Issue 70 November 2004  •  Contents  •  Home       


Who are the kids who
self-harm? An Australian self-report school survey

Diego De Leo and Travis S Heller

Introduction — Methods — Sampling — Survey instrument — Procedure — Data treatment — Results — Sample — Prevalence of DSH — Factors associated with DSH — Time to engage in DSH — Help-seeking behaviours — Discussion — Conclusion — Acknowledgements — Competing interests — References — Author details

Abstract Objective: To determine the prevalence and types of deliberate self-harm (DSH) in adolescents, and associated factors.

Design: A cross-sectional questionnaire study.

Participants and setting: 3757 of 4097 Year 10 and Year 11 students (91.7%) from 14 high schools on the Gold Coast, Queensland, during September 2002.

Main outcome measures: DSH behaviour, including descriptions of the last act, psychological symptoms, recent stressors, coping styles, help-seeking behaviour, lifestyle choices, and self-prescribing of medications.

Results: 233 students (6.2%) met the criteria for DSH in the previous 12 months, with DSH more prevalent in females than males (OR, 7.5; 95% CI, 5.1–10.9). The main methods were self-cutting (138 respondents; 59.2%) and overdosing with medication (69 respondents; 29.6%). Factors associated with DSH included similar behaviours in friends or family, coping by self-blame, and self-prescribing of medications. Most self-harmers did not seek help before or after their most recent action, with those who did primarily consulting friends.

Conclusions: DSH is common in Australian youth, especially in females. Preventive programs should encourage young people to consult health professionals in stressful situations.

Suicidal behaviour in adolescents is common,1 and suicide is the leading cause of death in young Australians.2 More startling is the number of young Australians who deliberately self-harm, with adolescent females engaging in substantially more acts of deliberate self-harm than males do.3

In recent years, community surveys have been conducted to assess suicidal behaviours in Australian youth.4 This method of data collection allows for population-based comparisons between people who report self-harm and those who do not. In a Queensland study, more than 60% of university-age students reported suicidal ideation, with 6.6% stating that they had performed at least one act of self-harm.5

Despite advances in research into suicidal behaviours, a lack of consistent and analytically useful data has inhibited knowledge. Moreover, varying definitions have prevented useful comparisons of studies. As part of the current investigation, the Child and Adolescent Self-harm in Europe (CASE) group ( has developed a definition consisting of predetermined criteria as to what constitutes an act of self-harm (Box 1). The Australian Institute for Suicide Research and Prevention is the only collaborator in this multisite study outside Europe.

The aim of this study was to examine the factors associated with adolescent self-harm in an Australian population and to better understand adolescents’ motivations for self-harm. Furthermore, we aimed to report on the extent to which self-harm is hidden within the community by identifying the proportion of self-reported self-harm episodes that are not presented at hospitals or mental health services. Coping strategies and help-seeking behaviour were also examined.6

Methods Sampling The Gold Coast, in Queensland, has a population of about 425 000.7 All 29 coeducational schools (15 government, 14 independent) on the Gold Coast were approached to participate in the study, and 14 schools (10 government, 4 independent) agreed to take part. Reasons for non-participation included only having students up to Year 9 enrolled, schools not wanting interruptions to their curriculum, and refusal to allow material regarding suicidal behaviours to be given to students. All students enrolled in Year 10 and Year 11 from these 14 schools were invited to participate (n = 4097). A consent form was developed to provide an “opt out” option whereby parents who did not wish their children to participate were required to return the consent form.

Survey instrument The 131-item instrument internationally developed for the CASE study (“Lifestyle and Coping Questionnaire”) investigates sociodemography, lifestyle choices, recent stressors, suicidal thoughts and behaviours, personality items, and coping techniques. The Australian version includes a set of questions investigating self-prescribing of medication. Three standardised scales were also included: the Hospital Anxiety and Depression Scale,8 Plutchik’s Impulsivity Scale,9 and an abridged version of the Self Concept Scale.10 The item of interest for the current study regarding deliberate self-harm (DSH) was “Have you ever deliberately taken an overdose (e.g. of pills or other medication) or tried to harm yourself in some other way (such as cut yourself)?”. Response options were “No”, “Yes, once”, and “Yes, more than once”. Students who indicated a history of at least one DSH episode completed a series of questions about their most recent DSH episode. These included an open-ended description of the DSH, when the last episode took place, motives for the actions, help-seeking behaviour, and hospitalisation after the act.

