paper and outline

Journal of Psychosomatic Research 66 (2009) 477–493

Review article

Risk factors and correlates of deliberate self-harm behavior: A systematic review

Herbert Fliege⁎, Jeong-Ran Lee, Anne Grimm, Burghard F. Klapp

Department of Psychosomatic Medicine and Psychotherapy, Charité Universitätsmedizin Berlin, Germany

Received 9 May 2008; received in revised form 20 October 2008; accepted 21 October 2008

Abstract

Objective: Deliberate self-harm behavior—without suicidal intent—is a serious health problem and may be studied as a clinical phenomenon in its own right. Empirical studies of sociodemographic and psychological correlates and risk factors are systematically reviewed. Methods: We searched Medline, PsycINFO, PSYNDEX (German psychological literature), and reference lists. We targeted self-induced bodily harm without conscious suicidal intent. Studies on suicidal behavior or self- poisoning were only included if they also assessed nonsuicidal self- harm. Results: Fifty-nine original studies met the criteria. Deliberate self-harm may occur at all ages, yet adolescents and young adults are at a higher risk. Evidence on gender is complex. Only 5 studies realize a prospective design (6 months to 10 years) and test predictors. The majority use cross-sectional and retro- spective methods. No longitudinal study (separately) examines

⁎ Corresponding author. Charité Universitätsmedizin Berlin, Depart- ment of Psychosomatic Medicine and Psychotherapy, Charitéplatz 1, D- 10117 Berlin, Germany. Tel.: +49 30 450 553097; fax: +49 30 450 553989.

E-mail address: [email protected] (H. Fliege).

0022-3999/08/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.jpsychores.2008.10.013

new incidence. Evidence of correlates encompasses distal/prox- imal, person/environment, and state/trait factors. Many studies report associations between current self-harm behavior and a history of childhood sexual abuse. Adolescent and adult self- harmers experience more frequent and more negative emotions, such as anxiety, depression, and aggressiveness, than persons who do not self-harm. Two studies yield specific interactions between childhood trauma and current traits and states such as low emotional expressivity, low self-esteem, and dissociation with respect to a vulnerability to self-harm. Conclusion: Evidence of distal, biographical stressors is fairly strong. Proximal stressors have rarely been investigated; protective factors, hardly at all. Despite many findings of correlates, the data do not yet justify terming them risk factors. Longitudinal studies are needed. © 2009 Elsevier Inc. All rights reserved.

Keywords: Deliberate self-harm; Self-injurious behavior; Risk factors; Correlates; Systematic literature review

Introduction

Deliberate self-harm behavior is a significant health problem that is increasingly being studied as a clinical phenomenon in its own right [1–3]. It is detrimental to the body and may impede social relations, medical treatment, and psychotherapy [4–9]. Some reports characterize aggres- sive acts against one's own body as indicative of especially severe psychopathological problems [6,10].

Definition of self-harm

The most accepted term for auto-destructive acts in the literature is self-harm [3] or, more specifically, deliberate self-harm behavior [11]. It is defined as the intentional self-induced harming of one's own body resulting in relevant tissue damage [5,6,11,12]. The term encompasses self-injurious behaviors and more indirect forms of bodily harm [13]. It is largely agreed upon to exclude the following from the definition: (a) phenomena that are explicit symptoms or classificatory criteria of other disorders, such as eating disorders or substance abuse; (b) everyday behaviors, such as unhealthy eating habits or lack of exercise; and (c) psychological self-harm, such as deliberately engaging in an abasing partnership. The latter is occasionally studied in the context of borderline personality disorder [14]. However, there is little consensus

478 H. Fliege et al. / Journal of Psychosomatic Research 66 (2009) 477–493

as to how to classify different levels of frequency, severity (e.g., delicate self-cutting vs. auto-mutilation), or specific forms of deliberate self-harm, such as self-poisoning. Another complex issue is how to distinguish between deliberate self-harm and suicidal behavior.

Prominent definitions of deliberate self-harm behavior exclude suicidal intention [5,12,15]. This understanding prevails in U.S. publications, whereas the term self-harm generally includes behaviors irrespective of a suicidal intention in the UK, thus encompassing self-harm behaviors with and without suicidal intention. This discrepancy limits cross-national study comparability. There is some comorbid- ity between suicidal and nonsuicidal self-harm, and deliber- ate self-harm is prognostic for suicide attempts [14,16–21]. However, there appear to be some important differences between self-harmwith the intent to die and self-harmwith no intent to die [22]. Studies have revealed a distinction between the psychological functions of suicidal and nonsuicidal self- harm behavior [13]. This distinction is based on the criterion of intention, which is difficult to operationalize. Suicidal intentions may be ambivalent, dissimulated, or concealed. Thus, intention is more difficult to measure reliably than observable behavior [16]. Skegg [3] proposes assessing self- harm behavior descriptively and many authors have adopted the approach of assessing deliberate self-harm as an act of intentional self-injury or self-poisoning “irrespective of the apparent purpose of the act” [23]. However, if we want to expand our knowledge on differences between suicidal and nonsuicidal self-harm, further efforts to develop reliable assessments of intent are required. Some advances have recently been made in this area [24].

Diagnostic classification

Deliberate self-harm behavior may occur in clinical as well as in nonclinical samples [9,10,25–27]. DSM-IV and ICD-10 F list self-harm behavior as a diagnostic criterion of borderline personality disorder. However, individuals who self-harm are diagnostically heterogeneous and may suffer from a spectrum of other psychological disorders [28,29]. Frequently reported co-occurring diagnoses—other than borderline personality disorder—are alcohol and substance abuse [29,30]; eating disorders [31–34]; dissociative, somatoform, or body dysmorphic disorders [35–37]; depres- sion and anxiety disorders [26,38,39]; posttraumatic stress disorder [30,40]; and several personality disorders and schizophrenia [2,29]. In DSM-IV, self-harm behavior may also be classified as a disorder of impulse control not otherwise specified and ICD-10 has an additional Z code for a personal history of self-harm. Concealed self-harm carried out with the aim of adopting the sick role is classified as factitious disorder [13]. Within factitious disorders, indirect forms of self-harm prevail [41]. Studies indicate certain overlaps between factitious disorder and open (or admitted) self-harm behaviors, making the diagnostic categories not absolutely distinct empirically [6,42–44].

Epidemiology

In school and college student samples, lifetime prevalence of nonsuicidal self-harm behavior ranges from 13% to 35% [11,45–48]. Several population-based samples from the United States and Canada yield a prevalence of 4% pertaining to the past 6 months and lifetime prevalence figures ranging from 2.2% to 6% [49–51]. Deliberate self-harm behavior is particularly prevalent in patients in psychosomatic medicine and psychiatric settings, including consultation–liaison patients [29,52–54].

Risk factors and protective factors

Generally, the relation between a factor and an illness is probabilistic, not deterministic. Therefore, the term risk factors is more often preferred over the term causal factors. Kraemer et al. [55] define risk factor as a measurable variable that must precede the outcome and be associated with a higher risk of developing the outcome. They distinguish between risk factors of first onset and those of a relapse and—inversely—factors raising the probability of a remission. Knowledge of these factors is relevant for the aims of primary, secondary, and tertiary prevention and for therapy. According to Kraemer et al. [55], a factor that has a correlation with an outcome, with both variables being assessed at the same time, should merely be termed a cor- relate. Kraemer argues that although a given factor may indeed be a risk factor, it might also be an accompanying condition or a consequence of the illness. To term a variable risk factor, it must be assessed before the outcome occurs.

Distal and proximal factors are distinguished depending on the proximity between factors and outcome. This is often operationalized along a time axis [56], with distal factors arising earlier in the biography and proximal factors nearer to illness onset. The terms distal and proximal are merely descriptive. Another aspect concerns the paths of effect. In a cause-and-effect chain, distal factors are supposed to have more indirect effects and proximal factors more direct effects on the outcome [57]. More specifically, diathesis-stress models define distal as predisposing factors, which are those that make the individual vulnerable to stress, such as parental problems or childhood traumata. Thus, in predisposed individuals, proximal factors may trigger the onset of mental or behavioral problems. Proximal factors are composed of daily hassles, conflicts, acute stress, life events, or traumata closely prior to illness onset.

Transactional models emphasize that risk factors can transact with each other as well as with the behavior itself [58]. While some risk factors are largely considered static, like a history of childhood trauma or personality traits, others are more often considered dynamic, like changing mood states or the current course of events. More recent approaches underscore the variability of risk factors over time, rather than assume that the risk factors remain valid indefinitely [58].

