Exploratory Study of the Clinical Characteristics of Adolescent Girls with a History of Physical or Sexual Abuse Consulting in a Mood Disorder Clinic

Abstract

Objective

To examine the clinical characteristics of adolescent girls consulting in a mood disorder clinic with a history of physical or sexual abuse.

Method

A retrospective review was conducted of the charts of 55 adolescent girls consulting in a mood disorder clinic. An analysis grid was used to gather data on demographics, personal antecedents, symptoms and diagnoses. Girls with a history of physical or sexual abuse were compared with girls without such a history. Univariate analyses and a logistic regression analysis were performed.

Results

Adolescent girls with a history of physical or sexual abuse did not differ statistically from those without such a history in terms of depressive symptoms or type and number of diagnoses. However, proportionally more girls with a history of physical or sexual abuse presented self-harm and relational problems with their parents and peers.

Conclusion

Both history of physical or sexual abuse and self-destructive behaviors are rooted in relational problems. The results show that these are related to one another among those adolescent girls. Clinically, these findings suggest that it is important for clinicians do a thorough exploration of self-destructive behaviors and family and peer relations when assessing depressed adolescent girls.

Introduction

Childhood abuse carries well known neurobiological and psychiatric consequences that persist into adulthood. Studies have shown that up to 82% of abused youths develop at least one psychiatric disorder later in life (Molnar, Buka, & Kessler, 2001). History of childhood or adolescent abuse has been found to be associated with a higher incidence of mood disorders (Finzi et al., 2001; Widom, DuMont, & Czaja, 2007) and other psychiatric disorders, particularly in women (Molnar et al., 2001). Moreover, depression in this population is characterized by earlier onset (Widom et al., 2007), more recurrences, lower remission rate (Nanni, Uher, & Danese, 2012), greater severity (Danielson, de Arellano, Kilpatrick, Saunders, & Resnick, 2005), and increased risk for chronic course (Nanni et al., 2012). History of abuse has also been associated with higher risk of attempting suicide in youths (Shaunesey, Cohen, Plummer, & Berman, 1993).

While physically or sexually abused youth evaluated in psychiatry have been repeatedly found to present more depressive symptoms (Lewis et al., 2010; Shamseddeen et al., 2011), as well as more suicidality and self-injurious behaviors (Shamseddeen et al., 2011) when compared to never abused youth, many studies conducted on psychiatric inpatients did not find any differences in the prevalence of mood symptoms (Y. Cohen et al., 1996; Monane, Leichter, & Lewis, 1984). To our knowledge, there are no studies investigating the impact of a history of physical or sexual abuse (P/SA) in adolescents attending a mood disorder clinic. Since abused adolescents have higher risk of future psychiatric disorders, poorer response to treatment and a poorer outcome, and since rate of abuse disclosure is very low, especially in young victims (Pipe, Lamb, Orbach, & Cederborg, 2007), clinicians will benefit from a better knowledge of clinical characteristics suggesting past victimization.

Objective

The psychiatric management of adolescent girls with a history of P/SA can be challenging for mental health professionals given the high prevalence of depressive disorders, suicidal ideation and suicidal attempts in this group of patients. How do adolescent girls with such a history of P/SA consulting in a mood disorder clinic differ from girls without such a history? What are the demographics, personal antecedents, symptoms and diagnoses associated with abuse among these adolescents? Answering these questions could help develop screening and intervention strategies adapted to these high-risk adolescents. This chart review being conducted in a clinical setting, on the basis of the information that these girls have shared with mental professionals during a clinical evaluation, it is clinically relevant and interesting as it is very close to information gather during daily “in the field” clinical work.

Methodology

A retrospective chart review of all adolescents consulting in a pediatric mood disorder clinic in Montreal from August 2005 to October 2007 was carried out to collect data regarding the demographics, personal and familial antecedents, symptoms and diagnoses of these adolescents. Access to medical records was authorized by the Director of Professional Services, as required by the hospital’s Ethics Committee.

Of the 75 adolescents 12 to 17 years of age (mean: 15.4), 55 (73%) were female. Of these, 22 (40%) reported a history of P/SA and 33 (60%) had no such history.