Procedure Ethics approval was gained from the Griffith University Human Ethics Committee and the Gold Coast Health District Ethics Committee before data collection began. Participants completed the questionnaire in group settings between 1 and 30 September 2002. Anonymity and confidentiality were assured by test administrators. The survey took about 30 minutes to complete, and researchers were available to discuss any questions at all times during the survey administration. Participants were advised of available clinical services, and a clinical psychologist was present at all testing sessions. Immediately following the survey administration, students were involved in focus-group discussions regarding the content and administration of the questionnaire.

Data treatment For students who indicated a history of DSH, a member of the research team (T S H) evaluated the open-ended description of the act to determine if the incident matched the CASE definition. Responses were coded as “Definitely DSH”, “Not DSH”, “Undecided”, or “No DSH information given”. An additional researcher independently examined cases that were deemed “Undecided”, and a consensus decision was made. Ingestion of any amount of illicit drugs (in answer to general lifestyle questions) was considered to be in excess. However, this was only considered “Definitely DSH” when the respondent indicated that he or she had used illicit drugs specifically in response to questions on their DSH episode.

Statistical techniques used included descriptive statistics, cross-tabulations, ?2 test of independence, and multiple logistic regression. For all analyses, significance was taken as P < 0.01, and 95% confidence intervals were calculated where necessary. Backward-selection multiple logistic regression was used to identify a list of associated factors (including recent life events, psychopathology, coping strategies, and lifestyle choices) at the P < 0.01 level.11 We analysed the data with SPSS.12

Factors that were adjusted for in the multiple logistic regression are listed in Box 2.

Results Sample On the day that their schools completed the questionnaire in September 2002, 3767 students were in attendance. Ten students did not participate because their parents refused consent. The 8% absenteeism rate was normal, so there appeared to be no effect on attendance levels because of the study taking place. Overall, the response rate was 91.7% (3757/4097).

Participants did not differ significantly from adolescents across Queensland or Australia in terms of sex distribution, proportion of population, or enrolments at types of educational institution, except that the Gold Coast had fewer Catholic school enrolments, and more independent school enrolments (Box 3).13

Given a conservative postulated prevalence of 4% for DSH (based on previous findings14), with a 95% confidence interval of 3.4% to 4.6%, the final sample approximated the expected sample of 4000.

Prevalence of DSH The average age of respondents was 15.4 years (95% CI, 15.37–15.42 years). A lifetime history of DSH was reported by 464 of 3757 students (12.4%). DSH in the previous 12 months was reported by 317 students (8.4%), of whom 233 (6.2% of total) described an act that satisfied our criteria. Forty-one per cent of males (23 of 56) who reported self-harm in the previous year did not include a description of their episode, and could not be matched against the study criteria, compared with 16% (41 of 259) of females.

Among those who described DSH episodes fitting our criteria, females (200 of 1800; 11.1%) were more likely than males (33 of 1943; 1.6%) to have self-harmed in the previous year (OR, 7.5; 95% CI, 5.1–10.9). For 24 respondents (10.3%), DSH had resulted in hospital presentation.

The main methods used for the 233 DSH episodes were cutting (138; 59.2%), overdose of medication (69; 29.6%), illicit drugs (7; 3.0%), self-battery (5; 2.2%), hanging (4; 1.7%), and sniffing/inhalation (4; 1.7%). For statistical purposes, all methods other than self-cutting and overdose were aggregated to form one category of “other” methods. There was no association between the methods of DSH and frequency of hospital presentation (?2 = 0.93, df = 2, P = 0.629).

Factors associated with DSH Insufficient numbers of males meeting the DSH criteria prevented a direct comparison of potential interactions between sex and other variables. Therefore, multivariate analyses were conducted for females alone, and for total respondents, to determine the contribution of adding males to the equation.

In multiple logistic regression, factors significantly associated with increased deliberate self-harm (P < 0.01) in the previous year in females were exposure to self-harm in friends or family members, sexual orientation worries, smoking (fewer than 5 cigarettes per week), low self-esteem, and “other” distressing events (Box 4). For the total sample, exposure to self-harm in friends or family members, smoking (fewer than 5 cigarettes per week), boyfriend/girlfriend problems, amphetamine use, self-prescribing medication, coping by blaming self for getting in distressing situations, and “other” distressing events were significantly associated with DSH (Box 4). Living with one parent was associated with lower rates of DSH for females and for the total sample.

Time to engage in DSH More than a third of respondents (36.1%; 84 of 233) engaged in self-harm less than an hour after first thinking of doing so, and an additional 30 respondents (12.9%) thought about harming themselves for between 1 and 24 hours before taking action. Conversely, 20.6% (48 of 233) had thought about harming themselves for more than 1 month before acting. The method of DSH did not vary according to the length of time that self-harmers thought about their actions (P = 0.402).