479H. Fliege et al. / Journal of Psychosomatic Research 66 (2009) 477–493

Risk factors may be inherent in the person or in the environment. Accordingly, protective factors can be personal resources (resilience) or social resources [59].

Protective factors According to the salutogenic approach, when a person

faces stress, protective factors diminish the risk of deve; loping a disease [60]. Identifying protective factors is relevant, provided they can be modified and enhanced by interventions.

Interaction between risk factors and protective factors Well-known models of the interaction between risk and

protective factors are the compensation model, which claims that a protective factor simply has an additive positive effect on health and thereby compensates the negative effect of risk factors [59], and the protective factors model, which claims that a protective factor must moderate the association between risk factor and outcome [61,62].

Literature review

Although there are some excellent reviews on deliberate self-harm behavior [1–3], we believe that we have added to this knowledge in several ways. In contrast to earlier reviews that give a comprehensive overview [3] or focus on the functions of self-harm [1,2], our review elucidates on the possible factors in the etiology of deliberate self- harm that have been empirically investigated. Since Gratz's 2003 review on risk factors and functions of deliberate self-harm [1], a considerable number of self- harm studies have been published, thus making an updated review expedient. Moreover, other than Gratz's 2003 review, which focused on childhood traumatic experiences and emotion regulation as possible factors [1], our review focuses on a broader range of empirically investigated sociodemographic and psychological factors. Additionally, we evaluate each risk factor or correlate in terms of the methodological strength of the available evidence and infer open research issues that need to be addressed.

Our review focuses on sociodemographic and psycholo- gical factors. Although several studies also yield associations between deliberate self-harm and alcohol or other substance abuse, binge eating, or restricted eating, we will not include this information. One reason for this is that substance abuse and disturbed eating may indicate co-occurring disorders. As a result, the question of correlates or risk factors would be strongly linked with the question of diagnostic comorbidity proper. A second reason is that substance abuse and disturbed eating are in themselves behavior with a strong inherent element of self-harm. This makes it a much more complex issue than just treating substance abuse and disturbed eating as correlates or risk factors. In our view, the relation between deliberate self-harm with alcohol or substance abuse and disturbed eating behaviors, in particular, deserves investigat- ing in more detail than this review would allow.

Methods

Research strategies

We searched Medline, PsycINFO, and PSYNDEX (Ger- man psychological literature) for the keywords [“self-harm” OR “self-injur⁎” OR “self-cut⁎” OR “self-destruct⁎” OR “auto-mutilat⁎” OR “auto-destruct⁎”] combined with [“risk facto⁎” OR “correlate⁎” OR “precurs⁎” OR “predict⁎” OR “causal facto⁎”]. We also searched reference lists. We included studies that adopt the definition of deliberate self- harm as self-induced bodily harm without conscious suicidal intent [1]. Studies on suicidal behavior or self-poisoning alone were generally excluded. However, in order not to omit relevant evidence, we included a small number of studies on suicidal behavior, provided that they also assessed non- suicidal self-harm. We considered original papers that report at least one correlate, risk factor, or protective factor. Case studies and studies without any group comparisons were not included.

Up to the end of 2007, a total of 59 original papers that met inclusion criteria emerged (Table 1). Table 1 gives an overview of included studies, sample characteristics, and assessment methods of deliberate self-harm. Assessment methods are described briefly and are then assigned to one of the following eight categories: (a) there is no information on how deliberate self-harm was assessed; (b) deliberate self- harm was established by an (unspecified) diagnostic case ascertainment at hospital admission or by inspecting medical records, chart reviews, or other; (c) deliberate self-harm was assessed by a standardized clinician rating based on clinical interview and medical documentation; (d) deliberate self- harm was established on the basis of interviews that were constructed ad hoc for the study; (e) deliberate self-harm was established on the basis of interviews for which validation data are available; (f) deliberate self-harm was assessed by single self-report (or interview) items; (g) deliberate self- harm was assessed by a self-report questionnaire constructed ad hoc for the study; (h) deliberate self-harm was assessed by a self-report questionnaire for which psychometrical valida- tion data are available.

Taxonomy of correlates and risk factors

We included sociodemographic factors, distal and proximal correlates, and risk factors. We distinguished correlates and risk factors from protective factors. A factor such as “self-esteem” may be reported as a protective factor or—inversely, as low self-esteem—as a risk factor. We presented factors according to the way they were treated in the studies.

Empirical evidence is summarized and presented in Table 2. For each factor, we listed all relevant studies by authors and indicated whether the study design was cross- sectional (CS), including retrospective assessment, or long- itudinal (LT). We rated the strength of the evidence for each

Table 1 Samples and self-harm assessment of the 59 studies

Study Sample Delibe e self-harm assessment

Authors (year of publication) Reference Kind/Setting N n/n Age Assess nt instruments Category

Psychometric validation data

Akyuz et al. (2005) [63] Population-based random sample of the city of Sivas, Central Turkey; females

628 14 with DSH 18–65 (34.5±11.5) Ad ho elf-report questio aire (frequency, duratio types)

7 None reported 614 without DSH

Boudewyn and Liem (1995a, 1995b)

[64,65] College psychology students

438 46 with DSH 392 without DSH

16–65 (mean=24.9) Questi about acts of deliber self-harm, not describ

(6) None reported

Briere and Gil (1998) [49] Population-based sample, clinical sample (inpatients, outpatients), and self-harmers via advertisement

1410 927 population based 18–90 (46±17) Single m from Trauma Sympt Inventory [66] [“Inten nally hurting yourse e.g., by scratch , cutting, or burnin even though you w n't trying to commi uicide”]

6 None reported 390 clinical 18–58 (36±10) 93 self-harmers ./. (35±9)

Brown et al. (2007) [67] College psychology students

223 23 with recent DSH 19.4±2.1 Standa zed self-report questio aire: DSHI [11] (type, quency, severit uration)

8 Available 39 with past DSH 161 without DSH

Brunner et al. (2007) [45] Students, 9th grade 5759 630 with occasional DSH

14.9±0.73 Single m from Schedu for Affective Disord

6 None reported

229 with repetitive DSH 4900 without DSH

Carroll et al. (1980) [68] Psychiatric patients, subgroup of patients with DSH and age-/ sex-matched sample without DSH

28 14 patients with DSH 21–41 (mean=28.1) Intervi , not specified (4) None reported 14 controls

Colman et al. (2004) [69] Patients presenting to 1 of 5 emergency departments of Edmonton, Alberta, Canada, for “parasuicide”

369 92 with recurrence within 1–2 years

16–55 Case a rtainment at hospita dmission; risk factors European Parasu de Study Intervi Schedule, Suicid tent Scale

2 None reported for the diagnostic code; available for interview and intent scale

8

Crowell et al. (2005) [70] “Parasuicidal” adolescents and age- matched controls; females

46 23 with DSH 14–18 (15.3±1.1) Semist tured/ structu interview; Lifetim Parasuicide Count erview [71] se /parent report, DSH g up only

(5) None reported 23 without DSH

Darche (1990) [72] Adolescent psychiatric inpatients and age-/sex- matched controls

96 48 patients 13–17 (mean=15.4) Review f medical record

2 None reported 48 controls

480 H . F liege

et al.

/ Journal

of P sychosom

atic R esearch

66 (2009)

477–493

rat

me

c s nn n,

ons ate ed ite om tio lf ( ing g) ere t s rdi nn fre y, d ite le ers

ew

sce l a by ici ew e In ruc red e Int lf- ro o

s

De Leo and Heller (2004)

[73] Students from 14 high schools on the Gold Coast, Queensland, Australia

3757 233 with DSH 10–11 Single-item from the Lifestyle and Coping Questionnaire [“Have you ever deliberately taken an overdose (…) or tried to harm yourself in some other way (such as cut yourself)?”] and further questions on most recent episode

6 None reported 3524 without DSH

Demitrack et al. (1990)

[74] Inpatients of an eating disorder unit and age-matched controls; females

60 30 patients 16–39 (23.3±4.4) Initial clinical interview at hospital admission

2 None reported 30 controls

Dubo et al. (1997) [75] Psychiatric inpatient with borderline or other personality disorder

59 33 with DSH 18–60 (mean=31.7) Ad hoc self-report questionnaire (duration, number of episodes, severity)

7 None reported 26 without DSH

Dulit et al. (1994) [76] Psychiatric inpatients with borderline personality disorder, with or without DSH

124 62 without DSH 32.8±9.2 Ad hoc self-report questionnaire (frequency); Suicide Intent Scale [77]

