FULL PAGE VIEW

Childhood Sexual and Physical Abuse and Adult Self-Harm and Suicidal Behaviour: A Literature Review

Elaine E Santa Mina, RN, BA, BAAN, MSc1, Ruth M Gallop, RN, PhD2


This article reviews the link between childhood sexual and/or physical abuse and adult self-harm, suicidal ideation, and suicidal attempt from 1988 to 1998 in the clinical literature. Despite the methodological and definitional concerns, empirical studies have generally demonstrated more reports of self-harm, suicidal ideation, and suicidal behaviour in clinical and community populations of adults who report sexual and/or physical abuse in childhood than in comparison groups who do not report abuse. Implications for practice include the identification of populations at risk for self-harm or suicide and routine inquiry of abuse histories on admission. Further research into universal definitions of the key terms, standardized measurements of the variables, and longitudinal studies is required.

(Can J Psychiatry 1998;43:793–800)

Key Words: childhood sexual physical abuse, adult self-harm, suicide

A large body of empirical research has linked childhood trauma and adult psychopathology. The nature of this link is not clear. Most research is correlational in nature, and the reality is that significant numbers of people experience childhood trauma without developing serious adult psychopathology. This paper reviews the current findings in the research literature in 1 specific area: the link between childhood sexual and physical abuse and adult self-harm, suicidal ideation, suicide attempt, and completed suicide. Not only do these self-destructive behaviours continue to pose clinical and research challenges for health care providers, but also the role of childhood sexual and physical trauma, as causal and/or contributing factors to these self-destructive acts in adults, is not well understood (1–3). Recommendations regarding future research and clinical practice are discussed.

Search

A search was conducted using the key words “childhood abuse/trauma,” “self harm,” and “suicide.” Research articles reported in the last 10 years are included in the review.

Criteria

Inclusion

In this literature review, trauma is defined as unwanted interpersonal physical and/or sexual abuse occurring in childhood (before the age of 18 years) with an adult (4–6). The identified population for study of the potential effects of childhood physical and/or sexual abuse on self-harm and suicide is adult males and females older than 16 years of age. The range of behaviours measured as the distal effects of abuse in adults are self-harm, which is defined as self-damaging acts causing bodily harm without the intent to die (5); suicidal ideation; suicide attempts, which may be of high or low intent and/or high or low lethality; and completed suicides.

Exclusion

The consequences of childhood trauma, such as natural disasters, severe neglect, illness, catastrophic events, or emotional abuse, and adult victimization and rape are not included in this review. Studies with samples of persons who reported self-harm, suicidal ideation, or suicidal attempts and who were under 16 years of age were excluded. Studies associated with long-term, indirect expressions of self-harm, such as alcohol or substance abuse, promiscuity, anorexia, and destructive interpersonal relationships, were also excluded.

Definitional Issues

The studies to be reviewed are subject to question because of the lack of clear definitions of variables and instruments of measure (7,8). The literature on childhood physical and sexual abuse and self-harm and suicide defines the terms inconsistently and unclearly.

Childhood Abuse

Many authors have noted that definitions of child sexual abuse vary and can cover a broad range of behaviours. Before the early 1990s, studies rarely reflected the extent of abuse, the relationship between the child and the perpetrator, the family background, the disclosure of abuse, or the treatment, if any, received by the victim. As well, the literature did not necessarily differentiate between the widely disparate types of abuse or the nature of abuse, such as exposure, molestation, intercourse, and rape. In the last few years, increasing attention to this problem has enabled readers generally to determine the age of victims, if the trauma involved genital contact (if sexual abuse), if abuse was a single or chronic occurrence, and if the perpetrator was familial or at least known to the victim. Without this minimal information, any conclusions drawn from findings are difficult to interpret.

The review revealed that multiple measures and techniques exist to assess the presence of trauma and the nature of the trauma, particularly in the area of childhood sexual trauma. Without some agreement or universal set of definitions, incomplete reporting of abuse and methodological problems in research occur (9). In some studies, the histories of victim abuse were recorded through chart reviews (6,10–16). In other studies, abuse was assessed in semistructured interviews (6,13,17–26) or in self-report scales (5,23,27–34).

In another approach to the definition issue, Carlin and Ward suggest that there are 2 definitions of abuse (35). One is an objective definition based on the occurrences of critical events prior to a chosen age. The second definition is subjective and is based on one defining oneself as abused. In their study of sexually abused females, 77.6% of the sample who acknowledged unwanted sexual experience more invasive than fondling before the age of 16 years labelled themselves as abused; 17% did not consider themselves abused; 19% had subjective definitions of abuse, which differed from the objective one.