Help-seeking behaviours Help-seeking behaviours of self-harmers were very similar before and after the most recent instance of DSH, with friends the preferred source of help (Box 5). There was no sex difference among respondents who sought help before DSH (89 females, 15 males; P = 0.802); however, there was a non-significant trend for females to seek help from telephone helplines (8/89 [9.0%] females, 0 males; P = 0.227) and teachers (6/89 [6.7%] females, 0 males; P = 0.300). Friends (61.4%) and mother (18.5%) were most likely to know that the respondent had self-harmed, and few GPs (2.6%) or mental health workers (7.3%) were made aware of these actions.

Discussion in this study, prevalence rates (6.2% of total sample; 11.1% of females) were consistent with those of a recent English study using the same criteria for DSH.6 As only 10.3% of the self-harmers in our study presented for hospital treatment, it appears that investigations of DSH based on monitoring studies of adolescents severely under-report the extent of the problem. Self-cutting and overdose were the most common methods of self-harm, consistent with previous community-based investigations.15,16

The factors that had the strongest association with DSH were exposure to self-harm in family and in friends. The Gold Coast Health District Ethics Committee did not permit the coding of individual schools, because schools felt their anonymity may be compromised. Therefore, it was not possible to determine if there was a school-based clustering of DSH. “Copycat” DSH has been shown to increase suicide17 and self-harm18,19 in adolescent populations; however, despite the lack of specific assessment in the current study, it is reasonable to assume that this was not the case because of the similarly weighted prevalence of DSH by family members, which would not have the same type of contagion effects.

Most students who had self-harmed did not seek help for the problems that preceded their act. However, those who did seek help (before and after the act) consulted their friends and family in preference to medical and mental health services. Telephone counsellors were used by a very small proportion of self-harmers, despite the attention and funding given to this service as a component of Australia’s National Youth Suicide Prevention Strategy since 1995.20 Given the lack of evidence of any effect on suicide rates for these services,21 perhaps a comprehensive evaluation of their accessibility and efficacy should be performed. Primary prevention of suicidal behaviours among adolescents should include educating young people in the use of professional services to deal with their problems.

Although absenteeism was not greater than normal at the schools when the study took place, those absent may be at increased risk of self-harming behaviour. Truancy is reported by more adolescent female self-harmers than non-self-harmers.22 Furthermore, those absent from school have elevated levels of psychopathology23 and engage in more frequent high-risk behaviours.24 Therefore, our results may underestimate the problems of DSH among adolescents.

The results of this study can only be generalised to coeducational school students. Future research should include students from all school types.

The definition of DSH did not differentiate between self-cutting and habitual self-mutilation. Self-mutilation is not considered a suicidal act,25 but respondents’ reporting of this behaviour could contribute to an over-representation of DSH, as self-cutting was the most common method of self-harm reported.

With many factors potentially related (eg, problems in making friends and depression), there is a risk of collinearity. However, multiple variables were retained because associated factors are invariably interrelated in studies of suicidality.

A further limitation is the low number of males who met the criteria for DSH, preventing multivariate analyses by sex. In addition, a large proportion of males who reported DSH did not describe their episode, and so had to be excluded from analyses because the CASE criteria could not be applied. The reluctance of male participants to complete this item should be addressed for future studies.

Conclusion A stringent definition of DSH can improve research into such behaviours and enable comparisons across different settings and countries. Future studies should pay attention to better identifying motivation or intention behind such behaviour for the most accurate representation of “true” DSH. Risk factors for DSH are many and varied and prevention programs should address these simultaneously, as well as educating young people on the use of mental health services.

  1. The definition of deliberate self-harm used by the Child and Adolescent Self-harm in Europe group6

An act with a non-fatal outcome in which an individual deliberately did one or more of the following:

Initiated behaviour (for example, self cutting, jumping from a height), which they intended to cause self-harm;

Ingested a substance in excess of the prescribed or generally recognised therapeutic dose;

Ingested a recreational or illicit drug that was an act that the person regarded as self-harm;

Ingested a non-ingestible substance or object.

