Typology of Self-injurious Behaviors Admitted to a Tertiary Care Hospital

Abstract

Background:

If we can classify self-injurious behaviors without a priori assumptions about related variables, it could contribute to the nosology of self-injurious behaviors with etiological and management underpinnings.

Methods:

A cross-sectional study was conducted among 90 adult subjects consecutively admitted to the medical and surgical wards of a tertiary care center. In addition to socio-demographic and relevant clinical variables, intent, lethality, impulsivity, social support, and stressful life events were measured. K-means cluster analysis was used to delineate groups.

Results:

Hierarchical cluster analysis followed by K-means cluster analysis yielded three groups and their characteristics. The first group consisted of older individuals with high lethality and intent, and low impulsivity. Mental illnesses such as depressive disorder and delusional disorder were more common in this group compared to the others. The second group showed some impulsivity and poor perceived social support. Their attempts were characterized by lower lethality and intent. The third group included the youngest individuals, with high impulsivity and high mean stress scores despite having high perceived social support. Both the second and third groups had a fair representation of borderline personality disorder.

Conclusions:

Adults with self-injurious behaviors could be divided into three distinct clusters, which has implications for nosology, etiological models, and management.

Keywords

Self-injurious behavior suicidal attempt cluster analysis typology

Key Messages:

Self-injurious behaviors vary in intent, lethality, and other variables. There are difficulties in the conceptualization and clinical application of categories of suicidal behaviors and self-injurious behaviors.

Using an unsupervised machine learning paradigm, classifying self-injurious behaviors without an a priori assumption may solve the problem.

The study could identify three types of self-injurious behaviors based on limited variables.

There are prospects for a more rigorous classification of self-injurious behaviors based on a larger sample size and more variables.

Suicide is a significant public health problem and a leading cause of mortality across all ages. The global burden of suicides is estimated to be more than seven lakhs per year.¹ Moreover, the burden of suicidal attempts is reported to be 10–20 times this estimate.² Distinguishing various suicidal behaviors has been a conundrum for both researchers and clinicians, and there is a constant search for evidence-based terminology.³ Fatality or completion of the act has been suggested as the cutting edge to differentiate suicide from suicide attempts. However, in addition to the lack of fatality, whether a suicide attempt is also characterized by the deliberate intention to die or by the deliberate intention not to die but only to inflict harm on oneself is a difficult question to answer.^4,5 To circumvent this issue, the broad term “deliberate self-harm” has been used by various authors for all self-injurious behaviors regardless of intent, which includes suicides, suicide attempts, and other non-suicidal self-harm behaviors.³ In this spectrum, non-suicidal self-injury (NSSI) has been defined by the International Society for the Study of NSSI as the deliberate, self-inflicted destruction of body tissue without suicidal intent and for purposes not socially sanctioned.⁶

Dimensions and degrees of lethality and intent are also challenging to measure, and they may vary among different types of suicide attempts.⁷ In addition to lethality and intent, demographic factors such as age, gender, impulsivity, social support, psychiatric morbidity, coping styles, and exposure to childhood trauma are distributed among different types of suicidal and self-injurious behaviors.^8–10 Thus, there are semantic, conceptual, and etiological imperfections regarding the terminology of types of suicidal behaviors, especially when applied to clinical practice. Hence, there is a need to define types of individuals with suicidal behaviors across many higher-order categories of variables or dimensions, including the nature of the attempt, the precipitants, psychological and behavioral mediators, and the purpose of the attempt.¹¹

There have been studies exploring the association of a range of factors and variables with suicide and self-injurious behaviors in diverse populations. There have also been many theory-based and empirical classification attempts. There have also been a number of statistical classification attempts, mainly using cluster analysis. However, there has been a lack of agreement between these studies, mainly because of the heterogeneity in the data collected.¹¹

There have been only a few recent attempts to classify suicidal and self-injurious behaviors inductively and without a priori concepts, using machine learning methods such as cluster analysis. An early review of studies on the classification of suicide attempters without a priori subgroups found a consensus of two groups, mild and severe. The variables measured were heterogeneous across these studies.¹² A subsequent review of cluster analysis research from 1993 to 2011, although identifying certain groups primarily based on the presence of psychiatric illness, noted the heterogeneity of patient groups and variables. Most recent studies have been conducted in special groups, such as military or Medicaid populations. They have not approached such a typology from a conceptual model of the severity of suicide attempts, the presence of psychiatric illness, impulsivity, social support, and the presence of stressful life events.^13–18 Moreover, these studies have used retrospective or chart-based data.