7 None reported for DSH items; available for intent scale

23 with infrequent DSH 8 39 with frequent DSH

Evren and Evren (2005)

[78] Substance dependents after inpatient detoxification; males

136 47 with DSH Mean=36.4 Single-item from the Childhood Abuse and Neglect Questionnaire

6 None reported 89 without DSH

Fliege et al. (2003) [79] Psychosomatic inpatients, consultation– liaison patients and outpatients

354 32 with DSH 18–65 (30.1±10.1) Standardized clinician- administered rating form, based on clinical interview and medical records

3 Available 322 without DSH

Fliege et al. (2004) [80] Psychosomatic medicine consultation–liaison patients and outpatients, general hospital

854 63 with DSH 17–65 (32.5±11.6) Standardized clinician- administered rating form, based on clinical interview and medical records

3 Available 63 matched controls out of 791 without DSH

Fliege et al. (2006) [52] Psychosomatic inpatients 361 109 with DSH 17–77 (41.9±14.9) Standardized self-report questionnaires: DSHI [11]; Self-Harm Behavior Questionnaire DSHQ [46]

8 Available 252 without DSH

Gladstone et al. (1999)

[81] Inpatients and outpatients with depressive disorder

269 26 with current DSH 18–77 Ad hoc semistructured/ structured interview containing questions on self-harm

4 None reported 29 with history of DSH

Gladstone et al. (2004)

[82] Outpatients with depressive disorder; females

125 33 with current DSH 17–68 (37.8±12.1) Ad hoc semistructured/ structured interview containing unspecified questions of self-harm

4 None reported 34 with history of DSH

(continued on next page) (continued on next page)

481 H . F liege

et al.

/ Journal

of P sychosom

atic R esearch

66 (2009)

477–493

Table 1 (continued)

Study Sample Deliberate self-harm assessment

Authors (year of publication) Reference Kind/Setting N n/n Age Assessment instruments Category

Psychometric validation data

Gratz (2001) [11] Undergraduate students of psychology

150 53 with DSH 18–64 (23.2±7.1) Standardized self-report questionnaire: DSHI [11]

8 Available 97 without DSH

Gratz (2006) [83] Undergraduate students of psychology; females

249 91 with DSH 18–55 (23.3±6.0) Standardized self-report questionnaire: DSHI [11]

8 Available 158 without DSH

Gratz et al. (2002) [84] Undergraduate students of psychology

133 51 with DSH 18–49 (22.7±6.2) Standardized self-report questionnaire: DSHI [11]

8 Available 82 without DSH

Haavisto et al. (2005) [85] Representative birth cohort sample at age 8 and again at age 18, males

2307 51 with DSH (at age 18) 8/18 Ad hoc single item (“I deliberately tried to hurt or kill myself”)

6 None reported 2256 without DSH

Haines and Williams (1997)

[86] Prison inmates and students of psychology; males

50 19 prisoners with DSH (22.3±6.17) Not reported 1 None reported 13 prisoners and 18 students without DSH

Hawton et al. (2002) [87] Pupils from England, survey

6020 398 with DSH 15–16 Ad hoc self-report questionnaire (type, consequences, functions, intent); a posteriori classification of open response categories

7 None reported 5622 without DSH

Herpertz et al. (1997) [88] Inpatients with any personality disorder and nonclinical controls

165 54 patients with DSH Mean=24 Ad hoc structured interview questions, in addition to the SCID Interview for DSM-III-R

4 None reported 66 patients without DSH Mean=24.3 45 nonclinical controls

Howat and Davidson (2002)

[89] Elderly psychiatric patients treated for “parasuicidal” behavior or depression, and controls

77 34 parasuicidal patients N65 Ad hoc structured interview questions on self-harm, method and intent

4 None reported 21 depressive patients 22 controls

Johnston et al. (2006) [90] Patients from an emergency department and psychiatric inpatients, treated for self-injury

4743 516 with history of recurrent DSH

N16 Medical records on occurrence (Manchester self-harm monitoring system MASH)

2 None reported

Joyce et al. (2006) [91] Psychiatric patients in a long-term treatment program for depression

195 46 with DSH N18 Ad hoc structured interview questions, in addition to the SCID Interview for DSM-IV

4 None reported 149 without DSH

Kapur et al. (2006) [23] Patients presenting to an emergency department with episodes of self-harm including self-injury

7723 1054 with recurrence within 12 months

15–98 (median=30)

Medical records on occurrence (Manchester and Salford self-harm project MASSH)

2 None reported

Kirkcaldy et al. (2006) [92] Patients from a hospital for child and adolescent psychiatry and psychosomatic medicine

3649 2153 with DSH 3–24 (13.4±3.3) Ad hoc self-report questionnaire

7 None reported 1496 without DSH

482 H . F liege

et al.

/ Journal

of P sychosom

atic R esearch

66 (2009)

477–493

Klonsky et al. (2003) [26] Military recruits, U.S. Air Force

1986 79 with DSH 20±5 Two items from Schedule of Nonadaptive and Adaptive Personality (“When I get very tense, hurting myself physically somehow calms me down”; “I have hurt myself on purpose several times”) (exclusion when item “I have tried to commit suicide” endorsed)

6 Available for the total questionnaire

1907 without DSH

Lipschitz et al. (1999) [93] Adolescent psychiatric inpatients

71 28 with DSH 12–18 (14.8±1.6) Ad hoc self-report questionnaire (type, frequency, age at onset, precipitants)

7 None reported 43 without DSH

Low et al. (2000) [35] Mentally disordered patients in high-security detention; females

50 37 with DSH (15 infrequent, 22 frequent)

Medical examination records on occurrence and frequency

2 None reported

13 without DSH Lundh et al. (2007) [94] Age-group cohort from

three schools in Southern Sweden

123 84 with DSH 15 Standardized self-report questionnaire: DSHI [11], modifiedàsuicidal intent was asked for each endorsed self-harm behavior

8 Available for original version44 without DSH

Marchetto (2006) [27] Patients presenting to an emergency department with five or more episodes of self-cutting

516 16–60 (28.8±8.74) Medical examination records on occurrence and frequency

2 None reported

Matsumoto et al. (2004)

[95] Adolescent inmates of a juvenile detention center

201 33 with DSH 14–19 (16.8±1.6) Ad hoc self-report questionnaire [“Have you cut your wrist and/or your arm more than once?”; “Have you ever attempted suicide with any ways (including self-cutting)?”] (plus items on burning, frequency, precipitants, functions)

7 None reported 168 without DSH

McAuliffe et al. (2005)

[96] Patients treated for “parasuicide”, several treatment centers (WHO multicenter study in 12 European regions)

836 Mean=36 Self-harm assessment not described; semistructured/ structured interview, European Parasuicide Study Interview Schedule

1 None reported 4

(continued on next page)(continued on next page) 483 H . F liege

et al.

/ Journal

of P sychosom

atic R esearch

66 (2009)

477–493

Table 1 (continued)

Study Sample Deliberate self-harm assessment

Authors (year of publication) Reference Kind/Setting N n/n Age Assessment instruments Category

Psychometric validation data

McKay et al. (2004) [97] Psychiatric inpatients with borderline personality disorder with or without DSH

48 30 with DSH 30.0±9.3 (with DSH) Semistructured/ structured interview: Self-Injury Interview (methods, location, pain perception, precipitants, urge, relief)

4 Available 18 without DSH 29.11±6.4

(without DSH)

Milnes et al. (2002) [98] Patients admitted to hospital following self-harm

150 N15 Ad hoc clinical interview with questions on self-harm (occurrence, type, frequency); Beck interview scale for suicidal intent

4 None reported

Newman and Bland (2007)

[99] Patients with “parasuicidal” behavior and age-/sex-matched controls treated at an emergency department

707 507 with DSH 16 to over 45 Medical examination records on occurrence, frequency, method

2 None reported 200 without DSH

Nijman et al. (1999) [100] Psychiatric inpatients with or without DSH

54 24 with DSH 37.5±12.4 Ad hoc interview on self-harm (type, frequency, age of onset, severity)

4 None reported 30 without DSH

Paivio and McCulloch (2004)

[101] Undergraduate students of psychology; females

100 41 with DSH 21±1.7 Ad hoc self-report questionnaire (type, frequency)

7 None reported 59 without DSH

Portzky et al. (2007) [102] Students from Belgium and the Netherlands

4431 and 4643 Belgium: 460 with DSH, 3969 without DSH

15–16 Ad hoc self-report questionnaire [“Have you ever deliberately taken an overdose (…) or other medication or tried to harm yourself in some other way (such as cut yourself)?”], with open description and a posteriori classification

7 None reported

The Netherlands: 183 with DSH, 4460 without DSH

Rodriguez-Srednicki (2001)

[103] College students; females

441 18 with DSH 18–23 (20.6±1.2) Ad hoc self-report questionnaire items (occurrence, frequency)

6 None reported 423 without DSH

Romans et al. (1995) [104] Random community sample; females

477 23 with DSH 18–65 Ad hoc interview, self-harm section not described

4 None reported 454 without DSH

Schaffer et al. (1982) [105] Psychiatric inpatients and outpatients with DSH, patients without DSH matched for age, sex, and setting

28 14 with DSH 21–41 (28.1) Ad hoc interview on self-harm; Diagnostic Interview for Borderline Patients [106]

4 None reported for self-harm interview, available for borderline interview

14 without DSH 5

484 H . F liege

et al.

