|Home | About | Journals | Submit | Contact Us | Français|
Intimate partner violence puts the victim at risk for substantial medical and psychiatric morbidity. As with other stress- and trauma-related experiences, intimate partner violence is associated with sleep disturbance, particularly insomnia and nightmares. This association, however, has not been well characterized in terms of general prevalence or its further relationship with depression, suicidality, and posttraumatic stress disorder (PTSD).
The present study used validated instruments to characterize insomnia and nightmares among 121 women exposed to intimate partner violence. Participants with and without depression were compared on demographic, abuse, and sleep characteristics as were those with and without suicidality. Logistic regression models were constructed to test sleep variables as independent predictors of depression controlling for demographic factors, abuse severity, and PTSD severity.
Clinically significant insomnia and nightmares were observed in 46% and 32% of participants, respectively. Depressed women had more severe PTSD and were more likely to have insomnia and to have nightmares than nondepressed women. In models controlling for PTSD severity, the presence of insomnia was associated with an approximately eightfold greater risk of being depressed; nightmares were associated with a twofold increase in risk.
Sleep disturbances were prevalent among women experiencing intimate partner violence, with both insomnia and nightmares predicting the presence of depression even after controlling for PTSD severity. In addition to the need to address common mental health issues such as depression, given that sleep problems are modifiable and potentially less stigmatizing than mental health problems, assessing and addressing insomnia and nightmares in survivors of interpersonal violence warrants strong clinical consideration and further investigation.
Intimate partner violence (IPV) is a public health issue. Approximately 20% of individuals who responded to a national telephone survey reported IPV.1 Abuse ranges have been reported to be greater than 44% nationally and abroad.2–5 The abuse can be physical, psychological, or sexual. However, all three types of abuse are known to result in negative health consequences.6–9 Mental health consequences of IPV include depression, anxiety, posttraumatic stress disorder (PTSD), and chronic fatigue.10,11 Suicidal ideation and suicide attempts are more prevalent for those who report being victimized than those who report no IPV.12 Approximately half of the women admitted for inpatient psychiatric care in the wake of suicide attempts, disclose IPV.13 Thus, both the epidemiology of and the clinical relationships between IPV and its mental health consequences have been studied.
One area that has received more limited attention in the IPV literature is the issue of sleep disturbance. Moreover, the intersection of IPV, mental health, and sleep disturbance has been largely ignored. This is an important intersection given the potential impact of sleep disturbance on decision-making and self-care, especially for victims navigating the court systems who need to make decisions related to safety, visitation, and custody.
There is a sizable literature demonstrating a strong relationship between sleep disturbances and mental health.14 The strongest and most striking of this evidence exists with respect to depression and insomnia, which frequently co-occur in all cohorts, but especially in older adults and in women (including in women of childbearing age who are also at risk for depression).15 Although, it is perhaps best viewed as a bidirectional relationship, a number of longitudinal studies support the designation of insomnia as a risk factor for the development or maintenance of depression with some indications that this risk is stronger in women (including in women of childbearing age who are at increased risk for depression).16 In addition, sleep disturbances have historically been considered a hallmark of PTSD.17 In the broader literature, ample data underscore a strong association between trauma exposure or PTSD and nightmares,18 whereas insomnia (difficulty initiating or maintaining sleep) is the most commonly endorsed symptom among PTSD patients.19 Insomnia occurs in 60–90% of patients with PTSD, whereas approximately 50% report nightmares.20 Nightmares and insomnia were each independently associated with PTSD in a trial used to establish DSM-IV PTSD criteria21 and have even been found to be risk factors for developing PTSD.22 Finally, both insomnia and nightmares have been identified as risk indicators for suicidality and suicide.23
Despite studies characterizing associations between IPV and mental health issues and between mental health issues and sleep disturbances, there are somewhat limited data examining sleep in the context of IPV victimization. Data that do exist suggest that IPV victims experience high rates of sleep disturbances.24,25 Uncontrolled studies report frequent complaints of nightmares26,27 and fear of sleep24; in population-based surveys victims report more difficulty getting enough sleep than IPV-negative women.28 In a controlled investigation that included objective measures of sleep, battered women in transitional housing had poorer sleep than healthy controls.29 These seminal publications alert us to the prevalence and importance of sleep disturbances in IPV victims and call for additional inquiry using validated measures. These studies also call for investigation of the sleep–IPV–mental health connection, given that sleep disturbances, depression, and PTSD are treatable conditions with evidence-based interventions.