  1. Factors adjusted for in the multiple logistic regression

Living situation

Healthy food consumption

Regularity of exercise

Use of alcohol, cannabis, “ecstasy”, heroin, amphetamines, other drugs, in the past year

Problems keeping up with schoolwork

Difficulty making friends

Arguments with friends

Serious boyfriend/girlfriend problems

Been bullied at school

History of physical or sexual abuse

Trouble with police

Parents separated/divorced

Serious arguments with parents

Fights between parents

Self, family member or close friend suffered serious illness

Death of immediate family member or close others

Suicide of family member or friend

DSH by family member or friend

Worries about sexual orientation

Other distressing events not specified

Psychological factors (impulsivity, self-esteem, depression, anxiety)

Coping style (talking to someone, blaming self, getting angry, staying in room, thinking how similar situations were dealt with, drinking alcohol, not thinking about what is worrying them, trying to sort things out)

Self-prescribing medication

  1. Characteristics of Gold Coast population (14–18-year-olds) compared with Queensland and Australia

Gold Coast Queensland Australia


Proportion of total population n = 396 588 n = 3 655 139 n = 18 972 350

Male 6.30% 7.29% 7.26%

Female 5.66% 6.80% 6.74%

Total 5.97% 7.04% 7.00%

Sex distribution among 14–18-year-olds n = 23 679 n = 257 438 n = 1 327 253

Male 51.31% 51.18% 51.21%

Female 48.69% 48.82% 48.79%

Proportion of students at types of secondary schools

Government 59.89% 62.94% 62.71%

Catholic 11.03% 18.86% 21.37%

Other 29.08% 18.20% 15.92%

4 Factors associated with deliberate self-harm on multivariate logistic regression

Females (200/1800)

Total sample (233/3757)

OR (95% CI) P

OR (95% CI) P


Self-harm by friends

No 1


Yes 2.68 (1.69–4.26) < 0.001 4.07 (2.64–6.26) < 0.001

Self-harm by family

No 1


Yes 3.25 (2.11–5.01) < 0.001 3.22 (2.17–4.78) < 0.001


High 1

Moderate 0.95 (0.45–1.99) 0.883

Low 2.58 (1.25–5.31) 0.010

Sexual orientation worries

No 1

Yes 2.22 (1.22–4.03) 0.009

Boyfriend/girlfriend problems




1.74 (1.16–2.61) 0.008

History of amphetamine use




2.47 (1.32–4.65) 0.005

Living situation

Both parents 1


One parent 0.44 (0.25–0.75) 0.003 0.47 (0.28–0.77) 0.003

One parent + step-parent 0.74 (0.42–1.29) 0.244 0.85 (0.52–1.39) 0.514

Other family member/s 1.34 (0.38–4.75) 0.655 1.23 (0.40–3.74) 0.718

Other 0.33 (0.10–1.20) 0.092 0.28 (0.09–0.84) 0.023

Cigarettes (per week)

Never 1


Given up 1.52 (0.88–2.64) 0.134 1.48 (0.90–2.46) 0.125

Less than 5 3.35 (1.56–7.19) 0.002 3.13 (1.62–6.06) < 0.001

6–20 0.99 (0.46–2.11) 0.977 1.06 (0.53–2.11) 0.876

21–50 1.74 (0.79–3.85) 0.168 1.95 (0.98–3.87) 0.058

More than 50 1.16 (0.35–3.86) 0.811 0.89 (0.34–2.35) 0.810

Blame self (coping)




1.33 (0.68–2.63) 0.404


2.98 (1.51–5.87) 0.002

Self-prescribing of medication




1.67 (0.83–2.24) 0.218


1.73 (1.07–2.80) 0.025


2.41 (1.29–4.51) 0.006

Other distressing events

No 1


Yes 2.45 (1.62–3.70) < 0.001 2.56 (1.73–3.79) < 0.001


— These factors were not significant in the particular group.

5 Sources of help before and after deliberate self-harm

Source Before (n = 105)* After (n = 99)†


Friend 85 (81.0%) 80 (80.8%)

Family member 14 (13.3%) 23 (23.2%)

Psychologist/ psychiatrist 10 (9.5%) 6 (6.1%)

Telephone helpline 8 (7.6%) 1 (1.0%)

Teacher 6 (5.7%) 5 (5.1%)

Social worker 6 (5.7%) 6 (6.1%)

General practitioner 2 (1.9%) 4 (4.0%)

Drop-in centre 2 (1.9%) 1 (1.0%)

Other 17 (16.2%) 7 (7.1%)


* 28 respondents (26.7%) sought help from more than one source. † 25 respondents (25.3%) sought help from more than one source.

The questionnaire was developed and the study conducted in collaboration with the Child and Adolescent Self-harm in Europe (CASE) Study, an international investigation of self-harm among young people. The Australian participation in the study was financially supported by the Australian Government Department of Health and Ageing, and Queensland Health, Mental Health Branch. We wish to express our gratitude to Ms Kate Swanton, Ms Kathy Harper, Dr Gayle Pollard, and Mr David Firman, all from Queensland Health. Thanks also to Ms Kym Spathonis, Ms Jacinta Hawgood, and Ms Erminia Colucci, all from AISRAP, for help in data gathering, and Dr Stefano Occhipinti for statistical guidance. Funding bodies were not involved in the study design, collection of data, or in the publication of relevant materials.