/ Journal

of P sychosom

atic R esearch

66 (2009)

477–493

Schmidtke et al. (1996) [107] Multicenter-study on “parasuicide” at 16 treatment centers in 13 European regions; self-harm included self-cutting behavior

3.968.125 6.310 with DSH 15 to over 55 Medical examination records on occurrence

2 None reported 3.961.815 without DSH

Simeon et al. (1992) [108] Patients with any personality disorder with or without DSH, age-matched groups

52 26 with DSH 31.0±9. Semistructured/ structured interview: Schedule for Interviewing Borderlines (occurrence, frequency)

5 Available 26 without DSH 31.4±9.4

Sourander et al. (2006) [109] Representative birth cohort assessed at age 12 and again at age 15

839 52 with DSH (at age 15) 12/15 Ad hoc single item (“I deliberately try to hurt or kill myself” response categories: not true, sometimes/somewhat, often/very true)

6 None reported 787 without DSH

Tuisku et al. (2006) [110] Adolescent psychiatric outpatients treated for depression

218 29 with DSH 13–19 Ad hoc structured interview questions on suicide attempts and deliberate self-harm without intent to die (none, occasional, frequent)

4 None reported 189 without DSH

Tulloch et al. (1997) [111] Adolescent patients presenting to an emergency department with self-harm and nonpsychiatric hospital controls, matched for age, sex and socioeconomic status

104 52 with DSH 14–19 (16.4±1.4) Medical examination records on self-harm occurrence and severity and ad hoc interview questions on suicide history

2 None reported 52 matched controls 4

van der Kolk et al. (1991)

[112] Patients with any personality disorder and/ or bipolar disorder, psychiatric hospital

74 28 with DSH/not suicide attempt

18–39 Ad hoc self-report questionnaire (occurrence, type)

7 None reported

Wright et al. (2004) [113] Adolescents with history of sexual abuse, pediatric hospital patients and students; students without sexual abuse

664 399 with wall-hitting as most frequent DSH

13–17 (15±1.2) Standardized self-report questionnaire: Self-Destructive Behaviors Inventory (unpublished manuscript) with 9 items on self-harm

7 Cronbach alpha reported

Young et al. (2006) [114] Male offenders receiving psychiatric treatment in prison

242 43 with DSH Not reported Semistructured/ structured ad hoc interview, not described; review of medical records

4 None reported 199 without DSH 2

(continued on next page)(continued on next page) 485 H . F liege

et al.

/ Journal

of P sychosom

atic R esearch

66 (2009)

477–493

Table 1 (continued)

Study Sample Deliberate self-harm assessment

Authors (year of publication) Reference Kind/Setting N n/n Age Assessment instruments Category

Psychometric validation data

Zlotnick et al. (1996) [115] Psychiatric inpatients; females

148 103 with DSH 33±9.2 Ad hoc self-report questionnaire (type, frequency, duration), including items on bingeing, reckless driving, etc.

7 None reported 45 without DSH

Zlotnick et al. (1999) [54] Psychiatric outpatients 256 85 with DSH 40.6±14.0 Ad hoc self-report questionnaire (type, frequency); semistructured/ structured interview for DSM-IV personality

7 Cronbach alpha reported171 without DSH 5

Zoroglu et al. (2003) [116] High-school students 839 175 with DSH 14–17 (15.9±1.8) Ad hoc self-report questionnaire (types)

7 None reported 664 without DSH

Zweig-Frank et al. (1994)

[117] Patients with borderline or other personality disorder

150 48 with DSH 18–48 Single item from Revised Diagnostic Interview for Borderlines [118] [“During the last two years, have you ever hurt yourself deliberately without intending to commit suicide (e.g., cutting your skin, burns, hitting yourself, breaking windows with your fists hitting walls, hitting your head)?”]

6 None reported 152 without DSH BPD: 28.3±6.3

Non-BPD: 29.7±7.2

DSH, deliberate self-harm; BPD, borderline personality disorder; single item's texts are included if reported in the paper or available. Assessment categories: 1=no information; 2=case ascertainment/medical records; 3=standardized clinician-administered rating; 4=ad hoc interviews; 5=validated interviews; 6=single items, self-report/interview; 7=ad hoc questionnaire; 8=validated questionnaire (assignment to assessment category is given in parentheses if information is insufficient to judge).

486 H . F liege

et al.

/ Journal

of P sychosom

atic R esearch

66 (2009)

477–493

Table 2 Overview of empirical evidence of correlates or risk factors of self-harm behavior

Variable Studies Design Strength of evidence Annotation

Sociodemographic factors Age [49,107] CS ++ Highest rates in adolescents and

young adults (15–24 years) [23] LT (12 months) +++ Highest recurrence of DSH for

age group 25–54 years Sex [11,26,27,49,52,67] CS 00/(++) 6 studies found no sex

differences in adults [107] 1 study found more females among

medical patients treated for self-harm [45,87,92,93,102,109,110] 6 studies on adolescents found a higher

prevalence in females (2:1 to 4:1); 1 study on adolescents found no sex differences

[69] LT (1–2 years) 000 Not predictive of DSH recurrence [23] LT (12 months)

Unemployment [78,99] CS +++ [90] LT (6 months) Predictive of DSH recurrence [23] LT (12 months)

No partnership [78] CS +++ [90] LT (6 months) Predictive of DSH recurrence

(except when widowed) [23] LT (12 months)

Distal factor Parental separation [87] CS +++

[109] LT (age 12/15) [85] LT (age 8/18)

Psychological problems of a parent

[109] LT (age 12/15) +++ Significant aspects: mother's nervousness and low well-being

Health problems in the family

[45] CS +++ [109] LT (age 12/15) [85] LT (age 8/18)

Experience of separation in childhood

[68,84,95,99,112] CS ++ [23] LT (12 months) +++ Not predictive of DSH recurrence

Childhood physical abuse

[35,63,68,83,84,87,95, 100,101,110,112,113]

CS ++

Childhood psychological abuse

[63,84,100,110] CS ++

Childhood emotional neglect

[63,75,84,92,93,100,112] CS ++

Childhood sexual abuse [35,49,63,65,72,75,81,82, 84,87,91,93,100,103,104, 110,112,113,115–117]

CS ++ Except: no association found by Rodriguez-Srednicki [103]

[23] LT (12 months) +++ Predictive of DSH recurrence Proximal factors Somatic complaints/ health problems

[45,72,85] CS +++/−−− [69] LT (6 months) Predictive of DSH recurrence [23] LT (12 months) [109] LT (age 12/15)

Negative affect intensity/reactivity

[83] CS +

Anxiety [26,35,67,72,87,102,108] CS +++ [109] LT (age 12/15)

Depression [26,45,52,64,67,70,72,76,87, 99,102,105,108,111,113]

CS +++ Except: no difference found by Schaffer et al. [105]

[109] LT (age 12/15) Predictive of DSH recurrence (2 studies) and incidence (1 study)

[85] LT (age 8/18) [23] LT (12 months)

Impulsivity [35,87,88,102,105,108] CS ++ Except: Portzky et al. [102] yielded significance in univariate analyses but not in multiple regression

(continued on next page)(continued on next page)

487H. Fliege et al. / Journal of Psychosomatic Research 66 (2009) 477–493

Table 2 (continued)

Variable Studies Design Strength of evidence Annotation

Aggressiveness/ hostility

[45,67,108] CS +++ [109] LT (age 12/15) [85] LT (age 8/18)

Psychopathy [114] CS Derealization/ Dissociation

[35,49,54,74,79,84, 95,100,112,115–117]

CS ++

Alexithymia/Lack of emotional expressivity

[83,97,101,115] CS ++

Low self-esteem [35,64,79,87,94,102] CS ++ Self-blame/ Self-derogation

[26,73,88] CS ++

Critical life events in previous year

[102] CS +

Perceived stress [52] CS + Low self-efficacy [80] CS + Hopelessness [98,108] CS ++

[23] LT (12 months) +++ Predictive of DSH recurrence Lack of coping skills, lack of problem-solving abilities, maladaptive coping

[67,86,89,96,98] CS ++

Protective factors Self-esteem [96] CS + Self-esteem buffers association between

passive-avoidant coping and self-harm Adaptive coping [67] CS 0 No differences found

CS, cross-sectional design/test correlates; LT, longitudinal design/test predictors; +, positive evidence; −, negative evidence; 0, no association.

