Neighborhood Disorder, Social Support, and Outcomes Among Violence-Exposed African American Women

Abstract

Intimate partner violence (IPV) against women, particularly those living in poverty who have multiple marginalized identities, is a significant public health issue. IPV is associated with numerous mental health concerns including depression, hopelessness, and suicidal behavior. The present study examined the ecological determinants of these mental health outcomes in a high-risk sample of 67 low-income, African American women survivors of IPV. Based on an ecological framework that conceptualizes individuals as nested in multiple, interactive systems, we examined, longitudinally, the main and interactive effects of self-reported neighborhood disorder and social support from family members and friends on participants’ mental health (i.e., self-reported depressive symptoms, hopelessness, and suicide intent). In multiple regression analyses, neighborhood disorder interacted with social support from family members to predict depressive symptoms and hopelessness over time. Neighborhood disorder also interacted with social support from friends to predict hopelessness and suicide intent over time. High levels of social support buffered against the dangerous effects of neighborhood disorder on depressive symptoms, hopelessness, and suicide intent; at low levels of social support, there was no significant association between neighborhood disorder and those mental health outcomes. Neighborhood disorder and social support did not yield significant main effects. These findings underscore the importance of interventions that target individuals, families, and communities (e.g., community empowerment programs). Group interventions may also be important for low-income, African American women survivors of IPV, as they can help survivors establish and strengthen relationships and social support.

There are mounting concerns about the prevalence and intractable nature of intimate partner violence (IPV) against women in the United States and worldwide (Breiding, Basile, Smith, Black, & Mahendra, 2015; Garcia-Moreno & Watts, 2012). More than one third of women in the United States have experienced one or more forms of IPV (Black et al., 2011), and prevalence rates are higher for African American women than for White women (Al’Uqdah, Maxwell, & Hill, 2016; Black et al., 2011). These elevated rates can be attributed to social class factors such as financial hardship, financial dependence on a partner, low educational attainment, and traditional gender beliefs (Cho, 2012; Golden, Perreira, & Durrance, 2013). Also relevant to African American women’s elevated risk are their multiple marginalized identities (i.e., being female and African American), cultural history and present experiences with oppression, and exposure to repeated interpersonal traumas (West, 2004). The additive stress of intersecting gender, racial, and social class factors contribute to low-income, African American women’s elevated risk for psychopathology following IPV, such as depression, hopelessness, and suicidality (Lacey et al., 2015).

Ecological models posit that mental health symptoms are best explained by a framework that considers individuals and their broader context (Bronfenbrenner, 1977). This theoretical framework can help identify risk factors for poor mental health outcomes in high-risk groups of women IPV survivors. Applied to IPV survivors, these models emphasize the interactive influence of women’s characteristics and histories, as well as the systems in which they are embedded, on their behavior and symptomatology. The systems include the microsystem, mesosystem, exosystem, and macrosystem. The microsystem encompasses interactions between individuals and their immediate settings, activities they engage in, and roles they play in those settings; the mesosystem is a system of microsystems or the interactions between the settings in which women are embedded. The exosystem consists of socially constructed structures that influence settings women must navigate (e.g., neighborhood, social services). The macrosystem is the set of broad sociocultural scripts and prototypes that govern women’s activity in other levels of the ecological model. Whereas much research has examined individual-level factors that elevate low-income, African American, IPV-exposed women’s risk for poor mental health outcomes (Bolland, Lian, & Formichella, 2005; Borges et al., 2010; Kaslow et al., 2005; Lacey et al., 2015; Lamis & Kaslow, 2014), less work has attended to system-level risk factors invoked by ecological models. Yet, neighborhood conditions and social factors may affect IPV survivors’ well-being as well, directly and in interaction with other system-level factors.

Neighborhood factors operate at the exosystem and include neighborhood disorder, or the lack of order and social control within one’s community (Skogan, 1990). Neighborhood disorder includes crime, vandalism, graffiti, danger, noise, dirt, and illicit substance use (Ross & Jang, 2000). Exposure to these conditions can lead to fear of victimization and mistrust (Ross & Jang, 2000). Neighborhood disorder is associated with trauma-related stress symptoms, increased risk for interpersonal trauma revictimization (Gapen et al., 2011; Obasaju, Palin, Jacobs, Anderson, & Kaslow, 2009), and depression (Blair, Ross, Gariepy, & Schmitz, 2014; Mair, Roux, & Galea, 2008), including among IPV survivors (Beeble, Sullivan, & Bybeee, 2011) and African American women (Cutrona et al., 2005). Perceived neighborhood disorder increases risk for hopelessness and suicide, controlling for individual-level sociodemographic variables (Denney, Wadsworth, Rogers, & Pampel, 2015; Mair, Kaplan, & Everson-Rose, 2012). Perceived neighborhood disorder may lead to psychopathology by increasing stress and eroding hope (Cutrona, Wallace, & Wesner, 2006). Despite evidence of a causal relation, the association between perceived neighborhood disorder and psychological symptomatology is not ubiquitous (Blair et al., 2014). As such, other individual- and system-level variables may amplify or attenuate the effects of neighborhood disorder on outcomes.

