- Email: [email protected]
Clinical characteristics of suicidal behavior in an intensive care unit at a university hospital in Japan: A 7-year observational study
Asian Journal of Psychiatry xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Asian Journal of Psychiatry journal homepage: www.elsevier.com/locate/ajp
Asian pearls
Clinical characteristics of suicidal behavior in an intensive care unit at a university hospital in Japan: A 7-year observational study ⁎
Takashi Takeuchia, , Yasuyuki Okumurab, Akihito Uezatoa, Toru Nishikawaa a
Section of Psychiatry and Behavioral Sciences, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan b Research Department, Institute for Health Economics and Policy, Association for Health Economics Research and Social Insurance and Welfare, 1-5-11 Nishi-Shinbashi, Minato-ku, Tokyo 105-0003, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords: Borderline personality disorder Drug overdose Interpersonal problems Psychiatrist Suicidal behavior
Background: Suicidal behavior accounts for at least 40,000 admissions per year to emergency departments in Japan; however, little is known about emergency admissions owing to suicidal behavior in metropolitan areas. Therefore, we examined the clinical characteristics of suicidal behavior using psychosocial assessments performed by experienced psychiatrists in an intensive care unit. Methods: Participants were 971 patients admitted to a university hospital’s intensive care unit for suicidal behavior between July 2006 and June 2013. Physicians and psychiatrists regularly assessed the participants using a standard data extraction form while the participants were in the intensive care unit. As suicidal behavior involving drug overdose is generally less fatal than other methods, we predicted that clinical characteristics would differ between patients with and without overdose. We classified participants into drug overdose or other method groups (ns = 732 and 239, respectively) to compare suicide methods. Results: In the overdose group, participants’ median age was approximately 5 years lower, and the following proportions were larger: female participants (77%) and participants with borderline personality disorders (21% vs. 10%), no clear suicidal ideation (30% vs. 15%), impulsively attempted self-harm (86% vs. 62%), and interpersonal problems (26% vs. 16%). Conclusion: Ameliorating interpersonal problems and improving stress coping skills would benefit people who attempt suicide via overdose.
1. Introduction In Japan, the annual number of people committing suicide exceeded 30,000 in 1998, and the suicide rate increased to 27.0 per 100,000 population in 2003, recording the highest rate in history. The annual total number fell below 30,000 in 2012 and has gradually decreased since then. However, according to the World Health Organization, the suicide rate in Japan is still as high as 21.7 (2012), which is the tenth highest rate in the world. The most common suicide method is by hanging, which accounts for approximately 60% of all suicides. This is followed by gassing oneself, jumping off a high place, and drug overdose; however, each of these methods accounts for 10% or less of the total number of suicides. Although some people drink alcohol before committing suicide, there are few cases of suicide by drinking. Moreover, because possession of guns is legally prohibited in Japan, suicide using a gun, which is common in the United States and other countries, is also rare (Specified Report of Vital Statistics, 2005).
⁎
Each year, suicidal behavior accounts for at least 40,000 admissions to emergency departments in Japan (Japan Organ Transplant Network, 2011). Suicidal behavior is a strong predictor of subsequent suicide and premature death (Finkelstein et al., 2015). Better understanding of region-specific suicidal behavior epidemiology may prevent suicide and reduce the burden of suicidal behavior on emergency medical systems; however, little research has examined suicidal behavior in metropolitan areas in Japan (Hori and Kinoshita, 2016; Kubota et al., 2015; Sugita et al., 2011). Additionally, most studies examining the clinical characteristics of suicidal behavior have not used specialized psychosocial assessment and, therefore, have not collected or assessed some important clinical information (e.g., psychiatric diagnosis) (Hori and Kinoshita, 2016; Kubota et al., 2015; Sugita et al., 2011). Therefore, we used psychosocial assessments by experienced psychiatrists to examine suicidal behavior’s clinical characteristics in an intensive care unit in a university hospital in Tokyo. As suicidal behavior involving drug overdose is generally less fatal than other methods,
Corresponding author. E-mail addresses: [email protected] (T. Takeuchi), [email protected] (Y. Okumura), [email protected] (A. Uezato), [email protected] (T. Nishikawa).
https://doi.org/10.1016/j.ajp.2017.11.005 Received 25 February 2017; Received in revised form 17 May 2017; Accepted 6 November 2017 1876-2018/ © 2017 Elsevier B.V. All rights reserved.
Please cite this article as: Takeuchi, T., Asian Journal of Psychiatry (2017), http://dx.doi.org/10.1016/j.ajp.2017.11.005
Asian Journal of Psychiatry xxx (xxxx) xxx–xxx
T. Takeuchi et al.
we predicted that clinical characteristics would differ between patients with and without overdose. We also compared clinical characteristics between individuals who attempted suicide via drug overdose and via other methods.
Table 1 Methods of suicide behavior. Methods
Total (n = 971) n (%)
Men (n = 273) n (%)
Women (n = 698) n (%)
2. Methods
Drug overdose (only overdose)
732 (75.4)
170 (62.3)
562 (80.5)
2.1. Design and setting
Non-overdose Self-cutting Chemical poisoning Jumping Use of charcoal briquettes Hanging Drowning Traffic Inhaling carbon monoxide Other
88 (9.1) 37 (3.8)
38 (13.9) 28 (10.3)
50 (7.2) 9 (1.3)
36 (3.7) 34 (3.5)
17 (6.2) 19 (7.0)
19 (2.7) 15 (2.1)
19 (2.0) 9 (0.9) 8 (0.8) 4 (0.4)
8 3 4 3
11 (1.6) 6 (0.9) 4 (0.6) 1 (0.1)
4 (0.4)
2 (0.7)
The protocol for this research project was approved by a suitably constituted Ethics Committee of the institution and conforms to the provisions of the Declaration of Helsinki. Committee of the Tokyo Medical and Dental University, Approval No. 1604. The requirement to obtain patients’ consent was waived as all the data assessed in the study were collected through our regular clinical care. We conducted a 7-year observational study of consecutive patients hospitalized for suicidal behavior (i.e., self-poisoning and self-injuries) in an intensive care unit at the Tokyo Medical and Dental Hospital University in Tokyo, Japan, between July 1, 2006 and June 30, 2013. At our hospital, among patients with suicidal behavior, those deemed to require hospitalization are always admitted to the intensive care unit. The unit provides tertiary emergency care to 10,000–15,000 patients annually.