488 H. Fliege et al. / Journal of Psychosomatic Research 66 (2009) 477–493

factor. “+” indicates positive evidence from only one cross- sectional study. “++” indicates positive evidence from at least two cross-sectional studies. “+++” indicates positive evidence from at least one prospective study. A zero means no association was found.

Results

Study design

Only 5 of the 59 original studies test predictors of deliberate self-harm in a longitudinal design [23,69,85,90,109]. They each include two time points of measurement, with intervals ranging from 6 months to 10 years. Three of the longitudinal studies [23,69,90] investigated patients who were medically treated for deliberate self-harm at t1. The predicted criterion at t2 was not new onset but recurrence of deliberate self-harm. In a Finnish epidemiological study [109], a representative birth cohort sample of adolescents was assessed at the age of 12 (t1) and again at the age 15 years (t2). Occurrence of deliberate self-harm at t2 was predicted. Among those who exhibited self-harm at t2, there were new cases of self- harm and repeaters. Self-harm rates at age 12 were 2.7% for girls and 3.1% for boys. At age 15, rates had risen to 12.6% for girls and 4.6% for boys. Thus, at age 15, the majority were new cases. However, no separate analyses for new cases and repeaters were reported. Another sample of the aforementioned Finnish epidemiological study [85]

included only male participants, assessed at age 8 (t1) and again at age 18 years (t2). Here, prevalence of deliberate self-harm at age 18 pertaining to the past 6 months was 2.2%. There was, however, no information offered concerning whether any or how many of these self- harmers had also exhibited self-harm behavior at the age of 8 years. It may well be assumed that, at age 18, the rate of new cases in the self-harm group was high but, again, no separate analyses were reported.

Consequently, we conclude that—with the possible exception of the latter study [85]—no study yielded evidence for risk factors of new incidences of deliberate self-harm behavior [55].

Sociodemographic correlates and risk factors

Deliberate self-harm behaviors may occur at all ages. Yet, notably high rates have been found in adolescents and young adults [49,107]. Patients who utilized medical treatment for self-harm were more often female [23,107]. In adolescents, six studies found higher self-harm rates for females [45,87,92,93,102,109]. Only one study found no such difference [110]. In three studies involving young adults [11,26,67] and in three studies involving adult samples of a large age range [27,43,49], there was no gender difference found. One study yielded no gender effect on recurrence of deliberate self-harm [23].

There were no studies on socioeconomic status or level of education and self-harm in adults. In adolescents, self-harm rates were associated with a lower level of education [45].

489H. Fliege et al. / Journal of Psychosomatic Research 66 (2009) 477–493

Prospective data yield that unemployment and having no partnership are predictive of the recurrence of self-harm within 12 months [23,90].

Distal correlates and risk factors

A large number of studies on deliberate self-harm investigate stressful traumatic experiences in childhood. Virtually all studies find associations with self-harm behavior. The following are significant factors: psychologi- cal problems on the part of a parent, parental separation, and early or prolonged separation from a parent. Most frequently found were associations between childhood experiences of emotional neglect, psychological or physical abuse, espe- cially sexual abuse, and adolescent or adult self-harm behavior. We identified 21 studies reporting associations between deliberate self-harm behavior and sexual abuse. Only one study did not report this particular association [103], a fact that its authors attribute to the small statistical power of the study.

However, with the exception of only one study, all responses were given retrospectively. Although the time sequence of early biographical events and self-harm in adolescence or adulthood is plausible in itself, retrospection is prone to various recall biases, which may lead to the overreporting or underreporting of biographical events.

In one longitudinal study [109], deliberate self-harm was measured at the age of 15, where parental separation and the psychological problems of a parent—as biogra- phical factors—had been assessed at age 12. Here, too, responses were given retrospectively, at least in part. Not living in a family with two biological parents, mother's health problems, somatic complaints, aggressiveness, and externalizing and internalizing problems at age 12 independently predicted deliberate self-harm 3 years later. In another, only male sample from the same longitudinal study, self-harm behavior at age 18 was more likely when, at age 8, the child's school performances had been poorer and when parents had lower educational levels [85].

Proximal correlates and risk factors

Individual factors There is good correlative and prognostic evidence for

the relevance of general psychopathology, in the way of anxiety, depression, and aggressiveness, in deliberate self- harm behavior. Self-harmers experience more frequent and more negative emotions in their daily lives than persons who do not self-harm. This heightened experience of negative emotion may be a principal reason for deliberate self-harm, as self-harm may acutely alleviate emotional distress. There is also an association between perceived health problems and deliberate self-harm. The evidence for those psychopathological factors deemed to be more specific to self-harm, such as derealization/dissociation and alexithymia/lack of emotional expressivity, is good but

correlative only. Self-harmers have more difficulties in identifying or understanding their emotions and in expressing their emotions than individuals who do not self-harm [83,97,101,115].

Stress and maladaptive coping One study showed a correlation between a higher level

of perceived stress in the last 4 weeks and deliberate self- harm behavior [52]; however, the time sequence was not assessed. A study in adolescents found an association between the number of critical life events and self-harm behavior, where both variables were measured retrospec- tively for the preceding year [102]. Again, no time sequence was assessed. Some cross-sectional results indicate associations between lower personal coping resources and deliberate self-harm. In several studies, deliberate self-harm was associated with low self-esteem and low problem-solving abilities. Self-harmers showed a lower belief in self-efficacy [80] and had a higher tendency towards a self-blaming coping style [73] and to self- derogation [26,88] than non-self-harmers.

Interaction between factors Distal and proximal factors may have more complex

than just additive effects. States and traits may also interact in the triggering of deliberate self-harm. Anxiety, depres- sion, aggressiveness, and low emotional expressivity have stronger trait aspects, whereas derealization and dissocia- tion have stronger state aspects. We found two cross- sectional studies that tested the paths of various factors. One study showed that it was not low emotional expressivity alone but childhood abuse and negative affect together with low emotional expressivity that made the individual prone to deliberate self-harm [83]. The other study found both a direct path from childhood abuse to current self-harm and indirect paths via low self-esteem and via dissociation [35].

Protective factors Some studies yielded associations between low self-

esteem and low coping resources, on the one hand, and deliberate self-harm, on the other. However, these studies adopted a pathogenic framework. In a study that did test within the salutogenic framework [96], high self-esteem buffered the negative effect of passive-avoidant coping on self-harm. Although the overall study was long- itudinal, the fact that the said variables were assessed at the same measurement time needs to be mentioned. In one study [67], maladaptive and adaptive coping strategies were tested. Although self-harmers relative to non-self-harmers showed more maladaptive coping, like behavioral disengagement and substance abuse, there were no group differences for adaptive coping, like active coping, acceptance, planning, or humor. Neither were there any differences for seeking instrumental or emotional support.

490 H. Fliege et al. / Journal of Psychosomatic Research 66 (2009) 477–493

Discussion

We reviewed empirical studies on correlates and potential risk factors of nonsuicidal, deliberate, self-harm behavior. We included studies that were in accordance with an accepted definition of deliberate self-harm behavior [1]. However, we considered some studies on suicidal behavior if they also included nonsuicidal self-harm behavior.

Sociodemographic and many other distal and proximal correlates of deliberate self-harm could be identified. There were few longitudinal studies, but such studies did not always test the associations that we are focusing on prospectively. Nevertheless, most of the correlates are at least suitable for inclusion in future prospective studies on risk factors.

Evidence of biographical stress and trauma in childhood was most frequent. Yet, the multitude of studies on childhood factors does not alter the fact that their validity is necessarily limited by their retrospective design.

With regard to proximal personal factors, evidence is strongest for anxiety, aggression, and depression, as indicators of general psychopathology, as well as for physical health problems.