The current study expands the nascent literature addressing sleep and IPV in four important ways by (1) incorporating community-based participatory research principles; (2) conducting the study among an IPV court-based sample; (3) utilizing a validated sleep instrument (the Insomnia Severity Index [ISI]30,31) to characterize sleep disturbances; and (4) assessing whether nightmares and insomnia were associated with either depression or suicidality controlling for socio-demographic and/or clinical factors. Policy and clinical implications for the findings are discussed.
The current study is grounded in principles of community-based participatory research.32,33 The tenets of this methodology suggest that along the spectrum of community work—community based, community partnered, and community participatory—the latter exists when a community partner poses the research question and the clients participate in the conceptualization and design of the study. Prior to the current study, the executive director of a large domestic violence agency, which included a hotline, 38-bed shelter, transitional services, court advocacy, and community-based support group, approached the authors with a dilemma. It appeared that many of the women served by her agency complained of sleep disturbance issues: insomnia, nightmares, restlessness, and waking from a full night's sleep feeling exhausted. She had combed the literature but was unable to find either an intervention to target this complaint among IPV victims or any literature characterizing sleep issues among IPV victims.
Contemporaneously, the agency's court-based advocates were observing judges' comments on their clients' court-based behavior. The judges expressed confusion over why the behavior of some victims testifying about horrific abuse seemed “flat and detached.” Following testimony, victims would often weep in the waiting room. When approached by advocates with assurances that all would be well, the women responded—“I am okay, I am just so tired.” These two issues spurred a partnership between our university and representatives from the shelter.
The resulting cross-sectional study recruited women from a private waiting room in a New York State family court from May 2007 to January 2008. After filing for a civil court protection order against an intimate partner and having the opportunity to meet with a domestic violence advocate for safety planning, women were approached with verbal study information. Following written informed consent, the women were enrolled in the study and completed initial study questionnaires. A second set of questionnaires was completed approximately 2 weeks later. This study was reviewed by and approved by the University of Rochester Research Subjects Review Board.
Of 380 women approached with information about the study for the purpose of recruitment, 190 women enrolled in the study (a 50% response rate). Our sample consists of women who completed the additional questionnaires including a validated insomnia instrument approximately 2 weeks after the initial assessment (n=121; a retention rate of 64%). These participants were ethnically diverse and young to middle-aged (see Table 1), although their ages ranged from 19 to 82 years old. Participants did not differ in age or race/ethnicity from those who declined participation or did not complete our sleep instrument.
Inclusion criteria included being a female, over the age of 18, and filing against an intimate partner for a protection order. The participants were English speaking and able to read at a fifth grade level. There were no inclusion criteria that focused on the length of the relationship, the severity of the violence, or resulting health problems.
Members of the research team approached participants in a locked, secure waiting room to inquire whether the participant was interested in learning more about a study. Once the study was explained and interest confirmed, the participant had the opportunity to read the informed consent and provide written consent. The participant received a grocery store gift certificate as a thank you. Interviews consisted of self-report surveys.
At the initial assessment participants completed a socio-demographic inventory and a number of self-report instruments described below.
The Modified PTSD Symptom Scale (MPSS)34 is a 17-item self-report measure that assesses both the frequency and severity of post-traumatic stress symptoms within the previous 2 weeks. The questionnaire uses a scale of 0–3 to rate frequency of symptoms and a 0–4 scale to rate severity of symptoms; total scores range from 0–119 with a validated cutoff of 46 used to identify clinically meaningful PTSD.
