The association between child sexual abuse and self-reported sleep in adolescent girls

Abstract

Study Objectives:

Child sexual abuse is associated with negative consequences on both physical and mental health. It has been found to influence child and adolescent sleep, which is an important developmental factor, which also influences mental and physical health. However, the literature examining this topic is plagued by methodological limitations, such as small sample sizes and unvalidated measures.

Methods:

Using the Pittsburgh Sleep Quality Index (PSQI), the present study examined 7 self-reported sleep dimensions, including sleep latency, efficiency, duration, disturbances, subjective sleep quality, daytime dysfunction, and use of sleep medication, in a sample of 707 adolescent girls aged 14–18 years old.

Results:

Statistical analyses revealed that child sexual abuse had a significant moderate association with the total PSQI score. Sexually abused adolescent girls, compared with their nonabused counterparts, reported more sleep difficulties. Significant differences also emerged on all sleep dimensions, with abused girls reporting increased difficulties. However, the magnitude of these associations differed from one dimension to another.

Conclusions:

Our findings support the continued study of sleep in adolescent girls who experienced sexual abuse. They also support the use of a fine-grained perspective when assessing various dimensions of sleep health in girls with a history of child sexual abuse for improved tertiary prevention and intervention.

Citation:

Langevin R, Pennestri M-H, Hershon M, Pirro T, Hébert M. The association between child sexual abuse and self-reported sleep in adolescent girls. J Clin Sleep Med. 2023;19(11):1933–1939.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Studies suggest a negative impact of child sexual abuse on youths’ sleep. However, this literature is plagued by methodological shortcomings, including a lack of differentiation between the various sleep dimensions and the use of unvalidated measures. The current study compares self-reported sleep difficulties of adolescent girls who experienced child sexual abuse with their nonabused peers using the Pittsburgh Sleep Quality Index.

Study Impact: Differences in subjective sleep quality, sleep latency, and the use of sleep medication appeared to be particularly salient when comparing abused and nonabused girls. Given the central role of sleep in healthy development, helping adolescent girls with a history of child sexual abuse sleep better should be the focus of sustained efforts in research and practice.

INTRODUCTION

Child sexual abuse (CSA) is conceptualized as nonconsensual sexual behaviors, with or without contact, inflicted on children under the age of 18 years.¹ CSA is a major social concern, with a worldwide prevalence of 8% for men and more than double this rate for women (20%).² In comparison to other maltreatment experiences, CSA is unique given the combined effect of 4 traumagenic dynamics—namely, traumatic sexualization, betrayal, stigmatization, and powerlessness.³ As such, victims of CSA report experiencing more negative consequences than victims of other forms of child maltreatment, including internalizing and externalizing behavior problems, neurodevelopmental (eg, altered neurogenesis), and psychological (eg, mood disorders) challenges.^4,5 Systematic reviews have concluded that CSA in children and adolescents is also associated with sleep problems, such as nightmares and nocturnal awakenings; this effect is even more pronounced for girls.⁶ Although there is an emerging body of literature on the topic of CSA and sleep in adolescent girls, these systematic reviews have outlined important methodological shortcomings in the current literature (eg, small sample sizes, use of unvalidated measures, and noncomprehensive assessment of sleep in terms of dimensions or difficulties).⁶ Therefore, the current investigation aims to bridge some of these gaps by comparing sexually abused and nonabused adolescent girls on various self-reported sleep dimensions.

Sleep is a fundamental component of one’s life, especially in highly sensitive developmental periods, such as early childhood and adolescence.⁷ It is associated with healthy brain and affective development, mental and physical health, academic outcomes, learning, and memory.^8–10 Medical guidelines recommend that typical adolescents sleep 8 to 10 hours to ensure adequate functioning.¹¹ Despite this recommendation, in a large Canadian sample, Michaud and Chaput¹² found that 30% of adolescents sleep less than the recommended guidelines for optimal health, and 42% of adolescents report having trouble falling or staying asleep. Furthermore, it has been shown that adolescent girls report more sleep problems than boys, including longer sleep-onset latency and poorer subjective sleep quality.^13,14 These sex differences remain in adulthood, as insomnia is more prevalent in women than men.¹⁵ Therefore, sleep difficulties pose a threat to healthy developmental trajectories, even in nonclinical populations of adolescent girls.

