Introduction

Suicide and different forms of sudden violent death (SVD) are the most common causes of death among young people worldwide. The phrase “young people,” the target group for our investigation, is generally used in Sweden and the Anglo-Saxon countries as a generic term for children, adolescents, and young adults. To better understand young people who kill themselves or expose themselves to risks resulting in SVD, we need to know more about the interplay between adverse life events, both in the far past and present, vulnerability factors, and coping deficits that may result in such lethal outcomes.

Suicide is defined by the World Health Organization [1] as “the act of deliberately killing oneself,” and a suicide attempt as “any non-fatal suicidal behaviour and refers to intentional self-inflicted poisoning, injury or self-harm which may or may not have a fatal intent or outcome.” The report warns that it can be difficult to distinguish between self-harm with or without suicidal intent, suggesting complexities in demarcating suicide and SVD.

In psychological terms, the process behind overt suicide and other forms of self-destructive behavior is often interpreted as hostile impulses winning over self-protecting forces. For example, Menninger [2] regarded self-destructive acts as inwards-directed aggressiveness. Furthermore, he described abortive, distorted, or attenuated forms of “latent suicide” as result of indirect or incomplete self-destructive behavior, such as addiction or accidents. According to Goldbladt’s [3] more recent psychoanalytic perspective, suicide takes place in the interpersonal and intrapsychic context of unbearable hostility towards the self. Accordingly, a French prospective epidemiological cohort study [4] found that increased risk of suicide is related to cognitive rather than behavioral hostility. The currently often cited Columbia—Suicide Severity Rating Scale (C-SSRS) [5] includes a definition of actual attempt, interrupted attempt, aborted or self-interrupted attempt or preparatory act or behavior to define different levels of suicidal behavior. Accordingly, several authors [6, 7] described a continuum of self-destructive behavior among younger age groups from covert expressions to overt suicidal behavior.

According to current reviews, common risk factors for suicide from childhood to young adulthood include genetic and epigenetic factors, early life adversities, lack of social support, life events, access to lethal means, effects of the media, severe mental illness, depression, personality disorder, substance misuse, economic factors, and physical health problems [8], as well as history of suicidal behavior and psychiatric care [9–11]. A systematic review [12] found evidence for an association between life stressors, particularly interpersonal stressors, and death by suicide. As summarized in an overview of psychological models of suicide [13], some models focus on vulnerability factors, such as “impulsive aggressive tendencies, maladaptive cognitive styles, problem solving deficits, attention bias, over-general memory, and acquired capability for self-harm,”, whereas other models emphasize the role of stressful life events, leading to “mental pain, hopelessness, entrapment, and interpersonal distress” (p. 306). However, few theoretical models and empirical studies try to integrate these two clusters of factors or grasp the interactions between different factors.

Major risk factors for SVD among young people, identified in several studies [14–17], include being of male sex, antisocial personality disorder, criminality, alcohol and drug abuse, adverse family psychosocial characteristics, as well as aggressive feelings and acts against oneself and others, health-compromising behavior, and risk-taking behavior. Thus, there are some similarities but also several differences between risk factors for suicide and SVD among youth. Accordingly, the two-stage model of suicide and violence [13] assumes that both suicide and SVD are expressions of the same underlying aggressive impulse. Hence, other intervening variables determine whether the aggression is directed inward or expressed in outward-directed behavior.

Major protective factors for youth suicidality, found in empirical studies [18, 19] and reviews [20, 21], include parental presence, connectedness to parents and peers, belongingness to community and social institutions, positive connection to school and academic achievement, social competence, coping and problem-solving skills, contacts with caregivers, and effective mental health care. Much less is known about protective factors against SVD. A systematic review [22] found that adaptive coping has a protective function against stress and is related to well-being in the transition to adulthood. Thus, adequate coping skills may be expected to have protective function against both suicide and SVD.

In a previous study [23], we examined associations between stressful life events and coping strategies in cases of youth suicide and in cases of youth SVD, as compared to general population control cases. We found that between-group differences in coping were partly accounted for by differences in negative life events, early and later in life. In the present study we explore the assumption that there may be both common factors, distinguishing children, adolescents, and young adults who died by suicide or SVD from the general population, and specific factors that are unique for the two groups, comparing the two target groups on variables significantly distinguishing them from control cases. A review [24] found that the term “sudden violent death” usually includes death by suicide, accident, homicide, or overdose. In the present study, suicide is defined as the act of deliberately killing oneself, and SVD as unintentional injury-related (that is, non-suicidal) death, which may still have occurred due to an underlying, hidden intention to die. As far as we know, there is a scarcity of studies and lack of reviews exploring common paths and dividing developments for suicide and SVD among children and young people.