The authors used an analysis grid from earlier studies of depressed children and adolescents in child psychiatric care (Breton et al., 2012) to review charts and collect demographic and clinical data. A large majority of the patients were evaluated by a child psychiatrist, a nurse, a psychologist, a social worker and an occupational therapist. The present study being part of a larger study on protective factors for depression and suicidal behavior among adolescents (Breton et al., 2015), the Schedule of Affective Disorders and Schizophrenia for School-Age Children – Present and Longitudinal Version (K-SADS-PL) (Kaufman et al., 1997), which is a semi-structured diagnostic interview, was also used to evaluate patients. Personal antecedents such as past psychiatric and medical history, past suicide attempts, alcohol and drug use and the presence or absence of all other variables were systematically inquired about during the multidisciplinary evaluation. Clinical diagnoses were established by the same child psychiatrist based on the DSM-IV-Text Revision criteria (American Psychiatric Association, 2000), using the best-estimate method and taking into account results from the K-SADS-PL. The identification of P/SA, was based on all the information available in the chart following the multidisciplinary team evaluation during which questions about past events and abuse were asked to the adolescent and her parents/tutor when available. However, no legal verification was done. Physical abuse ranged from being struck once with enough force to leave a mark to suffering much more severe trauma at the hands of someone else. Sexual abuse ranged from being forced to view sexual activities to being raped. Abuse could have happened only once or repeatedly. It was not necessarily related to the reason for consulting in the clinic. For power consideration, both type of abuse (physical and sexual) were put into one categorical variable. Unfortunately, the charts contained little information concerning psychological abuse and parental neglect. This information was therefore not considered for the study.

Statistical analyses

The adolescent girls with a history of P/SA were compared against those without such a history. Associations of demographic and clinical variables (see Table 1 and 2) with a reported history of P/SA were verified using Pearson’s chi-squared test or Fisher’s exact test, whichever was appropriate. Statistical significance was set at p < .05. An important effect is detected with a statistical power of 95% with this p value (Cohen, 1998). Given the exploratory nature of the study, tendencies with a p ≤ .10 are also reported in the tables with no correction for the multiplicity of tests. This p value allows for the detection of a medium effect with a power of 0.71% (Cohen, 1998). We report these statistical tendencies being aware that the differences indeed do not reach the accepted significance level but actually can be clinically meaningful. A logistic regression analysis was run on the significant variables to emerge from the univariate analyses. The model was build using a forward strategy with the inclusion threshold set at p < .05 for the likelihood ratio test (Hosmer & Lemeshow, 1989).

Table 1.

Demographics characteristics of the two groups

	History of physical or sexual abuse		Tests statistic
	Presence (n = 22) n (%)	Absence (n = 33) n (%)	p
Quebecer	16 (73)	27 (82)	NSb
Family structure other than with biological parents	19 (86)	22 (67)	.10a
No contact with one parent	8 (36)	6 (18)	NSb
Foster care	3 (14)	0 (0)	.06b
Major economic difficulties	6 (27)	6 (18)	NS b

NS: non-significant;

^aPearson’s chi-squared test;

^bFisher’s exact test

Table 2.

Personal antecedents and symptoms characteristics of the two groups

	History of physical or sexual abuse
Variables	Presence (n = 22) n (%)	Absence (n = 33) n (%)	Tests statistic p
Personal antecedents
Medical antecedents	14 (67)	12 (61)	NS a
Alcohol use	8 (36)	11 (33)	NS a
Drug use	11 (50)	9 (27)	.09 a
At least one suicide attempt	15 (68)	16 (49)	NS a
At least one grade repeated	4 (20)	6 (19)	NS b
Symptoms
Suicidal ideation	21 (96)	28 (85)	NS a
Depressive ideas	20 (91)	28 (85)	NS b
Low self-esteem	16 (73)	21 (64)	NS a
Hopelessness	9 (41)	9 (27)	NS a
Self-harm	19 (86)	19 (58)	.02 a
Loss of interest	7 (32)	8 (24)	NS a
Rapid mood swings	8 (36)	11(33)	NS a
Sleep problems	14 (64)	21 (64)	NS a
Aggressiveness	13 (59)	12 (36)	.09 a
Opposition	9 (41)	7 (21)	NS a
Parent-adolescent relational problems	17 (77)	17 (55)	.05 a
Academic underachievement	16 (73)	15 (46)	.04 a
Rejected by peers	4 (18)	11 (33)	NS a
Isolated/only few friends	12 (55)	13 (39)	.02 a
Intimidates or physically abuses others	3 (14)	5 (15)	NS b
Intimidated by peers	3 (14)	7 (21)	NS b

NS: non-significant

^aPearson’s chi-squared test

^bFisher’s exact test

Results

Demographics

No significant differences between the two groups were found regarding demographic data, although tendencies were observed for family structure other than with biological parents and foster care (Table 1).

