This is the first study to examine the characteristics of serious suicide attempters among adolescents in Japan, focusing on psychiatric disorders, methods, motives and socio-demographic factors compared with adults.
We first examined the differences between 59 adolescent suicide attempters and 102 adults. The results revealed significant differences in BPD among psychiatric disorders, in financial problems and school problems among motives for suicide attempts, and in parent loss experience among socio-demographic characteristics.
As for the gender ratio among suicide attempters, the ratio of females was higher than that of males both in adolescents and adults. The subsequent gender-specific analysis revealed different characteristics between male adolescents and female adolescents, highlighting the fact that, if we compare adolescent suicide attempters with adult suicide attempters, we should also consider their characteristics according to gender.
A previous study reported that 95% of suicide attempters met the DSM-IV criteria for either axis I or axis II psychiatric diagnosis or both
. In the present study, psychiatric disorders were present in 98% of female adults, 93% of male adolescents, and 89% of both male adults and female adolescents. Therefore, among not only adults but also adolescents, psychiatric disorders have a close relation with attempted suicide, and there may be psychiatric symptoms or social dysfunction with psychiatric disorders behind suicide.
Depressive disorders are consistently the most prevalent disorders among adolescent suicide victims
. In the present study, mood disorders were also most common among adolescents. However, according to the analysis by gender, schizophrenia was the most prevalent psychiatric disorder among male adolescents and was also seen more often in male adolescents than in male adults. The onset of schizophrenia is earlier in males than in females
 and schizophrenia patients committing suicide are more often men and tend to be young
. According to a recent meta-analysis, 4.9% of patients with schizophrenia die of suicide, most of them soon after illness onset
. Additionally, those with schizophrenia tend to use highly lethal suicide methods
. The reason for male adolescents to have a high rate of schizophrenia in our study might be related to the gender difference at onset age of the condition. Additionally, as all participants in our study were in a medically serious and potentially fatal state, there may have been a selection bias. It is important that we identify male patients with schizophrenia as possible adolescent suicide attempters, because schizophrenia among adolescents increased the risk of suicide
. We also need to provide comprehensive initial treatment that improves adherence to early treatment via a psycho-education for patients and their families, as 2-11% of patients with schizophrenia attempt suicide at least once a year after starting treatment
In females, BPD was more common among adolescents than adults. We especially focused on the diagnosis of personality disorders in adolescent participants, as adolescence is the developmental stage of personality. We investigated psychiatric symptoms in medical records from psychiatrist-conducted interviews of all participants or their parents as well as from information about any prior psychiatric treatments. Previous studies have reported that the frequency of BPD in female adolescents who committed suicide was 26%
 and that it was 33% in individuals 15–29 years old
. The result of the present study, that BPD was the second most common diagnosis for female adolescents, supported previous studies reporting that BPD was relatively frequently associated with suicide attempts and completed suicides among adolescents. Therefore, it is important to assess interpersonal relationships and emotional control of female adolescent suicide attempters with BPD. Likewise, it is necessary to examine possible developmental histories that might affect suicide attempts, focusing on the development of attachment, mother-child relationship problems, and traumatic episodes in childhood strongly associated with BPD. Indeterminate for psychiatric disorders also was more often seen among adolescents than adults. Some female adolescents may attempt suicide without any psychiatric symptoms and prior indications of emotional or behavioral problems. It has been pointed out that adolescents may be more impulsive than adults, and they may tend to focus on proximal consequences of behavior
[38, 39]. Therefore, we have to pay attention to impulses among female adolescents.
Motives for suicide attempts and socio-demographic characteristics
As for motives, family problems had the highest incidence in both male and female adolescents. The relation between family functioning and adolescent suicide attempts has been explored in many lines of research
[18, 23]. Our findings supported those previous studies. On the other hand, we found that adolescents, and particularly females, had more parent loss experience and school problems. This finding may indicate that female adolescents are more vulnerable to stresses in the immediate environment and interpersonal relationships than male adolescents and adults. Therefore, in regard to female adolescents, it is important to adequately assess family functioning and interpersonal conflict, and furthermore, to reach out to the patients and their families.
Male adults had significantly more financial problems than male adolescents, while there was no difference between females. A previous study in Japan
 reported that unemployment is an important suicidal risk factor in males. They may suffer more severe effects from failures on the job or struggles to survive from such as debts. For male adult suicide attempters with financial problems, they should be paid close attention, and adequate intervention with a view to solving their financial problems should be undertaken.
In terms of previous psychiatric history, female adults had significantly more than female adolescents, while there was no difference in males. This may indicate that the frequency of psychiatric care is greater in females than males and that the tendency becomes stronger with increasing age. Still, despite receiving psychiatric care in the past, more than half of all participants attempted suicide. If suicide attempters discontinue their psychiatric treatment, it is vital that their treatment be re-established. Even if they are undergoing psychiatric treatment, they may have poor medication adherence, severe psychiatric symptoms, and psychosocial impairments. Therefore, for the prevention of subsequent suicide attempts, we have to determine their psychosocial issues and remedy psychosocial dysfunctions in parallel with improving adherence and psychiatric symptoms as part of their psychiatric treatment.
This study has several limitations. First, it used a relatively small sample size, meaning that our results can only be generalized with caution. However, this study sample does present meaningful data, in consideration of measures to combat suicide, because all participants were serious suicide attempters with characteristics similar to complete suicide attempters. Another limitation is that the study period was different between adolescents and adults, and we could not control the differences between the time periods. A third limitation is that the psychiatric diagnoses of all participants were reached without structured interviews and assessment tools. It is difficult for psychiatrists to conduct structured and systematic interviews for suicide attempters because of the shorter hospitalization and temporal constraints in CCMC. We did our best to enhance the reliability of the psychiatric diagnoses, holding discussions in the case of disagreement with the psychiatric diagnoses. However, there may be an issue that the diagnoses in the present work are more studied, as they were based more on mental status records than diagnoses that are obtained by direct and structured questionnaires.
We did not divide adolescents or adults into subgroups according to age, such as younger adolescents or elderly adults, as our small sample size made it difficult to compare the different age generations in detail. Thus, in the present study, we just aimed to explore the robust differences between adolescents and adults. However, in a future study based on a larger study population we will have to investigate the differences among the various age groups from adolescents to adults. Moreover, it will be necessary to conduct a prospective longitudinal study by means of structured assessment tools and develop interventions tailored to adolescent characteristics.