Virtually all variables were investigated in a cross- sectional design. Thus, following Kraemer et al. [55], they can only be considered “correlates”. Even though the temporal distance of biographical factors to the occurrence of deliberate self-harm suggests that they are indeed risk factors, there are methodological limits to this conclusion. Precision of retrospective assessment is limited by various factors, such as recall biases, which may result in the overreporting or underreporting of events.

With the exception of the person's sex, all variables were modifiable variables [55]. This is a prerequisite for considering them in interventions.

The pathogenic approach dominates the literature on correlates and risk factors of deliberate self-harm. Studies report many potential risk factors, most frequently distal or biographical factors, such as emotional neglect, maltreatment, and abuse in childhood. More proximal factors that are relevant in stress-coping models were investigated in few instances. Although one study measured stress in the past 4 weeks, it is not clear whether this was a precursor, an accompanying condition, or even a consequence of self-harm. With respect to possible proximal factors, the self-harm literature is particularly incomplete. This state of knowledge is unsatisfactory, especially when we compare it with other areas, such as depression, where there are comprehensive studies that successfully integrate distal and proximal factors [56]. To our knowledge, there was only one study that tested interaction effects between different risk factors for the onset of deliberate self-harm [83].

Despite these analyses, the evidence on protective factors against deliberate self-harm is incomplete. While some studies investigated the lack of a resource as a correlate or

precursor of deliberate self-harm, buffering or moderating effects were neglected for the most part.

Although our review could identify a number of correlates and several probable risk factors of deliberate self-harm, a résumé of the evidence is limited by some methodological issues. Most importantly, the definition of deliberate self-harm varies across studies. Although we put the focus on studies of self-harm without conscious suicidal intent, a clear-cut distinction could not be made throughout. Even if deliberate self-harm acts were defined by the studies' investigators as having been carried out “without the intent to die”, some overlaps between nonsuicidal and suicidal behaviors may still have occurred. As a rule, suicidal intent was not assessed more closely; thus, it is virtually impossible to determine the extent to which the registered behaviors encompassed nonsuicidal or suicidal intent. Therefore, we recommend that future studies should include psychometric assessments of self-harm and intent in order to distinguish nonsuicidal self-harm as precisely as possible.

Other limitations result from the variability in the measurements. More specifically, deliberate self-harm is assessed by diverse methods. Although some advances in the development and psychometric validation of instruments [11,46,52] have been made, the assessments performed in many of the reviewed studies were constructed ad hoc for the study. Only 15 studies used validated instruments or reported any information on the psychometric properties of their assessments of self-harm. Properly validated instruments, like Gratz' Deliberate Self-Harm Inventory (DSHI) [11], were rarely used. In some instances, the assessments were not even described at all. Finally, small and unequal group sizes limit comparability.

Conclusions for research

Knowledge on correlates and risk factors is important in order to better predict, understand, and treat deliberate self- harm behavior. Future research should include (a) long- itudinal studies; (b) a psychometrically sound assessment of deliberate self-harm; (c) proximal stress factors (life events, stress, daily hassles, or situational triggers) that occur prior to the onset of deliberate self-harm; (d) the coping with stress dimension; (e) the potential role of social resources; (f) models that test interactions or transactional relations between risk factors, including mediating and moderating effects, as well as interactions between dynamic risk factors and target behavior, including unidirectional and bidirec- tional effects; and finally (g) models that test moderating effects of protective factors.

References

[1] Gratz KL. Risk factors for and functions of deliberate self-harm: an empirical and conceptual review. Clin Psychol 2003;10:192–205.

[2] Klonsky ED. The functions of deliberate self-injury: a review of the evidence. Clin Psychol Rev 2007;27:226–39.

491H. Fliege et al. / Journal of Psychosomatic Research 66 (2009) 477–493

[3] Skegg K. Self-harm. Lancet 2005;366:1471–83. [4] Clarke L, Whittaker M. Self-mutilation: culture, contexts, and nursing

responses. J Clin Nurs 1998;7:129–37. [5] Favazza AR. The coming of age of self-mutilation. J Nerv Ment Dis

1998;186:259–68. [6] Feldman MD. The challenge of self-mutilation: a review. Compr

Psychiatry 1988;3:252–69. [7] Linehan MM. Cognitive–behavioral treatment of borderline person-

ality disorder. New York: Guilford, 1993. [8] Ramon S. Attitudes of doctors and nurses to self-poisoning patients.

Soc Sci Med 1980;14:317–24. [9] Tantam D, Whittaker J. Personality disorder and self-wounding. Br J

Psychiatry 1992;161:451–64. [10] Herpertz S. Self injurious behaviour. Acta Psychiatr Scand 1995;91:

57–68. [11] Gratz KL. Measurement of deliberate self-harm: preliminary data on

the deliberate self-harm inventory. J Psychopathol Behav 2001;23: 253–63.

[12] Winchel RM, Stanley M. Self-injurious behavior: a review of the behavior and biology of self-mutilation. Am J Psychiatry 1991;148: 306–17.

[13] Claes L, Vandereycken W. Self-injurious behavior: differential diagnosis and functional differentiation. Compr Psychiatry 2007;48: 137–44.

[14] Sansone RA, Wiederman MW. The Self-harm Inventory (SHI): development of a scale for identifying self-destructive behaviors and borderline personality disorder. J Clin Psychol 1998;54: 973–83.

[15] Pattison EM, Kahan J. The deliberate self-harm syndrome. Am J Psychiatry 1983;140:867–72.

[16] Cooper J, Kapur N, Webb R, Lawlor M, Guthrie E, Mackway-Jones K, et al. Suicide after deliberate self-harm: a 4-year cohort study. Am J Psychiatry 2005;162:297–303.

[17] Suominen K, Isometsa E, Suokas J, Haukka J, Achte K, Lonnqvist J. Completed suicide after a suicide attempt: a 37-year follow-up study. Am J Psychiatry 2004;161:562–3.

[18] Foster T, Gillespie K, McClelland R. Mental disorders and suicide in Northern Ireland. Br J Psychiatry 1997;170:447–52.

[19] Hickey L, Hawton K, Fagg J, Weitzel H. Deliberate self-harm patients who leave the accident and emergency department without a psychiatric assessment: a neglected population at risk. J Psychosom Res 2001;50:87–93.

[20] Canetto S. Meanings of gender and suicidal behavior during adolescence. Suicide Life-Threat 1997;27:339–51.

[21] Zahl D, Hawton K. Repetition of deliberate self-harm and subsequent suicide risk: long-term follow-up study of 11583 patients. Br J Psychiatry 2004;185:70–5.

[22] Brown MZ, Comtois KA, Linehan MM. Reasons for suicide attempts and nonsuicidal self-injury in women with borderline personality disorder. J Abnorm Psychol 2002;111:198–202.

[23] Kapur N, Cooper J, King-Hele S, Webb R, Lawlor M, Rodway C, et al. The repetition of suicidal behavior: a multicenter cohort study. J Clin Psychiatry 2006;67:1599–609.

[24] Linehan MM, Comtois KA, Brown MZ, Heard HL, Wagner A. Suicide Attempt Self-Injury Interview (SASII): development, relia- bility, and validity of a scale to assess suicide attempts and intentional self-injury. Psychol Assess 2006;18:303–12.

[25] Croyle KL, Waltz J. Subclinical self-harm: range of behaviors, extent, and associated characteristics. Am J Orthopsychiatry 2007;77: 332–42.

[26] Klonsky M, Oltmanns T, Turkheimer E. Deliberate self-harm in a nonclinical population: prevalence and psychological correlates. Am J Psychiatry 2003;160:1501–8.

[27] Marchetto M. Repetitive skin-cutting: parental bonding, personality and gender. Psychol Psychother 2006;79:445–9.

[28] Klonsky ED, Muehlenkamp JJ. Self-injury: a research review for the practitioner. J Clin Psychol 2007;63:1045–56.

[29] Langbehn DR, Pfohl B. Clinical correlates of self-mutilation among psychiatric inpatients. Ann Clin Psychiatry 1993;5:45–51.

[30] Harned MS, Najavits LM,Weiss RD. Self-harm and suicidal behavior in women with comorbid PTSD and substance dependence. Am J Addict 2006;15:392–5.

[31] Favazza AR, DeRosear L, Conterio K. Self-mutilation and eating disorders. Suicide Life-Threat 1989;19:352–61.

[32] Truglia E, Mannucci E, Lassi S, Rotella CM, Faravelli C, Ricca V. Aggressiveness, anger and eating disorders: a review. Psychopathol- ogy 2006;39:55–68.