(2.9) (1.1) (1.5) (1.1)
2 (0.3)
admitted, there were, on an average, 1 or 2 participants per year who died after admission and before psychosocial assessment. Moreover, because no participant died after psychosocial assessment, the annual mean number of participants who died by suicide was 1 or 2, which was also the number of participants who died before psychosocial assessment. The most common method of suicidal behavior was drug overdose (75.4%). Among individuals who overdosed, 88.5% used psychotropic drugs prescribed by their physicians. The second most common method was self-cutting that required surgical procedures in an operating room (9.1%) followed by chemical poisoning (3.8%), jumping (3.7%), and use of charcoal briquettes (3.5%; Table 1).
2.2. Data collection Since 2006, physicians and psychiatrists in the unit had routinely assessed clinical information for all patients who survived self-harm episodes using a standard data extraction form that examines sex, age, psychiatric diagnosis, psychosocial factors (person’s predisposing distress, characteristics of attempts, suicidal ideation), and history of outpatient treatment. Two psychiatrists with at least 5 years’ experience conducted psychosocial assessment for all patients based on the International Classification of Diseases, 10th Revision. We included patients who died after psychosocial assessment and excluded those who died of critical physical conditions before the psychosocial assessment.
3.2. Cumulative probability of suicidal behavior as a function of age Participants in the overdose group were more commonly women (76.8% vs. 56.9%; p < 0.01). Fig. 1 illustrates the cumulative probability of suicidal behavior as a function of age. The median age was approximately 5 years lower in the overdose group (median age: 34 years for men and 32 years for women) than in the non-overdose group (median age: 39 years for men and 36 years for women,).
2.3. Statistical analysis First, we created the following categories of suicidal behavior: drug overdose, self-cutting, jumping, chemical poisoning, use of charcoal briquettes, hanging, drowning, traffic, inhaling carbon monoxide, and others. We calculated proportions of these methods by sex. Second, we divided participants according to major method of suicidal behavior: drug overdose (n = 732) and other methods (i.e., nonoverdose; n = 239). Participants who used a drug overdose in addition to other methods were classified as non-overdose, because such individuals typically used an overdose to encourage themselves to undertake more lethal methods of suicide. The cumulative probability of admissions owing to suicidal behavior as a function of age was estimated by sex and major method of suicidal behavior. Additionally, we compared the clinical characteristics of suicidal behavior between overdose and non-overdose. We used the standardized difference (d) as a categorical variable to measure balances of clinical characteristics between the groups—a standardized difference of > 10% indicated imbalance (Austin, 2011). We used chi-squared tests to evaluate differences between the groups. SPSS version 20 was used. Values of < 0.05 were considered significant.
3.3. Diagnosis The most common diagnosis was major depressive disorder (28.9%) followed by adjustment disorder (22.6%), borderline personality disorder (18.3%), schizophrenia (12.4%), and bipolar disorder (7.8%; Table 2). Participants in the overdose group more commonly had borderline personality disorder (21.0% vs. 10.0%; d = 30.7; p < 0.05) and less commonly had major depressive disorder (25.7% vs. 38.9%; d = 28.6; p < 0.05) or schizophrenia (11.1% vs. 16.3%; d = 15.3; p < 0.05). 3.4. Psychosocial factors The most frequent predisposing distress was family issues (32.5%) followed by interpersonal (23.6%), occupational (11.7%), financial (6.3%), and health problems (5.8%; Table 3). Participants in the overdose group were more likely to have interpersonal issues (26.1% vs. 15.9%; d = 25.2; p < 0.05) and less likely to have financial problems (4.8% vs. 10.9%; d = 22.8; p < 0.05). The most common type of suicidal behavior was impulsive action (80.3%) followed by planned action (12.0%) and action induced by delusion or hallucinations (4.1%; Table 3). In this study, no assessment scale was used to assess the impulsiveness underlying suicidal behavior. This impulsive action was defined as a case in which a participant impulsively attempted suicide without planning or the influence of
3. Results 3.1. Methods of suicidal behavior During the study period, 971 individuals were eligible for analysis. Although there were occasional cases in which some individuals had already died before arrival at our emergency center and thus were not 2
Asian Journal of Psychiatry xxx (xxxx) xxx–xxx
T. Takeuchi et al.
Fig. 1. The cumulative probability of suicidal behavior as a function of age.
Table 2 Diagnosis. Diagnosis
Total (n = 971) n (%)
Overdose (n = 732) n (%)
Non-overdose (n = 239) n (%)
Standardized difference, %
Depressive disorders Bipolar disorders Adjustment disorders Borderline personality disorders Schizophrenia Anxiety disorders
281 (28.9) 76 (7.8) 219 (22.6) 178 (18.3) 120 (12.4) 77 (7.9)
188 (25.7) 62 (8.5) 168 (23.0) 154 (21.0) 81 (11.1) 63 (8.6)
93 14 51 24 39 14
28.6* 10.1 3.9 30.7* 15.3* 10.6
(38.9) (5.9) (21.3) (10.0) (16.3) (5.9)
* p < 0.05.
hallucination or delusion, according to the participants’ statements. Participants in the overdose group more commonly impulsively attempted suicidal behavior (86.3% vs. 61.9%; d = 58.1; p < 0.05) and had no suicidal ideation (29.8% vs. 14.6%; d = 37.0; p < 0.05).
Table 3 Psychosocial factors. Psychosocial factors
3.5. Treatment status Participants in the overdose group more commonly received psychiatric treatment (85.4% vs. 69.0%; d = 39.7; p < 0.05; Table 4); 43.0% of patients with psychiatric treatment history had received psychiatric treatment at least once every two weeks. Frequency of psychiatric treatment did not differ significantly between the overdose and non-overdose groups.