In the reviewed studies, childhood sexual abuse has generally incorporated the concepts of unwanted sexual experience occurring before the age of 16 (4,6) or 18 years (11,12,15,30,31), excluding consensual sexual activity between children of similar ages. Specifics of the nature and extent of that abuse were incorporated in the measurements of some of the studies (5,17,22,24,26,27,30,31). Aspects of abuse that have been measured include the types of physical sexual involvement, the use of force, the duration of abuse, the age of onset, and the relationship of the perpetrator to the victim. Peters and Range expanded the definition and measurement of childhood sexual abuse in their study (33). The nature of the abuse and the perpetrator–victim relationship was broadened to incorporate peer sexual exploitation.

Childhood physical abuse was not consistently defined or measured in the studies. Either the study did not define the term (39) or recorded its absence or presence through chart reviews (14), interviews (19,26,32), or questionnaires (19,29,34). Studies that provided definitions generally incorporated the concepts of deliberate striking, hitting, punching, or burning of a child less than 18 years of age by an adult that resulted in physical injury such as bruises or fractures requiring medical intervention. The definitions distinguished these events from single slaps (15) or intrafamilial fights (11).

Self-Harm

Definitions and terminology for self-harm, self-mutilation, suicidal behaviours, parasuicidal behaviours, partial suicide, and suicidal gestures overlap and conflict (36,37). The term self-harm has been used to describe everything from overeating and substance abuse to severe body mutilation and suicide (5,19,24). Connors speaks of a continuum of self-harm behaviours including 1) chosen “body alterations” such as tattoos and cosmetic surgery; 2) indirect self-harm such as substance abuse; 3) failure to care for self, for example, excessive risk-taking; and 4) direct actions resulting in self-injury (7). Self-harm is also termed self-mutilation, self-injurious behaviour, self-inflicted violence, and auto-aggression (7). Self-harm is commonly defined as an individual’s intentional damage to a part of his or her own body, without a conscious intent to die, although the result may be fatal (37,38). Another group of researchers chose to use the term “deliberate self-harm” rather than suicide, because self-harm did not imply the conscious or unconscious intent to die (25). Other authors suggest that mutilators cognitively distinguish their mutilation activities from suicide attempts (34,37,39). Some authors argue that the purpose of trauma victims’ self-harm behaviours, which are often chronic in nature, is to manage the aftermath of the trauma and provide individual relief in order to avoid suicide (7,34,39).

The concept of self-harm was defined and measured in studies through semistructured interviews (26,31,32), questionnaires (19), and chart reviews (10,16,24) and by scales (5,34,39).

Suicidal Behaviour

Most empirical work distinguishes self-harm from suicidal ideation and suicide attempt. However, the phrase “suicidal behaviour” has been used broadly to include self-mutilation behaviours with or without differentiation of the intent to die, suicidal ideation with or without associated behaviours, and an ill-defined range of low to high levels of intent, motivation, and medical lethality. Suicidal behaviours and attempts are also termed suicidal gestures and parasuicide. The term parasuicide is referenced from the World Health Organization in a study by Bland and colleagues (1). The World Health Organization defines parasuicide as “an act of nonfatal outcome, in which an individual deliberately initiates a nonhabitual behaviour, that without intervention by others, will cause self-harm . . . which is aimed at realizing changes that he/she desired via the actual or expected physical consequences”  (40). As Bland and colleagues point out, this definition “does not presume a failed intent to produce a fatal outcome” (1). The confusion lies in the inappropriate occurrence of the word “suicide” in “parasuicide.” The word “suicide” inherently implies a death wish and, as such, is misleading and confusing. Studies of parasuicide should be classified under self-harm. Leenaars and colleagues suggest abandoning the use of the term parasuicide to describe suicidal behaviours (9).

Similarly, the term “suicidal gesture” to describe self-harm without the intent to die can be confusing (27). Dubo and colleagues focused on the definition of self-harm to differentiate it from suicidal behaviour (39). The frequency of occurrences of either self-harm or suicidal behaviour led to another classification. “Super” self-mutilators were defined as those who had made 50 or more self-mutilation efforts; and “super” suicide attempters were those who had made 5 or more suicide efforts.

There is no consensus on the meaning and scope of the terms suicidal ideation and suicidal behaviours. Most studies do not clearly define the terms. Angst and colleagues conducted a prevalence study on suicide attempts without a specified definition of suicide attempt (18). However, a semi-structured interview that incorporated thoughts of self-harm without specification of intent to die was the instrument chosen to measure suicidality. The inclusion of self-harm questions in a suicide interview may be misleading for the subject and for the interpretation of results.

Suicidal ideation and suicide attempt were defined through clinical documentation in chart reviews in some studies (6,10). Other instruments of measurement for suicidal ideation and behaviour did not separate the concepts into discretely different phenomena. Suicidal ideation and attempts were measured in combination through interviews (11,12,15,23,33) or by self-report suicidality severity scales (14,28,30). Other studies used scales to measure one or the other of the concepts. The concept of suicidal ideation was measured by questionnaires in a study by Bagley and colleagues (27), and suicidal behaviours were measured by different questionnaires in other studies (17,21,22,33). Self-harm, suicidal ideation, and suicidal behaviours were measured in the same interviews in one study (20) and within the same instrument, the impulse anger checklist, in another study (34). The number of suicide attempt events within an institution was the measurement chosen by other authors (13,20).

