1. Introduction
Suicide is the second leading cause of death among people between 10 and 24 years [
1,
2] and is considered one of the main public health issues [
3]. In particular, the literature shows that adolescents generally present suicidal thoughts and attempt suicide more frequently than adults [
4], thus highlighting the relevance of this specific developmental stage to investigate and pinpoint risk factors for suicidality. In addition, a study reported that more than a third of adolescents with suicidal ideation have attempted suicide [
5], and a Canadian work—based on a sample of approximately 2000 students—revealed that the risk of suicide was 25.5% among those who reported suicidal ideation [
6]. However, it often happens that signals of psychological suffering in young people are not recognized; for instance, requests for assistance due to general psycho-physical problems (e.g., psychological interview, psychotherapy, access to a psychiatric unit, eating disorders center, previous hospitalizations, etc.) should not be underestimated because they may be based on a profound malaise that can foster suicidal thoughts and acts [
7,
8,
9,
10,
11].
The transition from suicidal ideation to suicidal behavior usually occurs within 1–2 years from the onset of suicidal ideation [
12]. Over the years, several theories within the ideation-to-action framework have been developed; these theories try to explain the transition from suicidal ideation to suicide attempt by differentiating the variables involved in mere suicidal ideation from those underlying suicide attempt. In particular, Joiner’s theory [
13] highlighted the importance of the variable “capability for suicide” as a risk factor for suicide attempts among people with suicidal ideation [
14,
15].
Generally speaking, different studies have pinpointed the variables that can be associated with suicidality in developmental age, showing that most of them are common to both suicidal ideation and attempt (e.g., [
16]).
Regarding the family context, the literature reports that young people with suicidal behavior often have relatives with current or previous psychiatric problems [
17,
18]. Moreover, the relationship between parents and their children plays a crucial role in influencing the social, emotional, and psychological development of children [
19]. A dysfunctional family environment, characterized by low support and high conflict, has been associated with an increasing prevalence of suicidal ideation and suicidal attempt in developmental age [
20,
21]; as a consequence, the evaluation of the child in his family environment is fundamental [
22,
23], especially in cases of preexisting psychological difficulties.
About the extra-family environment, the school context should not be underestimated. In fact, school can be an adverse place for youth if bullying occurs, and this is another risk factor for suicidal behavior [
24,
25,
26]; in particular, the association of suicide attempt with being bullied seems greater than with suicidal ideation [
27]. Moreover, difficulties in peer relationships are also considered possible “precipitating factors” for suicide, particularly in young people [
28]; in fact, being accepted by peers is a fundamental value in adolescence, hence victimization by peers is often perceived as a devastating experience that could lead to depression and suicidal ideation [
3].
In addition, traumatic experiences during childhood and a history of childhood abuse and maltreatment were found to contribute to the onset of suicidality in adolescence, both directly and indirectly, through the role of mediators such as post-traumatic stress disorder, depression, emotion dysregulation, low self-esteem, and dissociative symptoms. The diathesis–stress model asserts that stressful life events interact with vulnerability factors and increase the probability of suicidal behavior [
27,
28]. Moreover, other situations perceived as particularly stressful by the child, such as change of residence, legal issues or sentimental relationship breakdown, can precede suicidal behavior, probably through the development of adaptive problems [
29]. In particular, it emerged that adolescents with a history of suicidal behavior have generally been more exposed to stressful life events than those with only ideation [
28,
30].
Among psychological factors, a meta-analysis found an association between alexithymia and suicidality; specifically, the association between alexithymia (particularly in forms of difficulty in identifying and describing one’s own emotional states) and suicidal ideation resulted in being stronger than that between alexithymia and suicidal attempt; however, the impact of depression on this correlation remained unclear [
31].
Another important risk factor for suicidal behavior, particularly in young people under 14 years of age, was found to be impulsiveness [
32,
33]; more specifically, in a meta-analysis by Liu and colleagues [
34], cognitive impulsiveness and motor impulsiveness were shown to have, respectively, a moderate-to-important and mild-to-moderate effect on suicide attempts. However, higher scores on motor impulsiveness emerged when impulsiveness was assessed shortly after the suicide attempt; this result thereby highlights that the assessment of different forms of impulsiveness could enable prompt recognition of subjects more at risk of acting out. However, the literature seems controversial with regard to the association between impulsiveness and suicidality. In a meta-analysis [
35], impulsiveness levels were found to be similar between subjects who attempted suicide and those with ideation only, and, at the same time, those who presented more impulsive attempts did not score significantly higher on specific tests [
36]. Millner et al. also found similar values of impulsiveness between the two groups, although those who attempted suicide showed higher values of impulsiveness when in a negative situation [
37].