Nightmares were assessed by the MPSS nightmare frequency and severity items. These two items were summed to derive a nightmare severity score. In addition, using the “1–2 rule” (>1 nightmare per week and a severity of >2),35 the sample was dichotomized into those with and without nightmare disturbance.
The Conflict Tactics Scale-2 Short (CTS2)36,37 is a 20-item measure modified from a 39-item self-report scale that assesses relationship negotiation, psychological aggression, physical assault, sexual coercion, and injury. The original CTS2 39 pair-item questionnaire demonstrated high reliability coefficients, which were retained in the current version.36,37 The CTS2 total score ranges from 0 to 60.
The Center for the Epidemiologic Studies Depression Scale (CESD)38 is a 20-item scale assessing criteria for major depression. Each item is coded from 0 (“not at all or less than one day”) to 3 (“five to seven days” or “nearly every day for two weeks”) so that scores range from 0 to 63. A cutoff of 16, which has been validated as an indicator of significant depression, was used to dichotomize the sample for analyses. The CESD was re-administered 2 weeks later and only scores from this latter administration were used in analyses.
The ISI30,31 is a validated seven-item scale that assesses difficulty initiating and maintaining sleep, the daytime consequences of lack of sleep, worry about sleep, and satisfaction with sleep quality on a scale of 0–4. The summed score on the instrument can be categorized into no, mild, moderate, and severe insomnia or a score of >10 (as used in many insomnia clinical trials) can capture clinically significant insomnia. The latter cutoff was used to dichotomize the sample into those with and without insomnia. The ISI was administered approximately 2 weeks after the initial assessment.
Suicidality was assessed by asking once about the frequency of suicidal thoughts with a range of 0–3. Any level of endorsement of suicidality (i.e., >1) was used to dichotomize the sample into those with and without suicidality.
Group differences on demographic, trauma, and sleep characteristics between participants with and without depression, were assessed using t-tests for continuous variables and chi-square tests for categorical variables. Depression was defined as a score of >16 on the second administration of the CESD.
In order to assess the association of specific sleep disturbances on the presence of depression, regression models were first built with the presence of insomnia and then the presence of nightmares as independent variables. Each model controlled for age, abuse severity, and PTSD severity (with sleep items removed) as continuous variables and employment and race (white/minority) as dichotomous variables. Adjusted odds ratios (OR) were calculated with 95% confidence intervals (CI). Similar analyses were planned for suicidality as the dependent variable, but these were abandoned due to the low base rate of suicidality (only four participants endorsed suicidality at the 2-week mark).
Since the validated cutoff for depression on the CESD is based on the full scale score, the full CESD score was used in all analyses. For item-level missing data the sample mean for the item was used to impute values. All data were analyzed by using SAS version 9.2 (SAS Institute, Inc.) and all alphas were two-tailed and set at 0.05.
In this sample of women with exposure to IPV, 46% had clinically significant insomnia based on the ISI cutoff of 10 and 32% met our definition of nightmare disturbance. Based on the ISI severity categories, 53% of the subjects reported some form of insomnia (24.1% mild, 19.0% moderate, and 9.9% severe). There were no differences between those with and without insomnia or with or without nightmares relative to age, race, or poverty (using U.S. Census Bureau levels), although participants with nightmares had a higher rate of unemployment. In terms of insomnia subtype, approximately 51% had moderate to severe sleep initiation insomnia, 57% had middle of the night insomnia, and only 30% had moderate to severe complaints of early morning awakening (totals do not sum to 100% because some individuals reported more than one subtype of severe insomnia).
Table 1 provides the results of bivariate analyses for participant characteristics (demographic, trauma, and sleep characteristics) and the presence or absence of depression. Compared with those without depression, participants who met criteria for depression were slightly older and had higher PTSD severity scores but did not differ with respect to employment, poverty status, or severity of violence they had experienced. They were also far more likely to have clinically significant insomnia and nightmare disturbance. Only four participants endorsed suicidality (3% of sample), and these data are not presented in Table 1; the results here are limited to the observation that all four of these subjects also had insomnia.