Classifying self-injurious behaviors into distinct groups can advance our understanding of these behaviors. Moreover, such a classification will be helpful in risk stratification among individuals with self-injurious behaviors, allowing for more focused attention and tailored interventions to prevent and manage suicidal as well as other self-injurious behaviors.

Hence, this proof-of-concept study was conducted to classify people with self-injurious behaviors based on explicitly and prospectively measured factors, including lethality, intent, clinical variables, impulsivity, social support, and stressful life events.

Methods

A cross-sectional study was conducted in the medical and surgical wards of a teaching hospital among adult subjects (18–65 years of age) admitted for suicidal or non-suicidal self-injurious behaviors. Those not physically stable enough to undergo assessment were excluded from the study. Since the identification of subgroups involved cluster analysis and the expected groups were approximately three, with vast differences between them, it was assumed that each subgroup would include 30 participants; therefore, a sample size of 90 was considered.¹⁹ Data were collected from 90 subjects. Patients with suicide or self-injurious behaviors who were consecutively admitted to the medical and surgical wards of the study center on two randomly chosen days per week during the study period were selected for the study.

In addition to the socio-economic variables and identification data, the study variables included were psychiatric morbidity diagnosed by a consultant based on Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, family history of suicide, past history of suicidal attempts, alcohol and other substance use disorders, and history of imprisonment. Other variables studied were impulsiveness, perceived social support, stressful life event score, suicidal intent, and lethality scores. Since many patients had features of adjustment disorder and this diagnosis did not fit the general framework of the study, they were grouped with the no diagnosis group.

Impulsiveness was measured by the Barratt Impulsiveness Scale-Revised 21-item (BIS-R-21). BIS-R-21 is a 21-item self-report scale commonly used to measure impulsiveness.²⁰ This is an abbreviated version of the original 30-item scale.²¹

Social support was measured using a multidimensional scale for perceived social support measured social support. It is a self-report measure of subjectively assessed social support. The scale consists of 12 items, and each item is rated on a Likert scale from 0 to 5.²²

The Presumptive Stressful Life Event Scale was used to measure the stressful life event score. This scale was adapted by Gurmeet Singh et al. for the Indian setting and is based on the original English version assessing social adjustment after life events. It consists of fifty-one items, which were further classified according to whether they were personal or impersonal, and as desirable, undesirable, or ambiguous events. Respondents were asked to list the life events that occurred in the year before their admission, and the mean stress score was calculated. This scale was standardized on a sample of 200 subjects from the community.²³

Lethality of suicide attempt was measured by the Scale for Assessment of Lethality of Suicide Attempt (SALSA).²³ The SALSA has two components: the first component has four items indicating the seriousness of the attempt and its likely consequences, and the second component is the global impression of lethality. All items are scored from 1 to 5, with higher scores suggesting increased lethality. An estimate of Cronbach’s alpha suggested an excellent level of internal consistency for SALSA.

The Beck Suicide Intent Scale²⁴ measures suicide intent associated with a previous attempt. The scale comprises 15 items, each rated on an ordinal scale of 0–2, with the total score ranging from 0 to 30. The scale consists of two sections: Section 1 contains items dealing with the objective circumstances related to the suicide attempt, while Section 2 includes items based on the patient’s self-report of their internal concept of intent. The tool was administered by the researcher.

All the instruments were translated into the vernacular by using the standard translation–back translation method. A training session was held to ensure proper administration of the tools and inter-rater reliability among the investigators involved in data collection. The researchers approached participants admitted on two randomly selected days of the week, Mondays and Thursdays, and administered the questionnaire. Data were collected just before discharge, and mostly within a week, once the patient was stable and confidentiality of the collected data was ensured.