Primal distress Family Interpersonal Occupational Financial Health Other No problem Unknown
4. Discussion
Total (n = 971) n (%)
Overdose (n = 732) n (%)
Nonoverdose (n = 239) n (%)
Standardized difference, %
316 (32.5) 229 (23.6) 114 (11.7) 61 (6.3) 56 (5.8) 72 (7.4) 78 (8.0) 45 (4.6)
249 (34.0) 191 (26.1) 79 (10.8) 35 (4.8) 36 (4.9) 57 (7.8) 50 (6.8) 35 (4.8)
67 38 35 26 20 15 28 10
13.0 25.2* 11.6 22.8* 13.9 5.9 16.9* 2.9
632 (86.3) 57 (7.8) 20 (2.7)
148 (61.9) 60 (25.1) 20 (8.4)
58.1* 48.0* 24.8*
9 (1.2) 14 (1.9)
2 (0.8) 9 (3.8)
3.9 11.2
425 (58.1) 218 (29.8) 89 (12.2)
173 (72.4) 35 (14.6) 31 (13.0)
30.4* 37.0* 2.5
Types of suicidal behavior Impulsive 780 (80.3) Planned 117 (12.0) Delusions or 40 (4.1) hallucinations Other 11 (1.1) Unknown 23 (2.4)
This study examined suicidal behavior’s clinical characteristics in an intensive care unit in Tokyo. We found that drug overdose was the most common method of self-harm, supporting previous research (Lilley et al., 2008; Perry et al., 2012). In Japan, there are few detailed reports on suicidal behavior. According to a report on emergency transports resulting from self-harm in Osaka, one of the Japanese metropolises (Matsuyama et al., 2016), the most common cause was an overdose of psychotropic and other drugs, followed by cutting of the wrist and arm and hanging. The mortality due to hanging was reported to be as high as approximately 80%. Moreover, suicidal behavior of men was more likely to result in death, whereas that of women was
Suicidal ideation Yes No Unknown * p < 0.05.
3
598 (61.6) 253 (26.1) 120 (12.4)
(28.0) (15.9) (14.6) (10.9) (8.4) (6.3) (11.7) (4.2)
Asian Journal of Psychiatry xxx (xxxx) xxx–xxx
T. Takeuchi et al.
individuals with suicidal behavior for 6 months. This effect was more prominent in women, participants younger than 40 years, and those with a history of previous suicide attempts (Kawanishi et al., 2014). At our hospital, when patients admitted to our emergency center for suicidal behavior are examined, psychiatrists assess the presence or absence of suicidal ideation, the state of hallucination and delusion, and support systems in order to determine the need for inpatient psychiatric treatment. In this study, because approximately 80% of the participants had a history of outpatient psychiatric treatment as is shown in Table 4, when we decided that they could be discharged, we instructed them to see their primary psychiatrists immediately. We informed their primary psychiatrists of the course of the participants’ attempts to commit suicide in writing. Those without a history of outpatient treatment were treated at our outpatient clinic or referred to neighborhood clinics. Although we did not check whether the participants were treated by their primary psychiatrists after discharge in this study, many participants seriously considered the recommendation for further treatment.
Table 4 Treatment status. Treatment
Total (n = 971) (%)
Outpatient treatment Never 179 (18.4) Clinics 507 (52.2) General 149 (15.3) hospitals Psychiatric 121 (12.5) hospitals Unknown 15 (1.5) Frequency of treatmenta Weekly 85 (13.5) Biweekly 185 (29.5) Monthly 94 (15.0) Discontinued 79 (12.6) Irregular 40 (6.4) Other 41 (6.5) Unknown 104 (16.6)
Overdose (n = 732) (%)
Nonoverdose (n = 239) (%)
Standardized difference, %
107 (14.6) 409 (55.9) 113 (15.4)
74 (31.0) 96 (40.2) 36 (15.1)
39.7* 31.8* 1.0
93 (12.7)
28 (11.8)
3.0
10 (1.4)
5 (2.1)
5.6
69 (13.9) 152 (30.6) 72 (14.5) 58 (11.7) 34 (6.8) 30 (6.0) 82 (16.5)
16 (12.2) 33 (25.2) 22 (16.8) 21 (16.0) 6 (4.6) 11 (8.4) 22 (16.8)
5.0 12.0 6.4 12.6 9.8 9.1 0.8
5. Conclusion This study examined the characteristics of suicidal behavior in a metropolitan area of Tokyo. The most frequent method of suicidal behavior was drug overdose. Individuals who overdosed were younger than those who use other methods. Individuals who overdosed more commonly had borderline personality disorder, no clear suicidal ideation, impulsively attempted self-harm, and interpersonal problems. These individuals would benefit from interventions that ameliorate interpersonal problems and improve stress coping skills. Thus, the support program (case management) that had been built using the framework of methods including regular face-to-face (or telephonic) interviews with individuals with suicidal behavior, promotion of psychiatric treatment, provision of information on public social programs and private support organizations, arrangements for using these services, and provision of information regarding psychological education, which were also applied in the ACTION-J trial, appeared to be useful. Because the reduction in the number of people committing suicide in Japan seems to be largely attributable to this case management, we consider that this aspect should also be learned in other countries.
a
Sample sizes were 628 for total, 497 for overdose, and 131 for non-overdose. * p < 0.05.
