Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Jul 1.
Published in final edited form as: Arch Suicide Res. 2019 Jul 26;24(3):327–341. doi: 10.1080/13811118.2019.1635931

Characteristics of a First Suicide Attempt that Distinguish Between Adolescents Who Make Single Versus Multiple Attempts

Annamarie B Defayette 1, Leah M Adams 1, Emma D Whitmyre 1, Caitlin A Williams 1, Christianne Esposito-Smythers 1
PMCID: PMC6982559  NIHMSID: NIHMS1533273  PMID: 31248348

Abstract

Background:

Although a history of a suicide attempt is the strongest predictor of future suicide attempts, not all adolescents who make an attempt engage in repetitive suicidal behavior. The present study sought to determine whether certain characteristics of a first suicide attempt (e.g., age of first attempt, method of attempt used, intent seriousness, medical lethality, and receipt of treatment after attempt) can distinguish between adolescents who make single versus multiple suicide attempts.

Methods:

Adolescents (N = 95) who were psychiatrically hospitalized and their guardian completed a diagnostic interview to gather information on all lifetime suicide attempts. A multivariate hierarchical logistic regression was conducted, predicting single attempt (SA) versus multiple attempt (MA) status.

Results:

Of the first-attempt characteristics examined, only age of first attempt, OR = 0.33, 95% CI [0.17–0.63], p = .001, and receipt of treatment following attempt, OR = 0.28, 95% CI [0.09–0.88], p = .028, significantly distinguished SA vs. MA status, even after controlling for current age and depression at the time of first attempt.

Limitations:

Female and White participants were overrepresented in this sample, which limits generalization to more heterogenous and diverse samples. The cross-sectional nature of data introduces the potential for retrospective recall bias.

Conclusions:

Younger age of first attempt and lack of receipt of mental health treatment following a first attempt were associated with MA status. These findings highlight the importance of early mental health screening, parental psychoeducation, and linkage to mental health care after a suicide attempt.

Keywords: adolescent, suicide, suicide attempt, mental health treatment

Suicide is the second leading cause of death among adolescents and is on the rise. In 2017, 1,719 adolescents aged 12–17 died by suicide in the U.S. (Centers for Disease Control and Prevention [CDC], 2019). To date, a history of a suicide attempt is the strongest predictor of future suicide attempts (Ribeiro et al., 2016). Indeed, up to 42% of individuals with an attempt go on to make more than one lifetime attempt (Pagura, Cox, Sareen, & Enns, 2008). Those with a prior attempt are also more likely to die by suicide (Christiansen & Frank Jensen, 2007). Efforts are sorely needed to improve understanding of those at greatest risk for repetitive suicidal behavior. The primary purpose of the present study is to determine whether certain characteristics of a first suicide attempt (e.g., age of first attempt, method of attempt used, intent seriousness, medical lethality, and receipt of treatment after attempt), examined concurrently, can distinguish between adolescents who make a single versus multiple suicide attempts.

The qualitative distinction between those who make a single suicide attempt, herein referred to as single suicide attempters (SAs), and those who make more than one attempt, herein referred to as multiple suicide attempters (MAs), has been well studied. According to Goldston et al. (2016), multiple suicide attempts represent a separate, more severe dimension of suicidal behavior than one suicide attempt. Relative to SAs, MAs have been found to have more psychiatric disorder(s) (mood, anxiety, and substance use disorders), greater diagnostic comorbidity, more frequent histories of childhood trauma, poorer overall functioning, greater deficits in social problem-solving, greater affect dysregulation, and poorer cognitive control (Esposito, Spirito, Boergers, & Donaldson, 2003; Forman, Berk, Henriques, Brown, & Beck, 2004; Pagura et al., 2008; Stewart, Glenn, Esposito, Cha, & Nock, 2017). Overall, this literature suggests that those with greater psychological and functional impairment may be more likely to make more than one attempt. Less well studied are more immediate factors associated directly with the suicidal behavior.

The characteristics of an individual’s first suicide attempt may be of importance in differentiating between those who will go on to make more than one attempt. This includes characteristics such as age at the time of first attempt, method of attempt used, seriousness of suicidal intent (or intent seriousness), medical lethality, and receipt of treatment after the attempt. To date, only two older studies have examined this question, but few characteristics were included in analyses. Specifically, in a clinical sample of an unreported age, Worden and Sterling-Smith (1973) found that MAs, relative to SAs, reported a younger age at the time of their first attempt as well as lower medical lethality of their first attempt. Moreover, MAs who received “higher levels” of treatment versus “minimal or no” treatment after their first attempt, had less medically lethal second suicide attempts. In a study conducted by Michaelis et al. (2003), with a sample of adults diagnosed with Bipolar Disorder, MAs relative to SAs, reported lower intent seriousness of their first attempt. To date, no studies have explored this question in an adolescent sample or examined multiple characteristics of a first attempt concurrently.

Though there is minimal research on characteristics of a first attempt, more research has examined differences in characteristics of the most recent, or index, suicide attempts across SAs and MAs. In a longitudinal study conducted with high school students with a history of suicidal ideation or suicide attempt(s), Miranda et al. (2008) found that MAs reported a greater wish for death than SAs at baseline assessment. Further, at follow-up, a relative greater wish for death at baseline distinguished SAs who became MAs from SAs who never re-attempted. In a longitudinal study that followed adolescents who had been psychiatrically hospitalized over the course of 15 years, the highest lifetime rating of intent seriousness and medical lethality were each predictive of future attempts, though this association was lost when accounting for psychiatric disorders (Saptya et al., 2012). Relatedly, in a study conducted with adults recruited from an emergency room following a suicide attempt, no difference in intent seriousness or medical lethality of their most recent suicide attempts across SAs and MAs was found (Forman et al., 2004).