In addition to the data description, hierarchical cluster analysis was conducted to identify the possible subgroups, whose characteristics were then discerned using K-means cluster analysis. Cluster analysis is a multivariate analytic technique included under unsupervised machine learning methods. Only continuous measures such as age, impulsiveness, social support, stressful life event score, lethality, and intent were considered for cluster analysis. The elbow method in the scree plot and the silhouette method discerned the suitability of the groups. The groups that emerged were then compared for categorical variables using the chi-square or Fisher’s exact test, depending on the expected cell size. Numerical measures were also compared using one-way analysis of variance or the Wilcoxon’s rank-sum test. Statistical analysis was performed using R software Version 4.4.1, with the packages Rcmdr (2.9–5), FactoMineR (2.1–1), and factoextra (1.0–7) (Bell Laboratories, Murray Hill, New Jersey, USA).

Results

Between February 2024 and July 2024, 120 subjects were admitted to the medical and surgical wards of the study center for suicidal or non-suicidal self-injurious behaviors. We were able to collect data from 90 out of the 120 patients. Data from 30 patients could not be collected; this included one death, eight absconding, 12 discharges against medical advice, and nine patients who were not willing to give consent for the study. Data from 90 patients were analyzed. The background data are summarized in Table 1. The sample population assessed had a slight preponderance of females. Most of the patients belonged to a lower socio-economic class, were married, employed, and were in the third or fourth decade of their life. Approximately 15% had either a family history of suicide or suicide attempts. Nearly a quarter of patients had a past history of self-injurious behavior. Seventy percent of patients had no psychiatric diagnosis other than adjustment disorder, and the remaining had either borderline personality disorder or depressive disorder.

Table 1.

Socio-demographic and Baseline Clinical Characteristics of Participants.

Baseline Characteristics	n (%)
Gender (female)	50 (55.56)
Marital status
Single	22 (24.44)
Married	58 (64.44)
Widowed/separated	10 (11.11)
Occupation (employed)	53 (58.89)
Socio-economic status (BPL)	76 (84.44)
History of suicide in first- or second-degree relative (yes)	16 (17.78)
History of suicide attempt in first- or second-degree relative (yes)	13 (14.44)
Past history of suicide attempts (yes)	25 (27.78)
Legal history (yes)	5 (5.56)
Alcohol and substance use	25 (27.78)
Lethal mode of attempt	39 (43.33)
Psychiatric diagnosis (DSM-5)
No specific psychiatric diagnosis or adjustment disorder	63 (70.00)
Borderline personality disorder	15 (16.67)
Delusional disorder	2 (2.22)
Depressive disorder	10 (11.11)
	Mean (SD)
Age	34.52 (13.56)
Education	11.60 (2.63)
Current Attempt Intent Score	22.81 (5.97)
Current Attempt Lethality Score	12.99 (5.56)
Impulsivity Score	47.14 (8.25)
Perceived Social Support Score	56.98 (15.77)
Stressful Life Event Score	258.72 (281.25)^a

BPL, below poverty line; DSM-5, Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition; ^aMedian (IQR).

Our main objective was to determine whether the subjects with self-injurious behaviors fall into distinct groups. Since the sample size was small and considering the appropriateness of parametric data for K-means cluster analysis, only continuous variables were considered for cluster analysis. These continuous variables were standardized before cluster analysis. The sample could be practically split into three groups based on hierarchical cluster analysis using the Ward method and Euclidean distance as the measure of difference. Figure 1 shows the dendrogram of the cluster analysis.

K-means cluster analysis for three groups identified the characteristics of each group. Table 2 shows the details of the results of the cluster analysis. Group 1 consisted of individuals in a higher age group with high intent and lethality, as evidenced by positive cluster mean scores on these measures, and lower impulsivity with negative cluster mean scores. Group 2 belonged to a comparatively lower age group with low intent and low lethality. Group 3 was the youngest age group with high impulsivity and high stress scores, but low intent and lethality. Interestingly, perceived social support was higher in this group. Figure 2 shows the K-means clusters and corresponding vectors by principal component analysis. Figure 3 shows the results of the adequacy of the three groups by the elbow method and silhouette method. Both methods demonstrated the adequacy of the three groups.