more likely to be non-fatal. There are some reports on drug overdose in Tokyo. Sugita et al. reported that 60% of the suicides were by women, and that, although approximately half of the patients needed hospitalization, all of them had physically improved and were discharged (Sugita et al., 2011). Hori et al. reported that approximately 80% of the patients were women, while the mortality rate for women was 0.4% (Hori and Kinoshita, 2016). Thus, the results of our study appeared to be nearly consistent with those of other Japanese reports. Participants’ median age was younger in the overdose group; this contrasts with previous research examining an emergency department (Lilley et al., 2008). This may be because we examined inpatients rather than emergency department patients. Most participants who overdosed ingested psychotropic medications. Prior to admission for suicide attempts, approximately 60% of patients had received psychiatric treatment at least once a month. This finding suggests a need for more careful selection of treatment options for psychiatrists. Overdose-group participants more commonly had borderline personality disorders and interpersonal problems and less commonly had suicidal ideation; therefore, psychiatrists should choose suicide prevention strategies according to the type of self-harm observed. Regarding individuals who overdose, psychiatrists should aim to improve patients’ coping skills for interpersonal issues using psychosocial interventions and take judicious care when prescribing psychotropic medications for borderline personality disorders. Individuals who engage in other forms of suicidal behaviors should aim to ameliorate any physical or financial problems in collaboration with social workers. In Japan, a multicenter randomized controlled trial titled PostSuicide Attempt Case Management for the Prevention of Further Attempts in Japan (ACTION-J) was conducted. In this trial, a new support program (case management) was built using the framework of regular face-to-face (or telephonic) interviews with individuals with suicidal behavior, promotion of psychiatric treatment among them, provision of information on public social programs and private support organizations and arrangements for using these services, provision of information on psychological education, and so on. Then, a nationwide study group consisting of 17 institutions where the departments of emergency medicine and psychiatry were already cooperating (ACTION-J Group) was organized, and the effectiveness of this program was evaluated by a multicenter randomized controlled trial. The results revealed that the case management failed to achieve long-term prevention of repeated suicide attempts but strongly prevented them in
5.1. Limitations The data were collected from one hospital in a metropolitan area; therefore, possible geographical trends cannot be excluded from our findings. Regarding emergency admissions, psychiatrists sometimes did not diagnose participants for several days, and diagnoses were not made using structured clinical interviews. Suicidal ideation and types of suicide were examined from participants’ history rather than using a standardized scale. Contributors Data were collected by TT, analyzed by YO, and interpreted by TT and YO. The work was drafted by TT and AU, and the manuscript was reviewed and edited by TT, YO, and TN. Funding and conflict of interest TT has received research grants from Grants-in-Aid for Scientific Research. TT has also received personal fees from Otsuka, GlaxoSmithKline, MSD, Dainippon-Sumitomo, Janssen, Eisai, Eli Lilly, Daiichi-Sankyo, Shionogi, and Tanabe-Mitsubishi. However, these companies had no role in this study’s design or execution. YO received research grants from the Japan Agency for Medical Research and Development; Ministry of Health, Labour and Welfare; Japan Society for the Promotion of Science; Institute for Health 4
Asian Journal of Psychiatry xxx (xxxx) xxx–xxx
T. Takeuchi et al.
Economics and Policy; and Mental Health and Morita Therapy. YO also served as a member of an advisory board for Janssen Pharmaceuticals, Inc., and the chairman of the Reporting Quality Initiative of Researchers in Clinical Epidemiology. AU received research grants from the Ministry of Education, Culture, Sports, Science and Technology and from the SONY corporation. However, these grants had no role in the current study. TN has received research grants from the Otsuka, MSD, Eisai, Astellas, Tsumura, Pfizer, Shionogi, Yoshitomi, and Takeda and personal fees for lectures on psychiatry and brain research from the Otsuka, MSD, Dainippon-Sumitomo, Meiji Seika Pharma, Shiseido, Mochida, Glaxo-Smith-Kline, and Takeda. However, these companies had no role in the design or execution of the present study.
Network (CDSERN), 2015. Risk of suicide following deliberate self-poisoning. JAMA Psychiatry 72, 570–575. Hori, S.I., Kinoshita, K., 2016. Clinical characteristics of patients who overdose on multiple psychotropic drugs in Tokyo. J. Toxicol. Sci. 41, 765–773. Japan Organ Transplant Network, 2011. Japan Organ Transplant Network Homepage. Donors and Transplants Data. 1 January–31 December 2011. ([cited 23 Nov 2014]. Available from:). http://www.jotnw.or.jp/english/2011data.html. Kawanishi, C., Aruga, T., Ishizuka, N., Yonemoto, N., Otsuka, K., Kamijo, Y., Okubo, Y., Ikeshita, K., Sakai, A., Miyaoka, H., Hitomi, Y., Iwakuma, A., Kinoshita, T., Akiyoshi, J., Horikawa, N., Hirotsune, H., Eto, N., Iwata, N., Kohno, M., Iwanami, A., Mimura, M., Asada, T., Hirayasu, Y., ACTION-J Group, 2014. Assertive case management versus enhanced usual care for people with mental health problems who had attempted suicide and were admitted to hospital emergency departments in Japan (ACTION-J): a multicentre, randomised controlled trial. Lancet Psychiatry 1, 193–201. Kubota, Y., Hasegawa, K., Taguchi, H., Kitamura, T., Nishiyama, C., Iwami, T., Nishiuchi, T., Hiraide, A., 2015. Characteristics and trends of emergency patients with drug overdose in Osaka. Acute Med. Surg. 2, 237–243. Lilley, R., Owens, D., Horrocks, J., House, A., Noble, R., Bergen, H., Hawton, K., Casey, D., Simkin, S., Murphy, E., Cooper, J., Kapur, N., 2008. Hospital care and repetition following self-harm: multicentre comparison of self-poisoning and self-injury. Br. J. Psychiatry 192, 440–445. Matsuyama, T., Kitamura, T., Kiyohara, K., Hayashida, S., Kawamura, T., Iwami, T., Ohta, B., 2016. Characteristics and outcomes of emergency patients with self-inflicted injuries: a report from ambulance records in Osaka City, Japan. Scand. J. Trauma Resusc Emerg. Med. 24, 68. http://dx.doi.org/10.1186/s13049-016-0261-0. Perry, I.J., Corcoran, P., Fitzgerald, A.P., Keeley, H.S., Reulbach, U., Arensman, E., 2012. The incidence and repetition of hospital-treated deliberate self-harm: findings from the world's first national registry. PLoS One 7, e31663. Specified Report of Vital Statistics, 2005. Ministry of Health, Labour and Welfare Homepage. (Available from:). http://www.mhlw.go.jp/toukei/saikin/hw/jinkou/ tokusyu/suicide04/. Sugita, M., Nomura, T., Sekii, H., 2011. An analysis of 611 patients with drug overdose. Juntendoigaku 617–623.
