Coping with Suicidal Urges: An Important Factor for Suicide Risk Assessment and Intervention

Abstract

Suicide-related coping refers to strategies for adaptively managing suicidal urges and can be important an important factor for assessing risk and targeting intervention. The current study evaluated whether suicide-related coping predicted a suicidal event within 90-days, independently of other known risk factors. Veterans (N=64) were evaluated shortly after a suicidal crisis and completed several assessments, including a measure of suicide-related coping. Multivariate analyses showed that suicide-related coping remained protective of a suicidal event (OR = 0.93; p = .047) after adjusting for suicidal ideation, previous suicide attempts, mood disorder, distress tolerance, and gender. Suicide-related coping may augment commonly assessed clinical factors in prediction of a suicidal event and is a suitable target for suicide prevention efforts.

Introduction

In the US, age-adjusted suicide rates increased by 30% between 2000 and 2016 (Hedegaard, Curtin, & Warner, 2018). The US military population, whose members were historically at lower risk of suicide (Eaton, Messer, Wilson, & Hoge, 2006), has also been facing increasing rates, exceeding that of the civilian population in 2008 (Kuehn, 2009; Nock et al., 2013). This has led to suicide prevention becoming the top clinical priority for the Veterans Health Administration. Research is desperately needed to improve our understanding of the problem and identify viable suicide prevention targets that point to effective interventions. With this goal in mind, the current study focuses on the specific coping skills that individuals can utilize when experiencing suicidal urges, in order to prevent escalation into a suicidal crisis. These skills are referred to as suicide-related coping, which has been defined as, “knowledge of and perceived self-efficacy in using internal coping strategies and external resources to manage suicidal thoughts with the goal of decreasing imminent risk and averting suicidal crises” (Stanley, Green, Ghahramanlou-Holloway, Brenner, & Brown, 2017, pg. 190). Examples of suicide-related coping include engaging in distracting activities, seeking social and professional support, and limiting access to lethal means. The importance of such skills is illustrated by a recent study finding that individuals who report some ability to control their suicidal thoughts were less likely to make a first suicide attempt (Nock et al., 2018). Thus, assessment of suicide-related coping may serve as a useful tool for predicting suicide risk and informing intervention efforts.

Studies have started to provide a glimpse into suicide-related coping. One study examined the coping responses that mental health consumers most commonly utilized when experiencing suicidal thoughts (Alexander, Haugland, Ashenden, Knight, & Brown, 2009). The most common ones included turning to spirituality, talking to someone, and positive thinking, though other common responses included engaging in pleasurable activity and using professional resources. A subsequent study of consumers with a mood disorder showed that over 80% of survey respondents described a similar range of coping responses (Simon, Specht, & Doederlein, 2016). Importantly, this study found that not only were these coping responses common, but also that over 70% of participants found these techniques to be at least somewhat helpful. The benefit of such coping responses was also illustrated by a study that assessed in real-time the coping techniques used by adolescents with urges to engage in non-suicidal self-injury (Nock, Prinstein, & Sterba, 2009). Among the adolescents who experienced the urge to self-injure but did not do so, their coping involved trying to change their thoughts, talking to someone, or engaging in other distracting activities (e.g., movie, computer). These studies illustrate that suicide-related coping is distinct from more general forms of coping, in that it is applied specifically in response to suicidal urges. The aim is to prevent escalation to a suicidal crisis, where coping skills are overwhelmed, distress becomes intolerable, and the risk for suicidal behavior is acute (Wenzel & Beck, 2008).

Given that suicide-related coping is commonly employed by individuals experiencing suicidal thoughts or urges (Alexander et al., 2009; Simon et al., 2016) and that it is often perceived as helpful, it is necessary to evaluate the degree to which this type of coping is effective for preventing a suicidal crisis. Although some suicide-specific psychotherapies tend to target suicide-related coping and are effective at reducing suicidal behavior (Brown et al., 2005; Jobes, 2012; Linehan et al., 2006), these psychotherapies also target a range of other coping skills to address mood, resilience, and functioning more broadly. As noted, suicide-related coping is distinct from more general forms of coping. Thus, it remains unknown whether suicide-related coping skills in and of themselves are effective at preventing a suicidal crisis. This question has strong implications for suicide prevention efforts, especially for the range of brief, single-session interventions that focus directly on increasing suicide-related coping (Bryan et al., 2017; Stanley & Brown, 2012; Stanley et al., 2015; Stanley et al., 2016). An additional implication is that assessment of an individual’s ability to cope with their suicidal urges can inform approaches to suicide risk assessment.