Pertaining to psychiatric disorders, the relationship between Borderline Personality Disorder (BPD) and suicidal behavior is also relevant: this disorder is often associated with experiences of negative emotions, perceived as intolerable, and impulsive attempts to regulate them [
38]. Yen and colleagues conducted a longitudinal study in patients with BPD and showed that approximately a fifth of the patients exhibited a suicidal act during the two-year follow-up [
39].
Other psychopathologies to be considered when studying suicidality in developmental age are affective disorders. In particular, epidemiological studies and a meta-analysis confirmed the role of depression as a strong risk factor for suicidal thoughts and for the transition from suicidal ideation to suicide attempt [
5,
40]. According to Cash and Bridge, in clinical samples of adolescents, approximately 85% of patients with major depressive disorder or dysthymia are at a high risk of presenting suicidal ideation: 32% will attempt suicide during adolescence or early adulthood and 20% will make multiple attempts [
41]. Bipolar disorders may be associated with an increased risk of suicidal behavior, too. A review on suicidality in children and adolescents with bipolar disorder found that suicidal ideation and suicide attempts have incidence rates of 25% and 15%, respectively, with a prevalence of 50% for suicidal ideation, and 25% for suicide attempts [
42].
Finally, when investigating suicidal ideation and behavior, nonsuicidal self-injury (NSSI)—which is characterized by self-injurious acts without suicidal intent—should also be considered. The prevalence of NSSI in adolescents varies between 7.5% and 46.5% [
43,
44,
45,
46] and has increased during the COVID-19 pandemic [
47,
48,
49]; its onset is generally between 12 and 14 years of age [
50], but cases of younger people have been reported [
51]. Several theories have been proposed on the relationship between different self-injurious phenomena. For example, NSSI and suicide have been considered as two expressions of the same spectrum (i.e., self-harming), with the main difference that self-injurious behaviors in NSSI are not motivated by suicidal intention (e.g., [
14]). Moreover, some authors consider NSSI as a risk factor for suicidal ideation and suicide attempt, while others consider it as a protective factor [
52,
53,
54]; in particular, some authors posited that NSSI could be a direct risk factor for suicide, or its influence on suicide could be mediated by other factors, such as depression, suicidal ideation, personality disorders, low self-esteem, or low family support, which can facilitate acting out [
55,
56,
57]. Nevertheless, the role of NSSI in the transition from suicidal ideation to suicide attempt remains controversial.
NSSI, suicidal ideation, and suicide attempt share several risk factors, psychiatric comorbidity, and high prevalence in young people. At a symptomatological and psychopathological level, the main risk factors associated with suicidal and nonsuicidal self-harm are impulsiveness [
32,
33,
58,
59], alexithymia [
60,
61,
62,
63,
64]—considered a transdiagnostic risk factor [
65,
66,
67,
68]—somatic problems [
69,
70,
71,
72,
73,
74,
75], emotion dysregulation [
76,
77,
78,
79,
80], affective disorders [
81,
82,
83,
84,
85], borderline functioning and borderline personality disorder [
86,
87,
88,
89]. However, it seems still unclear what differences exist in the psycho-behavioral profiles of patients with both suicidal ideation/attempt and NSSI.
To date, a certain body of evidence has been accumulated on the characteristics of suicidal ideation and suicide attempt in different populations; however, the literature on this topic has some limitations. First, most studies have considered nonclinical samples and examined risk factors for suicidal ideation in general, without differentiating suicidal attempters from ideators [
15,
90]. Nevertheless, according to the ideation-to-action framework, it is necessary to expand the research on the specific factors associated with suicide attempt or mere suicidal ideation, as several different variables could be involved in the path from suicidal thoughts to suicidal acts [
15,
91].
Furthermore, most works have focused on the adult population, but young people should be carefully kept in mind since nowadays more and more pre-adolescents and adolescents have access to inpatient and outpatient services for suicidal behavior and thoughts; this notwithstanding, only 20% of the literature has investigated risk factors for suicide in such a population [
92]. In particular, to the best of our knowledge, a few studies have examined the risk factors specifically associated with suicidal ideation and suicide attempt in Italian youth [
93,
94]. The triggers or precipitants of suicidal behavior were found to vary widely across cultures (e.g., [
95,
96]); hence the importance of exploring the specific risk factors associated with suicidal ideation and suicide attempt in different cultural contexts.