Methods

Study design and sample

The present study is based on archival interview data collected from May 28, 2001, to January 28, 2008, in an investigation of suicide and SVD (murder, accident, unclear accident) up to the age of 25 years in Stockholm County, Sweden, and including a control group. Consequently, the sample and some of the data presented here have already been used in previous publications [23, 25]. The project was approved by the Regional Ethical Review Board, Karolinska Institutet, Stockholm (reference number 96–204 and 2005/530-32), and for the control group also the Swedish State Personal Address Register (reference number 2004/0146). All informants (parents and relatives of the deceased, as well as all participants in the control group and their parents) gave written informed consent.

In order to compute the required sample size a priori power analysis was conducted applying G*Power [26]. For a one-way ANOVA with three groups, 0.05 alpha level, effect size f = 0.25 (medium effect size corresponding to Cohen’s d = 0.5), and power 0.8, the required sample size per group is 53. With 60 persons per group, the achieved power will be 0.85.

Consecutive cases of non-natural death among children, adolescents, and young adults were identified at the Department of Forensic Medicine in Stockholm, which is responsible for all forensic autopsies in the Stockholm Region. Information on causes of death was based on autopsy protocols and police reports. Consecutive cases of suicide were collected from October 6, 2000, through December 30, 2004, and the consecutive cases of SVD from October 1, 2000, through September 11, 2002. Such a long time was required to achieve the target number of at least 60 cases in both groups. The control cases were collected from a randomized sample obtained from the population registry in Stockholm County (data collected from January 18, 2006, through January 28, 2008). Looking for risk factors for suicide and SVD, the control group was originally matched with the two target groups taken together on two variables: gender and age, according to a review [27] the most common procedure in psychological autopsy studies. No other matching criteria were used, as we wanted to include a wide range of potential risk factors, such as strains in life and socio-demographic characteristics.

In the present study, the anonymized database was accessed from July 6, 2019. All interview data were collected by the first author who unavoidably had access to information that could identify individual participants during the data collection. None of the other authors had access to such information during or after data collection.

Two target samples and a control group were included (for sociodemographic characteristics and study variables, see Table 1). Of the 63 cases of suicide (aged 12–25 years), 41 were males and 22 females. Seven of the cases of suicide (11%) were younger than 18 years. Of the 62 cases of SVD (aged 10–25 years), 55 were males and 7 females. Ten of the cases of SVD (16%) were younger than 18 years. Both target groups, taken together, included 125 cases of premature unnatural death; 96 males (77%) aged 10–25 years (M = 21.0; SD = 3.2; Md = 22) and 29 females (23%) aged 14–24 years (M = 19.8; SD = 3.0; Md = 22). This can be compared with the 104 control cases, of which 76 were males (73%) aged 10–25 years (M = 20.7; SD = 3.4; Md = 21) and 28 were females (27%) aged 14–24 years (M = 19.7; SD = 3.0; Md = 20). In all, 229 families took part in the study (among the control families, the individual young person was included in the interview).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sociodemographic, psychosocial and psychiatric data, and univariate effects of the potential risk factors on the dependent variable.

https://doi.org/10.1371/journal.pone.0313673.t001

Results

Approximately half of those who died by suicide (50.8%; 54.5% females and 48.8% males) had made previous suicide attempts, whereas none of the females and only 10.9% of the males attempted suicide in the SVD group. Looking at distinguishing risk factors for the two target groups, identified in univariate analyses, we found 21 significant differences on variables significantly different from the control cases (yellow-highlighted in Table 1). Distinguishing for the suicide group was lower frequency of living in a steady relationship (28.6%), adult outpatient (55.6%) and inpatient (39.7%) psychiatric care, depression (66.7%), autism spectrum disorder (17.5%), being sexually assaulted (20.6%), higher frequency of recent stressful life events (98.4%), and lower levels of Planful Problem-Solving (M = -0.56). Distinguishing for the SVD group was a predominance of males (88.7%), younger mother (M = 27.1), lower elementary school results (30.6%), abuse of psychoactive drugs (32.3%), being investigated or sentenced for criminal acts (58.1%), conduct disorder (30.6%) or antisocial personality disorder (36.5%).

Looking at similarities between the suicide group and the SVD group, we found 16 risk factors common for both groups, but distinguishing both of them from the control group (yellow-highlighted in Table 1). Common risk factors included lower educational level, absence of work or studies, different forms of addiction, child and adolescent psychiatric care, borderline personality disorder, adverse childhood experiences, low levels of Planful Problem-Solving, and high levels of Escape-Avoidance and Confrontive Coping.

In regression analyses with group as a dichotomous predictor there was a significant difference between the suicide and the SVD group in Planful Problem-Solving (b = 0.416, p = 0.026) but not on the other three coping factors (p = 0.058, 0.168, and 0.523 for Escape-Avoidance, Seeking Social Support, and Confrontive Coping, respectively). Taken together, these analyses indicate that the suicide group had lower levels of the more adaptive coping strategy Planful Problem-Solving than the SVD group, but both groups had similarly high levels of the more maladaptive coping strategies (Escape-Avoidance and Confrontive Coping).