The current study therefore examined whether suicide-related coping was protective against a subsequent suicidal crisis in a sample of Veterans at high-risk for suicide. Suicidal crisis was operationalized as a suicidal event (Brent et al., 2009; Oquendo, Stanley, Ellis, & Mann, 2004). Suicidal events involved suicidal ideation that resulted in suicidal behavior or resulted in acute hospitalization to ensure safety. The current study first examined whether suicide-related coping was predictive of a suicidal event during the subsequent 90 days. Next, the study examined whether suicide-related coping predicted a suicidal event during the subsequent 90 days, independently of other known suicide risk factors. Notably, these factors included a more general capacity to tolerate negative emotions (i.e., distress tolerance; Zvolensky, Vujanovic, Bernstein, & Leyro, 2010), given the intervention implications of targeting suicide-related coping instead of a broader capacity to deal with emotional difficulty. Finally, the current study also examined the relationship between suicide-related coping and other key measures of suicide risk, including depression, and hopelessness, as well as suicidal ideation and attempts.

Methods

Participants

Participants were Veterans (N=64) who were considered high-risk for suicide. Nearly all participants (90.1%) were recruited from two Veterans Health Administration (VHA) acute inpatient psychiatric facilities, where they were receiving treatment after suicidal ideation or attempt. Participants were recruited for a randomized controlled trial examining a mindfulness-based intervention to reduce suicidal behavior (Kline et al., 2016). Because the trial was examining a suicide-specific intervention, the current study utilized only participants randomized to the control condition, which included treatment-as-usual for high-risk Veterans. This consisted of enhanced monitoring by suicide prevention case managers, who tracked and facilitated engagement with usual mental health care. Inclusion criteria were a) significant suicide risk during the previous 30 days and b) designation by the VHA as high-risk for suicide or a 12-month history of actual, aborted or interrupted suicide attempt. Exclusion criteria were: a) clinically significant cognitive deficits; b) severe hallucinations or delusions; c) disorganized or disruptive behaviors; d) medical instability; or e) previous year receipt of a mindfulness-based psychotherapy.

Procedures

Participants were presented with information about the study and those who were interested underwent an informed consent procedure. If consent was provided, a subsequent baseline research assessment was scheduled, where interview and self-report measures were administered. During the follow-up period of 90 days, participants were reassessed for suicidal behaviors and VHA electronic medical records were reviewed to capture instances of suicidal behavior and suicidal-related hospitalizations. Study assessments were conducted by research staff, who ranged from supervised mental health trainees to licensed mental health providers. Participants received monetary compensation for completing each study assessment. Although this study focused on a 90-day follow-up period, follow-up assessments were conducted according to the randomized clinical trial design, which is described in detail elsewhere (Kline et al., 2016). All study visits included assessment of suicide risk, with appropriate clinical intervention provided when indicated to enhance safety. All research procedures were approved by the facility’s Institutional Review Board.

Measures

Suicidal behavior measures

Suicidal thoughts and behaviors were assessed with the Columbia Suicide Severity Rating Scale (C-SSRS; Posner et al., 2011), which has received support for its criterion-validity and sensitivity to change (Brown, Currier, Jager-Hyman, & Stanley, 2015; Posner et al., 2011). The C-SSRS was administered via interview at the baseline and follow-up assessments. Data collected using the C-SSRS at baseline was used to determine the number of lifetime suicidal attempts (i.e., actual, aborted, interrupted attempts). Suicidal ideation intensity during the worst point in the previous year was also assessed with five items on the C-SSRS (scores range from 5–25). The C-SSRS at follow-up was used to determine the occurrence suicidal behavior occurring during the follow-up 90-day period. Suicidal ideation was also evaluated using the Scale for Suicide Ideation (SSI). The SSI assessed the severity of suicidal ideation (scores range from 0 – 38) during the week prior to baseline (Beck, Kovacs, & Weissman, 1979). The SSI has long been used in suicide research, with demonstrated concurrent and predictive validity for suicidal behaviors (Brown, Beck, Steer, & Grisham, 2000; Mann, Waternaux, Haas, & Malone, 1999). Internal consistency was .58 for C-SSRS intensity and .79 for SSI, in the current sample.