Acknowledgements We wish to thank the department psychiatrists who evaluated and treated the participants and the emergency/ICU staff who provided the participants’ information. We also thank Michio Itasaka for his help in statistical analysis. References Austin, P.C., 2011. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav. Res. 46, 399–424. Finkelstein, Y., Macdonald, E.M., Hollands, S., Sivilotti, M.L., Hutson, J.R., Mamdani, M.M., Koren, G., Juurlink, D.N., Canadian Drug Safety and Effectiveness Research
5
Contents lists available at ScienceDirect
Asian Journal of Psychiatry journal homepage: www.elsevier.com/locate/ajp
Asian pearls
Clinical characteristics of suicidal behavior in an intensive care unit at a university hospital in Japan: A 7-year observational study ⁎
Takashi Takeuchia, , Yasuyuki Okumurab, Akihito Uezatoa, Toru Nishikawaa a
Section of Psychiatry and Behavioral Sciences, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan b Research Department, Institute for Health Economics and Policy, Association for Health Economics Research and Social Insurance and Welfare, 1-5-11 Nishi-Shinbashi, Minato-ku, Tokyo 105-0003, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords: Borderline personality disorder Drug overdose Interpersonal problems Psychiatrist Suicidal behavior
Background: Suicidal behavior accounts for at least 40,000 admissions per year to emergency departments in Japan; however, little is known about emergency admissions owing to suicidal behavior in metropolitan areas. Therefore, we examined the clinical characteristics of suicidal behavior using psychosocial assessments performed by experienced psychiatrists in an intensive care unit. Methods: Participants were 971 patients admitted to a university hospital’s intensive care unit for suicidal behavior between July 2006 and June 2013. Physicians and psychiatrists regularly assessed the participants using a standard data extraction form while the participants were in the intensive care unit. As suicidal behavior involving drug overdose is generally less fatal than other methods, we predicted that clinical characteristics would differ between patients with and without overdose. We classified participants into drug overdose or other method groups (ns = 732 and 239, respectively) to compare suicide methods. Results: In the overdose group, participants’ median age was approximately 5 years lower, and the following proportions were larger: female participants (77%) and participants with borderline personality disorders (21% vs. 10%), no clear suicidal ideation (30% vs. 15%), impulsively attempted self-harm (86% vs. 62%), and interpersonal problems (26% vs. 16%). Conclusion: Ameliorating interpersonal problems and improving stress coping skills would benefit people who attempt suicide via overdose.
1. Introduction In Japan, the annual number of people committing suicide exceeded 30,000 in 1998, and the suicide rate increased to 27.0 per 100,000 population in 2003, recording the highest rate in history. The annual total number fell below 30,000 in 2012 and has gradually decreased since then. However, according to the World Health Organization, the suicide rate in Japan is still as high as 21.7 (2012), which is the tenth highest rate in the world. The most common suicide method is by hanging, which accounts for approximately 60% of all suicides. This is followed by gassing oneself, jumping off a high place, and drug overdose; however, each of these methods accounts for 10% or less of the total number of suicides. Although some people drink alcohol before committing suicide, there are few cases of suicide by drinking. Moreover, because possession of guns is legally prohibited in Japan, suicide using a gun, which is common in the United States and other countries, is also rare (Specified Report of Vital Statistics, 2005).
⁎
Each year, suicidal behavior accounts for at least 40,000 admissions to emergency departments in Japan (Japan Organ Transplant Network, 2011). Suicidal behavior is a strong predictor of subsequent suicide and premature death (Finkelstein et al., 2015). Better understanding of region-specific suicidal behavior epidemiology may prevent suicide and reduce the burden of suicidal behavior on emergency medical systems; however, little research has examined suicidal behavior in metropolitan areas in Japan (Hori and Kinoshita, 2016; Kubota et al., 2015; Sugita et al., 2011). Additionally, most studies examining the clinical characteristics of suicidal behavior have not used specialized psychosocial assessment and, therefore, have not collected or assessed some important clinical information (e.g., psychiatric diagnosis) (Hori and Kinoshita, 2016; Kubota et al., 2015; Sugita et al., 2011). Therefore, we used psychosocial assessments by experienced psychiatrists to examine suicidal behavior’s clinical characteristics in an intensive care unit in a university hospital in Tokyo. As suicidal behavior involving drug overdose is generally less fatal than other methods,
Corresponding author. E-mail addresses: [email protected] (T. Takeuchi), [email protected] (Y. Okumura), [email protected] (A. Uezato), [email protected] (T. Nishikawa).
https://doi.org/10.1016/j.ajp.2017.11.005 Received 25 February 2017; Received in revised form 17 May 2017; Accepted 6 November 2017 1876-2018/ © 2017 Elsevier B.V. All rights reserved.
Please cite this article as: Takeuchi, T., Asian Journal of Psychiatry (2017), http://dx.doi.org/10.1016/j.ajp.2017.11.005
Asian Journal of Psychiatry xxx (xxxx) xxx–xxx
T. Takeuchi et al.
we predicted that clinical characteristics would differ between patients with and without overdose. We also compared clinical characteristics between individuals who attempted suicide via drug overdose and via other methods.
Table 1 Methods of suicide behavior. Methods
Total (n = 971) n (%)
Men (n = 273) n (%)
Women (n = 698) n (%)
2. Methods
Drug overdose (only overdose)
732 (75.4)
170 (62.3)
562 (80.5)
2.1. Design and setting
Non-overdose Self-cutting Chemical poisoning Jumping Use of charcoal briquettes Hanging Drowning Traffic Inhaling carbon monoxide Other
88 (9.1) 37 (3.8)
38 (13.9) 28 (10.3)
50 (7.2) 9 (1.3)
36 (3.7) 34 (3.5)
17 (6.2) 19 (7.0)
19 (2.7) 15 (2.1)
19 (2.0) 9 (0.9) 8 (0.8) 4 (0.4)
8 3 4 3
11 (1.6) 6 (0.9) 4 (0.6) 1 (0.1)
4 (0.4)
2 (0.7)
The protocol for this research project was approved by a suitably constituted Ethics Committee of the institution and conforms to the provisions of the Declaration of Helsinki. Committee of the Tokyo Medical and Dental University, Approval No. 1604. The requirement to obtain patients’ consent was waived as all the data assessed in the study were collected through our regular clinical care. We conducted a 7-year observational study of consecutive patients hospitalized for suicidal behavior (i.e., self-poisoning and self-injuries) in an intensive care unit at the Tokyo Medical and Dental Hospital University in Tokyo, Japan, between July 1, 2006 and June 30, 2013. At our hospital, among patients with suicidal behavior, those deemed to require hospitalization are always admitted to the intensive care unit. The unit provides tertiary emergency care to 10,000–15,000 patients annually.