Bearing all this in mind, it seems fundamental to thoroughly investigate the characteristics that could constitute specific risk factors for suicidal ideation or suicide attempt in the pediatric population, in order to better understand and prevent suicide mortality; to obtain the fullest picture possible of this important phenomenon, the differences in terms of psychopathological symptoms according to the concomitant presence of NSSI should also be considered.
Aims of the Study and Hypotheses
On the basis of the above premises, the present study aimed to further expand the research within the ideation-to-action framework in the Italian context. In particular, our main objectives were:
1. To identify the socio-demographic, clinical, and psychopathological variables specifically associated with suicidal ideation or suicide attempt. To date, the studies investigating this topic have found that most predictors of suicidality seem to be common to both suicidal ideation and attempted suicide (e.g., [
16]); however, as stated above, it is important to distinguish the risk factors peculiar to each suicidal phenomenon. In this regard, including some of the variables thought to underlie suicidality in the same model could be particularly useful to establish the statistical significance of each variable over and above the effect of the others. Among the different risk factors common to both suicidal ideation and suicide attempt, we hypothesized that the most relevant for the purpose of our study could be: intra-family problems (e.g., parental conflict, conflictual separation, conflicts among family members; [
20]), psychiatric familiarity (e.g., [
17]), history of being bullied [
27], previous hospitalizations [
7], previous requests for assistance or previous access to child mental health services [
9], borderline personality functioning (e.g., [
93]), general psychopathological difficulties, affective problems (e.g., [
92]), impulsivity (e.g., [
60]), and alexithyimic traits (e.g., [
69]). However, to our knowledge, no previous research has included such variables in the same model yet; therefore, in light of the exploratory nature of this investigation, we did not formulate any definite hypothesis as to which of the above variables could be a specific predictor of suicidal ideation or suicide attempt.
2. To compare patients with suicidal ideation alone and with attempted suicide according to the presence of NSSI. We hypothesized to find greater clinical severity and impairment—expressed in terms of significantly higher scores on all the scales considered—in patients with NSSI, since this could be underpinned by a more severe psychopathology, especially when present in conjunction with other suicidal phenomena.
4. Discussion
The main aim of the present work was to investigate the socio-demographic, clinical-symptomatological, and psychopathological features associated with suicidal phenomena in a group of Italian pre-adolescent and adolescent inpatients. Specifically, this research sought to identify the specific risk factors for suicidal ideation or suicide attempt, and to investigate the differences in terms of psychopathological symptoms between patients with only suicide attempt/ideation and those who also presented with NSSI. Indeed, given that suicidality in adolescence is considered a public health problem [
3], it seems paramount to pinpoint the possible variables associated with this phenomenon, in order to promptly intervene and prevent suicide mortality in such a vulnerable population.
As regards the description of our sample, a general increase in admissions for suicidality in association with NSSI was observed, in line with the literature that has highlighted a progressive raising in suicidal and nonsuicidal self-harm in young people [
2,
47]. In particular, some authors have suggested that the exponential increase in NSSI and emergency admissions for suicidal ideation and behavior among young people in the last year may be related to the psychosocial impact of the COVID-19 pandemic [
47,
48,
49]. Our sample was mainly composed of pediatric inpatients hospitalized before the outbreak of the pandemic (i.e., 2015–2019), and the rate of inpatients hospitalized in 2019 is almost equal to that of 2020; therefore, we did not observe an increase in admissions for suicidality from the year immediately before the pandemic to the year of the outbreak of the pandemic. This finding is probably due to the imposed restriction on territorial mental health services and scheduled hospitalizations during the COVID-19 pandemic [
49], but it may also be related to the fact that for some young people the effects of the pandemic on mental health have not been immediately observed; in fact, previous studies on children and adolescents with preexisting neuropsychiatric disorders showed a general good adjustment of young patients to the pandemic situation [
126,
127,
128]. However, the long-term negative impact of the pandemic on the psychological well-being of youth could emerge later [
129], so we would expect an increase in suicidality rate in the following years. A noteworthy aspect is that most of the patients hospitalized in 2015–2019 reported suicidal ideation and NSSI, while the majority of those hospitalized in 2020–2021 presented suicide attempts and NSSI, thus showing a more severe clinical picture. Future studies in this direction are warranted to better clarify the long-term consequences of the pandemic in terms of suicidality among young people.