Participants’ VHA electronic medical records were also reviewed during the follow-up period. This allowed for identification of suicidal behavior that was not captured with the C-SSRS due to the participant being lost-to-contact during follow-up. Electronic medical record review also was used to capture suicidal-related acute hospitalizations during the follow-up period. Occurrence of these hospitalizations outside of the VHA health system were captured via interview during follow-up.

The study’s primary outcome was a suicidal event at 90 days, a period associated with elevated suicidal risk after hospitalization or emergency services for suicidal thoughts or behavior (Olfson, Wall, Wang, & et al., 2016; Shneidman & Farberow, 1956; Simon et al., 2018). Suicidal event was defined as the occurrence of any suicidal behavior or any acute psychiatric hospitalization involving suicidal ideation (Brent et al., 2009; Oquendo et al., 2004). Suicidal behavior included actual, aborted, or interrupted suicide attempts, as well as suicidal preparatory behavior (e.g., acquiring means to suicide, taking initial steps; Interian et al., 2018; Posner, Brodsky, Yershova, Buchanan, & Mann, 2014). The occurrence of a suicidal event was identified using information obtained from the C-SSRS and VHA medical record review during the follow-up period.

Suicide-Related Coping

The Suicide-Related Coping Scale (SRCS) was administered at baseline and contained 17 items, with 5 items scored in reverse (Stanley et al., 2017). SRCS items were summed to generate a total score (17–85), as well as separate subscale scores for internal and external coping. Each subscale contained seven items yielding a score range between 7–35. Higher scores indicate greater suicide-related coping. The internal and external scores were based on factor analysis of the scale (Stanley et al., 2017). Examples of internal coping strategies include recognizing when suicidal ideation is emerging and engaging in activities that provide distraction and self-soothing. Examples of external strategies include seeking social support and formal mental health assistance when suicidal. Stanley et al. (2017) found that the SRCS correlated negatively with suicidal ideation and previous suicide attempts. The same study showed that the SRCS demonstrated good internal consistency, factor structure, and sensitivity to change. In the current sample, SRCS internal consistencies were .90 for the total score, .77 for the external subscale, and .82 for the internal subscale.

Secondary measures

DSM-IV psychiatric diagnoses were obtained using the Mini International Neuropsychiatric Interview (MINI), which is a semi-structured diagnostic interview that is commonly used and has been shown to correlate with the gold standard semi-structure diagnostic interview (Sheehan et al., 1997). Substance use was assessed via structured interview questions that determined lifetime, past year, and past month use of a comprehensive list of substances. Several additional self-report measures were administered at baseline. Perceived capacity to tolerate negative emotions was assessed with the Distress Tolerance Scale (DTS), which has received support for reliability and validity (Simons & Gaher, 2005). The DTS is a 15-item scale with items rated between 1–5 and one item scored in reverse. Lack of distress tolerance has been conceptualized as an incapacity to tolerate negative emotions that leads to maladaptive coping (Zvolensky et al., 2010) and has been a target of psychotherapeutic intervention (McMain, Guimond, Barnhart, Habinski, & Streiner, 2017). The DTS has shown convergence with maladaptive coping responses, such as substance use (Gorka, Ali, & Daughters, 2012), self-injury (Anestis, Pennings, Lavender, Tull, & Gratz, 2013), and general impulsivity (Anestis et al., 2012). The Beck Depression Scale (BDI; 21 item; Beck, 1961) was used to assess depression and is a well-established self-report measure of depression with ample studies supporting its reliability and validity (Beck, Steer, & Carbin, 1988). Finally, hopelessness was assessed using the Beck Hopelessness Scale (BHS) (Beck, Weissman, Lester, & Trexler, 1974), which is similarly widely used and has been shown to correlate with suicidal behavior (Beck, Brown, Berchick, Stewart, & Steer, 2006; Brown et al., 2000). In the current sample, internal consistency scores were .87 for DTS, .91 for BDI, and .91 for BHS.