Method of attempt may also play a role in distinguishing between those who make one versus multiple suicide attempts. Notably, the primary classification of method of attempt used in this literature is “violent” versus “non-violent.” However, there is no standard system for classifying suicide attempt methods into “violent” versus “non-violent” categories. Further, many studies do not fully define methods (Ajdacic-Gross et al., 2003; Ajdacic-Gross et al., 2008; Lin, Chen, Xirasagar, & Lee, 2008) or will identify methods in one category but not clearly define the other (Dumais et al, 2005; Isometsä et al., 1994; Stenbacka & Jokinen, 2015). Generally, non-violent methods are often considered to be poisoning, overdosing, and drowning, while violent methods are often considered to be cutting, jumping, hanging, suffocation/strangulation, traffic accident, electrocution, and use of a firearm (Denning, Conwell, King, & Cox, 2000; Dumais et al., 2005; Isometsä et al., 1994; Stenbacka & Jokinen, 2015; but see Stenbacka & Jokinen, 2015 for classification of ‘drowning’ as violent).

To date, method of attempt has not been examined in relation to SAs and MAs. It has only been examined in adult samples who have died by suicide via psychological autopsy studies and use of mortality databases. For example, Anestis (2016) found that those who died via firearm were more likely to be SAs relative to those who used another method of attempt in their suicide death. Other studies have found a positive association between younger age and use of a violent method in suicide deaths (Dumais et al., 2005). Suicidal intent, defined as the will to die, was not found to differentiate between use of a violent versus non-violent method in suicide deaths in one study (Denning et al., 2000).

Receipt of treatment immediately following a first attempt may also distinguish SAs from MAs. Though clearly indicated, there are multiple barriers to receipt of mental health treatment. Parents are often unaware of suicide attempts made by their teenagers (Breton, Tousignant, Bergeron, & Berthiaume, 2002; Klaus, Mobilio, & King, 2009) and thus fail to seek treatment. Further, treatment services are often underutilized for reasons such as fear of stigma, cultural perceptions of therapeutic services, and structural barriers, among others (Esposito-Smythers et al., 2011a). This is unfortunate given that evidence-based treatment can lead to a lower incidence of future suicidal behavior (see Glenn, Franklin, & Nock, 2015 for review; Asarnow, Berk, Hughes, & Anderson, 2015; Esposito-Smythers, Spirito, Kahler, Hunt, & Monti, 2011b). No studies to date have examined whether receipt of treatment immediately following a first attempt influences the likelihood of repetitive suicidal behavior.

As suggested above, few studies have examined whether characteristics of a first suicide attempt can aid in the distinction between those who make single versus multiple attempts. Moreover, the few studies that have addressed this question did not examine multiple characteristics in the context of one comprehensive model. More commonly, characteristics of index attempts or deaths by suicide have been examined, and most studies have been conducted with adults. The present study builds upon this literature by including multiple characteristics of a first suicide attempt in a single model predicting the odds of SA versus MA status in a sample of adolescents who were psychiatrically hospitalized. Specifically, based on available literature, we hypothesized that younger age, higher intent seriousness, lower medical lethality, use of a non-violent (versus violent) method of attempt, and lack of mental health treatment immediately following a first attempt would each be associated with greater odds of MA status.

Methods

Participants

Participants included 95 adolescents hospitalized on a psychiatric inpatient unit in the Northeast and a parent. They were drawn from a larger study that examined the relationship between psychopathology, cognitive processes, and suicidal thoughts and behaviors in a sample of 186 adolescents. All participants selected for the present study had history of one or more actual suicide attempts. There were an additional nine participants identified in the parent study sample who reported only an aborted and/or interrupted attempt (as defined by Posner et al., 2011) but were not included in the present study. The 95 participants ranged in age from 13–18 years (M = 15.18, SD = 1.37) and were predominantly female (81.1%) and White (83.2%; 3.2% African American, 2.1% Native American, 2.1% Asian, 9.5% Other).

Inclusion criteria for the parent study was (1) fluency in English, and (2) Verbal IQ estimate of 70 or greater (assessed via the K-BIT; Kaufman & Kaufman, 1990). Exclusion criteria were (1) adolescents in full legal custody of Child Protective Services (per hospital admission records), and (2) active symptoms of psychosis.

Procedure

Adolescents and their parents were approached for recruitment during family visits or following family meetings on an adolescent inpatient unit. Trained masters- and doctoral-level clinicians conducted diagnostic interviews with adolescents and a parent. Trained research assistants administered the remaining self-report assessment battery. Upon completion of the full assessment battery, a summary of responses to clinical measures was added to each adolescent’s file to help inform treatment and discharge. All participants provided informed consent/assent, and both university and hospital Institutional Review Boards approved study procedures prior to the start of data collection.

Measures

Schedule of Affective Disorders and Schizophrenia for School-Aged Children, Present and Lifetime Version (K-SADS-PL; Kaufman, Birmaher, Brent, Rao, & Ryan, 1996)

The K-SADS-PL is a semi-structured diagnostic interview that is widely used and provides a reliable and valid measurement of DSM-IV-TR diagnoses in children and adolescents (Kaufman, Birmaher, Brent, Rao, & Ryan, 1997). Parents and adolescents completed the K-SADS-PL. Interviews were conducted and scored by one of six trained masters or postdoctoral degree level clinical psychology trainees. As part of their training, trainees were required to rate training audiotapes, observe and rate live interviews, and administer the K-SADS-PL while being observed. Upon completing this training, all interviews were audiotaped and 10% were randomly selected and rated for reliability. Kappa coefficients reflected strong agreement for major depressive and depressive NOS disorders (Kappa = .89–1.0), anxiety and post-traumatic stress disorders (kappa= .92–1.0), disruptive behavior and attention-deficit hyperactivity disorders (Kappa = 1.0), and substance use disorders (Kappa = .79–1.0). There was fair agreement for dysthymia and bipolar disorders (Kappa = .48-.65, respectively).