Subsequently, we wanted to investigate in more detail the characteristics that can constitute risk factors for suicidal ideation or suicide attempt. The final model showed that the socio-demographic variables considered (i.e., age, intra-family problems, psychiatric familiarity, and being victim of bullying) were not significant predictors of suicidality in general, although it is possible to assume that they may influence the phenomenon indirectly by affecting the environment where pre-adolescents and adolescents live. Among the variables related to the patient’s clinical history and psychopathological condition, previous access to child mental health services and the presence of psycho-behavioral problems were significant predictors of suicidal ideation. Therefore, it may be useful, for preventive purposes, to pay attention to young people who seek some form of assistance (e.g., psychological interview, psychotherapy, access to a psychiatric unit, eating disorders center, etc.) for general psycho-physical, relational, and/or behavioral problems since such a suffering condition could foster suicidal thoughts [
7]. This result is in line with the literature highlighting that people who attempt suicide tend to seek help by confiding in family/friends or contacting mental health services [
9], and the request often occurs close to the act [
8]. Specifically, a recent study has noted that, in the year before suicide, 25.3% of young people had contacted a mental health service and in 9.7% of cases the reason had been NSSI [
10]. Therefore, improving the ability to intercept them, listen to them, and give them a reason to trust could act as a barrier against the transition to suicidal action.
Then, previous hospitalizations, regardless of the cause, affective disorders, and borderline personality functioning (vs. neurotic personality functioning) were found to be predictors of an increased risk of suicide attempt. In our sample, one-third of the individuals have previously been hospitalized at least once (for abdominalgia, infection, fever, road trauma, neuropsychiatric disorders). This result is consistent with most published studies showing that suicidal behavior is associated with a higher number of hospitalizations, mainly due to suicidality or affective disorders: these data may indicate greater severity of the clinical condition, fewer personal and family resources, and probably a lower response to therapy [
11]. Moreover, the finding about borderline personality functioning aligns with the literature that supports the association between borderline personality functioning/disorder and suicidal attempt in adolescence [
86,
87,
88,
89].
Considering the abovementioned results together, it is reasonable to surmise that more specific risk factors, such as affective disorders [
81,
84], borderline personality functioning, and severe clinical conditions that resulted in previous hospitalizations, may be characteristic of suicidal behavior, while more general and cross-cutting emotional issues and previous access to outpatient services may be related to suicidal ideation.
Then, to investigate the psychopathological differences between patients with suicidal ideation and suicide attempt according to the presence or absence of NSSI, we divided the sample into four groups: (1) suicidal ideation alone, (2) attempted suicide, (3) suicidal ideation and NSSI and 4) attempted suicide and NSSI.
First, the groups with suicidal ideation + NSSI and suicide attempt + NSSI obtained higher scores on all the TAS-20 scales, the YSR internalizing problems and total problems scales, and the DESR profile compared to inpatients with suicidal ideation only. Generally speaking, these findings seem to show that NSSI significantly deteriorates the overall psychopathological profile compared to the mere presence of suicidal ideation.
To be more specific, alexithymia is characterized by the inability to identify and communicate emotions with appropriate words and was found to be a risk factor for several problems in adolescent age, such as risk behaviors (e.g., [
65]), social withdrawal (e.g., [
68]) somatization (e.g., [
66,
67]), and suicidal behavior (e.g., [
63]). On the basis of the aforementioned results, we assume that general difficulties in emotion regulation and, in particular, in identifying and communicating feelings may be involved in NSSI, thus differentiating adolescents with mere suicidal ideation from those with suicidal ideation/attempt together with NSSI. In particular, people with emotion dysregulation and alexithyimic traits may be more likely to use maladaptive coping strategies; therefore, as also noted in the literature [
79], suicidal and nonsuicidal self-harm could be interpreted as a dysfunctional attempt to cope with emotions in the face of an inability to find more functional strategies. To further support this, recent literature has highlighted that improving the ability to identify and describe emotions in individuals who attempted suicide could improve mood and problem solving strategies and reduce states of hopelessness [
63].
With regard to internalizing problems, instead, NSSI could be considered a way to recover from negative emotional or cognitive states [
80,
93], especially if the person has difficulty regulating emotions [
64], as suggested above.