Pre-1988 Studies

Two major reviews of the pre-1988 literature on the impact of childhood sexual and/or physical abuse on adults were conducted by Beitchman and colleagues (2) and Browne and Finklehor (41). In these 2 reviews, 8 studies were found in which self-harm and suicide were identified as adult symptoms of psychopathology following childhood abuse. The impact of abuse variables such as the age of abuse, onset of abuse, gender of victim, relationship between victim and offender, and use of force or penetration were described. However, contextual issues about the experience of abuse, such as the parent reaction to the disclosure of abuse, child participation, and possible interventions provided at the time of disclosure, were generally not researched (41–43).

The authors of the reviews critiqued the studies for the lack of nonclinical samples; small sample sizes; the lack of control and comparison groups; the lack of control for correlated variables; the lack of clear definitions of childhood abuse, self-harm, and suicide attempt; and the lack of standardized measures.

The reviewed studies, which included the variables of self-harm and suicide, linked adults with histories of childhood abuse to self-harm and/or suicidality in adulthood (2,41). Briere and Runtz, in a study with a sample size of 133 females, reported that 56% of the women who had a history of sexual abuse had attempted suicide compared with 23% of the control group (44). Bryer and colleagues found that women with a history of suicidal ideation, gestures, and/or attempts (N = 36) were 3 times more likely to have been abused in childhood than were patients without symptoms (45). Bagley and Ramsey documented a 5% frequency of suicide plans, attempts, and/or deliberate self-harm in female victims of abuse compared with 0% in a nonabused group (abused N = 82) (46).

Briere reported that 51% of sexual abuse victims versus 34% of nonabused participants had a history of suicide attempts, and 31% of victims versus 19% nonabused individuals had self-harm ideation (N = 153) (47). Sedney and Brooks also reported that 39% of a female college sample with history of sexual abuse (N = 301) had thoughts of self-harm versus 16% in the nonabused group (48). Two reported studies did not find a link between a history of childhood sexual abuse and suicidality (48,49).

Post-1988 Studies

Twenty-nine studies of childhood sexual and physical abuse, conducted between 1988 and 1998, that included self-harm and/or suicidal behaviours as the study variables were reviewed. Table 1 cites the studies reviewed. There were no studies that related childhood sexual or physical abuse with completed suicides. However, it is noteworthy that in Stone’s text on the fate of borderline patients, 8 of the 19 borderline patients who committed suicide had a positive history of incest (50).

Summary

The number of studies has increased from 1988 to 1998. The majority of the studies have comparison groups. Few studies define self-harm and suicide as the major variables of interest in relation to childhood trauma. One descriptive study by Arnold was conducted on women and self-injury (19). The majority of studies were cross-sectional and identified research questions rather than theory-testing hypotheses.

Samples

Twelve studies had female samples only, while 16 studies sampled female and male subjects. Only 1 study (27) addressed just males. Twenty studies had clinical samples, and 9 studies had community samples (Table 1).

Table 1.  Studies on childhood abuse and self-harm and suicide: 1988–1998

Study

Sample

N (abused)

Gender

Comparison

Outcome

Anderson and others (17)

CL

(51)

F

No

SA

Angst and others (18)

CY

407 (57)

17 MA, 40 FA

Yes

SI, SA

Arnold (19)

CY

76 (34)

76 F

No

SH

Bagley and others (27)

CY

750 (117)

M

Yes

SA

Boudewyn and Huser Liem (20)

CY

438 (91)

28 MA, 63 FA

Yes

SH, SI, SA

Briere and Zaidi (10)

CL

100 (35)

F

Yes

SH, SI, SA

Brodsky and others (  )

CL

60 (36)

F

Yes

SH

Brown and Anderson (11)

CL

947 (166)

71 FA, 25 MA

Yes

SI, SA

Bryant and Range (28)

CY

114 (54)

F

Yes

SI, SA

De Wilde and others (29)

CL

157

33 MA, 124 FA

Yes

SA

Dubo and others (39)

CL

59

19 MA, 40 FA

Yes

SH, SA

Kaplan and others (30)

CL

251 (128)

68 MA, 183 FA

Yes

SI, SA

Gould and others (21)

CL

292 (152)

29 MA, 133 FA

Yes

SA

McCauley and others (12)

CL

1931 (424)

F

Yes

SA

Links and others (32)

CL

69 (22)

M, F

Yes

SH

Modestin and others (22)

CL

90 (84)

39 MA, 45 FA

Yes

SA

Mullen and others (23)

CY

1376 (252)

F

Yes

SI, SA

Otnow Lewis and others (13)

CL

42 (21)

F

Yes

SA, SC

Pettigrew and Burcham (24)

CL

146 (73)

F

Yes

SH

Peters and Range (33)

CY

266 (69)

26 MA, 43 FA

Yes

SI, SA

Read (14)

CL

100 (29)