[33] Whitlock J, Eckenrode J, Silverman D. Self-injurious behaviors in a college population. Pediatrics 2006;117:1939–48.

[34] Paul T, Schroeter K, Dahme B, Nutzinger DO. Self-injurious behavior in women with eating disorders. Am J Psychiatry 2002;159:408–11.

[35] Low G, Jones D, MacLeod A, Power M, Duggan C. Childhood trauma, dissociation, and self-harming behaviour: a pilot study. Br J Med Psychol 2000;73:269–78.

[36] Phillips KA, Taub SL. Skin picking as a symptom of body dysmorphic disorder. Psychopharmacol Bull 1995;31:279–88.

[37] Simeon D, Stein DJ, Hollander E. Depersonalization disorder and self-injurious behavior. J Clin Psychiatry 1995;56:36–40.

[38] Andover MS, Pepper CM, Ryabchenko KA, Orrico EG, Gibb BE. Self-mutilation and symptoms of depression, anxiety, and borderline personality disorder. Suicide Life-Threat 2005;35:581–91.

[39] Ross S, Heath N. A study of the frequency of self-mutilation in a community sample of adolescents. J Youth Adolesc 2002;31: 67–77.

[40] Gershuny BS, Keuthen NJ, Gentes EL, Russo AR, Emmott EC, Jameson M, et al. Current posttraumatic stress disorder and history of trauma in trichotillomania. J Clin Psychol 2006;62:1521–9.

[41] Kocalevent R-D, Fliege H, Rose M, Walter M, Danzer G, Klapp BF. Autodestructive syndromes: a literature review. Psychother Psycho- som 2005;74:202–11.

[42] Feldman MD, Eisendrath SJ, Stuart J. The spectrum of factitious disorders. Clinical practice No. 40. Washington DC: American Psychiatric Press, 1996.

[43] Fliege H, Scholler G, Rose M, Willenberg H, Klapp BF. Factitious disorders and pathological self-harm in a hospital population: an interdisciplinary challenge. Gen Hosp Psychiatry 2002;24:164–71.

[44] Sachsse U. Overt self-injury. Psychother Psychosom 1994;62: 82–90.

[45] Brunner R, Parzer P, Haffner J, Steen R, Roos J, Klett M, et al. Prevalence and psychological correlates of occasional and repetitive deliberate self-harm in adolescents. Arch Pediatr Adolesc Med 2007; 161:641–9.

[46] Gutierrez PM, Osman A, Barrios FX, Kopper BA. Development and initial validation of the self-harm behavior questionnaire. J Pers Assess 2001;77:475–90.

[47] Hawton K, Hall S, Simkin S, Bale E, Bond A, Codd S, et al. Deliberate self-harm in adolescents: a study of characteristics and trends in Oxford, 1990–2000. J Child Psychol Psychiatry 2003;44: 1191–8.

[48] Pearce C, Martin G. Predicting suicide attempts among adolescents. Acta Psychiatr Scand 1994;90:324–8.

[49] Briere J, Gil E. Self-mutilation in clinical and general population samples: prevalence, correlates, and functions. Am J Orthopsychiatry 1998;68:609–20.

[50] Kessler R, Borges G, Walters E. Prevalence of and risk factors for lifetime suicide attempts in the National Comorbidity Survey. Arch Gen Psychiatry 1999;56:617–26.

[51] Ramsay R, Bagley C. The prevalence of suicidal behaviors, attitudes and associated social experiences in an urban population. Suicide Life-Threat 1985;15:151–67.

[52] Fliege H, Kocalevent R-D, Walter M, Beck S, Gratz KL, Gutierrez PM, et al. Three assessment tools for deliberate self-harm and suicide behavior: evaluation and psychopathological correlates. J Psychosom Res 2006;61:113–21.

492 H. Fliege et al. / Journal of Psychosomatic Research 66 (2009) 477–493

[53] Schmid Ch, Hochstrasser DH, Hell D. Selbstbeschädigiung als Symptom psychischer Krankheiten in der PUK Zürich zwischen 1987 und 1995. Schweiz Arch Neurol Psychiat 2000;151:184–9.

[54] Zlotnick C, Mattia J, Zimmerman M. Clinical correlates of self- mutilation in a sample of general psychiatric patients. J Nerv Ment Dis 1999;187:296–301.

[55] Kraemer H, Kazdin A, Offord D, Kessler R, Jensen P, Kupfer D. Coming to terms with the terms of risk. Arch Gen Psychiatry 1997;54: 337–43.

[56] Kendler KS, Gardner CO, Prescott CA. Toward a comprehensive developmental model for major depression in women. Am J Psychiatry 2002;159:1133–45.

[57] Bendersky M, Lewis M. Environmental risk, biological risk, and developmental outcome. Dev Psychol 1994;30:484–94.

[58] Douglas KS, Skeem JL. Violence risk assessment—getting specific about being dynamic. Psychol Public Policy Law 2005; 11:347–83.

[59] Masten A, Reed M. Resilience in development. In: Snyder C, Lopez S, editors. The handbook of positive psychology. Oxford, UK: Oxford University Press, 2002. p. 74–88.

[60] Antonovsky A. Unraveling the mystery of health. How people manage stress and stay well. San Francisco: Jossey-Bass, 1987.

[61] Garmezy N, Masten AS, Tellegen A. The study of stress and competence in children: a building block for developmental psychopathology. Child Dev 1984;55:97–111.

[62] Rutter M. Psychosocial resilience and protective mechanisms. In: Rolf J, Masten AS, Cichetti KH, Nuechterlein KH, Weintraub S, editors. Risk and protective factors in the development of psychopathology. Cambridge: University Press, 1990. p. 181–91.

[63] Akyuz G, Sar V, Kugu N, Dogan O. Reported childhood trauma, attempted suicide and self-mutilative behavior among women in the general population. Eur Psychiatry 2005;20:268–73.

[64] Boudewyn AC, Liem JH. Psychological, interpersonal, and beha- vioral correlates of chronic self-destructiveness: an exploratory study. Psychol Rep 1995;77:1283–97.

[65] Boudewyn AC, Liem JH. Childhood sexual abuse as a precursor to depression and self-destructive behavior in adulthood. J Trauma Stress 1995;8:445–59.

[66] Briere J. Trauma Symptom Inventory (TSI). Odessa (FL): Psycho- logical Assessment Resources, 1995.

[67] Brown S, Williams K. Past and recent deliberate self-harm: emotion and coping strategy differences. J Clin Psychol 2007;63:791–803.

[68] Carroll J, Schaffer C, Spensley J, Abramowitz SI. Family experiences of self-mutilating patients. Am J Psychiatry 1980;137:852–3.

[69] Colman I, Newman SC, Schopflocher D, Bland RC, Dyck RJ. A multivariate study of predictors repeat parasuicide. Acta Psychiatr Scand 2004;109:306–12.

[70] Crowell SE, Beauchaine TP, McCauley E, Smith CJ, Stevens AL, Sylvers P. Psychological, autonomic, and serotonergic correlates of parasuicide among adolescent girls. Dev Psychopathol 2005;17: 1105–27.

[71] Linehan MM, Comtois KA. Lifetime parasuicide count. Unpublished manuscript. Seattle: University of Washington 1996.

[72] Darche MA. Psychological factors differentiating self-mutilating and non-self-mutilating adolescent inpatient females. Psychiatr Hosp 1990;21:31–5.

[73] De Leo D, Heller TS. Who are the kids who self-harm? An Australian self-report school survey. Med J Aust 2004;181:140–4.

[74] Demitrack M, Putnam F, Brewerton T, Brandt H, Gold P. Relation of clinical variables to dissociative phenomena in eating disorders. Am J Psychiatry 1990;148:1184–8.

[75] Dubo ED, Zanarini MC, Lewis RE, Williams AA. Childhood antecedents of self-destructiveness in borderline personality disorder. Can J Psychiatry 1997;42:63–9.

[76] Dulit RA, Fyer MR, Leon AC, Brodsky BS, Frances AJ. Clinical correlates of self-mutilation in borderline personality disorder. Am J Psychiatry 1994;151:1305–11.

[77] Beck AT, Kovacs M, Weissman A. Assessment of suicidal intention —the scale for suicide ideation. J Consult Clin Psych 1979;47: 343–52.

[78] Evren C, Evren B. Self-mutilation in substance-dependent patients and relationship with childhood abuse and neglect, alexithymia and temperament and character dimensions of personality. Drug Alchol Depend 2005;80:15–22.