(2.9) (1.1) (1.5) (1.1)
2 (0.3)
admitted, there were, on an average, 1 or 2 participants per year who died after admission and before psychosocial assessment. Moreover, because no participant died after psychosocial assessment, the annual mean number of participants who died by suicide was 1 or 2, which was also the number of participants who died before psychosocial assessment. The most common method of suicidal behavior was drug overdose (75.4%). Among individuals who overdosed, 88.5% used psychotropic drugs prescribed by their physicians. The second most common method was self-cutting that required surgical procedures in an operating room (9.1%) followed by chemical poisoning (3.8%), jumping (3.7%), and use of charcoal briquettes (3.5%; Table 1).
2.2. Data collection Since 2006, physicians and psychiatrists in the unit had routinely assessed clinical information for all patients who survived self-harm episodes using a standard data extraction form that examines sex, age, psychiatric diagnosis, psychosocial factors (person’s predisposing distress, characteristics of attempts, suicidal ideation), and history of outpatient treatment. Two psychiatrists with at least 5 years’ experience conducted psychosocial assessment for all patients based on the International Classification of Diseases, 10th Revision. We included patients who died after psychosocial assessment and excluded those who died of critical physical conditions before the psychosocial assessment.
3.2. Cumulative probability of suicidal behavior as a function of age Participants in the overdose group were more commonly women (76.8% vs. 56.9%; p < 0.01). Fig. 1 illustrates the cumulative probability of suicidal behavior as a function of age. The median age was approximately 5 years lower in the overdose group (median age: 34 years for men and 32 years for women) than in the non-overdose group (median age: 39 years for men and 36 years for women,).
2.3. Statistical analysis First, we created the following categories of suicidal behavior: drug overdose, self-cutting, jumping, chemical poisoning, use of charcoal briquettes, hanging, drowning, traffic, inhaling carbon monoxide, and others. We calculated proportions of these methods by sex. Second, we divided participants according to major method of suicidal behavior: drug overdose (n = 732) and other methods (i.e., nonoverdose; n = 239). Participants who used a drug overdose in addition to other methods were classified as non-overdose, because such individuals typically used an overdose to encourage themselves to undertake more lethal methods of suicide. The cumulative probability of admissions owing to suicidal behavior as a function of age was estimated by sex and major method of suicidal behavior. Additionally, we compared the clinical characteristics of suicidal behavior between overdose and non-overdose. We used the standardized difference (d) as a categorical variable to measure balances of clinical characteristics between the groups—a standardized difference of > 10% indicated imbalance (Austin, 2011). We used chi-squared tests to evaluate differences between the groups. SPSS version 20 was used. Values of < 0.05 were considered significant.
3.3. Diagnosis The most common diagnosis was major depressive disorder (28.9%) followed by adjustment disorder (22.6%), borderline personality disorder (18.3%), schizophrenia (12.4%), and bipolar disorder (7.8%; Table 2). Participants in the overdose group more commonly had borderline personality disorder (21.0% vs. 10.0%; d = 30.7; p < 0.05) and less commonly had major depressive disorder (25.7% vs. 38.9%; d = 28.6; p < 0.05) or schizophrenia (11.1% vs. 16.3%; d = 15.3; p < 0.05). 3.4. Psychosocial factors The most frequent predisposing distress was family issues (32.5%) followed by interpersonal (23.6%), occupational (11.7%), financial (6.3%), and health problems (5.8%; Table 3). Participants in the overdose group were more likely to have interpersonal issues (26.1% vs. 15.9%; d = 25.2; p < 0.05) and less likely to have financial problems (4.8% vs. 10.9%; d = 22.8; p < 0.05). The most common type of suicidal behavior was impulsive action (80.3%) followed by planned action (12.0%) and action induced by delusion or hallucinations (4.1%; Table 3). In this study, no assessment scale was used to assess the impulsiveness underlying suicidal behavior. This impulsive action was defined as a case in which a participant impulsively attempted suicide without planning or the influence of
3. Results 3.1. Methods of suicidal behavior During the study period, 971 individuals were eligible for analysis. Although there were occasional cases in which some individuals had already died before arrival at our emergency center and thus were not 2
Asian Journal of Psychiatry xxx (xxxx) xxx–xxx
T. Takeuchi et al.
Fig. 1. The cumulative probability of suicidal behavior as a function of age.
Table 2 Diagnosis. Diagnosis
Total (n = 971) n (%)
Overdose (n = 732) n (%)
Non-overdose (n = 239) n (%)
Standardized difference, %
Depressive disorders Bipolar disorders Adjustment disorders Borderline personality disorders Schizophrenia Anxiety disorders
281 (28.9) 76 (7.8) 219 (22.6) 178 (18.3) 120 (12.4) 77 (7.9)
188 (25.7) 62 (8.5) 168 (23.0) 154 (21.0) 81 (11.1) 63 (8.6)
93 14 51 24 39 14
28.6* 10.1 3.9 30.7* 15.3* 10.6
(38.9) (5.9) (21.3) (10.0) (16.3) (5.9)
* p < 0.05.
hallucination or delusion, according to the participants’ statements. Participants in the overdose group more commonly impulsively attempted suicidal behavior (86.3% vs. 61.9%; d = 58.1; p < 0.05) and had no suicidal ideation (29.8% vs. 14.6%; d = 37.0; p < 0.05).
Table 3 Psychosocial factors. Psychosocial factors
3.5. Treatment status Participants in the overdose group more commonly received psychiatric treatment (85.4% vs. 69.0%; d = 39.7; p < 0.05; Table 4); 43.0% of patients with psychiatric treatment history had received psychiatric treatment at least once every two weeks. Frequency of psychiatric treatment did not differ significantly between the overdose and non-overdose groups.