M, F

Yes

SA

Romans and others (25)

CY

477 (252)

F

Yes

SH

Silk and others (31)

CL

55 (41)

5 MA, 36 FA

Yes

SH

Stepakoff (43)

CY

393

M, F

Yes

SI, SA

van der Kolk and others (34)

CL

74

39 F, 35 M

No

SH, SA

Windle and others (15)

CL

802

481 M, 321 F

Yes

SA

Yellowlees and Kaushik (6)

CL

707 (44)

F

Yes

SA

Yeo and Yeo (16)

CL

178 (14)

4 M, 10 F

Yes

SH

Zweig-Frank and others (26)

CL

150

F

Yes

SH

CL = clinical; CY = community; F = female; FA = female abused; M = male; MA = male abused; SA = suicide attempt; SH = self-harm; SI = suicidal ideation.

Study sample sizes varied considerably. Although generally the total sample sizes per study were greater than 50, the comparison groups based on the presence or absence of abuse were considerably smaller (Table 1).

Childhood Abuse and Self-Harm and Suicide

Sixteen studies focused on the target variables of sexual abuse alone, (5,6, 10,13,16–18,20,22–25,27,31,33,43). Thirteen studies focused on sexual and physical abuse combined (11,12,14,15, 19,21,26,28–30,32,34,39). There were no studies specific to childhood physical abuse.

Self-harm with or without the intent to die was among the identified outcome variables of childhood abuse in 8 studies (5,16,19,24–26,31,32). Suicidal ideation and/or suicide attempt were among the target variables in 17 studies (6,11–15,17,18,21–23,27–30,33,43). Both self-harm and suicide were the variables of interest in 4 studies (10,20,34,39).

Major Findings of Post-1988 Literature

Childhood Abuse and Self-Harm

Arnold recorded narratives of 76 women who engaged in self-harm behaviours (19). Forty-nine percent of these women reported a history of childhood sexual abuse. According to Arnold, 25% of the participants reported a history of physical abuse, and “most of those who had been physically abused also reported sexual abuse” (p 10).

Four studies compared self-harm in abused and nonabused samples. Links and colleagues found that 77.3% of the abused group participated in self-mutilation compared with 47.6% of the nonabused group (P < 0.05) (32). The abused group engaged in more physically self-damaging acts based on the BPD items in the Diagnostic and Statistical Manual of Mental Disorders (DSM-III) (95.2%) compared with the nonabused group (69.5%) (P < 0.05). Yeo and Yeo found that 93% of the abused group took overdoses compared with 49% of the nonabused (P < 0.01) and 50% (P < 0.001) engaged in wrist-cutting compared with 12% of the nonabused group (16). In a study by Pettigrew and Burcham, the abused group differed from the nonabused group in the following aspects of self-harm: deliberate self-harm, 27% abused versus 15% nonabused; repeated self-harm, 18% abused versus 7% nonabused; overdoses, 25% abused versus 14% nonabused (all P < 0.01); and self-mutilation, 10% abused versus 4% nonabused (P < 0.05) (24). In a study by Romans and colleagues (25), 8.7% of the abused group versus 0.4% of the nonabused comparison group engaged in self-harm (abused N = 252, P < 0.0005).

Zweig-Frank and colleagues examined the parameters of childhood sexual abuse and the diagnosis of BPD to differentiate self-mutilators from non–self-mutilators (N = 150) (26). Penetration was the only significant variable (P < 0.05) in the model when the parameters of sexual abuse were regressed on self-mutilation. This significance disappeared, however, when diagnosis (BPD) was added to the regression equation. Silk and colleagues also addressed the dimensions of sexual abuse and found that ongoing abuse was a significant variable in differentiating individuals with and without parasuicidal behaviour (P < 0.05) (31).

Childhood Abuse and Suicidal Ideation and Attempt

Seventeen studies addressed childhood sexual and/or physical abuse with suicidal ideation and/or attempts (6,11–15,17,18,21–23,27–30,33,43). Five studies comparing abused and nonabused groups showed the following: Brown and Anderson found that 75% of the abused group were suicidal versus 57% of the nonabused group (P < 0.0001) (11). Angst found that 19% of the abused group made multiple suicide attempts versus 3.2% of the nonabused (P < 0.001) (18 ). In a male population, Bagley and colleagues reported that 7.9% of victims of short-term abuse (N = 65) and 30.8% of long-term abuse victims (N = 52) made a suicidal gesture in their lifetime compared with 4.1% of the nonabused group (N = 633) (P < 0.01) (27). Bryant and Range found that college women in the sexually and/or physically abused group were more suicidal compared with the control group (abuse N = 40, P < 0.05) (28). Gould and colleagues found that 18% of the abused group had a history of suicide attempts compared with 3% of the nonabused group (N = 292, P < 0.00001) (21).