The presence of suicidal behavior was also assessed with the K-SADS-PL. For the suicide item in the K-SADS-PL, interviewers asked the following questions: “Have you ever tried to kill yourself? How many times?” Then for each individual attempt they asked: “How old were you the first time? What did you do? Was anyone in the room, apartment, house? Did you tell anyone in advance? Did you ask for any help after you did it? What happened to you after you tried to kill yourself? Were you unconscious? Did you get sick? Did you go to a doctor for medical treatment or an evaluation?” These data were used to code suicide attempt characteristics. A clinical consensus team, comprised of doctoral-level child psychologists and K-SADS interviewers, also reviewed all K-SADS symptoms and suicide data, including the coding of characteristics of each attempt (see coding of suicide attempts below). A best-estimate clinical consensus procedure was used to resolve discrepancies between adolescent and parent report as well as to confirm psychiatric diagnoses and suicide ratings. In instances of parent-adolescent disagreement, a suicide attempt and/or psychiatric diagnosis was coded as present if either the parent or adolescent endorsed it. A best-estimate clinical consensus procedure is commonly used to reconcile discrepancies (Cantwell et al., 1997, Klein et al., 1994; 2001) and yields good to excellent reliability (Klein et al., 1994; 2001).

Coding of Suicide Attempt Characteristics

Single versus multiple attempt status (SA vs. MA).

Participants were coded as single attempters (SA) if they had made only one lifetime suicide attempt and multiple attempters (MA) if they had made two or more lifetime suicide attempts.

Age of first attempt.

Participant age at the time of their first suicide attempt was recorded on a continuous scale.

Intent seriousness of suicidal acts.

The intent subscale from the Longitudinal Interval Follow-up Evaluation (LIFE; Keller et al., 1987) was used to code suicide attempts. To assess seriousness of intent expressed in suicidal acts, factors such as the likelihood of being rescued, actions taken to avoid discovery, efforts made to obtain help during or after the attempt, degree of planning, and likely purpose of attempt were taken into account. This rating is based on circumstances surrounding the attempt, not the adolescent’s expressed intent. Intent seriousness was coded on a six-point scale ranging from (1) “obviously no intent, purely manipulative” to (6) “extreme, every expectation of death.”

Medical lethality of suicidal acts.

The medical threat subscale from the LIFE (Keller et al., 1987) assesses actual medical threat to life or physical condition following a suicidal act. Variables such as method selected, degree of consciousness at time of rescue, seriousness of any physical injury, toxicity of ingested items, ability to reverse the effects of the attempt, and the amount of time and medical treatment needed for full recovery are taken into account. Guides are provided to aid in coding of lethality for different methods of attempt. Medical lethality is coded on a six-point scale ranging from (1) “no danger” to (6) “extreme danger.”

Receipt of treatment following attempt.

Coding for receipt of any mental health treatment (yes/no) immediately following the first suicide attempt was based on adolescent and parent responses provided to questions from the K-SADS interview. This categorization is consistent with previous methods of classifying treatment immediately following suicide attempts (e.g., Prinstein, Nock, Spirito, & Grapentine, 2001).

Method of attempt.

Method of attempt was coded using the LIFE (Kellar et al., 1987) Primary Method subscale. Consistent with past literature, various methods were then collapsed into “violent” or “non-violent” categories. The present study coded suicide attempt methods based on those most commonly reported in the literature (Ajdacic-Gross et al., 2008; Denning et al., 2000; Dumais et al., 2005; Isometsä et al., 1994; Stenbacka & Jokinen, 2015). Violent methods were cutting, jumping, hanging, shooting, drowning, explosion, choking, hit by traffic/train, suffocation, and electrocution. Non-violent methods were overdose, ingestion of a toxic substance, and dehydration. Of note, variability exists in the literature regarding whether drowning is considered violent (Stenbacka & Jokinen, 2015) or non-violent (Denning et al., 2000; Dumais et al., 2005; Isometsä et al., 1994).

Data Analytic Plan

All analyses were conducted using SPSS Version 19. Descriptive statistics were assessed to examine the distributional properties of variables and characterize the study sample. Correlations were run to examine the bivariate associations between variables. A multivariate hierarchical logistic regression analysis was conducted to examine hypotheses. The presence of a major depressive episode at the time of the first attempt, to control for depressive status, and age at assessment, to control for current age, were entered in the first block, and all other predictors were included in the second block. The decision to include all predictors, as opposed to just those that were significant at the bivariate level, was made to investigate the predictive ability of each attempt characteristic within the context of all other attempt characteristics.

Results

Sample Characteristics

Descriptive statistics for sample characteristics appear in Tables 1 and 2 and descriptives for all model variables appear in Table 3. All 95 participants who met eligibility criteria for this study were included in all analyses. More than half of participants (N = 50; 52.63%) had made multiple lifetime suicide attempts. A total of 218 suicide attempts (M = 2.10, SD = 2.05) were made in this sample. The mean age of first attempt was 14.19 years (SD = 1.84, range = 7–17). The method of attempt varied across the 218 total attempts. The most common methods were overdose, cutting, hanging, and drowning. Other attempt methods included choking, ingesting a toxic substance, suffocation, explosion, jumping, dehydration, and shooting. More than half of adolescents (N = 55; 57.9%) did not receive treatment immediately following their first attempt. For those adolescents who received some form of treatment post-attempt, the type and extent of treatment received varied. Treatment included: emergency psychiatric evaluation or psychiatric hospitalization with minimal medical treatment (N = 26); outpatient psychiatric evaluation with no medical evaluation (N = 8); or psychiatric hospitalization with significant medical treatment (N = 6). Approximately 37.9% of first suicide attempts were reported by both adolescent and parent, 50.5% reported by adolescents only, and 11.6% by parent only.