In response to experiences of trauma, it is common to experience sleep difficulties.^6,16,17 Specifically, after experiencing CSA, posttraumatic stress disorder, anxiety, and depression symptoms often develop and are associated with and even include sleep difficulties as part of the core symptomatology.^16–20 As a result, CSA has the potential to interfere with several sleep dimensions, particularly the capacity to fall asleep, stay asleep, and get restful sleep, directly and indirectly through mental health difficulties. These sleep difficulties have all been identified in systematic reviews on the topic.⁶ However, these reviews covered a large age range, with only a few studies focusing more specifically on adolescents, despite the age-specific developmental aspects of sleep. For instance, Mignot et al²¹ found that adolescent victims of CSA reported lower sleep satisfaction and more frequent nightmares and nocturnal awakenings in comparison to their nonabused counterparts. However, this was documented in a mixed sample of 1,719 girls and boys using unvalidated measures of sleep and CSA. Moreover, in a small sample of 24 adolescent girls using an unvalidated sleep measure, Keeshin and colleagues²² found that victims of sexual abuse reported experiencing more sleep difficulties, measured as either having difficulty falling asleep, nightmares, or more awakenings, than their nonabused counterparts.

Taken together, these studies suggest a negative impact of CSA on sleep in adolescent girls. However, this literature is limited by the reliance on small samples, little consideration for developmental specificities as well as sex and gender differences related to sleep and CSA, a lack of differentiation between the various sleep health dimensions, and most importantly, the use of unvalidated measures to document sleep and CSA.⁶ In this context, the current study aimed to compare self-reported sleep difficulties of adolescent girls who experienced CSA with their nonabused peers using a gold-standard questionnaire, the Pittsburgh Sleep Quality Index (PSQI). Given that sleep is a multidimensional concept²³ and that CSA could impact some sleep dimensions more strongly than others,⁶ the present study also examined self-reported sleep latency, efficiency, duration, disturbances, and quality, as well as daytime dysfunction, and use of sleep medication. It was hypothesized that sexually abused girls would report having overall worse sleep than nonabused girls. Given the sparse empirical literature on the topic, no hypotheses were formulated regarding specific sleep dimensions that may be more impacted by CSA.

METHODS

Participants

A sample of 776 adolescent girls aged 14 to 18 years (mean = 16.04, standard deviation [SD] = 1.28) from the Province of Quebec, Canada, was recruited online using the secure platform Qualtrics (Provo, UT). Most participants self-identified as Canadians (88.1%). They were allowed to select more than 1 ethnicity and 3.4% identified as Indigenous/Metis, 4.0% as Latinx, 3.7% as Black, 2.8% as Asian, 8.8% as West European, 2.1% as East European, 2.5% as Caribbean, and 2.5% as North African/Middle Easterner. Using their postal codes,²⁴ participants were classified in quintiles based on their neighborhood’s material deprivation index. One-quarter of participants lived in very privileged neighborhoods; about one-fifth of participants lived in either privileged, average, or underprivileged neighborhoods; and about 15% of participants lived in very underprivileged neighborhoods. After the data were cleaned to exclude noneligible participants, a sample of 707 adolescent girls remained: 54 were excluded because they did not self-identify as girls, 13 because they were not aged between 14 and 18 years old, 1 because she did not live in Quebec, and 1 because she did not answer the question documenting CSA.

Procedure

Participants wanting to partake in the study were recruited by means of posting on social media and were asked to give their informed consent electronically. In accordance with the regulation in effect in the Province of Quebec, parental consent was not required since participants were 14 years and older. Participants were then asked to fill out an anonymous online survey on Qualtrics (Provo, UT), which took approximately 25 minutes to complete. Participants who completed the survey were then entered in a drawing to win a $50 gift card. The current procedure was approved by the Research Ethics Board of the Université du Québec à Montréal.

Measures

Sociodemographics

Participants were asked to complete a sociodemographic questionnaire. Of relevance to the current analysis, ethnicity/cultural background and date of birth were obtained.