Analyses

Since the occurrence of suicidal events was also identified via VHA electronic medical record review, all 64 participants were included in the analyses regardless of whether a follow-up interview was conducted. To examine whether suicide-related coping was predictive of a subsequent suicidal event, separate ANOVA’s examined whether baseline SRSC total, internal coping, and external coping scores were significantly related to a suicidal event within 90 days. Square root transformations were utilized to normalize SRSC scores. To examine whether baseline suicide-related coping was predictive of a suicidal event independently of other known factors at baseline, a logistic regression evaluated whether SRCS scores predicted a suicidal event during the follow-up period of 90 days, adjusting for established risk factors. Established risk factors were chosen a priori based on previous research conducting similar analyses for suicidal behavior (Cha, Najmi, Park, Finn, & Nock, 2010; Nock et al., 2010) and included number of previous suicide attempts, suicidal ideation at baseline (SSI), and presence of a mood disorder (i.e., Major Depressive or Bipolar Disorders). To assess whether suicide-related coping predicted a suicidal event independently of distress tolerance, an additional logistic regression examined SRCS scores as predictors, adjusting for DTS total score and the established factors. Given established variations on suicidal behavior by gender, both logistic regression models adjusted for these effects (Hawton, 2000). Finally, relationship to other measures was evaluated using Spearman rank correlations, due to non-normal distributions.

Results

Out of the 64 participants, 53 (82.8%) completed a follow-up assessment covering the 90-day period. Sample characteristics are presented in Table 1. Participants were mostly male and diverse in terms of race/ethnicity. The majority were unemployed and the most common psychiatric diagnosis was major depressive disorder. A total of 17 (26.6%) participants experienced a suicidal event within 90 days. This included three participants who did not complete a follow-up assessment, but had a suicidal event identified through electronic medical record review. Suicidal crises by type were as follows: actual attempt - 9 (52.9%); aborted/interrupted attempt - 2 (11.7%); preparatory behavior - 3 (17.6%); and suicidal ideation resulting in acute psychiatric hospitalization - 3 (17.6%).

Table 1.

Demographic and clinical characteristics

Age (m, sd)	46.4	14.3
Female (n, %)	9	14.1
Race/Ethnicity
White	30	46.9
Black	19	29.7
Asian American	1	1.6
Latino	14	21.9
Marital Status
Married/Living as married	15	23.4
Never married	18	28.1
Separated/Divorced	28	43.8
Widowed	3	4.7
Education
High School or less	26	40.6
Some College	28	43.8
College Degree	10	15.6
Employment Status
Employed	18	28.1
Unemployed	44	68.8
Therapeutic employment	1	1.6
Full Time Student	1	1.6
DSM-IV Diagnosis
Major Depressive Disorder	45	70.3
Post-Traumatic Stress Disorder	36	56.3
non-alcohol substance use past 30 days	29	45.3
Alcohol binge drinking past 30 days	17	26.6
Bipolar Disorder	10	15.6
Panic Disorder	5	7.8
Generalized Anxiety Disorder	4	6.3
Schizoaffective Disorder	2	3.1
Lifetime Number of Suicide Attempts
0	7	10.9
1	15	23.4
≥2	42	65.6

Note. DSM-IV diagnoses are not mutually exclusive. Two Bipolar Disorder diagnoses were Bipolar II Disorder.

Baseline SRCS total scores were significantly lower among participants having a suicidal event within the subsequent 90 days (Figure 1a), compared to those without (F [1, 63] = 8.85, p = .004). These results indicated that those experiencing a suicidal crisis had lower self-reported suicide-related coping at baseline. As shown in Figure 1b, results for the internal (F [1, 60] = 8.31, p = .006) and external (F [1, 60] =6.05, p = .017) scales were also statistically significant and showed the same pattern.

Figure 1a.

SRCS total score mean and suicide event at 90 days

Figure 1b.