Table 1.

Descriptive statistics for sample characteristics (N = 95).

Variables N (%) M (SD)
Demographics
Gender
 Male 18 (18.90) .
 Female 77 (81.10) .
Age (current) . 15.18 (1.37)
Ethnicity
 Hispanic 14 (14.70)
 Non-Hispanic 81 (85.30)
Race
 White 79 (83.20) .
 African-American 3 (3.20) .
 Native American 2 (2.10) .
 Asian 2 (2.10) .
 Other 9 (9.50) .
Current DSM-IV Psychiatric Diagnoses
Major Depressive Disorder 65 (68.40) .
Dysthymia 4 (4.20) .
Depressive Disorder NOS 9 (9.50) .
Bipolar I Disorder 6 (6.30) .
Panic Disorder 12 (12.60) .
Agoraphobia 2 (2.10) .
Generalized Anxiety Disorder 28 (29.50) .
Social Anxiety Disorder 35 (36.80) .
Post-Traumatic Stress Disorder 29 (30.50) .
Acute Stress Disorder 2 (2.10) .
Bulimia 2 (2.10) .
Eating Disorder NOS 13 (13.70) .
Attention Deficit/Hyperactivity Disorder 31 (32.60) .
Conduct Disorder 27 (28.40) .
Oppositional Defiant Disorder 16 (16.80) .
Adjustment Disorder 4 (4.20) .
Alcohol Abuse Disorder 14 (14.70) .
Alcohol Dependence Disorder 10 (10.50) .
Substance Abuse Disorder 15 (15.80) .
Substance Dependence Disorder 19 (20.00) .
Substance-Induced Mood Disorder 1 (1.10) .
Anxiety Disorder NOS 5 (5.30) .
No Psychiatric Diagnosis 2 (2.10) .
Open in a new tab

Note: No participants in the sample met criteria for Bipolar II or NOS disorders or Anorexia Nervosa.

Table 2.

Methods used across all suicide attempts (N = 218).

Method of Attempt N (%)
Violent 124 (56.88)
  Cutting 75 (34.40)
  Hanging 25 (11.47)
  Drowning 9 (4.13)
  Choking 7 (3.21)
  Suffocation 3 (1.38)
  Explosion 2 (0.92)
  Jumping 2 (0.92)
  Shooting 1 (0.46)
Non-Violent 94 (43.12)
  Overdose 89 (40.83)
  Ingestion of toxic substance 4 (1.83)
  Dehydration 1 (0.46)
Open in a new tab

Table 3.

Mean values and frequencies for predictor variables as a function of SA vs. MA status.

Variable Full sample (N = 95) SA status (n = 45) MA status (n = 50)
Age at time of first attempt 14.19 (1.84) 14.84 (1.41) 13.60 (1.98)
Intent seriousness 2.69 (1.15) 2.87 (1.24) 2.54 (1.05)
Lethality 2.07 (1.05) 2.31 (1.18) 1.86 (0.88)
Receipt of treatment
  No 55 (57.9%) 17 (37.8%) 38 (76.0%)
  Yes 40 (46.3%) 28 (62.2%) 12 (24.0%)
Method type
  Nonviolent 51 (53.7%) 31 (68.9%) 20 (40.0%)
  Violent 44 (46.3%) 14 (31.1%) 30 (60.0%)
Attempt within depressive episode
  No 26 (27.4%) 11 (24.4%) 15 (30.0%)
  Yes 69 (72.6%) 34 (75.6%) 35 (70.0%)
Age (current) 15.18 (1.37) 15.20 (1.39) 15.16 (1.36)
Open in a new tab

Note: Means and standard deviations reported for age of attempt, intent seriousness, lethality, and age at assessment. Frequency and percentage reported for all other variables.

Bivariate correlations among model and outcome variables appear in Table 4. As can be seen, younger age at time of the first attempt, lower lethality of the first attempt, use of a violent method in the first attempt, and no receipt of treatment immediately following the first attempt, were associated with attempter status at the bivariate level. Seriousness of suicidal intent of the first suicide attempt was not associated with attempter status. Examination of variance inflation factors (VIFs) of predictor variables indicated that level of multicollinearity did not preclude inclusion of all predictors in the same logistic regression model.

Table 4.

Bivariate correlations for all model variables (N = 95).

Variable 1. 2. 3. 4. 5. 6. 7.
1. Single vs. multiple attempt status --
2. Age of attempt −0.34** --
3. Intent seriousness −0.14 0.05 --
4. Lethality −0.22* 0.04 0.35** --
5. Receipt of treatment −0.39** 0.13 0.12 0.25* --
6. Method type 0.29** −0.14 −0.23** −0.01 −0.28** --
7. Attempt within depressive episode −0.06 0.14 −0.06 0.002 0.05 0.10 --
8. Age (current) −0.02 0.61** 0.08 −0.02 0.11 −0.12 −0.20
Open in a new tab
*

p < 0.05.

**

p < 0.01.

Note: Reference categories are: single attempt; “no” receipt of treatment; nonviolent method; “no” depressive episode.

Multivariate Analyses

In the multivariate hierarchical logistic regression predicting SA vs. MA status, current major depressive episode and age at assessment were simultaneously entered into block 1, and age of first attempt, intent seriousness, medical lethality, receipt of treatment following attempt, and method of attempt were simultaneously entered into the second block. Results are presented in detail in Table 5. The overall model significantly predicted SA vs. MA status, χ2 (7) = 42.01, p < .001. Of these predictors, only age of first attempt, OR = 0.33, 95% CI [0.17–0.63], p = .001, and receipt of treatment following attempt, OR = 0.28, 95% CI [0.09–0.88], p = .028, significantly predicted SA vs. MA status. These findings suggest that the odds of being of MA status are higher for adolescents whose first attempt was at an earlier age, as well as for those who did not receive some form of treatment immediately following their first attempt.