Figure 1.

Hierarchical Cluster Analysis—Dendrogram.

Table 2.

Results of K-means Cluster Analysis Indicating the Average Standardized Value of Each Variable in the Identified Clusters.

		Age	Current Attempt Intent Score	Current Attempt Lethality Score	Current Attempt Impulsivity Score	Perceived Social Support Score	Stressful Life Event Score
Cluster means	Group 1 (n = 45)	0.354	0.754	0.825	−0.26	0	−0.128
	Group 2 (n = 34)	−0.307	−0.757	−0.892	0.06	−0.069	−0.481
	Group 3 (n =11)	−0.5	−0.745	−0.619	0.875	0.215	2.013

Total within sum of squares = 343.77; between cluster sum of squares = 122.38; between cluster/total within sum of squares = 35.6%. Bold characters indicate the cluster means of variables delineating the clusters

Table 3.

Comparison of Means of Various Numerical Measures Used for Cluster Analysis.

	Group 1 (n = 45)		Group 2 (n = 34)		Group 3 (n = 11)		F Value	p Value
	Mean	SD	Mean	SD	Mean	SD	F Value	p Value
Age	39.33^a	14.34	30.35	11.34	27.73	9.70	6.54	.002
Current Attempt Intent Score	27.31^a	4.17	18.29	3.98	18.36	2.41	58.93	<.001
Current Attempt Lethality Score	17.58^a	3.06	8.03	2.72	9.54	4.25	99.53	<.001
Current Attempt Impulsivity Score	45.00	9.11	47.65	5.31	54.36^a	8.24	6.518	.002
Perceived Social Support Score	56.98	12.94	55.88	19.88	60.36	12.52	0.33	.72
Stressful Life Event Score	191^b	252.00^c	107^b	146.25^c	613^b	302.50^c	36.67^d	<.001

^aGroup significantly different from others with adjusted p value (Tukey’s correction) for multiple comparisons; ^bmedian; ^cinterquartile range; ^dKruskal–Wallis chi-squared.

Figure 2.

Clusters of People with a Suicide Attempt and Their Corresponding Variable Vectors.

Figure 3.

Adequacy of Clusters by Elbow Method in Scree Plot and Silhouette.

In Table 3, the numerical measures were also compared using the independent samples t-test or Wilcoxon’s rank-sum test. Perceived social support did not distinguish the groups, although it was higher in group 3. When comparing the categorical variables between clusters, it was found that gender, alcohol and substance use, history of suicide in first- or second-degree relatives, past history of self-injurious behaviors, history of imprisonment, marital status, occupation, and socio-economic status did not distinguish the groups. However, psychiatric illness and a history of self-injurious behaviors in first or second-degree relatives were higher in Group 1 and Group 3, respectively (Table 4).

Table 4.

Comparison of Categorical Variables Between Clusters.

	Group 1 (n = 45)		Group 2 (n = 34)		Group 3 (n = 11)		X²	p Value
	n	%	n	%	n	%	X²	p Value
Gender (female)	21	46.7	23	67.6	6	54.5	3.46	.180
Marital status
Single	9	20.0	10	29.4	3	27.3	2.08	.722
Married	32	71.1	19	55.9	7	63.6
Widowed/separated	4	8.9	5	14.7	1	9.1
Occupation (employed)	31	68.9	17	50	5	45.5	3.79	.150
Socio-economic status (BPL)	38	84.4	28	82.4	10	90.9	0.46	.793
Legal history	2	4.4	2	5.9	1	9.1	0.37	.829
Alcohol and substance use disorder	16	35.6	6	17.6	3	27.3	3.098	.212
Lethal means of attempt	27	60^a	8	23.5	4	36.4	10.74	.005
Psychiatric diagnosis (DSM-5)
Major mental illness	11	24.4^a	1	2.9	0	0	10.69	.030
Borderline personality disorder	5	11.1	7	20.6^a	3	27.3^a
No psychiatric diagnosis or adjustment disorder	29	64.4	26	76.5	8	72.7
Past history of suicide attempts (yes)	12	26.7	8	23.5	5	45.5	2.05	.359
History of suicide attempt in first- or second-degree relative	5	11.1	3	8.8	5	45.5^a	9.83	.007
History of suicide in first- or second-degree relative	7	15.6	8	23.5	1	9.1	1.49	.474

DSM-5, Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition;^aGroup significantly different from others.