Seven studies compared the parameters of abuse history: De Wilde and colleagues grouped 48 suicide attempters by severity of suicide risk and reported the high-risk group to have more sexual and physical abuse than the lower-risk groups (29). Mullen and colleagues’ study demonstrated a greater number of suicide attempts with sexual abuse in conjunction with penetration (N = 252, P < 0.001); 8% of the abuse without penetration group attempted suicide, compared with 25% of the abuse with intercourse group (23). Anderson and colleagues found that of 51 abuse victims, 49% had attempted suicide at least once (17). Peters and Range found that the abused group with physical contact, such as touching, reported more suicide attempts than the abused group with exploitation, regardless of gender of the victim, or the age of the perpetrator (adult versus peer) (abused N = 42, P < 0.001) (33). McCauley and colleagues reported that 11.3% of the abused group attempted suicide versus 2.1% of the nonabused group (12). Kaplan and colleagues found that childhood sexual abuse alone was not associated with suicidal behaviour, but the age of onset of abuse (0–6 years) and the perpetrator of abuse (parent perpetrator of sexual or physical abuse) demonstrated a significant association by univariate analysis (30); but when abuse in adulthood was controlled for, the relationship was no longer significant; Stepakoff found that adult and childhood groups who had experienced sexual abuse with severity and force made more suicidal attempts than the nonabused group (P < 0.04) (43).

Six studies compared gender: Read found that 100% of the sexually or physically abused males reported suicidality compared with 30% of the nonabused males (N = 57, P < 0.00001) (14). Yellowlees and Kaushik, with a female sample, found that 34.1% of the sexually abused group made suicide attempts compared with 11.4% of the nonabused group (abuse N = 44, P < 0.0059) (6). Modestin and colleagues found significantly more suicidal behaviour in the female patients diagnosed with personality disorders who reported a history of childhood sexual abuse than in the male group (N = 90, P < 0.001) (22); Brown and Anderson found that abused male patients (88%) were more likely to be suicidal upon admission to hospital than nonabused males (57%, P < 0.004) (11). Windle and colleagues compared suicide attempts in abused and nonabused groups: sexual and physical abuse 32.0% in abused males versus 11.6% nonabused males and 52.7% in abused females versus 18.0% nonabused (15). In a longitudinal study by Otnow Lewis and colleagues of 21 women in a correctional facility, 10 had a history of childhood sexual abuse, and 19 had attempted suicide (13). In this study, 1 of the subjects with a positive history of abuse completed suicide within the time of the longitudinal study. The authors stated that this 1 completed suicide represented 180 times the expected rate of mortality for that population.

Childhood Abuse and Self-Harm and Suicide

Four studies examined abuse with self-harm and suicide as variables. Briere and Zaidi found that there was significantly more self-harm, suicidal ideation, and suicide attempts in the abused group than the nonabused group; and severity of abuse was positively correlated with suicide attempts (r = 0.36) (10). In the study by van der Kolk and colleagues, sexual abuse was significantly correlated with self-harm (r = 0.049) and suicide attempts (0.41), and physical abuse was positively correlated with self-harm (0.22) and suicide attempts (0.31) (34). Dubo and colleagues found that victims of sexual abuse perpetrated by a caretaker were more self-destructive (N = 59, P < 0.05) and made more suicide attempts (P < 0.005) than the nonabused group (39). In Boudewyn and Huser Liem’s study of depressed adults, abused females and males differed from nonabused females and males on the following dimensions: self-harm ideation, 45% versus 21% (females) and 42% versus 15% (males); self-harm, 33% versus 10% (females) and 18% versus 7% (males); suicidal ideation, 68% versus 42% (females) and 60% versus 32% (males); and attempted suicide, 29% versus 7% (females) and 30% versus 4% (males) (P < 0.001) (20).

Some studies did not demonstrate that a history of abuse was significantly related to the presence of self-harm, suicidal ideation, and/or suicide attempts. Brodsky and colleagues found that there was no significant difference in self-mutilation in the abused group compared with the nonabused group (P < 0.58) (5). Read found that sexually abused females (N = 43) did not report more suicidality than the nonabused comparison group (14). Stepakoff found that child sexual abuse was not significantly predictive of suicidal ideation (43).

Limitations of Post-1988 Studies

The number of studies within the past 10 years remains relatively small. Apart from 1 qualitative study (19), all studies were quantitative. Leenaars and colleagues state that both qualitative and quantitative research are essential for the understanding of suicide (9). Kral (42) and Stepakoff (43) argue that a lack of the study of contextual factors surrounding sexual abuse and suicide, such as the experience of sexual violence, gender oppression, and the social psychological issues of feminist research, limits the understanding of a complex relationship.

All studies reviewed were essentially atheoretical. Leenaars and colleagues argue that theory as a base for research in the area of suicide is essential for hypothesis-testing and the development and expansion of existing theories (9). There was no specific theory of childhood abuse or suicide incorporated in the studies and thus no theory-testing.