Self-reported sleep

Self-reported sleep was measured with the PSQI.^25,26 The PSQI examines 7 dimensions of sleep using participant reports on a 4-point Likert scale ranging from 0 to 3 (eg, “not during the past month” to “three or more times a week”; “very good” to “very bad”). Documented sleep dimensions include the following: (1) subjective sleep quality (1 item), (2) sleep latency (2 items), (3) sleep duration (1 item), (4) sleep efficiency (time spent asleep/time in bed), (5) sleep disturbances (9 subitems), (6) use of sleeping medication (1 item), and (7) daytime dysfunction (2 items). In line with the original coding instructions, answers were recoded into a 0–3 score for each of the 7 dimensions,²⁵ with higher scores reflecting greater difficulties. A total aggregate score was obtained by summing all the PSQI components, resulting in a total score ranging from 0 to 21, with increasing scores reflecting increasing sleep difficulties. In the original validation study, the PSQI had a Cronbach’s alpha of .83.²⁵ The PSQI was also validated for use with community-based adolescents.²⁷ In the current study, Cronbach’s alpha is .67. This questionnaire was validated in a French sample.²⁷

Child sexual abuse

Having a history of CSA was self-reported using an item adapted from previous studies: “Has anyone touched you sexually when you did not want it or has anyone coerced or manipulated you into having sex.”

Statistical analysis

Analyses were conducted using SPSS version 28 (IBM Corporation, Armonk, NY). A visual examination was done to identify anomalies in answers and scores, and anomalies were recoded as missing (some PSQI scores for 3 participants). Spearman’s correlations between dependent variables were computed. A t test was run to compare CSA and non-CSA participants on age, and a chi-square analysis was performed to compare them on nationality. Since there was no age or nationality difference between groups, further analyses were performed without covariates. As scores for all sleep dimensions are ordinal and not normally distributed, Mann-Whitney U tests were deemed suitable to compare groups on sleep dimensions. A t test was used for the total sleep score comparison. Preliminary analyses were done to compare the distributions of scores in the CSA and no-CSA groups and to determine if the Mann-Whitney U test should be used comparing the medians or the mean ranks between groups.

RESULTS

Preliminary analyses

In our final sample, 62% of adolescent girls reported no history of CSA (n = 439) and 38% of them did (n = 268). All PSQI scores were correlated with one another in the complete sample (Table 1). Table 2 displays means, SDs, medians, and sample sizes for each PSQI dimension and the total PSQI score separated by groups. Between 0.43% and 4.66% of participants had missing data on a PSQI dimension and 6.78% were missing for the total score; listwise deletion was used with these cases. Participants in the CSA and no-CSA groups did not significantly differ in terms of age (mean_CSA = 16.16 ± 1.24; mean_no-CSA = 15.97 ± 1.31; t[705] = –1.875, P = .061) and nationality (CSA = 89.9% Canadians; no-CSA = 87.0% Canadians; χ² [1, n = 706] = 1.370, P = .242). Since sleep efficiency, subjective sleep quality, sleep duration, use of medication, and daytime dysfunction were distributed similarly across groups, the Mann-Whitney U test was used to compare medians. The Mann-Whitney U test was used to compare mean ranks for sleep latency and sleep disturbances, considering that their distributions were different in both groups.

Table 1.

Spearman’s rho correlations between study variables.

	1	2	3	4	5	6	7
1. Subjective sleep quality	–
2. Sleep latency	.384***	–
3. Sleep duration	.452***	.236***	–
4. Habitual sleep efficiency	.230***	.255***	.377***	–
5. Sleep disturbances	.287***	.287***	.152***	.183***	–
6. Use of sleep medication	.189***	.232***	.098*	.136***	.217***	–
7. Daytime dysfunction	.325***	.167***	.252***	.110**	.223***	.152***	–
8. Total sleep score	.669***	.671***	.550***	.502***	.506***	.524***	.538***

*P < .05; **P < .01; ***P < .001.

Table 2.

Description of PSQI scores.

Scores	CSA	No-CSA	CSA	No-CSA	CSA	No-CSA	Statistical Test: Mann-Whitney U/t Test
	Mean (SD)		Median		n
Subjective sleep quality	1.64 (0.73)	1.39 (0.71)	2	1	267	437	68,788.00***
Sleep latency	1.94 (0.97)	1.51 (1.00)	2	1	260	422	41,840.00***
Sleep duration	0.60 (0.83)	0.34 (0.69)	0	0	252	427	62,680.00***
Habitual sleep efficiency	0.60 (0.90)	0.42 (0.76)	0	0	250	424	58,233.00**
Sleep disturbances	1.52 (0.61)	1.30 (0.54)	1	1	267	437	46,858.00***
Use of sleep medication	0.93 (1.25)	0.55 (0.99)	0	0	266	437	66,650.50***
Daytime dysfunction	1.91 (0.87)	1.70 (0.85)	2	2	267	437	66,221.50**
Total sleep score (t test)	9.11 (3.50)	7.18 (3.16)	9	7	246	413	−7.287***

The scores displayed here are from the 0-3 score for each sleep dimension except for the total score that can range from 0 to 21. **P < .01; ***P < .001. CSA = child sexual abuse, SD = standard deviation.

