Three assessment tools for deliberate self-harm and suicide behavior: evaluation and psychopathological correlates

Abstract

Objective

The aims of this study are to adapt two validated self-report questionnaires of deliberate self-harm and suicidal behavior to German, to investigate their psychometric properties and agreement with clinician-administered ratings, and to examine their psychopathological correlates.

Methods

The Deliberate Self-Harm Inventory [Gratz KL. Measurement of deliberate self-harm: preliminary data on the deliberate self-harm inventory. J Psychopathol Behav 2001;23:253–263] and the Self-Harm Behavior Questionnaire [Guttierez PM, Osman A, Barrios FX, Kopper BA. Development and initial validation of the self-harm behavior questionnaire. J Pers Assess 2001;77:475–490] were completed by 361 patients hospitalized for depressive, anxiety, adjustment, somatoform, and/or eating disorders. A clinician-administered rating scale of self-destructive behavior was included. Psychopathological variables were assessed by standardized questionnaires.

Results

The self-report questionnaires demonstrated good reliability (α=.81–.96, split-half r=.78–.98, test–retest r=.65–.91). Reliability of the clinician-administered ratings was acceptable (interrater κ=.46–.77, test–retest κ=.35–.48). Intraclass correlations (ICC=.68) for all three instruments were satisfactory. Rates of self-harm and associations between self-harm and suicidal behaviors are reported. The findings support the hypotheses of a higher degree of psychiatric symptomatology in patients with self-harm behavior compared to those without.

Conclusion

The two questionnaire adaptations are reliable and valid self-report scales for the assessment of self-harm and past suicidal behavior.

Introduction

Deliberate self-harm behavior is frequent in a variety of clinical and nonclinical groups, including psychiatric patients (4.3% to 17%) [1], [2], college students (14% to 35%) [3], [4], and the general population (4%) [5]. In a general hospital setting, open or disguised forms of self-harm must be expected among patients in all medical disciplines [6], [7]. Although deliberate self-harm is a common symptom of borderline personality disorder, it is not specific to that disorder occurring across a variety of disorders, as well as among nonclinical samples [8]. Consequently, self-harm may be studied as a behavioral phenomenon “in its own right” [8].

Deliberate self-harm behavior has been defined as the intentional and direct destruction or alteration of body tissue, resulting in tissue damage [1], [4], [9], [10]. Although many definitions of deliberate self-harm explicitly exclude behaviors with conscious suicidal intent, some researchers have argued that intent may not always be reliably measured, as patients may be ambivalent about their intent to die and/or may dissimulate their suicidal intention [11]. Moreover, given the high comorbidity of self-harm and suicidal behaviors, as well as evidence that a history of self-harm increases the risk of suicide [11], [12], [13], [14], it may be important to assess for the presence of suicidal behaviors in studies on self-harm [15], [16]. To ensure the comparability and replicability of studies on deliberate self-harm, we required standardized assessment instruments. Published approaches range from the use of single self-harm items [8] to more lengthy and elaborate instruments. Although several self-report questionnaires have undergone proper psychometric evaluation [4], [15], [17], [18], German versions have not been available. The aims of our study are (1) to translate empirically supported self-report measures of self-harm into German, (2) to test their psychometric properties in clinical samples, (3) to cross-validate these instruments with each other and with a clinician-administered measure of self-harm, (4) to determine the rates of self-harm in mentally/behaviorally disordered (nonpsychotic) patients, and (5) to examine psychopathological correlates of self-harm.

A language-adapted instrument should have similar psychometric properties to the original version. We expected to replicate the original questionnaires' dimensional structures and to obtain comparable reliability values. Furthermore, given the past findings of gender differences in rates of suicidal behaviors but not in rates of nonsuicidal deliberate self-harm (for which an absence of gender differences has repeatedly been found), we expected suicidal behaviors to be reported at higher rates among women and deliberate self-harm to be reported at comparable rates among women than men [4], [16], [17]. Among this sample of psychosomatic patients, we expected to find rates of self-harm at least as high as those observed among psychiatric patients [1], [2].

Deliberate self-harm behavior has been found to be associated with higher levels of depression, hopelessness, anxiety, hostility, impulsivity, aggression, and narcissistic and paranoid personality traits, as well as lower levels of self-esteem [5], [8], [19], [20], [21], [22], [23], [24], [25], [26]. We expected to replicate those findings in our sample. Moreover, given the evidence of reduced levels of optimism and self-efficacy among patients with overt self-harm (compared to those who concealed their self-harm [27]), we expected to find lower optimism and self-efficacy in patients with self-harm compared to those without. Finally, given that self-harm is considered to be a coping mechanism that functions to alleviate distress, we expected to find heightened levels of perceived stress among self-harming patients.