Primal distress Family Interpersonal Occupational Financial Health Other No problem Unknown
4. Discussion
Total (n = 971) n (%)
Overdose (n = 732) n (%)
Nonoverdose (n = 239) n (%)
Standardized difference, %
316 (32.5) 229 (23.6) 114 (11.7) 61 (6.3) 56 (5.8) 72 (7.4) 78 (8.0) 45 (4.6)
249 (34.0) 191 (26.1) 79 (10.8) 35 (4.8) 36 (4.9) 57 (7.8) 50 (6.8) 35 (4.8)
67 38 35 26 20 15 28 10
13.0 25.2* 11.6 22.8* 13.9 5.9 16.9* 2.9
632 (86.3) 57 (7.8) 20 (2.7)
148 (61.9) 60 (25.1) 20 (8.4)
58.1* 48.0* 24.8*
9 (1.2) 14 (1.9)
2 (0.8) 9 (3.8)
3.9 11.2
425 (58.1) 218 (29.8) 89 (12.2)
173 (72.4) 35 (14.6) 31 (13.0)
30.4* 37.0* 2.5
Types of suicidal behavior Impulsive 780 (80.3) Planned 117 (12.0) Delusions or 40 (4.1) hallucinations Other 11 (1.1) Unknown 23 (2.4)
This study examined suicidal behavior’s clinical characteristics in an intensive care unit in Tokyo. We found that drug overdose was the most common method of self-harm, supporting previous research (Lilley et al., 2008; Perry et al., 2012). In Japan, there are few detailed reports on suicidal behavior. According to a report on emergency transports resulting from self-harm in Osaka, one of the Japanese metropolises (Matsuyama et al., 2016), the most common cause was an overdose of psychotropic and other drugs, followed by cutting of the wrist and arm and hanging. The mortality due to hanging was reported to be as high as approximately 80%. Moreover, suicidal behavior of men was more likely to result in death, whereas that of women was
Suicidal ideation Yes No Unknown * p < 0.05.
3
598 (61.6) 253 (26.1) 120 (12.4)
(28.0) (15.9) (14.6) (10.9) (8.4) (6.3) (11.7) (4.2)
Asian Journal of Psychiatry xxx (xxxx) xxx–xxx
T. Takeuchi et al.
individuals with suicidal behavior for 6 months. This effect was more prominent in women, participants younger than 40 years, and those with a history of previous suicide attempts (Kawanishi et al., 2014). At our hospital, when patients admitted to our emergency center for suicidal behavior are examined, psychiatrists assess the presence or absence of suicidal ideation, the state of hallucination and delusion, and support systems in order to determine the need for inpatient psychiatric treatment. In this study, because approximately 80% of the participants had a history of outpatient psychiatric treatment as is shown in Table 4, when we decided that they could be discharged, we instructed them to see their primary psychiatrists immediately. We informed their primary psychiatrists of the course of the participants’ attempts to commit suicide in writing. Those without a history of outpatient treatment were treated at our outpatient clinic or referred to neighborhood clinics. Although we did not check whether the participants were treated by their primary psychiatrists after discharge in this study, many participants seriously considered the recommendation for further treatment.
Table 4 Treatment status. Treatment
Total (n = 971) (%)
Outpatient treatment Never 179 (18.4) Clinics 507 (52.2) General 149 (15.3) hospitals Psychiatric 121 (12.5) hospitals Unknown 15 (1.5) Frequency of treatmenta Weekly 85 (13.5) Biweekly 185 (29.5) Monthly 94 (15.0) Discontinued 79 (12.6) Irregular 40 (6.4) Other 41 (6.5) Unknown 104 (16.6)
Overdose (n = 732) (%)
Nonoverdose (n = 239) (%)
Standardized difference, %
107 (14.6) 409 (55.9) 113 (15.4)
74 (31.0) 96 (40.2) 36 (15.1)
39.7* 31.8* 1.0
93 (12.7)
28 (11.8)
3.0
10 (1.4)
5 (2.1)
5.6
69 (13.9) 152 (30.6) 72 (14.5) 58 (11.7) 34 (6.8) 30 (6.0) 82 (16.5)
16 (12.2) 33 (25.2) 22 (16.8) 21 (16.0) 6 (4.6) 11 (8.4) 22 (16.8)
5.0 12.0 6.4 12.6 9.8 9.1 0.8
5. Conclusion This study examined the characteristics of suicidal behavior in a metropolitan area of Tokyo. The most frequent method of suicidal behavior was drug overdose. Individuals who overdosed were younger than those who use other methods. Individuals who overdosed more commonly had borderline personality disorder, no clear suicidal ideation, impulsively attempted self-harm, and interpersonal problems. These individuals would benefit from interventions that ameliorate interpersonal problems and improve stress coping skills. Thus, the support program (case management) that had been built using the framework of methods including regular face-to-face (or telephonic) interviews with individuals with suicidal behavior, promotion of psychiatric treatment, provision of information on public social programs and private support organizations, arrangements for using these services, and provision of information regarding psychological education, which were also applied in the ACTION-J trial, appeared to be useful. Because the reduction in the number of people committing suicide in Japan seems to be largely attributable to this case management, we consider that this aspect should also be learned in other countries.
a
Sample sizes were 628 for total, 497 for overdose, and 131 for non-overdose. * p < 0.05.