PSQI total sleep score

The total sleep problems score was significantly higher in sexually abused adolescents than in nonabused adolescents (see Table 2). The Cohen’s d value was –0.587, representing a robust effect size of CSA on the PSQI total score. Using total scores of more than 5 as the clinical cutoff,²⁵ chi-square analyses revealed that girls in the CSA group were more likely than girls in the no-CSA group to report clinical levels of sleep difficulties (86.6% vs 67.1%: χ² [1, n = 659] = 30.79, P < .001).

PSQI sleep dimensions

Results displayed in Table 2 show that the subjective sleep quality dimension differed between groups, with sexually abused girls reporting lower sleep quality than nonabused girls. More than half of nonabused girls (58.8%) compared to 43.4% of nonabused girls reported good or very good sleep quality. Significant but small differences also emerged with the sleep duration dimension. Specifically, when looking at the mean hours of sleep per night, girls in the CSA group reported 6.9 hours per night (SD = 1.34 hours; median = 7, 25th percentile = 6, 75th percentile = 8) and nonabused girls reported sleeping 7.3 hours per night (SD = 1.28 hours; median = 7, 25th percentile = 7, 75th percentile = 8). The CSA group also had a statistically lower sleep efficiency dimension than the no-CSA group. Sexually abused adolescent girls reported slightly lower sleep efficiency than nonabused girls; the mean percent efficiency was 89% (SD = 10.4%; median = 91%, 25th percentile = 83%, 75th percentile = 91%) for nonabused girls and 87% (SD = 12.4%; median = 89%, 25th percentile = 80%, 75th percentile = 100%) for abused girls. Girls in the CSA group reported using more sleep medication than those in the no-CSA group. An examination of the distributions of scores shows that nonabused girls reported not using sleep medication in the past month in 71.6% of cases, while this was found to be in 59.4% of cases for abused girls. Conversely, 9.8% of nonabused girls reported using sleep medication ≥ 3 times per week in comparison to 22.6% in the abused group. Sexually abused adolescent girls reported greater daytime dysfunction compared with nonabused girls. More specifically, while examining the distributions, one-quarter to one-third (28.5%) of sexually abused girls reported frequent difficulties staying awake and maintaining their enthusiasm for activities, while this proportion was one-fifth for nonabused girls (19.2%).

Furthermore, results showed a significant difference between the CSA and no-CSA groups in the sleep latency dimension, with abused girls showing longer latency than nonabused girls. An examination of the distributions in minutes revealed that sexually abused adolescent girls reported a mean sleep latency of 43 minutes (SD = 33 minutes; median = 30, 25th percentile = 20, 75th percentile = 60), while their non–sexually abused counterparts had a mean latency of 32 minutes (SD = 28 minutes; median = 20, 25th percentile = 15, 75th percentile = 45). Similarly, adolescent girls in the CSA group reported greater sleep disturbances than those in the no-CSA group. Thus, by examining the distributions of scores, non–sexually abused girls reported no or few sleep disturbances in 69.8% of cases, while this was only true for 50.9% of sexually abused girls.

DISCUSSION

This study aimed to compare sexually abused and nonabused adolescent girls on various self-reported sleep dimensions using a large sample and a gold-standard measure of sleep. In line with our hypotheses, results showed that sexually abused adolescent girls, compared with their nonabused counterparts, reported more sleep difficulties both in terms of their continuous total scores and clinical levels of difficulties. CSA had a robust effect size on the total PSQI score. While significant differences emerged on all sleep dimensions, differences in terms of sleep duration and efficiency were small and likely not clinically significant. However, more important differences emerged for subjective sleep quality, sleep latency, and use of sleep medication. While most nonabused girls were satisfied with their sleep quality, more than half of abused girls were dissatisfied. Sexually abused girls reported taking approximately 30% more time to fall asleep (43 vs 32 minutes), and twice as many of them, despite their young age, reported a frequent use of sleep medication. Relatedly, moderate differences were apparent in the amount of sleep disturbances and of daytime dysfunction reported in both groups. These findings support the relevance of adopting a more fine-grained perspective when assessing sleep in sexually abused adolescent girls for research or intervention purposes.