Table 5.

Multivariate hierarchical logistic regression analysis predicting single versus multiple attempt status (N = 95).

Block and variable b SE Wald statistic Odds ratio (95% confidence interval) χ2
Block 1 χ2 (2) = 0.44
  Depressive episode −0.31 0.47 0.42 0.74 (0.29–1.86)
  Age (current) −0.04 0.16 0.07 0.96 (0.71–1.30)
Block 2 χ2 (5) = 41.57**
  Age of first attempt −1.11 0.33 11.48 0.33 (0.17–0.63)**
  Intent seriousness −0.04 0.25 0.03 0.96 (0.59–1.56)
  Lethality −0.49 0.32 2.41 0.61 (0.33–1.14)
  Receipt of treatment −1.26 0.58 4.80 0.28 (0.09–0.88)*
  Method of attempt 1.05 0.57 3.34 2.84 (0.93–8.72)
Open in a new tab

Note:

*

p < 0.05.

**

p < 0.01.

Non-significant predictors have odds ratio confidence intervals that cross a value of 1.

Consistent with recommendations offered by Rogers et al. (2018), multivariate analyses were re-run with major depressive disorder removed as a covariate from the model. The pattern of results remained unchanged.

Discussion

Characteristics of a first suicide attempt hold the potential to aid in differentiating between adolescents who make single versus multiple suicide attempts. Yet, only two older studies have examined this question (Michaelis et al., 2003; Worden & Sterling-Smith, 1973) and the scope of these studies was limited. Other studies have examined characteristics of index (i.e., most recent) suicide attempts across SAs and MAs, and this work has only been conducted using adult samples or psychological autopsy methods (Anestis, 2016; Denning et al., 2000; Dumais et al., 2005). The present study builds upon the existing literature by examining multiple characteristics of a first suicide attempt as predictors of suicide attempt status, SA versus MA, within a single model. Moreover, to provide a conservative test of study hypotheses, age at assessment and major depressive disorder at time of attempt were controlled for in study analyses. This question was examined in a sample of psychiatrically hospitalized adolescents at significant risk for repetitive suicidal behavior.

Of the first attempt characteristics examined concurrently, only age at time of attempt and receipt of mental health treatment post-attempt were significantly associated with SA vs. MA status. Specifically, as age of first attempt decreased, adolescents were more likely to be of MA status. Additionally, adolescents who did not receive treatment immediately following their first attempt were more likely to be of MA status. No other attempt characteristics (i.e., intent seriousness, lethality, violent vs. non-violent method) significantly predicted SA vs. MA status in the multivariate model.

The finding that younger age at time of first attempt was associated greater likelihood of multiple attempts is consistent with prior research (Worden & Sterling-Smith, 1973). It is possible that these younger youth have more severe clinical histories. For example, prior research suggests that younger age of first attempt has been linked to familial history of suicidal behavior (Roy, 2004) and childhood sexual abuse (Enns, et al., 2006). Younger youth may also be less likely to disclose their suicidal behavior. Younger age has been associated with lower odds of parental knowledge of youth suicide attempts (Klaus, Mobilio, & King, 2009), thus decreasing the likelihood of receipt of mental health treatment.

Consistent with prior research (Worden & Sterling-Smith, 1973), lack of mental health treatment post-attempt was also associated with increased odds of MA status. Only one-quarter of adolescents who made multiple attempts received any form of treatment immediately following their first suicide attempt. Notably, in the present study, parents were unaware of over half of the first suicide attempts reported by adolescents. These findings are consistent with prior research which suggests that parent and adolescent report of adolescent suicidal behaviors are often discrepant, with adolescents reporting more suicide attempts than their parents (Berk & Asarnow, 2015; Breton et al., 2002; Klaus et al., 2009). Thus, it is possible that lack of parental knowledge of adolescent first attempt played a role in treatment receipt.

In the present study, method (violent vs. non-violent) and lethality of a first attempt were found to be associated with MA status only at the bivariate level. Specifically, use of a violent method (i.e., cutting, hanging, drowning, choking, suffocation, explosion, jumping, shooting) and a less medically severe first attempt, respectively, were associated with MA status. With regard to violent attempts, the large majority were made via cutting and secondarily hanging. These results may suggest that adolescents who are able to inflict and withstand bodily pain in their first attempt may be more likely to engage in repetitive suicidal behavior relative to those who use less painful non-violent means such an as overdose. Indeed, higher pain tolerance is associated with a greater acquired capability for suicide (Franklin, Hessel, & Prinstein, 2011). With regard to lethality, one prior study also found an association between lower medical lethality of a first attempt and MA status (Worden & Sterling-Smith, 1973). These findings may reflect that adolescents inaccurately estimate lethality of methods of attempts. Indeed, adolescent suicidal intent and medical lethality of attempt are often discrepant (Sapyta et al., 2012). It is also possible that lower lethality attempts serve as a means of “behavioral rehearsal” for more severe attempts. Most importantly, these findings suggest that attempts of low medical lethality should not be discounted when conducting youth suicide risk assessments.

Though associated with attempter status, neither method nor medical lethality of a first suicide attempt retained significance in a multivariate model with other attempt characteristics. In the present sample, method of a first suicide attempt and lethality appeared to be relatively less useful in differentiating between high-risk adolescents who make single versus multiple suicide attempts than earlier age of first attempt and receipt of immediate treatment post-attempt. It is possible that shared variance between attempt characteristics may have attenuated the strength of these relationships with attempter status when included in the same model. Variability reflected in the method of attempt variable may have also been attenuated by adolescents’ limited access to the most lethal methods (e.g., firearms). Further, in the present study, medical lethality of attempts was low in general. Approximately 90% of first attempts were classified as “no,” “minimal,” or “mild” danger, which restricted the range in this variable.