Discussion

Among 90 people admitted to a general hospital psychiatry teaching hospital for self-injurious behaviors, the study identified three groups differing in their characteristics. The first group consisted of older individuals with high lethality and intent and low impulsivity. The reported mean stress scores were very low in this group. Mental illnesses such as depressive disorder and delusional disorder were more common in this group compared to others. Intuitively, they may be referred to as “serious attempters.” The second group showed indications of some impulsivity and poor perceived social support. Their attempts were of lower lethality and intent. They had a fair representation of borderline personality disorder diagnoses. They are the “Desolate attempters.” The third group consisted of the youngest individuals, with high impulsivity and high mean stress scores despite having high perceived social support. Intent and lethality were also very low in these individuals. They also had a fair representation of borderline personality disorder. The history of suicidal behavior in first- and second-degree relatives was also high in this group. They can be called “Impulsive attempters.”

Thus, impulsivity, intent, lethality of attempts, and stressful life events seem to be significant factors determining the clusters, and the other variables are perceived social support and stressful life events. Based on secondary analysis of data from a nationwide survey of attempted suicide in Korea, the authors were able to divide the sample into two clusters primarily based on impulsivity. The non-impulsive cluster had high intent and lethality. Although the sample was of a higher age group in that study, our findings align with these findings.¹⁶ The distinction between impulsivity vs lethality and intent was also evident from another study using cluster analysis in a group of people with serious suicide attempts.²⁵ Impulsivity and intent-deciding subgroups are also evident from another study, which found another subset of severe, violent, and frequent attempts.²⁶ However, a study from Hong Kong identified two clusters, one with a high intent score associated with chronic stress, and another with low intent associated with psychiatric illness and acute stress, among people who have died by suicide.²⁷ Our study population, with self-injurious behaviors without fatality and selected irrespective of severity, may be the reason for the difference in the findings.

The inverse relationship between impulsivity and lethality, intent, and age has been explored in studies with a priori hypotheses. A study from China on youth suicides reported that those with high intent are likely to be older, whereas those with low intent are more impulsive.²⁸ Baca-Garcia et al. reported that impulsivity and lethality were inversely associated during the attempt, suggesting that impulsive suicide attempts are less lethal, which is similar to the findings in this study.^29,30 Gade et al. also reported that the mean impulsivity scores are inversely proportional to intent.³¹

Personality disorders, especially borderline personality disorder, fared significantly more in both the “impulsive” and “desolate” clusters. Comparable results have been reported in other studies as well. Personality dimensional scores were negatively correlated with intent and lethality in Indian research on suicide attempts.³² Mental illness, especially depression, fared more in “serious attempters” with higher age, intent, and lethality, as reported in many previous studies.^33,34

One of the unexpected findings from our study is that perceived social support was higher in the “impulsive” cluster. They had high stressful life event scores. Family history of suicide or suicide attempts was also high in this group. Comparable results have been reported in other studies. It has been reported that both stressful life events and perceived social support are related to suicide attempts; however, there was no moderating or mediating effect of perceived social support in the association between suicide attempts and stressful life events.³⁵ Contradictory findings have also been reported.³⁶ It may also be that perceived social support does not reflect received social support.³⁷ Moreover, perceived social support is often dynamic and may change after an attempt.³⁸ Although the cluster mean size is low, poor perceived social support alone delineates a “Desolate” cluster with very low lethality and attempts. The results should be replicated in a larger sample, and the possible reasons must be explored further.

