American teenagers: Risk factors and coping mechanisms. Suicide and Life-Threatening Behavior, 36(5), 553–568. 10.1521/suli.2006.36.5.553 [DOI] [PubMed] [Google Scholar]
  33. Spirito A, Williams CA, Stark LJ, & Hart KJ (1998). The Hopelessness Scale for Children: Psychometric properties with normal and emotionally disturbed adolescents. Journal of Abnormal Child Psychology, 16, 445–458. 10.1007/bf00914174 [DOI] [PubMed] [Google Scholar]
  34. Stumper A, Olino TM, Abramson LY, & Alloy LB (2019). A factor analysis and test of longitudinal measurement invariance of the Children's Depression Inventory (CDI) across adolescence. Journal of Psychopathology and Behavioral Assessment, 41, 692–698. 10.1007/s10862-019-09746-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Wang Y, Jiang N, Cheung EFC, Sun H, & Chan RCK (2015). Role of depression severity and impulsivity in the relationship between hopelessness and suicidal ideation in patients with major depressive disorder. Journal of Affective Disorders, 183, 83–89. 10.1016/j.jad.2015.05.001 [DOI] [PubMed] [Google Scholar]
  36. Wolff J, Esposito-Smythers C, Frazier E, Stout R, Gomez J, Massing-Schaffer M, Nestor B, Cheek S, Graves H, Yen S, Hunt J, & Spirito A (2020). A randomized trial of an integrated cognitive behavioral treatment protocol for adolescents receiving home-based services for co-occurring disorders. Journal of Substance Abuse Treatment, 116, 108055. 10.1016/j.jsat.2020.108055 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Zhenggang B, Shiga L, Luayo Z, Wu S, & Iris C (2020). Acceptance and commitment therapy (ACT) to reduce depression: A systematic review and meta-analysis. Journal of Affective Disorders, 260, 728–737. 10.1016/j.jad.2019.09.040 [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Table
NIHMS1763315-supplement-Supplementary_Table.docx (19KB, docx)

Data Availability Statement

Research data are not shared.

RESOURCES