In contrast to other attempt characteristics examined, seriousness of suicidal intent evidenced in suicidal acts was not associated with attempter status at the bivariate or multivariate level. In the present study, circumstances surrounding the attempt, not the adolescent’s expressed intent, was measured. One prior study that did find a positive association between intent seriousness and subsequent suicide attempts (Sapyta et al., 2012) used a measure of participant “expressed” intent. Moreover, most attempts made in this sample (75%) were coded as having relatively low intent (i.e., no, minimal intent, or definite but still ambivalent intent) which limited variability.

Clinical Implications

Findings from the present study hold a number of important clinical implications. First, the association between younger age of first attempt and MA status highlights the importance of early mental health screening and intervention. In the present sample, several adolescents made first attempts as early as seven years old. Relatedly, as more than half of all first attempts reported by adolescents were unknown to their parents, and lack of receipt of immediate mental health treatment post-attempt was associated with MA status, parent psychoeducation may also play an important role in youth suicide prevention. Parent gatekeeper and psychoeducation programs may help parents identify risk factors and warning signs for suicidal behavior, learn how to talk with their children about and monitor for suicidality, understand the importance of seeking mental health treatment, and obtain instruction in how to navigate the mental health system to acquire treatment when concerns arise. Last, while method of first attempt and lethality are important to consider in risk assessments and safety planning, they may be less useful, relative to other attempt characteristics, in determining the likelihood of a repetitive attempt.

Limitations and Future Research Directions

Several limitations exist within the present study. First, the sample used in this study included an overrepresentation of female and White participants. It is also comprised of high-risk adolescents recruited from a psychiatric inpatient unit. Though youth psychiatrically hospitalized for suicidal events are at significantly heightened risk for future attempts, the results of this study may not generalize to adolescents with less severe clinical presentations. Therefore, future studies should target a more diverse sample and include a greater range of clinical presentations. Second, for those adolescents who received treatment immediately following their first attempt, the type of psychiatric treatment received varied (e.g., outpatient, inpatient, etc.). However, there was not enough power to examine differences in outcomes based on treatment type. Third, the relatively small number of multiple attempters included in the study sample and cross-sectional research design negated our ability to examine potential moderators and mediators of outcomes. Future research in this area should employ larger samples, prospectively follow adolescents after a first attempt, and explore factors that influence outcomes to further increase our knowledge and understanding in this area. Finally, adolescents were asked to retrospectively recall prior suicide attempts which introduced the potential for retrospective recall bias. However, adolescents were able to provide ample details around each attempt when probed, parent as well as adolescent report was obtained, and all attempts were staffed during clinical consensus meetings, which bolsters the validity of suicide attempt ratings.

Conclusions

The present study demonstrates the relative importance of age of first suicide attempt and receipt of treatment immediately following a first attempt in differentiating between adolescents who make single versus multiple attempts. Though method and medical lethality of a first attempt were associated with attempt status at the bivariate level, these characteristics may not be as useful in determining attempter status relative to other first attempt characteristics. Study findings highlight the importance of early mental health screening, parental psychoeducation on youth suicidal behavior, and linkage to mental health services following a suicide attempt.