Certain theoretical speculations, although imperfect, are possible based on the findings. The identified cluster of “serious attempters” represents the true suicide attempt, characterized by an intent to die but an inability to complete the act. They may represent a group for which appropriate and optimal psychopharmacological treatment could be an important preventive strategy. The “impulsive” subtype represents a personality pathology, mainly the ICD-10 emotionally unstable personality disorder, impulsive subtype. The “desolate” attempters also represent personality pathology, primarily the classical borderline subtype of emotionally unstable personality disorder.³⁹ The “impulsive” cluster probably represents the affective and impulsive subtypes, while the “desolate” cluster is the dependent subtype of borderline personality disorder in the classification of borderline subtypes by Oldham.⁴⁰ It had been proposed that the former is associated with repeated suicide attempts, which unfortunately could not be examined in the current study.⁴¹ A family history of suicidal behavior in this group also supports this speculation.

Some of the early classification attempts also align with the proposed classification from this study. Paykel and Rassaby classified parasuicides into severe with violent attempts, low lethality with interpersonal motivation, and multiple attempts with high hostility.⁴² Henderson et al. have identified three primary types: depressed with high lethality, operant non-alienated repeater, and operant alienated.⁴³

The study was able to identify three clusters that are distinctly different in terms of suicide-related variables with some similarity to the existing literature. Such an approach could help identify different risk levels and etiological models, allowing for tailored interventions to predict and prevent further attempts. Since the study sample included all self-injurious behaviors without an a priori labeling of the severity of the attempt in terms of intent and lethality, a potential contribution could also be made to unravel the confusing classification of self-injurious behaviors.

Limitations and Future Implications

The extent to which the results can be generalized to other populations is unknown. The study sample was relatively small. Hence, only continuous variables were considered for cluster analysis for statistical robustness, although an attempt was made to examine how the categorical variables fared among the resultant clusters. Studies with larger sample sizes could include more variables in the clustering and further delineate clusters more effectively. Another limitation is that although a past history of attempts was collected and included in the comparison between clusters, the number of suicide attempts was not considered, which intuitively could differ among clusters.

Conclusions

The study showed that adults with self-destructive behaviors can be divided into three distinct clusters based on impulsivity, lethality, intent, psychiatric morbidity, stressful life events, and perceived social support. The possibility of these clusters is consistent with the existing literature. Such a classification of self-injurious behaviors can contribute to a more precise nosology of suicidal and other self-injurious behaviors, with implications for risk stratification and the development of causative models. However, multiple other variables could influence the delineation of clusters, including demographic, clinical, biological, imaging, and genetic variables. This proof-of-concept study could not address many issues due to a limited sample size. A study with a large number of participants measured for many variables, using unsupervised machine learning models with rigorous model validation techniques, could potentially contribute further to the nosology of suicidal behaviors. Geographical clustering could also identify hot spots so that interventions can be prioritized. Moreover, the variables such as lethality and intent could help predict possible risk stratification among these clusters. The resultant clusters may differ in terms of interventions, such as pharmacological treatment of psychiatric disorders, management of impulsivity, crisis intervention, etc.

Supplemental Material

Supplemental material for this article is available online.

Footnotes

Acknowledgements

We thank the faculty members and residents of the Department of Psychiatry, Government T.D. Medical College, Alappuzha: Dr Deenu Chacko, Dr Ganga G. Kaimal, Dr Shalima S., Dr Aswathy M., Dr Gopika S. S., and Dr Sarath S.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Declaration Regarding the Use of Generative AI

The authors attest that there was no use of generative artificial intelligence (AI) technology in the generation of text, figures, or other informational content of this manuscript.

Ethical Approval

This article was approved by the Institutional Ethics Committee of the Government T.D. Medical College, Vandanam, Alappuzha (01/2024/ ECR/122/Inst/KL/2013/RR-19/01/2024, dated 02/02/2024).

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Informed Consent

Informed consent was obtained from all individuals who participated in the study.

Previous Presentation

The study was presented during IPSOCON-2024, held in Bengaluru on October 19, 2025, and secured the D. S. Raju Memorial Award for the Best PG Paper.