Then, our results showed that the group with suicidal attempt and NSSI reported more somatic problems than the group with only suicidal ideation. This finding is consistent with the literature reporting the association between somatic symptoms and both suicidality [
69] and NSSI [
71]. In particular, somatoform symptoms should be carefully considered because, if not properly investigated and recognized as psychological distress, they can lead to suicidal behavior [
72,
73]. Not to be forgotten are serious chronic diseases, such as epilepsy, asthma, or the consequences of head trauma, spinal cord injury, which have been associated with an increased risk of suicide [
74]. In addition, both NSSI and somatization use the body to express pain or psychological distress [
62,
66,
67,
75]. To date, studies on this topic are still too few; nevertheless, there is evidence of an association between these two phenomena: both somatic symptoms and NSSI are often correlated with depressive and anxiety disorders, and their coexistence could be mediated by comorbidity with internalizing problems and alexithymia, thus becoming risk factors for the development of suicidal and nonsuicidal self-injury [
85].
Finally, significant differences between patients with SI alone and patients with SA + NSSI emerged in all the above-mentioned variables, too. Generally speaking, we can hypothesize that patients with both SA (and therefore also SI) and NSSI have a more severe clinical pattern than patients with SI alone, thus assuming that a history of NSSI and other clinical factors may play a role in fostering acting out; however, longitudinal studies are needed to corroborate such a hypothesis and assess how said variables interact in the evolution of this phenomenon.
Pertaining to total, attentional, and nonplanning impulsiveness, significant differences emerged between the group with suicide attempts only and the group with suicide attempts + NSSI. This result seems to indicate that a greater level of impulsiveness characterizes patients with NSSI who also attempted suicide. Lockwood and colleagues [
59] highlighted that cognitive impulsiveness was associated with the maintenance of NSSI over time, thus leading to a greater risk of suicide attempts. In this scenario, it is possible to surmise that inpatients with NSSI who attempt suicide exhibit a greater tendency to make sudden decisions, without considering different consequences and options; therefore, this behavior could encourage the transition from self-harming to suicidal action. To clarify the relationship between impulsiveness, NSSI, and subsequent suicide attempt, longitudinal studies on clinical and nonclinical samples should be undertaken.
Finally, we would like to read our overall findings in light of the Joiner’s interpersonal theory of suicidal behavior [
57]. He stated that social isolation and depressive symptoms may be relevant factors for the transition from NSSI to suicidal action. The results of the present study indicated that, although affective disorders seem to play a crucial role in predicting suicide attempt, the SI and SA groups did not differ significantly in internalizing problems when inpatients with NSSI were excluded from such two groups; therefore, it is possible that NSSI itself contributes to increasing the severity of internalizing disorders. Furthermore, according to Joiner, repetitive NSSI could facilitate the acquisition of the capability for suicide, since the individual becomes used to pain and no longer considers NSSI as a method of regulating emotions [
14,
91]. Consistently, our results showed greater difficulty regulating emotions in patients with SI + NSSI and SA + NSSI compared to those with SI alone.
The current study presents some limitations. First, the sample size did not enable a homogeneous comparison between different types of suicidal phenomena; moreover, in light of the relatively small sample size, caution must be applied in interpreting and generalizing the results. Second, due to the retrospective observational nature of the study, some data are missing. In addition, self-report questionnaires may lead to some biases linked to social desirability or non-understanding of questions. Data were collected from only one neuropsychiatric unit in northern Italy, thus limiting the generalizability of our results to the whole Italian adolescent population. Further longitudinal studies are required to corroborate our findings, in order to pinpoint the variables associated with suicidal ideation and suicidal attempt in the long-term and investigate the process underlying the transition from suicidal ideation to suicide attempt. In particular, NSSI should be further analyzed to clarify whether it is a specific or aspecific risk factor for suicidal behaviors, as well as its role and that of the other related psychopathological characteristics in fostering suicide attempt. Moreover, protective factors should also be taken into account to develop primary and secondary preventive interventions. Another limitation is related to the wide age range considered in the present study; future investigations should analyze the differences in terms of suicidality according to the different stages of adolescence (i.e., pre-adolescence, early adolescence, middle adolescence, and late adolescence). Moreover, as previously said, the family environment plays a crucial role in the onset of suicidality in developmental age; therefore, further work should describe and distinguish in more detail different family variables associated with suicidality given that we only considered intra-family problems. In conclusion, further studies should better investigate suicidal phenomena in the male population; in fact, the composition of our sample, which consisted mainly of girls, did not enable us to examine the differences between boys and girls in suicidal behavior. However, given that male adolescents manage to complete suicide more frequently, it is relevant to examine the correlates of suicidality in this population for preventive purposes.