The empirical study designs are predominantly comparison studies and have little focus on hypothesis-testing. They are predominantly cross-sectional. Longitudinal studies would address the mediating factors that may protect against the outcomes of self-harm and suicide. While several studies report large sample sizes, the actual number of abused or suicidal subjects in the sample was in some cases quite small (13,16,18,31).

The definitions of the variables of childhood abuse, self-harm, and suicidal behaviour are inconsistent throughout the literature. Any conclusion regarding differences among these variables must consider that the operational terms across studies may not be completely comparable. Peters and Range identify that researchers disagree about the age that defines a child, the age that defines an abuser, and the acts that constitute sexual abuse (33). Leenaars and colleagues identified the lack of consensus in nomenclature and classification of suicidal behaviour in suicide research (9). They argue that the definitions are Cartesian in nature and ignore the human elements of motivation, intention, and severity.

The authors of the reviewed studies identify as a limitation that all reports of abuse were retrospective victim reports and were not validated. Leenaars and colleagues also concurred that much research about people and trauma is based on introspective accounts rather than objective reports (9). The results from clinical samples represented a traumatized population that had already presented itself for care. Preselection of an already traumatized sample may exclude a stable population who have also experienced childhood abuse (27). Browne and Finkelhor also identified this concern in the pre-1988 literature (41).

Sheldrick argued that abuse occurs in the context of multiple problems (8). The cause and effect relationship of traditional empirical research does not consider the context in which the abuse occurred. Multiple variables may occur in tandem with other stressors for the long-term effects of childhood abuse in adults. Some theoretical and empirical work in this area has been done in other areas of trauma. For example, Zweig-Frank and colleagues have tried to identify the relationship between trauma and the diagnosis of BPD (26). The path from childhood sexual abuse to a diagnosis of BPD clearly is not linear but rather is mediated by many factors, including genetics and resilience. Adams and Lehnert recommend an interactional paradigm of stress that would incorporate environmental and individual factors to understand the multicausal phenomena of childhood trauma and suicidal behaviour (51).

Implications for Clinical Practice

Despite the methodological limitations of the research in this area, there is strong evidence of a link between childhood sexual and physical abuse and adult self-harm, suicidal ideation, and suicidal behaviours. Clinicians must be able to identify clinical populations who are at risk for these negative sequelae through a thorough assessment of childhood trauma, self-harm behaviours, suicidal ideation, and suicide attempts.

The link between trauma and self-harm or suicide is strongest when the abuse has been of long duration, the perpetrator has been known to the victim, and force and penetration have occurred. Many of the studies found a positive association between trauma and self-harm or suicide, but the odds ratios (ORs) varied considerably. Yellowlees and Kaushik demonstrated an OR of 3.39 for sexually abused women to evidence suicide attempt and compared them with nonabused women (6). The OR for suicide attempt was 4.1 in the sexually abused group and 1.2 in the physically abused group in a study by Gould and colleagues (21). Mullen and colleagues found that abused subjects were 20 times more likely to engage in suicidal attempts than nonabused subjects and 70 times more likely if the abuse involved penetration (23). Zweig-Frank and colleagues demonstrated that the OR for victims of abuse with penetration to be more likely to attempt suicide was 1.3 (26). There is something specific about abuse when its nature is sexual and physically invasive that increases the risk. These contextual aspects of the abuse history must be incorporated in the assessment to determine the risk of suicide. There is a need to develop clinical measures and research measures that assess the sexual abuse severity risk factors such as type, severity, duration, relationship of the victim to perpetrator, violence, treatment, and disclosure.

Assessment instruments that identify the overlapping and differentiating aspects of self-harm and suicidal behaviours are also required. Scales must be developed that incorporate assessment of the meaning of the self-harm for the patient with particular discrimination of the intent to die. Different treatment modalities may be required to provide safety and alternative coping strategies based on the careful discernment of intent, motivation, and meaning of the behaviour for the patient. Child victims of sexual and physical abuse require careful monitoring over time for potential negative outcomes that may present during adulthood. Recognition by clinicians of the distal outcomes of suicidality and self-harm in a history of childhood abuse may support routine inquiry about abuse within admission procedures and training for staff about effective follow-up (52).

Implications for Research

Universal definitions and instruments are required to measure the independent variable childhood sexual and physical abuse and the dependent variables self-harm, suicidal ideation, and suicide behaviours. Further qualitative work may help identify the contextual aspects of these variables. Additional studies of male populations and specifically physical abuse from both community and clinical subgroups are required. Longitudinal, prospective studies that objectively verify episodes of abuse would also help evaluate the risk of these negative outcomes over time and develop the complex theoretical models of contributing factors. Given the evidence that there is a link between trauma and self-harm/suicide, research must focus on developing and testing intervention strategies for high-risk populations. It is not known how abuse relates to completed suicides, and this work is also necessary.

Conclusion

Despite methodological limits, there is a relationship between childhood sexual and physical abuse and adult self-harm and suicide. It is not a simple linear relationship. Longitudinal studies are required to understand both the contributing and mediating factors. However, persons with abuse histories are at greater risk for self-harm and suicide.