[79] Fliege H, Becker J, Weber C, Schoeneich F, Klapp BF, Rose M. Disordered regulation of self-esteem in patients with overt versus covert self-destructive behaviour [Störungen der Selbstwertregulation bei Patienten mit offen versus heimlich selbstschädigenden Verhal- ten]. Z Psychosom Med Psychoanal 2003;49:151–63.

[80] Fliege H, Kocalevent R-D, Rose M, Becker J, Walter M, Klapp BF. Patients with overt or covert self-harm: differences in optimism and self-efficacy. Dermatol Psychosom 2004;5:54–60.

[81] Gladstone G, Parker G, Wilhelm K, Mitchell P, Austin M-P. Characteristics of depressed patients who report childhood sexual abuse. Am J Psychiatry 1999;156:431–7.

[82] Gladstone G, Parker G, Mitchell P, Malhi G, Wilhelm K, Austin M-P. Implications of childhood trauma for depresses women: an analysis of pathways from childhood sexual abuse to deliberate self-harm and revictimization. Am J Psychiatry 2004;161:1417–25.

[83] Gratz KL. Risk factors for deliberate self-harm among female college students: the role and interaction of childhood maltreatment, emotional inexpressivity, and affect intensity/reactivity. Am J Orthopsychiatry 2006;76:238–50.

[84] Gratz KL, Conrad SD, Roemer L. Risk factors for deliberate self- harm among college students. Am J Orthopsychiatry 2002;72: 128–40.

[85] Haavisto A, Sourander A, Multimaki P, Parkkola K, Santalahti P, Helenius H, et al. Factors associated with ideation and acts of deliberate self-harm among 18-year-old boys. Soc Psychiatry Psychiatr Epidemiol 2005;40:912–21.

[86] Haines J, Williams C. Coping and problem-solving of self-mutilators. J Clin Psychol 1997;53:177–86.

[87] Hawton K, RodhamK, Evans E, Weatherall R. Deliberate self harm in adolescents: self report survey in schools in England. Br Med J 2002; 325:1207–11.

[88] Herpertz S, Sass H, Favazza A. Impulsivity in self-mutilative behavior: psychometric and biological findings. J Psychiatr Res 1997;31:451–65.

[89] Howat S, Davidson K. Parasuicidal behaviour and interpersonal problem solving performance in older adults. Br J Med Psychol 2002; 41:375–86.

[90] Johnston A, Cooper J, Webb R, Kapur N. Individual- and area-level predictors of self-harm repetition. Br J Psychiatry 2006;189:416–21.

[91] Joyce PR, Mckenzie JM, Mulder RT. Genetic, developmental and personality correlates of self-mutilation in depressed patients. Aust N Z J Psychiatry 2006;40:225–9.

[92] Kirkcaldy BD, Brown J, Siefen R. Disruptive behavioural disorders, self harm and suicidal ideation among German adolescents in psychiatric care. Int J Adolesc Med Health 2006;18:597–614.

[93] Lipschitz D, Winegar R, Nicolaou A, Hartnick E, Wolfson M, Southwick S. Perceived abuse and neglect as risk factors for suicidal behavior in adolescent inpatients. J Nerv Ment Dis 1999; 187:32–9.

[94] Lundh L, Karim J, Quilisch E. Deliberate self-harm in 15-year-old adolescents: a pilot study with a modified version of the Deliberate Self-Harm Inventory. Scand J Psychol 2007;48:33–41.

[95] Matsumoto T, Yamaguchi A, Chiba Y, Asami T, Iseki E, Hirayasu Y. Patterns of self-cutting: a preliminary study in differences in clinical implications between wrist- and arm-cutting using a Japanese juvenile detention center sample. Psychiatry Clin Neurosci 2004;58:377–82.

[96] McAuliffe C, Corcoran P, Keeley HS, Arensman E, Bille-Brahe U, De Leo D, et al. Problem-solving ability and repetition of deliberate self-harm: a multicentre study. Psychol Med 2005;35:1–11.

493H. Fliege et al. / Journal of Psychosomatic Research 66 (2009) 477–493

[97] McKay D, Gavigan CA, Kulchycky S. Social skills and sex-role functioning in borderline personality disorder: relationship to self- mutilating behavior. Cogn Behav Ther 2004;33:27–35.

[98] Milnes D, Owens D, Blenkiron P. Problems reported by self-harm patients. Perception, hopelessness, and suicidal intent. J Psychosom Res 2002;53:819–22.

[99] Newman SC, Bland RC. Case-control study of unemployment and parasuicide. Compr Psychiatry 2007;48:511–5.

[100] Nijman HLI, Dautzenberg M, Merkelbach HLGJ, Jung P, Wessel I, del Campo JA. Self-mutilating behaviour of psychiatric inpatients. Eur Psychiatry 1999;14:4–10.

[101] Paivio SC, McCulloch CR. Alexithymia as a mediator between childhood trauma and self-injurious behaviors. Child Abuse Negl 2004;28:339–54.

[102] Portzky G, De Wilde E-J, Van Heeringen K. Deliberate self-harm in young people: differences in prevalence and risk factors between The Netherlands and Belgium. Eur Child Adolesc Psychiatry 2007 (Epub ahead of print).

[103] Rodriguez-Srednicki O. Childhood sexual abuse, dissociation, and adult self-destructive behavior. J Child Sex Abus 2001;10:75–90.

[104] Romans SE, Martin JL, Anderson JC, Herbison GP, Mullen PE. Sexual abuse in childhood and deliberate self-harm. Am J Psychiatry 1995;152:1336–42.

[105] Schaffer C, Carroll J, Abramowitz SI. Self-mutilation and the borderline personality. J Nerv Ment Dis 1982;170:468–73.

[106] Gunderson JG, Kolb JE, Austin V. The diagnostic interview for borderline patients. Am J Psychiatry 1981;138:896–903.

[107] Schmidtke A, Bille-Brahe U, DeLeo D, Kerkhof A, Bjerke T, Crepet P, et al. Attempted suicide in Europe: rates, trends and socio- demographic characteristics of suicide attempters during the period 1989–1992. Results of the WHO/EURO Multicentre Study on Parasuicide. Acta Psychiatr Scand 1996;93:327–38.

[108] Simeon D, Stanley M, Frances AJ, Mann JJ, Winchel R, Stanley M. Self-mutilation in personality disorders: psychological and biological correlates. Am J Psychiatry 1992;149:221–6.

[109] Sourander A, Aromaa M, Pihlakoski L, Haavisto A, Rautava P, Helenius H, et al. Early predictors of deliberate self-harm among adolescents. A prospective follow-up study from age 3 to age 15. J Affect Disord 2006;93:87–96.

[110] Tuisku V, Pelkonen M, Karlsson L, Kiviruusu O, Holi M, Ruuttu T, et al. Suicidal ideation, deliberate self-harm behaviour and suicide attempts among adolescent outpatients with depressive mood disorders and comorbid axis I disorders. Eur Child Adolesc Psychiatry 2006;15:199–206.

[111] Tulloch AL, Blizzard L, Pinkus Z. Adolescent–parent communication in self-harm. J Adolesc Health 1997;21:267–75.

[112] van der Kolk BA, Perry JC, Herman JL. Childhood origins of self- destructive behavior. Am J Psychiatry 1991;148:1665–71.

[113] Wright J, Friedrich W, Cinq-Mars C, Cyr M, McDuff P. Self- destructive and delinquent behaviors of adolescent female victims of child sexual abuse: rates and covariates in clinical and nonclinical samples. Violence Vict 2004;19:627–43.

[114] Young MH, Justice JV, Erdberg P. Risk of harm: inmates who harm themselves while in prison psychiatric treatment. J Forensic Sci 2006; 51:156–62.

[115] Zlotnick C, Shea T, Pearlstein T, Simpson E, Costello E, Begin A. The relationship between dissociative symptoms, alexithymia, impulsiv- ity, sexual abuse, and self-mutilation. Compr Psychiatry 1996;37: 12–6.

[116] Zoroglu S, Tuzun U, Sar V, Tutkun H, Savas HA, Ozturk M, et al. Suicide attempt and self-mutilation among Turkish high school students in relation with abuse, neglect and dissociation. Psychiatry Clin Neurosci 2003;57:119–26.

[117] Zweig-Frank H, Paris J, Guzder J. Psychological risk factors for dissociation and self-mutilation in female patients with borderline personality disorder. Can J Psychiatry 1994;39:259–64.

[118] Zanarini MC, Gunderson JG, Frankenburg FR, Chauncey DL. The revised diagnostic interview for borderlines: discriminating BPD from other Axis II disorders. J Personal Disord 1989;3: 10–8.