References

World Health Organisation. Suicide [Internet]. 2024 [cited 2024 Nov 20]. Available from: https://www.who.int/news-room/fact-sheets/detail/suicide

World Health Organization. Figures and facts about suicide. Geneva: World Health Organisation, 1999.

Jakobsen

, Nielsen

, Larsen

, . Definitions and incidence rates of self-harm and suicide attempts in Europe: A scoping review. J Psychiatr Res, 2023; 164: 28–36.

Suicidal behaviour in Europe: The situation in the 1990s / by Unni Bille-Brahe [Internet]. World Health Organization, Regional Office for Europe. 2024 [cited 2025 Apr 15]. Available from: https://iris.who.int/handle/10665/108117

Kerkhof

AJFM

. Suicide and attempted suicide. World Health, 1994; 47(2): 18–20.

Halicka

and Kiejna

. Non-suicidal self-injury (NSSI) and suicidal: Criteria differentiation. Adv Clin Exp Med, 2018; 27(2): 257–261.

Beck

, Beck

and Kovacs

Classification of suicidal behaviors: I. Quantifying Intent and medical lethality. Am J Psychiatry, 1976; 132(3): 285–287.

Liang

, Kõlves

, Lew

, . Coping strategies and suicidality: A cross-sectional study from China. Front Psychiatry, 2020; 11: 514376.

Zatti

, Rosa

, Barros

, . Childhood trauma and suicide attempt: A meta-analysis of longitudinal studies from the last decade. Psychiatry Res, 2017; 256: 353–358.

10.

Fox

, Franklin

, Ribeiro

, . Meta-analysis of risk factors for non-suicidal self-injury. Clin Psychol Rev, 2015; 42: 156–167.

11.

Ellis

. Classification of suicidal behavior: A review and step toward integration. Suicide Life Threat Behav, 1988; 18(4): 358–371.

12.

Arensman

and Kerkhof

AJFM

. Classification of attempted suicide: A review of empirical studies, 1963–1993. Suicide Life Threat Behav, 1996; 26(1): 46–67.

13.

Bonkalo

, Shmeleva

, Sabanchieva

, . Typology and factorial conditioning of suicidal behavior of adolescents. Propósitos y Represent. 2021; 9(SPE3): e1194. DOI: 10.20511/pyr2021.v9nSPE3.1194.

14.

Martin

, Novak

, Perera

, . Suicide typologies in the United States air force: A hierarchical cluster analysis. Suicide Life Threat Behav, 2019; 49(6): 1707–1720.

15.

Caves

Sivaraman JJ

, Ranapurwala

, Proescholdbell

, . Suicide typologies among medicaid beneficiaries, North Carolina 2014–2017. BMC Psychiatry, 2022; 22(1): 104.

16.

Kim

, Kim

, . Classification of attempted suicide by cluster analysis: A study of 888 suicide attempters presenting to the emergency department. J Affect Disord, 2018; 235: 184–190.

17.

Kerkhof

AJFM

, van der Wal

, Hengeveld

. Typology of persons who attempted suicide with predictive value for repetition: A prospective cohort study. In Möller

Hans-Jürgen

and Armin

Schmidtke RW

(eds) Current Issues of Suicidology. 1st ed. Springer, Berlin, Heidelberg; 1988. 193–203 p.

18.

O’Connor

, Sheehy

and O’Connor

. The classification of completed suicide into subtypes. J Ment Health, 1999; 8(6): 629–637.

19.

Dalmaijer

, Nord

and Astle

. Statistical power for cluster analysis. BMC Bioinformatics, 2022; 23(1): 205.

20.

Kapitány-Fövény

, Urbán

, Varga

, . The 21-item Barratt Impulsiveness Scale Revised (BIS-R-21): An alternative three-factor model. J Behav Addict, 2020; 9(2): 225–246.

21.

Patton

, Stanford

and Barratt

. Factor structure of the Barratt Impulsiveness Scale. J Clin Psychol, 1995; 51(6): 768–774.

22.

Zimet

, Dahlem

, Zimet

, . The multidimensional scale of perceived social support. J Pers Assess, 1988; 52(1): 30–41.