Clinical Implications

Limitations

References

1. Bland RC, Newman SC, Dyck RJ. The epidemiology of parasuicide in Edmonton. Can J Psychiatry 1994;39:391–6.

2. Beitchman JH, Zucker KJ, Hood JE, DaCosta GA, Akman D, Cassavia E. A review of the long-term effects of child sexual abuse. Child Abuse Negl 1992;16:101–18.

3. Wagner BM. Family risk factors for child and adolescent suicidal behavior. Psychol Bull 1997;121:246–98.

4. Jacobson A, Herald C. The relevance of childhood sexual abuse to adult psychiatric inpatient care. Hospital and Community Psychiatry 1990;41:154–8.

5. Brodsky BS, Cloitre M, Dulit RA. Relationship of dissociation to self-mutilation and childhood abuse in borderline personality disorder. Am J Psychiatry 1995;152:1788–92.

6. Yellowlees PM, Kaushik AV. A case-control study of the sequelae of childhood sexual assault in adult psychiatric patients. Med J Aust 1994;160:408–11.

7. Connors R. Self-injury in trauma survivors: functions and meanings. American Orthopsychiatric Association 1996;66:197–206.

8. Sheldrick C. Adult sequelae of child sexual abuse. Br J Psychiatry 1991;158:55–62.

9. Leenaars AA, De Leo D, Diekstra RW, Goldney RD, Kelleher MJ, Lester D, and others. Consultations for research in suicidology. Archives of Suicide Research 1997;3:139–51.

10. Briere J, Zaidi LY. Sexual abuse histories and sequelae in female psychiatric emergency room patients. Am J Psychiatry 1989;146:1602–6.

11. Brown GR, Anderson B. Psychiatric Morbidity in adult inpatients with childhood histories of sexual and physical abuse. Am J Psychiatry 1991;148:55–61.

12. McCauley J, Kern DE, Kolodner K, Dill L, Schroeder AF, DeChant HK, and others. Clinical characteristics of women with a history of childhood abuse. JAMA 1997;277:1362–8.

13. Otnow Lewis D, Yeager CA, Cobham-Portorreal CS, Klien N, Showalter C, Anthony A. A follow-up of female delinquents: maternal contributions to the perpetuation of deviance. J Am Acad Child Adolesc Psychiatry 1991;30:197–201.

14. Read J. Child abuse and severity of disturbance among adult psychiatric inpatients. Child Abuse Negl 1998;22:359–68.

15. Windle M, Windle RC, Scheidt DM, Miller GB. Physical and sexual abuse and associated mental disorders among alcoholic inpatients. Am J Psychiatry 1995;152:1322–8.

16. Yeo HM, Yeo WW. Repeat deliberate self-harm: a link with childhood sexual abuse? Archives of Emergency Medicine 1993;10:161–6.

17. Anderson G, Yasenik L, Ross CA. Dissociative experiences and disorders among women who identify themselves as sexual abuse survivors. Child Abuse Negl 1993;17:677–86.

18. Angst J, Degonda M, Ernst C. The Zurich study: XV. Suicide attempts in a cohort from age 20 to 30. Eur Arch Psychiatry Clin Neurosci 1992;242:135–41.

19. Arnold L. Women and self injury: a survey of 76 women. Bristol: The Mental Health Foundation, Bristol Crisis Service for Women; 1995.

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

21. Gould DA, Stevens NG, Ward NG, Carlin AS, Sowell HE, Gustafson B. Self-reported childhood abuse in an adult population in a primary care setting: prevalence, correlates and associated suicide attempts. Arch Fam Med 1994;3:252–6.

22. Modestin J, Oberson B, Erni T. possible correlates of DSM-III-R personality disorders. Acta Psychiatr Scand 1997;96:424–30.

23. Mullen PE, Martin JL, Anderson JC, Romans SE, Herbison GP. Childhood sexual abuse and mental health in adult life. Br J Psychiatry 1993;163:721–32.

24. Pettigrew J, Burcham J. Effects of childhood sexual abuse in adult female psychiatric patients. Aust N Z J Psychiatry 1997;31:208–13.

25. 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.

26. 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.

27. Bagley C, Wood M, Young L. Victim to abuser: mental health and behavioral sequels of child sexual abuse in a community survey of young adult males. Child Abuse Negl 1994;18:683–97.

28. Bryant SL, Range LM. Suicidality in college women who report multiple versus single types of maltreatment by parents: a brief report. Journal of Child Sexual Abuse 1995;4:87–93.

29. De Wilde EJ, Kienhorst CWM, Diekstra RFW, Wolters WHG. Social support, life events, and behavioral characteristics of psychologically distressed adolescents at high risk for attempting suicide. Adolescence 1994;29(113):49–60.

30. Kaplan ML, Asnis GM, Lipschitz DS, Chorney P. Suicidal behavior and abuse in psychiatric outpatients. Compr Psychiatry 1995;36:229–35.