References

  1. Ajdacic-Gross V, Wang J, Bopp M, Eich D, Rossler W, & Gutzwiller F (2003). Are seasonalities in suicide dependent on suicide methods? A reappraisal. Social Science & Medicine, 57, 1173–1181. 10.1016/S0277-9536(02)00493-8 [DOI] [PubMed] [Google Scholar]
  2. Ajdacic-Gross V, Weiss MG, Ring M, Hepp U, Bopp M, Gutzwiller F, & Rossler W (2008). Methods of suicide: international suicide patterns derived from the WHO mortality database. Bulletin of the World Health Organization, 86(9), 726–732. doi: 10.2471/BLT.07.043489 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Anestis MD (2016). Prior suicide attempts are less common in suicide decedents who died by firearms relative to those two died by other means. Journal of Affective Disorders, 189, 106–109. doi: 10.1016/j.jad.2015.09.007 [DOI] [PubMed] [Google Scholar]
  4. Asarnow JR, Berk M, Hughes JL, & Anderson NL (2015). The SAFETY Program: A Treatment-Development Trial of a Cognitive-Behavioral Family Treatment for Adolescent Suicide Attempters. Journal of Clinical Child and Adolescent Psychology, 44(1), 194–203. doi: 10.1080/15374416.2014.940624 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Berk MS, & Asarnow JR (2015). Assessment of Suicidal Youth in the Emergency Department. Suicide and Life-Threatening Behavior, 45(3), 345–359. 10.1111/sltb.12133 [DOI] [PubMed] [Google Scholar]
  6. Breton J, Tousignant M, Bergeron L, & Berthiaume C (2002). Informant-Specific Correlates of Suicidal Behavior in a Community Survey of 12-to 14-Year-Olds. Journal of the American Academy of Child and Adolescent Psychiatry, 41(6), 723–730. 10.1097/00004583-200206000-00012 [DOI] [PubMed] [Google Scholar]
  7. Cantwell DP, Lewinsohn PM, Rohde P, & Seeley JR (1997). Correspondence between adolescent report and parent report of psychiatric diagnostic data. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 610–619. 10.1097/00004583-199705000-00011 [DOI] [PubMed] [Google Scholar]
  8. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. (2019). Web-based Injury Statistics Query and Reporting System (WISQARS) Retrieved from www.cdc.gov/injury/wisqars
  9. Christiansen E, & Frank Jensen B (2007). Risk of repetition of suicide attempt, suicide or all deaths after an episode of attempted suicide: a register-based survival analysis. Australian and New Zealand Journal of Psychiatry, 41(3), 257–265. [DOI] [PubMed] [Google Scholar]
  10. Clement S, Schauman O, Graham T, Maggioni F, Evans-Lack S, Bezborodovs N, … Thornicroft G (2015). What is the impact of mental health-related stigma on help-seeking? A systematic review of quantitative and qualitative studies. Psychological Medicine, 45, 11–27. doi: 10.1017/S0033291714000129 [DOI] [PubMed] [Google Scholar]
  11. Corrigan PW, Druss BG, & Perlick DA (2014). The Impact of Mental Illness Stigma on Seeking and Participating in Mental Health Care. Psychological Science in the Public Interest, 15(2), 37–70. doi: 10.1177/1529100614531398 [DOI] [PubMed] [Google Scholar]
  12. Denning DG, Conwell Y, King D, & Cox C (2000). Method Choice, Intent, and Gender in Completed Suicide. Suicide and Life-Threatening Behavior, 30(3), 282–288. 10.1111/j.1943-278X.2000.tb00992.x [DOI] [PubMed] [Google Scholar]
  13. Dumais A, Lesage AD, Lalovic A, Senguin M, Tousignant M, Chawky N, & Turecki G (2005). Is Violent Method of Suicide a Behavioral Marker of Lifetime Aggression? American Journal of Psychiatry, 162, 1375–1378. doi: 10.1176/appi.ajp.162.7.1375 [DOI] [PubMed] [Google Scholar]
  14. Enns MW, Cox BJ, Afifi TO, De Graaf R, Ten Have M, & Sareen J (2006). Childhood adversities and risk for suicidal ideation and attempts: a longitudinal population-based study. Psychological Medicine, 36(12), 1769–1778. 10.1017/S0033291706008646 [DOI] [PubMed] [Google Scholar]
  15. Esposito C, Spirito A, Boergers J, & Donaldson D (2003). Affective, Behavioral, and Cognitive Functioning in Adolescents with Multiple Suicide Attempts. Suicide and Life-Threatening Behavior, 33(4), 389–399. 10.1521/suli.33.4.389.25231 [DOI] [PubMed] [Google Scholar]
  16. Esposito-Smythers C, Miller A, Weismoore J, Doyle O, & Goldston D (2011a). Suicide. In Bradford Brown B and Prinstein M (Eds.), Encyclopedia of Adolescence New York, NY: Academic Press. [Google Scholar]
  17. Esposito-Smythers C, Spirito A, Kahler CW, Hunt J, & Monti P (2011b). Treatment of Co-Occurring Substance Abuse and Suicidality Among Adolescents: A Randomized Trial. Journal of Consulting and Clinical Psychology, 79(6), 728–739. doi: 10.1037/a0026074 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Franklin JC, Hessel ET, & Prinstein MJ (2011). Clarifying the role of pain tolerance in suicidal capability. Psychiatry Research, 189(3), 362–367. [DOI] [PubMed] [Google Scholar]
  19. Forman EM, Berk MS, Henriques GR, Brown GK, & Beck AT (2004) History of Multiple Suicide Attempts as a Behavioral Marker of Severe Psychopathology. The American Journal of Psychiatry, 161, 437–443. doi: 10.1176/appi.ajp.161.3.437 [DOI] [PubMed] [Google Scholar]
  20. Glenn CR, Franklin JC, & Nock MK (2015). Evidence-Based Psychosocial Treatments for Self-Injurious Thoughts and Behaviors in Youth. Journal of Clinical Child and Adolescent Psychology, 44(1), 1–29. doi: 10.1080/15374416.2014.945211 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Goldston DB, Erkanli A, Daniel SS, Heilbron N, Weller BE, & Doyle O (2016). Developmental Trajectories of Suicidal Thoughts and Behaviors from Adolescence Through Adulthood. Journal of the American Academy of Child and Adolescent Psychiatry, 55(5), 400–407. doi: 10.1016/j.jaac.2016.02.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hughes JL, & Asarnow JR (2013). Enhanced Mental Health Interventions in the Emergency Department: Suicide and Suicide Attempt Prevention in the ED. Clinical Pediatric Emergency Medicine, 14(1), 28–34. doi: 10.1016/j.cpem.2013.01.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Isometsä ET, Henriksson MM, Aro HM, Heikkinen ME, Kuoppasalmi KI, Lönnqvist JK (1994). Suicide in Major Depression. The American Journal of Psychiatry, 151(4), 530–536. doi: 10.1176/ajp.151.4.530 [DOI] [PubMed] [Google Scholar]