23.

Singh

, Kaur

and Kaur

Presumptive Stressful Life Events Scale (PSLES): A new stressful life events scale for use in India. Indian J Psychiatry, 1984; 26(2): 107.

24.

Beck

, Schuyler

and Herman

Development of suicidal intent scales. In: Beck

, Resnik

and Lettieri

(eds) The prediction of suicide. Oxford: Charles Press Publishers, 1974, p.xii, 249–249, xii.

25.

Rapeli

and Botega

. Clinical profiles of serious suicide attempters consecutively admitted to a university-based hospital: A cluster analysis study. Braz J Psychiatry, 2005; 27(4): 285–289.

26.

Lopez-Castroman

, Nogue

, Guillaume

, . Clustering suicide attempters: Impulsive-ambivalent, well-planned, or frequent. J Clin Psychiatry, 2016; 77(6): e711–e718.

27.

Chen

EYH

, Chan

WSC

, Chan

SSM

, . A cluster analysis of the circumstances of death in suicides in Hong Kong. Suicide Life Threat Behav. 2007; 37(5): 576–584.

28.

Sun

and Zhang

Characteristics of Chinese rural young suicides: Who did not have a strong intent to die. Compr Psychiatry. 2015; 57: 73–78.

29.

Baca-Garcia

, Diaz-Sastre

, García

Resa E

, . Suicide attempts and impulsivity. Eur Arch Psychiatry Clin Neurosci, 2005; 255(2): 152–156.

30.

Baca-Garcia

, Diaz-Sastre

, Basurte

, . A prospective study of the paradoxical relationship between impulsivity and lethality of suicide attempts. J Clin Psychiatry, 2001; 62(7): 15824.

31.

Gade

, Kalasapati

, Kadari

, . Impulsivity in suicide attempters: A cross-sectional study from a teaching hospital in India. J Clin Diagn Res, 2018; 12(10): VC15–VC20.

32.

Sahoo

, Swain

and Kar

Personality dimension, suicidal intent, and lethality: A cross-sectional study of suicide attempters with or without personality disorders. Indian J Psychol Med. Epub ahead of print 15 November 2024. DOI: 10.1177/02537176241287161.

33.

Woo

, Lee

, . Characteristics of high-intent suicide attempters admitted to emergency departments. J Korean Med Sci. 2018; 33(41): e259.

34.

Yaqoob

and Ahsan

Impact of mood disorders and personality disorders on suicide intent among self-harm patients. Pak J Med Sci, 2023; 39(2): 491–496.

35.

Panesar

, Rosic

, Rodrigues

, . The role of perceived social support in the association between stressful life events and suicidal behavior. Front Psychiatry, 2021; 12: 699682.

36.

Kleiman

, Riskind

, Schaefer

, . The moderating effect of social support. Crisis, 2012; 33(5): 273–279.

37.

Uchino

, Bowen

and Kent

Social support and mental health. In: Friedman

(ed) Encyclopedia of mental health. 2nd ed. New York: Academic Press, 2016, pp. 189–195.

38.

Coppersmith

DDL

, Kleiman

, Glenn

, . The dynamics of social support among suicide attempters: A smartphone-based daily diary study. Behav Res Ther, 2019; 120: 103348.

39.

World Health Organization. The ICD-10 Classification of Mental and Behavioral Disorders: clinical descriptions and diagnostic guidelines. Geneva: World Health Organization, 1992.

40.

Oldham

. Borderline personality disorder and suicidality. Am J Psychiatry, 2006; 163(1): 20–26.

41.

Rebok

, Teti

, Fantini

, . Types of borderline personality disorder (BPD) in patients admitted for suicide-related behavior. Psychiatr Q, 2015; 86(1): 49–60.

42.

Paykel

and Rassaby

Classification of suicide attempters by cluster analysis. Br J Psychiatry, 1978; 133(1): 45–52.

43.

Henderson

, Lance

. Types of attempted suicide (parasuicide). Acta Psychiatr Scand. 1979; 59(1): 31–39.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.09 MB

0.00 MB