SRCS subscale means and suicidal crisis at 90 days

Table 2 summarizes two logistic regression models examining whether SRCS total score predicted a suicidal event within the subsequent 90 days. Model 1 (R²=.13) displays results for SRCS total score, adjusting for other established suicide risk factors: number of lifetime suicide attempts, suicidal ideation severity (SSI), and presence of a mood disorder. This model shows that a lower SRCS total score significantly predicted a suicidal event within the subsequent 90 days (p = .031), independently of the established suicide risk factors and gender. Model 2 (R² = .23) shows results for SRCS total score, adjusted by DTS and the established risk factors. This model again shows that SRCS significantly (p=.047) predicted a suicidal event within 90 days, above and beyond the effects for DTS and the established suicide factors and gender. Suicidal ideation at baseline was not significantly predictive of a suicidal event. The odds ratio for the SRCS total scores is essentially the same in both models, indicating that each unit increase in SRCS total score is associated with a 6–7% lower likelihood of a subsequent suicidal event. The above analyses were repeated with the SRCS external and internal subscales. The external SRCS subscale significantly predicted a suicide event in the subsequent 90 days, adjusting for the established risk factors and gender (OR = 88, p = .04). In all others instances, SRCS subscale scores were related in the same direction with suicidal event, but were subthreshold for statistical significance (all p-values > .06 to < .11).

Table 2.

Multivariate analyses of suicide-related coping as a predictor of suicidal crisis within 90 days

	Model 1					Model 2
	OR	95% CI			p-value	OR	95% CI			p-value
Suicide-related coping	0.94	0.88	-	0.99	0.031	0.93	0.87	-	1.00	0.047
Distress tolerance	-	-		-	-	0.23	0.07		0.77	0.017
Lifetime # of attempts	1.10	0.91	-	1.33	0.315	1.19	0.97	-	1.45	0.095
Suicidal Ideation	0.99	0.92	-	1.05	0.672	0.95	0.88	-	1.02	0.172
Mood Disorder	1.11	0.10	-	12.12	0.933	0.55	0.04	-	7.63	0.654
Female Gender	1.84	0.34	-	9.88	0.478	2.59	0.45	-	15.00	0.288

Note. Suicide-related coping measured with SRCS total score; Distress Tolerance assessed with DTS; Suicidal ideation assessed with SSI; Mood disorder includes Major Depressive Disorder and Bipolar Disorder.

Finally, analyses examined the relationship between the SCRS total score and subscales with key suicide-related measures at baseline (Table 3). These results show that the SCRS total score correlates strongly with depression and hopelessness, while correlating moderately with past year worst-point suicidal ideation intensity (C-SSRS) and suicidal ideation at baseline (SSI). The SCRS total score shows a small correlation with lifetime number of attempts and distress tolerance. A similar pattern is observed for the internal and external subscales, except for nonsignificant correlations with lifetime number of attempts and DTS.

Table 3.

Spearman rank correlations between SRCS scores and key suicide measures

	SSI	C-SSRS Intensity	Lifetime # of attempts	BDI	BHS	DTS
SRCS Total	−0.48***	−0.38**	−0.25*	−0.56***	−0.68***	0.29*
SRCS External	−0.41***	−0.33**	−0.18	−0.52***	−0.56***	0.24
SRCS Internal	−0.38**	−0.26*	−0.19	−0.51***	−0.66***	0.14

Note.

^*p<.05

^**p<.01

^***p<.001

Discussion

The current study examined whether suicide-related coping predicted a subsequent suicidal crisis. Current results show that participants at high-risk for suicide were less likely to experience a suicidal event within 90 days if they endorsed greater ability to use suicide-related coping. Illustrating the importance of this outcome, the vast majority of these suicidal crises were comprised of suicidal behaviors. Findings also showed that the total score for SRCS remained predictive of a future suicidal event, independently of level of suicidal ideation, number of previous attempts, and presence of a mood disturbance. Suicidal ideation is ubiquitous in suicide risk assessment, yet was not found in the current results to predict a suicidal event within the 90-day period. This may be due to its known fluctuations (Kleiman et al., 2017) or the fact that it may not offer enough specificity in this high-risk sample (Franklin et al., 2017). Thus, current results point to the unique prediction provided by the SRCS and suggest that assessment of an individual’s knowledge of and ability to engage in suicide-related coping can add to existing suicide risk assessment methods. Finally, correlational analyses also showed that the SRCS converges with key measures of suicide risk.

Review of the types of coping assessed by the SRCS illustrates its utility in suicide risk assessment and clinical intervention. SRCS items evaluate whether attempts are made to cope with suicidal thoughts, availability of social supports to turn to, willingness to talk to others and/or seek professional resources, willingness to restrict means, and even knowledge of how to seek professional resources. Based on the current results, indications that high-risk individuals express difficulty in these areas can serve as red flags for elevated short-term risk. Such difficulties also point to specific targets for brief or more ongoing intervention efforts. Given that suicidal thoughts have been shown to fluctuate in a matter of hours (Kleiman et al., 2017), suicide-related coping likely leads to alternative behaviors that allow periods of high-risk to subside.