  24. Kaufman J, Birmaher B, Brent D, Rao U, & Ryan N (1996). Diagnostic Interview: Kiddie-SADS-Present and Lifetime Version (K-SADS-PL). Kaufman, Birmaher, Rao, & Ryan
  25. Kaufman J, Birmaher B, Brent DA, Rao U, & Ryan ND (1997). Schedule for affective disorders and schizophrenia for school age children, present and lifetime version (K-SADS-PL): Initial reliability and validity data. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 980–988. 10.1097/00004583-199707000-00021 [DOI] [PubMed] [Google Scholar]
  26. Kaufman AS, & Kaufman NL (1990). K-BIT: Kaufman Brief Intelligence Test: Manual American Guidance Service. [Google Scholar]
  27. Keller MB, Lavori PW, Friedman B, Nielsen E, Endicott J, McDonald-Scott P, & Andreasen NC (1987). The Longitudinal Interview Follow-up Evaluation: A Comprehensive Method for Assessing Outcome in Prospective Longitudinal Studies. Archives of General Psychiatry, 44(6), 540–548. doi: 10.1001/archpsyc.1987.01800180050009 [DOI] [PubMed] [Google Scholar]
  28. Klaus NM, Mobilio A, & King CA (2009). Parent-Adolescent Agreement Concerning Adolescents’ Suicidal Thoughts and Behaviors. Journal of Clinical Child and Adolescent Psychology, 38(2), 245–255. doi: 10.1080/15374410802698412 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Klein DN, Lewinsohn PM, Seeley JR, & Rohde P (2001). A family study of major depressive disorder in a community sample of adolescents. Archives of General Psychiatry, 58, 13–20. 10.1001/archpsyc.58.1.13 [DOI] [PubMed] [Google Scholar]
  30. Klein DN, Ouimette PC, Kelly HS, Ferro T, & Riso LP (1994). Test-retest reliability team consensus best-estimate diagnoses of Axis I and II disorders in a family study. American Journal of Psychiatry, 151, 1043–1047. 10.1176/ajp.151.7.1043 [DOI] [PubMed] [Google Scholar]
  31. Lin H, Chen C, Xirasagar S, & Lee H (2008). Seasonality and Climatic Associations with Violent and Nonviolent Suicide: A Population-Based Study. Neuropsychobiology, 57, 32–37. 10.1159/000129664 [DOI] [PubMed] [Google Scholar]
  32. Michaelis BH, Goldberg JF, Singer TM, Garno JL, Ernst CL, & Davis GP (2003). Characteristics of First Suicide Attempts in Single Versus Multiple Suicide Attempters with Bipolar Disorder. Comprehensive Psychiatry, 44(1), 15–20. doi: 10.1053/comp.2003.50004 [DOI] [PubMed] [Google Scholar]
  33. Miranda R, Scott M, Hicks R, Wilcox HC, Munfakh JLH, & Shaffer D (2008). Suicide Attempt Characteristics, Diagnoses, and Future Attempts: Comparing Multiple Attempters to Single Attempters and Ideators. Journal of the American Academy of Child and Adolescent Psychiatry, 47(1), 32–40. doi: 10.1097/chi.0b013e31815a56cb [DOI] [PubMed] [Google Scholar]
  34. Nock MK (2009). Why do people hurt themselves? New insights into the nature and functions of self-injury. Current Directions in Psychological Science, 18, 78–83. doi: 10.1111/j.1467-8721.2009.01613.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Pagura J, Cox BJ, Sareen J, & Enns MW (2008). Factors Associated with Multiple Versus Single Episode Suicide Attempts in the 1990–1992 and 2001–2003 United States National Comorbidity Surveys. The Journal of Nervous and Mental Disease, 196(11), 806–813. doi: 10.1097/NMD.0b013e31818b6a77 [DOI] [PubMed] [Google Scholar]
  36. Posner K, Brown GK, Stanley B, Brent DA, Yershova KV, Oquendo MA, … Mann JJ (2011). The Columbia-Suicide Severity Rating Scale: Initial Validity and Internal Consistency Findings From Three Multisite Studies With Adolescents and Adults. American Journal of Psychiatry, 168, 1266–1277. doi: 10.1176/appi.ajp.2011.10111704 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Prinstein MJ, Nock MK, Spirito A, & Grapentine WL (2001). Multimethod Assessment of Suicidality in Adolescent Psychiatric Inpatients: Preliminary Results. Journal of the American Academy of Child and Adolescent Psychiatry, 40(9), 1053–1061. 10.1097/00004583-200109000-00014 [DOI] [PubMed] [Google Scholar]
  38. Ribeiro JD, Franklin JC, Fox KR, Bentley KH, Kleiman EM, Chang BP, & Nock MK (2016). Self-injurious thoughts and behaviors as risk factors for future suicide ideation, attempts, and death: a meta-analysis of longitudinal studies. Psychological Medicine, 46(2), 225–236. doi: 10.1017/S0033291715001804 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rogers ML, Stanley IH, Hom MA, Chiurliza B, Podlogar MC, & Joiner TE (2018). Conceptual and empirical scrutiny of covarying depression out of suicidal ideation. Assessment, 25(2), 159–172. doi: 10.1177/1073191116645907 [DOI] [PubMed] [Google Scholar]
  40. Roy A (2004). Family history of suicidal behavior and earlier onset of suicidal behavior. Psychiatry Research, 129(2), 217–219. doi: 10.1016/j.psychres.2004.08.002 [DOI] [PubMed] [Google Scholar]
  41. Sapyta J, Goldston DB, Erkanli A, Daniel SS, Heilbron N, Mayfield A, & Treadway SL (2012). Evaluating the Predictive Validity of Suicidal Intent and Medical Lethality in Youth. Journal of Consulting and Clinical Psychology, 80(2), 222–231. doi: 10.1037/a0026870 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Stenbacka M, & Jokinen J (2015). Violent and nonviolent methods of attempted and completed suicide in Swedish young men: the role of early risk factors. BMC Psychiatry, 15. doi: 10.1186/s12888-015-0570-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Stewart JG, Glenn CR, Esposito EC, Cha CB, & Nock MK (2017). Cognitive Control Deficits Differentiate Adolescent Suicide Ideators from Attempters. Journal of Clinical Psychiatry doi: 10.4088/JCP.16m10647 [DOI] [PubMed] [Google Scholar]
  44. Van Orden KA, Witte TK, Cukrowicz KC, Braithwaite SR, Selby EA, & Joiner TE (2010). The Interpersonal Theory of Suicide. Psychological Review, 117(2), 575–600. doi: 10.1037/a0018697 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Worden JW, & Sterling-Smith RS (1973). Lethality Patterns in Multiple Suicide Attempts. Life-Threatening Behavior, 3(2), 95–104. 10.1111/j.1943-278X.1973.tb00977.x [DOI] [Google Scholar]

RESOURCES