Personal antecedents

No significant differences between the two groups were found regarding personal antecedents although a tendency was observed for drug use (Table 2).

Symptoms

No significant differences between the two groups were found regarding depressive symptoms (e.g., suicidal ideation, depressive ideas, low self-esteem) and type and number of Axis 1 diagnoses (e.g., depressive disorders, bipolar disorders, disruptive disorders, etc) or Axis 2 pathological personality traits. No statistical differences were observed for family psychiatric history as well (data not shown). A significantly higher proportion of girls with a history of P/SA had engaged in self-harm (86% vs. 58%, p < .02), had parent-adolescent relational problems (77% vs. 55%, p < .05), underachieved at school (73% vs. 46%, p < .04), and were isolated socially (55% vs. 39%, p < .02) (see Table 2).

According to our logistic regression model (Table 3), into which the above mentioned significant variables were introduced, two variables were associated with a history of P/SA among adolescent girls: self-harm and parent-adolescent relational problems. The odds ratios for these were 6.4 and 4.5, respectively.

Table 3.

Correlates of history of physical or sexual abuse among adolescent girls in a mood disorder clinic: Multiple logistic regression analysis (n = 55)

Variables	n (%)	History of physical or sexual abuse		Likelihood ratio test p
		Odds ratio	95% CI
Self-harm	38 (69)	6.4	1.5 to 28.4	.007
Parent-adolescent relational problems	34 (62)	4.5	1.2 to 16.5	.018

Discussion

A study based on the retrospective review of clinical files presents limitations, as the information collected by clinicians may be incomplete. However, all the patients in this study were thoroughly evaluated by the same child psychiatry team. Unfortunately, we could not carry out specific analyses by abuse type.

In our study, the prevalence of mood disorder or depressive symptoms was not significantly higher among girls with a history of P/SA compared with those with no such history. This runs counter to the findings of an earlier study (Finzi et al., 2001). The lack of difference between the two groups might in part be explained by the fact that all patient were referred to this specialized outpatient clinic because they presented with mood symptoms (selection bias).

Regarding suicide, 68% of girls with a history of P/SA had made at least one attempt compared with 49% of girls without such a history. Though the difference is not statistically significant (p = .14), it is clinically meaningful and should be of concern to clinicians. The association between history of abuse and attempted suicide is well documented in the literature (Finzi et al., 2001; Shaunesey et al., 1993) and this behavior is strongly associated with further suicide attempts and completed suicide (Owens, Horrocks, & House, 2002).

Furthermore, proportionally more adolescent girls with a history of P/SA than girls without such a history engaged in self-harm, as reported in other studies (Shamseddeen et al., 2011). Moreover, girls with a history of abuse displayed statistically significant relational problems with parents and with peers and although not statistically significant, came more often from broken families (see Table 1) as reported by others (Carey et al., 2008; Trocmé et al., 2005). Although no significant differences were observed between the two groups in terms of depressive symptoms and diagnoses, what clearly emerges from this exploratory study is a picture of self-destructive behaviors and relational problems.

Conclusion

Self-destructive behaviors and P/SA are both destructive and occur in the context of relationships. It is not possible to know from such an exploratory study which of the two destructive phenomenon comes first but results suggest that they may be associated in abused adolescent girls. This study emphasizes the relevance for clinicians to have a better knowledge of clinical characteristics suggesting a history of abuse in adolescent, especially since rate of abuse disclosure is known to be very low. Identification of a history of abuse also has therapeutic implications. It has been associated with poorer outcomes and resistance to treatment in youths suffering from depressive symptoms particularly when psychotherapy is used (Lewis et al., 2010; Shamseddeen et al., 2011).