Our findings mostly align with the previous body of literature on sleep and CSA. Xiao and colleagues²⁸ found that adolescent victims of CSA had heightened sleep problems as measured by the PSQI when compared with nonabused adolescents. Similar findings were reported with a sample of Turkish adolescents²⁹ as well as with younger children in foster care.³⁰ While our results regarding sleep duration and efficiency were likely not clinically significant, past studies also have found significant differences on these sleep variables both with female-only and mixed samples.^21,22,31,32 However, in a mixed child and adolescent sample of boys and girls, Sadeh and colleagues³³ reported no associations between CSA and sleep efficiency, as measured by nocturnal awakenings.

Additionally, consistent with our findings, previous literature has shown that adolescent girl and boy victims of CSA were less satisfied with their sleep than their nonabused peers.²¹ In contrast, Wamser-Nanney and Chesher³⁴ found no difference on sleep satisfaction; this difference may be attributed to the mixed girls and boys sample used, given that the current investigation used a girls-only sample. Furthermore, Mignot et al²¹ and Turner et al³² reported findings similar to ours where CSA victims reported greater difficulties falling asleep than nonvictims in mixed samples of boys and girls. However, this was also found in a sample of boys only.³⁵ Nonetheless, studies did not find an association between CSA and sleep-onset latency in samples of children and adolescents.^33,34 Overall, our findings are consistent with the current body of literature, emphasizing that sleep difficulties are more prevalent in victims of CSA.

Lower subjective quality and longer sleep latency for CSA victims compared with nonvictims were particularly salient in the present study. This finding is consistent with the hyperarousal theory underlying both insomnia and trauma etiology. Specially, studies conducted in patients with insomnia show increased hyperarousal state as measured by several markers, such as higher frequency electroencephalographic activity, increased blood pressure, elevated heart rate, and increased cortisol level.^36–38 Low sleep quality in CSA individuals likely reflects high levels of vigilance and physiological arousal during sleep, symptoms that are integral to posttraumatic stress.^39,40 Moreover, fear of sleep may play an important role in reducing sleep quality and increasing sleep latency⁴¹ in sexually abused girls. Although present in several types of traumas, fear of sleep might be particularly associated with CSA, since these traumatic experiences often occur in a sleep environment (eg, in bed) at night or in the dark. Future studies could directly examine these potential mechanisms underlying the associations between CSA and sleep difficulties.

The increased use of sleep medication reported by sexually abused girls in our sample is also an important finding. In adults, insomnia is a highly prevalent disorder, with a clearly higher prevalence in women.¹⁵ Women are also more likely to use sleep medication or to report a need to treat their sleep difficulties.^42,43 While the psychophysiological mechanisms leading to insomnia are not completely understood, it is believed to be multifactorial and often the result of the presence of predisposing factors (eg, genetic vulnerability), precipitating factors (eg, acute stress), and perpetuating factors (eg, poor sleep habits). Therefore, early traumatic experiences such as CSA likely precipitate the development of insomnia in some women. Moreover, the use of sleep medication and poor sleep habits could act as perpetuating factors. While the PSQI does not provide details regarding the specific medication used, considering that nonpharmacological interventions are preferable as an effective and safe first-line therapy for insomnia,⁴⁴ it remains concerning to observe high rates of medication usage in this community sample of adolescent girls following CSA. Whether this finding is generalizable to other young vulnerable populations remains to be determined.

While our study did not aim to document sleep difficulties in a general population of adolescent girls, some of our findings to that effect warrant a short discussion. Consistent with past studies,^12–14 our results reveal significant sleep difficulties in both abused and nonabused girls. A striking result is that almost 68% of nonabused girls reached the clinical threshold of the PSQI for sleep difficulties. Many nonabused girls reported using sleep medication in the past month (28%), experiencing daytime dysfunction (19%), and experiencing some form of sleep disturbances (30%). Both abused and nonabused girls reported sleeping, on average, 8 hours per night, which is the minimum recommended for adolescents. Thus, our findings align with the “perfect storm” metaphor proposed by Carskadon.⁴⁵ This metaphor highlights that adolescents are faced with conflicting bioregulatory (ie, circadian phase delay), psychosocial (eg, self-selected bedtimes, social networking, and availability/use of technology delaying sleep), and societal (ie, early rise for school) pressures regarding their sleep timing that result in insufficient sleep and daytime dysfunction. Given the major implications of sleep for physical and psychological health and for healthy development in adolescence, these concerning results emphasize the need to more closely examine sleep in girls, sexually abused or not.