Differences in the association between coping and life events between the two groups were calculated through interaction effects (group × coping interaction effects on life events, and group × life events interaction effects on coping). We also calculated life events by coping interaction effects on group membership. None of these associations was significant, i.e., the association between coping and life events did not differ between the suicide and the SVD groups, and the association between group membership and coping (life events) did not differ depending on degree of life events (coping). Accordingly, pair-wise comparisons with the control group showed similar patterns for both target groups, distinguishing them from the control group. Differences in Planful Problem-Solving could be accounted to some degree by differences in ACEs, indicating similar mediated effects of ACE on the low levels of the more adaptive coping strategy Planful Problem-Solving in the suicide and the SVD group.

Discussion

Our findings are congruent with several psychological theories of suicide [13]. The examination of coping strategies confirmed social problem-solving vulnerability both in the suicide group and in the SVD group, consistent with the theory of cognitive rigidity in problem-solving [35]. The examination of risk factors suggested that psychiatric illness per se might constitute a severe strain in life but lead to suicide mainly if combined with other vulnerability factors, as previously postulated by the clinical-biological model of suicide [36]. According to this model, a common trait factor associated with suicidality is responding to stressful life experiences with hostility and aggression [21]. The present study showed that this factor is common for cases of suicide and SVD but is more prominent in cases of SVD. Thus, suicide and SVD might be consequences of underlying aggressive impulses that, in combination with other risk factors, determine whether the aggression is directed toward others or toward oneself (two-stage model of outward or inward directed aggression) [37, 38].

To summarize, the answer to our question “Two sides of the same coin?” must be: only partially. There are several common risk factors for both groups, distinguishing them from the control cases. But there are also several significant differences between the cases of suicide and SVD. In other words, there is a common ground of adverse childhood experiences, other vulnerabilities and strains in life, lower educational level and lack of work or studies, addiction, as well as less adaptive coping for both forms of premature violent death, but there is also a substantial divide for the two contrasting developments ending with suicide or with SVD, respectively. The pathway to death by suicide includes lack of a steady relationship, sexual traumatization, depression, autism, being psychiatric out- or inpatient, recent stressors in everyday life, lowest levels of adaptive coping, and use of alcohol and drugs in connection with suicide. The pathway to SVD includes male gender, younger mother, poor elementary education, acting out in form of criminal behavior, conduct disorder and antisocial personality disorder, together contributing to risk-taking behavior. The dividing line between the two pathways to premature unnatural death can be interpreted in terms of inward or outward directed aggression and internalizing versus externalizing psychopathology (Fig 1). However, it has to be noted that the path to sudden violent death seems to show some similarities to a third path: to destructivity and violence in interpersonal and intergroup relationships, criminal acts, and hurting others [39]. For example, young adult violent offenders might direct aggressive behaviors not only toward other people, but also toward themselves [40].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Common risk factors and distinct pathways to death by suicide and to SVD.

https://doi.org/10.1371/journal.pone.0313673.g001

In the present study, approximately 29% of those who died by suicide and 42% of those who died a SVD, respectively, had never sought or never received any help from mental health services. This might suggest barriers to help-seeking among suicidal decedents, studied in a systematic review [41], and even higher barriers in the SVD group, probably due to the marked male predominance in this group, more of outward directed aggression and externalizing psychopathology (Fig 1). However, it should be noted that barriers against help-seeking are not only related to the young persons themselves but also include professionals, organization of school system, labor market, and routine mental health care.

On a theoretical level, the common feature of paths to suicide and to SVD is the existence of destructive and self-destructive processes. Based on our findings (Fig 1), common goals for prevention of both suicide and SVD should include focus on consequences of adverse childhood experiences; facilitating school learning and finding a job or alternative forms of meaningful occupation; promoting more adaptive coping strategies; addiction prevention and treatment; as well as routine follow-up of child and adolescent psychiatric contacts. Specific targets for suicide prevention should comprise focus on young people with lowest levels of adaptive coping, depression prevention and treatment, and paying attention to young people with autism. Specific targets for prevention of SVD should involve early recognition and response to learning difficulties, abuse of psychoactive drugs and delinquent behavior, paying attention to expressions of outward directed aggression, and effective treatment of conduct disorder and antisocial personality disorder. Ultimately, prevention of suicide and other forms of life-threatening behavior among young people may be facilitated by social, educational, and therapeutic interventions addressing feelings of powerlessness and loss of control over their lives, hostile contempt, violence in interpersonal and intergroup relationships, and externalization and projection onto others of own shortcomings and weaknesses.