For the logistic regression models, assessments of multicolinearity included the largest condition index being 14.8 and the largest variance inflation factor being 1.25, indicating that multicolinearity was not an issue for the logistic regression models. Findings from the logistic regressions (Table 2) showed that the presence of insomnia was significantly associated with the presence of depression, controlling for demographic variables, abuse severity, and PTSD severity, with adjusted OR=8.64 (CI: 2.96, 25.20). The presence of nightmare disturbance was also significantly associated with the presence of depression, controlling for demographic variables and abuse severity, with adjusted OR=3.33 (CI: 1.00, 11.03).
Insomnia and nightmares were highly prevalent in this court-based sample of women experiencing recent IPV. The presence of insomnia and nightmares each conferred significantly increased risk for the presence of depression in these women, when controlling for socio-demographic characteristics, victimization experiences, and PTSD symptoms. Overall, these findings suggest that among the indicated population of court-based IPV survivors, in which PTSD and depression were well-documented, sleep disturbances were strongly associated with mental health burden.
The finding that suicidal thoughts were endorsed by 3% of the sample (with all cases endorsing insomnia) preclude interpretation but merit an observation. On the one hand, this rate is lower than in other IPV research in which rates of suicidality were revealed at their original interview and were considerably higher (e.g., 15%).9 This may have been due to such studies assessing the presence of suicidality over the past year, whereas our question was in the past month, a time after which protective orders had been received. On the other hand, the possible association between presence of suicidal thoughts and presence of insomnia is consonant with observed relationships between sleep disturbances and suicide outcomes.23
The full set of results are consistent with a growing body of literature suggesting that insomnia symptoms accompany psychiatric disorders and are associated with chronic illness, lower quality of life, increased probability of relapse and recurrence of depression, and increased suicide risk.39 These issues are all well-known comorbid conditions reported by IPV survivors. Despite the type of violence victims report, they often report physical and mental health burdens40–42 that last long after the visible wounds have healed and safety is achieved.
Insomnia may be common following IPV for several reasons. For IPV victims who are in actively violent or dangerous situations, such as those seeking protection orders, the sleep environment may be especially associated with insomnia if the violence and danger surround the sleep experience and occur in the home, where violence is most likely to happen.1 Hyperarousal and hypervigilance may become a conditioned response to the bedroom. Sleep may be perceived as dangerous if it restricts the ability to monitor one's surrounding, as reported by childhood sexual violence survivors.43
Treatment of insomnia in the context of IPV may be complicated by the possibility that the insomnia may be adaptive (or viewed as adaptive) for some patients (i.e., for patients who continue to be in unsafe environments, or continue to feel unsafe at night or in the sleep environment). The former may need to be directly addressed (methods for optimizing personal safety) and the latter (the belief that one is unsafe) may require direct treatment prior to or as part of any form of insomnia therapy.
Thus, a tension exists between reducing insomnia symptoms and foregoing safety. Some IPV advocates may voice concerns that victim safety may be complicated further by treating sleep disturbances with medications or approaches that may blunt vigilance. As noted earlier, for some victims, their hyperarousal, a symptom of both PTSD and sleep disorders, may serve as a mechanism by which to protect their safety. Several approaches, both pharmacologic and behavioral, exist to address trauma-related sleep disturbances in a manner that provides relief for individuals in this population while keeping them safe. These include cognitive–behavioral therapies for insomnia,44 for nightmares,45 and their combination,46 as well as the use of the nonsedating generic alpha-1 adrenoreceptor antagonist prazosin for trauma-related nightmares.47 It is important that insomnia be considered, assessed, and treated independent of the other mental health burdens of depression and PTSD because insomnia can persist following conventional treatments for depression and/or PTSD.48–50 Likewise, once safety is secured, victims may need assistance to “unlearn” sleep-related behaviors that lead to insomnia because the sleep disturbances may continue despite achieving safety. This is not to suggest that depression and PTSD be ignored or necessarily take less precedence in assessment or treatment planning, but rather that sleep disturbances be given equal consideration.