The current study also examined suicide-related coping in relation to distress tolerance, which has been shown in previous research to be related to maladaptive coping (Anestis et al., 2012; Anestis et al., 2013; Gorka et al., 2012; Zvolensky et al., 2010). Current results showed that each were independently predictive of a suicidal event and showed a significant, but small, correlation with one another. These results suggest that both provide important contributions to preventing suicide. Both are targeted in psychotherapies that have received support for reducing suicidal behavior (Brown et al., 2005; Jobes, 2012; Linehan et al., 2006). For example, Dialectical Behavior Therapy teaches skills to reduce life-threatening behaviors, as well as longer-term development of behavioral skills, including emotion regulation skills (Linehan, 1993). Not surprisingly, the current results are consistent with the evidence-base for these suicide-specific psychotherapies.

At the same time, however, suicide-related coping predicted a suicidal event independently of distress tolerance and showed only a small magnitude correlation with this variable, pointing to the unique role of this form of coping and its suitability as a target for suicide prevention efforts. This is notable for interventions that are brief by necessity, such as those deployed by emergency departments and crisis centers. As an example, the Safety Planning Intervention was designed to directly target suicide-related coping in a brief session (e.g., discharge planning from emergency department visit; Stanley & Brown, 2012; Stanley et al., 2015; Stanley et al., 2016). Other interventions relying on single-session or brief interactions, such as Crisis Response Planning (Bryan et al., 2017), similarly target suicide-related coping among individuals showing signs of risk. These interventions often seek to increase self-management skills and contact with social supports, as well as increase access to mental health and emergency services.

The current results have additional implications that are worth noting. First, the current results point to the utility of SRCS as an outcome measure to evaluate interventions that seek to reduce risk for suicide. One challenge in many suicide prevention studies is that suicidal behavior is rare, thereby posing power-based methodological challenges in detecting intervention effects. For example, one depression quality improvement study reported statistically nonsignificant effects for reduced suicide attempt, despite an approximately 50% decrease associated with the intervention, a finding partly attributable to the low base-rate of suicide attempt (Asarnow, Jaycox, Duan, & et al., 2005). Because the current results link suicide-relating coping to subsequent suicidal event, the SRCS may be a suitable alternative outcome that would be associated with decreased risk. Second, identification of factors that indicate short-term risk of suicide has been identified as a priority in suicide research (Glenn & Nock, 2014). Illustrating this need, a meta-analysis of 50 years of suicide risk factor research showed that studies employed a mean follow-up assessment window of 108.82 months (Franklin et al., 2017). Factors that point to longer-term risk, while informative, create the challenge of providing limited information for immediate clinical planning (e.g., discharge planning). It is therefore notable that suicide-related coping predicted suicidal crises within a relatively shorter period of 90 days, a timeframe, moreover, that corresponds with a typical course of short-term psychotherapy.

The current study had several limitations. First, by illustrating that higher suicide-related coping precedes a lower likelihood of suicidal event, the current results point to suicide-related coping as a protective factor. However, given that suicide-related coping was not experimentally manipulated, the current results do not establish a causal protective factor (Kraemer et al., 1997). It may be that factors unique to individuals with greater suicide-related coping explain the association between SRCS scores and subsequent suicidal crises. We did, however, control for other known risk factors associated with suicidal crises and suicide-related coping (e.g., psychiatric symptoms, suicidal ideation) and found that it remained a significant predictor of subsequent suicidal crises.

Second, the relatively small sample size likely limited power and impacted the analyses with the SRCS subscales, with several models showing only trends towards significance. It may be that these subscales provide similar utility as the total score, but this should be examined in future studies. The fact that the SRCS subscales can point to different suicide prevention strategies warrants future research on this issue. Third, the current sample provides some considerations for generalizability. It was a Veteran sample, consisting mostly of males and it is important to replicate these findings in samples with greater gender diversity. Also, while the focus on a high-risk for suicide sample likely represents a strength of this study, it is also important to understand suicide-related coping in a broader range of risk. For example, for suicide to be truly prevented, it is necessary to understand the effectiveness of suicide-related coping among those experiencing a first-episode of suicidal ideation (Kessler, Borges, & Walters, 1999).