Limitations and implications for research and practice

The current study has many strengths including the large sample of adolescent girls, with over 250 of them reporting a history of CSA, affording sufficient statistical power to conduct our analyses. The use of a gold-standard measure of subjective sleep difficulties is also a strength and so is the examination of various sleep health dimensions. Notwithstanding these contributions, some limitations must be acknowledged. There is some debate regarding the use of the PSQI dimension scores, and the factor structure with 7 dimensions was not replicated in the validation study with an adolescent sample.²⁷ We did not use a validated measure to assess CSA and only self-reports were considered; a multimethod assessment of CSA (eg, official records and validated self-report with several questions) and the examination of CSA characteristics (eg, severity and chronicity) that may be differentially associated with sleep dimensions are desirable in future research. We had limited information on the sociodemographic characteristics of our participants (eg, schooling, income, and sexual orientation) and their family of origin (eg, family income and parent education and occupation). Based on the information we had, our sample was not very ethnically/racially diverse; therefore, our findings might not be representative of more diverse populations of adolescent girls. The high percentage of our sample reporting experiencing sexual abuse (39%) compared with other studies^46,47 could reflect some biases that may have been introduced by our online recruitment strategy targeting adolescent girls who wished to share their experiences in romantic relationships. It could also be related to the item that we used that defines CSA rather broadly compared with measures used in other studies, or to the increased awareness brought upon by recent movements such as #MeToo.⁴⁸ Future studies could replicate our findings using a more diverse and representative population. Given possible sex specificities identified in past studies, exploring the associations between CSA and the different dimensions of sleep in boys would also be interesting and provide relevant cues for intervention.

Expanding on our study, future research could explore the interplay between CSA, sleep, and posttraumatic symptomatology and consider possible mediating and moderating factors of the association between CSA and sleep in adolescent girls (eg, psychopathology, CSA characteristics, and social and professional support received). Exploring the impact of sleep difficulties in CSA survivors’ recovery would be another worthy avenue for future research. Longitudinal studies of CSA survivors could also help determining the importance of early trauma in the development and maintenance of adulthood insomnia and other sleep difficulties, such as recurrent nightmares. Using a quantitative-qualitative mixed methodology could provide rich information on the lived experiences and individual perspectives of victims regarding potential mechanisms underlying sleep difficulties following CSA. This could also contribute to the development of a theoretical model of the impact of CSA on sleep that could orient future research and clinical practice.

Our findings have some implications for health practitioners, especially if replicated. Screening for sleep problems in adolescent girls with a history of CSA could be relevant and a fine-grained assessment of the sleep health dimensions impacted should follow to orient intervention. Cognitive behavioral therapy for insomnia is the recommended treatment for insomnia in adults, but also in adolescents, and should be prioritized over medication.⁴⁹ Adolescence is already a vulnerable developmental period regarding sleep, where biological, psychological, and sociocultural influences interact, leading to sleep loss and decreased sleep regulation.^45,50 Therefore, very special attention should be paid to vulnerable adolescents, such as sexually abused adolescent girls.

CONCLUSIONS

The current study contributed to the literature by documenting associations between various sleep dimensions and having a history of CSA in adolescent girls. Differences in subjective sleep quality, sleep latency, and the use of sleep medication appeared to be particularly salient when comparing abused and nonabused girls. More research is needed to deepen our understanding of the mechanisms underlying these associations. Given the central role of sleep in healthy developmental trajectories, helping adolescent girls with a history of CSA sleep better could foster their recovery and resilience; this could be the focus of sustained efforts in research and practice.

DISCLOSURE STATEMENT

All authors have seen and approved the manuscript. Work for this study was performed at the Université du Québec à Montréal. Drs. Langevin and Pennestri are supported by Chercheur-Boursier Awards from the Fonds de Recherche du Québec–Santé. They both hold a William Dawson Award from McGill University. This project was supported by a grant from Canada Research Chairs Program awarded to Dr. Hébert. Teresa Pirro and Malka Hershon have no financial support to declare. The authors report no conflicts of interest.

ABBREVIATIONS

CSA

child sexual abuse

PSQI

Pittsburgh Sleep Quality Index

SD

standard deviation