In this essentially cross-sectional study, no causal or directional inferences can be made with respect to sleep disturbances and depression or suicidality. We were not able to assess the association of nightmares or insomnia on PTSD as a dependent variable since the MPSS was administered only once and this was approximately 2 weeks prior to administration of the ISI. It is also possible that court-based IPV victims seeking civil protection orders may represent a different group of women than those who may avail themselves of alternative venues for help such as shelters, hospitals, or informal means via friends and family. This study was limited to IPV victims of female gender although many IPV victims are male, who also have mental health consequences following abuse.51,52 While victims were asked about their mental health medications, the participants responded verbally that they were unfamiliar with their medication names and dosages. Accordingly, future research might be designed to consider medication effects. Finally, including women who declined to participate, this study had a response rate of 50% and a retention rate of 64%, limiting generalizability. These limitations are offset by the first use of the ISI, a validated sleep instrument, in this population and the unique family court venue for recruitment adding to the literature on women exposed to IPV.
The current study highlights the need to evaluate sleep disturbance among IPV victims and to refer them for targeted treatments that consider both their victimization and comorbid mental health symptoms. Given that sleep disturbances reduce individuals' abilities to function,53 that women with IPV histories anecdotally identify poor sleep as a barrier to effective functioning, and that sleep disturbance is a potentially treatable risk factor for significant medical and psychiatric morbidity, it seems important to begin to treat these co-occurring issues with interventions designed for this population. Future work might include longitudinal intervention trials to effectively treat insomnia and/or nightmares while simultaneously addressing safety.
While speculative, it is possible that sleep interferes with a woman's ability to seek safety and navigate complex decision-making such as whether to remain in relationships or seek orders or other available care. Improved sleep may allow women to better cope with their situations, make informed decisions about safety planning, ameliorate depression and PTSD symptoms, and/or increase treatment-seeking for other mental health services. Whether sleep disturbance moderates future mental health morbidity in this group of women remains to be explored. What is known is that these burdens have evidence-based interventions, which now must be woven together to provide integrated treatment.
If a patient reports sleep disturbances in isolation of the IPV she is experiencing, and a provider prescribes a sleeping medication, the victim may put herself at risk for further harm by masking her hypervigilance, which viewed through an IPV lens is a necessity for safety. Viewed solely through a mental health lens, that same hypervigilance, leading to sleep disorders, is seen as a symptom of a mental health condition in need of treatment. Because sleep disturbance, PTSD, and depression may be inextricably woven together following IPV, providers need to assess for all three issues and consider alternative and/or adjuvant treatment approaches.
This study was supported by National Institutes of Health grants: K01MH075965-01, T32MH18911, and K23NR010408. The authors also wish to acknowledge the contributions they enjoy and benefit from their collaborations with the Monroe County Family Court and Alternatives for Battered Women, Inc., which are both in Rochester, New York.
The following disclosures of potential conflicts of interest and/or financial interests are made by WRP: participation in pharmaceutical multi-site clinical trials (Merck), consulting fees (Pennside Partners), authorship fee (Sanofi-Aventis), royalties (Sleep Manual, Barron's Educational Press) and by MLP: industry supported research (Sanofi-Aventis, Cephalon), participation in pharmaceutical multi-site clinical trials (Sanofi, Aventis, Sanofi-Aventis, Vanda), consulting fees (Gerson Lehman Group, Clinical Advisors, MedaCopr/Leerink Swann, L.E.K. Consulting, Sleep Easily, Actelion, Takeda), speaker honoraria (Sanofi-Aventis, Sleep Medicine Education Institute), authorship fee (Sanofi-Aventis), directorship and salary (Internet Didactic Services), free use of equipment (sstnet.com, Minimitter/Respironics, Lifelines, Inc.), and industry sponsorship of continuing medical education course (Sanofi-Aventis, Respironics, Sepracor, Cephalon, Merck, Neurocare, Phillips). All remaining authors report no financial interests, relationships, or affiliations relevant to the subject matter or materials discussed in this article.