more likely to be non-fatal. There are some reports on drug overdose in Tokyo. Sugita et al. reported that 60% of the suicides were by women, and that, although approximately half of the patients needed hospitalization, all of them had physically improved and were discharged (Sugita et al., 2011). Hori et al. reported that approximately 80% of the patients were women, while the mortality rate for women was 0.4% (Hori and Kinoshita, 2016). Thus, the results of our study appeared to be nearly consistent with those of other Japanese reports. Participants’ median age was younger in the overdose group; this contrasts with previous research examining an emergency department (Lilley et al., 2008). This may be because we examined inpatients rather than emergency department patients. Most participants who overdosed ingested psychotropic medications. Prior to admission for suicide attempts, approximately 60% of patients had received psychiatric treatment at least once a month. This finding suggests a need for more careful selection of treatment options for psychiatrists. Overdose-group participants more commonly had borderline personality disorders and interpersonal problems and less commonly had suicidal ideation; therefore, psychiatrists should choose suicide prevention strategies according to the type of self-harm observed. Regarding individuals who overdose, psychiatrists should aim to improve patients’ coping skills for interpersonal issues using psychosocial interventions and take judicious care when prescribing psychotropic medications for borderline personality disorders. Individuals who engage in other forms of suicidal behaviors should aim to ameliorate any physical or financial problems in collaboration with social workers. In Japan, a multicenter randomized controlled trial titled PostSuicide Attempt Case Management for the Prevention of Further Attempts in Japan (ACTION-J) was conducted. In this trial, a new support program (case management) was built using the framework of regular face-to-face (or telephonic) interviews with individuals with suicidal behavior, promotion of psychiatric treatment among them, provision of information on public social programs and private support organizations and arrangements for using these services, provision of information on psychological education, and so on. Then, a nationwide study group consisting of 17 institutions where the departments of emergency medicine and psychiatry were already cooperating (ACTION-J Group) was organized, and the effectiveness of this program was evaluated by a multicenter randomized controlled trial. The results revealed that the case management failed to achieve long-term prevention of repeated suicide attempts but strongly prevented them in
5.1. Limitations The data were collected from one hospital in a metropolitan area; therefore, possible geographical trends cannot be excluded from our findings. Regarding emergency admissions, psychiatrists sometimes did not diagnose participants for several days, and diagnoses were not made using structured clinical interviews. Suicidal ideation and types of suicide were examined from participants’ history rather than using a standardized scale. Contributors Data were collected by TT, analyzed by YO, and interpreted by TT and YO. The work was drafted by TT and AU, and the manuscript was reviewed and edited by TT, YO, and TN. Funding and conflict of interest TT has received research grants from Grants-in-Aid for Scientific Research. TT has also received personal fees from Otsuka, GlaxoSmithKline, MSD, Dainippon-Sumitomo, Janssen, Eisai, Eli Lilly, Daiichi-Sankyo, Shionogi, and Tanabe-Mitsubishi. However, these companies had no role in this study’s design or execution. YO received research grants from the Japan Agency for Medical Research and Development; Ministry of Health, Labour and Welfare; Japan Society for the Promotion of Science; Institute for Health 4
Asian Journal of Psychiatry xxx (xxxx) xxx–xxx
T. Takeuchi et al.
Economics and Policy; and Mental Health and Morita Therapy. YO also served as a member of an advisory board for Janssen Pharmaceuticals, Inc., and the chairman of the Reporting Quality Initiative of Researchers in Clinical Epidemiology. AU received research grants from the Ministry of Education, Culture, Sports, Science and Technology and from the SONY corporation. However, these grants had no role in the current study. TN has received research grants from the Otsuka, MSD, Eisai, Astellas, Tsumura, Pfizer, Shionogi, Yoshitomi, and Takeda and personal fees for lectures on psychiatry and brain research from the Otsuka, MSD, Dainippon-Sumitomo, Meiji Seika Pharma, Shiseido, Mochida, Glaxo-Smith-Kline, and Takeda. However, these companies had no role in the design or execution of the present study.
Network (CDSERN), 2015. Risk of suicide following deliberate self-poisoning. JAMA Psychiatry 72, 570–575. Hori, S.I., Kinoshita, K., 2016. Clinical characteristics of patients who overdose on multiple psychotropic drugs in Tokyo. J. Toxicol. Sci. 41, 765–773. Japan Organ Transplant Network, 2011. Japan Organ Transplant Network Homepage. Donors and Transplants Data. 1 January–31 December 2011. ([cited 23 Nov 2014]. Available from:). http://www.jotnw.or.jp/english/2011data.html. Kawanishi, C., Aruga, T., Ishizuka, N., Yonemoto, N., Otsuka, K., Kamijo, Y., Okubo, Y., Ikeshita, K., Sakai, A., Miyaoka, H., Hitomi, Y., Iwakuma, A., Kinoshita, T., Akiyoshi, J., Horikawa, N., Hirotsune, H., Eto, N., Iwata, N., Kohno, M., Iwanami, A., Mimura, M., Asada, T., Hirayasu, Y., ACTION-J Group, 2014. Assertive case management versus enhanced usual care for people with mental health problems who had attempted suicide and were admitted to hospital emergency departments in Japan (ACTION-J): a multicentre, randomised controlled trial. Lancet Psychiatry 1, 193–201. Kubota, Y., Hasegawa, K., Taguchi, H., Kitamura, T., Nishiyama, C., Iwami, T., Nishiuchi, T., Hiraide, A., 2015. Characteristics and trends of emergency patients with drug overdose in Osaka. Acute Med. Surg. 2, 237–243. Lilley, R., Owens, D., Horrocks, J., House, A., Noble, R., Bergen, H., Hawton, K., Casey, D., Simkin, S., Murphy, E., Cooper, J., Kapur, N., 2008. Hospital care and repetition following self-harm: multicentre comparison of self-poisoning and self-injury. Br. J. Psychiatry 192, 440–445. Matsuyama, T., Kitamura, T., Kiyohara, K., Hayashida, S., Kawamura, T., Iwami, T., Ohta, B., 2016. Characteristics and outcomes of emergency patients with self-inflicted injuries: a report from ambulance records in Osaka City, Japan. Scand. J. Trauma Resusc Emerg. Med. 24, 68. http://dx.doi.org/10.1186/s13049-016-0261-0. Perry, I.J., Corcoran, P., Fitzgerald, A.P., Keeley, H.S., Reulbach, U., Arensman, E., 2012. The incidence and repetition of hospital-treated deliberate self-harm: findings from the world's first national registry. PLoS One 7, e31663. Specified Report of Vital Statistics, 2005. Ministry of Health, Labour and Welfare Homepage. (Available from:). http://www.mhlw.go.jp/toukei/saikin/hw/jinkou/ tokusyu/suicide04/. Sugita, M., Nomura, T., Sekii, H., 2011. An analysis of 611 patients with drug overdose. Juntendoigaku 617–623.
Acknowledgements We wish to thank the department psychiatrists who evaluated and treated the participants and the emergency/ICU staff who provided the participants’ information. We also thank Michio Itasaka for his help in statistical analysis. References Austin, P.C., 2011. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav. Res. 46, 399–424. Finkelstein, Y., Macdonald, E.M., Hollands, S., Sivilotti, M.L., Hutson, J.R., Mamdani, M.M., Koren, G., Juurlink, D.N., Canadian Drug Safety and Effectiveness Research
5