31. Silk KR, Lee S, Hill EM, Lohr NE. Borderline personalty disorder symptoms and severity of sexual abuse. Am J Psychiatry 1995;152:1059–64.

32. Links PS, Boiago I, Huxley G, Steiner M, Mitton JE. Sexual abuse and biparental failure as etiologic models in borderline personality disorder. In: Links PS, editor. Family environment and borderline personality disorder. Washington (DC): American Psychiatric Press; 1990.

33. Peters DK, Range LM. Childhood sexual abuse and current suicidality in college women and men. Child Abuse Negl 1995;19:335–41.

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

35. Carlin AS, Ward NG. Subtypes of psychiatric inpatient women who have been sexually abuse. J Nerv Ment Dis 1992;180:392–7.

36. Faye P. Addictive characteristics of the behavior of self-mutilation. Journal of Psychosocial Nursing 1995;33:36–9.

37. Feldman MD. The challenge of self-mutilation: a review. Compr Psychiatry 1988;29:252–69.

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

39. Dubo ED, Zanarini MC, Lewis RE, Williams AA. Childhood antecedents of self- destructiveness in borderline personality disorder. Can J Psychiatry 1997;43:63–9.

40. Working group on preventive practices in suicide and attempted suicide. World Health Organization Summary Report. York, England; 1986 Sept 22–26.

41. Browne A, Finkelhor D. Impact of child sexual abuse: a review of the research. Psychol Bull 1986;99:66–77.

42. Kral MJ. Suicide as social logic. Suicide Life Threat Behav 1994;24:245–55.

43. Stepakoff S. Effects of sexual victimization on suicidal ideation and behavior in US college women. Suicide Life Threat Behav 1998;28:107–26.

44. Briere J, Runtz M. Suicidal thoughts and behaviors in former sexual abuse victims. Canadian Journal of Behavioural Science 1986;18:413–23.

45. Bryer JB, Nelson BA, Miller JB, Krol PA. Childhood sexual and physical abuse as factors in adult psychiatric illness. Am J Psychiatry 1987;144:1426–30.

46. Bagley C, Ramsay R. Sexual abuse in childhood: psychosocial outcomes and implications for social work practice. Journal of Social Work and Human Sexuality 1986;4:33–47.

47. Briere J. The effects of childhood sexual abuse on later psychological functioning: defining a post-sexual abuse syndrome. In: The Third National Conference on Sexual Victimization of Children; 1984; Washington (DC).

48. Sedney MA, Brooks B. Factors associated with a history of childhood experience in a nonclinical female population. Journal of the American Academy of Child Psychiatry 1984;23:215–8.

49. Peters SD. Child sexual abuse and later psychological problems. In: Wyatt GE, Powell GJ, editors. Lasting effects of child sexual abuse. Newbury Park (CA): Sage; 1988.

50. Stone MH. Suicide and suicidal behavior. In: Frances AF, editor. The fate of borderline patients: successful outcome and psychiatric practice. New York: Guilford Press; 1990. p 40–653.

51. Adams DM, Lehnart KL. Prolonged trauma and subsequent suicidal behavior: child abuse and combat trauma reviewed. J Trauma Stress 1997;10:619–34.

52. Read J, Fraser A. Abuse histories of psychiatric inpatients: to ask or not to ask? Psychiatr Serv 1998;49:355–9.


Résumé

Cet article examine le lien entre, d’une part, la violence sexuelle et/ou physique subie durant l’enfance et d’autre part, le comportement autonuisible, les idées et les tentatives de suicide chez l’adulte, de 1988 à 1998 dans la documentation clinique. Malgré les préoccupations méthodologiques et définitionnelles, les études empiriques ont démontré en général davantage de cas de comportements autonuisibles, d’idées de suicide et de comportements suicidaires chez des populations cliniques et communautaires d’adultes qui déclarent de la violence sexuelle et/ou physique durant l’enfance que chez les groupes témoins qui ne déclarent pas de violence. Les répercussions sur la pratique incluent l’identification des populations à risque de comportement autonuisible ou suicidaire et l’interrogation systématique sur les antécédents de violence à l’admission. Il faut plus de recherche sur les définitions universelles des termes clés, les mesures normalisées des variables et les études longitudinales.

Manuscript received August 1998.

1Clinical Leader Manager, Inpatient Mental Health Services, Wellesley Central Site, St Michael’s Hospital; PhD student, Faculty of Nursing and Arthur Sommer Rotenberg Chair in Suicide Studies, University of Toronto, Toronto, Ontario.

2Professor and Associate Dean of Research, Faculty of Nursing; Senior Researcher, Division of Society, Women and Health, University of Toronto, Toronto, Ontario.

Address for correspondence: Ms EE Santa Mina, 7th Floor Turner Wing Wellesley Central Site, St Michael’s Hospital, 160 Wellesley St East, Toronto, ON   M4Y 1J3

email: ruth.gallop@utoronto.ca

Can J Psychiatry, Vol 43, October 1998