Competing interests None identified.

Diekstra RW, Gulbinat W. The epidemiology of suicidal behaviour: a review of three continents. World Health Stat Q 1993; 46: 52-68.
Australian Bureau of Statistics. Deaths from external causes. Canberra: ABS, 2004. (ABS Catalogue no. 3320.0.)
Beautrais AL. Gender issues in youth suicidal behaviours. Emerg Med (Fremantle) 2002; 14: 35-42. <PubMed>
Goldney RG, Fisher LJ, Wilson DH, Cheok F. Suicidal ideation and health-related quality of life in the community. Med J Aust 2001; 175: 546-549. <PubMed>
Schweitzer R, Klayich M, McLean J. Suicidal ideation and behaviours among university students in Australia. Aust N Z J Psychiatry 1995; 29: 473-479. <PubMed>
Hawton K, Rodham K, Evans E, Weatherall R. Deliberate self-harm in adolescents: self report survey in schools in England. BMJ 2002; 325: 1207-1211. <PubMed>
Gold Coast City Council. Facts and Figures. Available at: (accessed Apr 2003).
Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983; 67: 361-370. <PubMed>
Plutchik R, van Praag HM, Picard S, et al. Is there a relation between seriousness of suicidal intent and the lethality of the suicide intent? Psychiatry Res 1989; 27: 71-79. <PubMed>
Robson P. Development of a new self-report questionnaire to measure self esteem. Psychol Med 1989; 19: 513-518. <PubMed>
Kleinbaum DG, Klein M. Logistic regression: a self-learning text. 2nd ed. New York: Springer, 2002.
SPSS [computer program]. Version 11.0. Chicago, Ill: SPSS, 2002.
Australian Bureau of Statistics. 2001 Census of population and housing. Canberra: ABS, 2003.
Bjarnason T, Thorlindsson T. Manifest predictors of past suicide attempts in a population of Icelandic adolescents. Suicide Life Threat Behav 1994; 24: 350-358. <PubMed>
Patton GC, Harris R, Carlin JB, et al. Adolescent suicidal behaviours: a population-based study of risk. Psychol Med 1997; 27: 715-724. <PubMed>
Schweitzer R, McLean J, Cvetkovich L, Rose D. Methods and consequences of suicide attempts among Australian students. Aust N Z J Psychiatry 2000; 34: 170-172. <PubMed>
Poijula S, Wahlberg KE, Dyregrov A. Adolescent suicide and suicide contagion in three secondary schools. Int J Emerg Ment Health 2001; 3: 163-168. <PubMed>
Gould MS, Petrie K, Kleinman MH, Wallenstein S. Clustering of attempted suicide: New Zealand national data. Int J Epidemiol 1994; 23: 1185-1189. <PubMed>
Taiminen TJ, Kellio-Soukainen K, Nokso-Koivisto H, et al. Contagion of deliberate self-harm among adolescent inpatients. J Am Acad Child Adolesc Psychiatry 1998; 37: 211-217. <PubMed>
Australian Department of Health and Ageing. National Youth Suicide Prevention Strategy [website]. Available at: (accessed Jul 2004). Burns JM, Patton GC. Preventative interventions for youth suicide: a risk-factor approach. Aust N Z J Psychiatry 2000; 34: 388-407. <PubMed>
Ribakoviene V, Puras D. Relationships between social factors and suicidal attempts of adolescent girls. Medicina (Kaunas) 2002; 38: 379-386. <PubMed>
Berg I. Absence from school and mental health. Br J Psychiatry 1992; 161: 154-166. <PubMed>
Handelsman DJ, Gupta L. Prevalence and risk factors for anabolic-androgenic steroid abuse in Australian high school students. Int J Androl 1997; 20: 159-164. <PubMed>
Favazza AR. The coming of age of self-mutilation. J Nerv Ment Dis 1998; 186: 259-268. <PubMed>
(Received 23 Sep 2003, accepted 27 Apr 2004)
Australian Institute for Suicide Research and Prevention, Griffith University, Mt Gravatt, QLD.

Diego De Leo, MD, PhD, FRANZCP, Director; Travis S Heller, BSc(Hons), BA, Researcher. Reprints: Professor D De Leo, Australian Institute for Suicide Research and Prevention, Griffith University, Messines Ridge Road, Mt Gravatt, QLD 4122.

©The Medical Journal of Australia 2004 ISSN: 0025-729X