Social support operates at the microsystem and refers to perceived or actual, instrumental, or expressive, resources supplied by family members, friends, and community social networks (Lin, 1986). Individuals reporting higher levels of social support show more resilience following adversity (Southwick et al., 2016). Unfortunately, IPV-exposed women report smaller, less supportive social networks than their nonexposed counterparts (Katerndahl, Burge, Ferrer, Becho, & Wood, 2013). Compared with low-income, African American women with no history of IPV exposure, IPV-exposed African American women report lower levels of social support from family and friends (Thompson, Kaslow, Short, & Wyckoff, 2002). Levels of social support impact survivors’ trajectories (Coker, Watkins, Smith, & Brandt, 2003; Teaster, Roberto, & Dugar, 2006) and increase risk for depression and suicidal behavior (Coker et al., 2002) in underserved groups (Kaslow et al., 2005; Kaslow et al., 2002). Importantly, high levels of social support mitigate risk for psychological distress (Escribà-Agüir et al., 2010), particularly among suicidal, low-income, African American women (Compton, Thompson, & Kaslow, 2005; Kaslow et al., 1998).

It is unclear whether the mitigating effects of social support on mental health symptoms vary as a function of the source of such support (family or friends). Most survivors disclose their IPV to at least one informal support person, but disclosure varies as a function of demographic, intrapersonal, and situational factors (Sylaska & Edwards, 2014). IPV-exposed African American women are less likely than their White counterparts to disclose IPV to support networks (Kaukinen, 2004) and friends (Flicker et al., 2011). Some studies have found no social support source-specific effects on outcomes among IPV survivors (Escribà-Agüir et al., 2010; Woodward et al., 2015), but this issue has not been studied in low-income, African American, IPV-exposed women.

Consistent with a central premise of ecological models that conditions of one system influence, and interact with, conditions of other systems, some (Kim & Ross, 2009), but not all (Chung & Docherty, 2011; Latkin & Curry, 2003), studies suggest that neighborhood disorder and social support interact to predict mental health outcomes (Schieman & Meersman, 2004). However, the nature of those interaction effects varies across studies. Some research suggests that high levels of social support and neighborhood social ties buffer against negative mental health outcomes at high, but not low, levels of neighborhood disorder (Kim & Ross, 2009; Ross & Jang, 2000). Other studies have found that social cohesion protects against poor physical health in low to moderately, but not highly, disordered neighborhoods (Bjornstrom, Ralston, & Kuhl, 2013). Additional research is needed to clarify the interactive impact of neighborhood disorder and social support on mental health outcomes, particularly in underserved populations.

Guided by an ecological framework, this study examined neighborhood disorder, social support (from family, friends), and mental health outcomes (depressive symptoms, hopelessness, suicidality) in low-income, African American, IPV-exposed women; participants were women who participated in a randomized controlled trial (RCT) comparing a culturally informed group intervention with treatment as usual (Kaslow et al., 2010). Based on ecological models and prior research, we hypothesized the following among this sample:

• Hypothesis 1: Higher levels of neighborhood disorder would predict higher levels of depressive symptoms, hopelessness, and suicidal intent.

• Hypothesis 2: Lower levels of social support from family and friends would predict higher levels of depressive symptoms, hopelessness, and suicidal intent.

• Hypothesis 3: Neighborhood disorder and social support would interact to predict mental health outcomes, such that high levels of support from both family and friends would buffer against the deleterious effects of highly disordered neighborhoods.

Method

Participants

African American women (N = 67) aged 18 to 56 years (M = 38.62, SD = 10.63) were recruited from a large, university-affiliated public hospital as part of a RCT for IPV-exposed women with a recent suicide attempt (Kaslow et al., 2010). Women who self-identified as African American or Black, were 18 to 65 years old, and reported a suicide attempt and exposure to IPV in the past year were eligible for study inclusion. Women were excluded for acute psychosis or cognitive impairment precluding reliable assessment completion and treatment engagement. Regarding relationship status, 43.8% of participants were single/never married, 39.1% were married or partnered, 15.7% were divorced or separated, and 1.6% were widowed. Most (70.1%) women reported having children, 53.7% were homeless, 86.4% were unemployed, 60.3% used food stamps, 20.6% reported Supplemental Security Income (SSI)/Disability, 94.4% had a monthly household income less than US$2,000, and 44.8% reported highest level of education as less than 12th grade.

Procedure

Recruitment methods and study procedures were approved by the hospital’s Research Oversight Committee and the University’s Institutional Review Board. Participants were recruited from hospital inpatient and outpatient psychiatric and medical clinics. After obtaining informed consent and determining study eligibility, women completed a 2- to 3-hr face-to-face assessment (T1). They then were randomized to the 10-session intervention or control conditions. The intervention addressed reducing intrapersonal, social/situational, and cultural/environmental risk factors and increasing protective factors in these domains. Two sessions focused on social support. Participants in the control group were welcome, but not required, to attend drop-in, ongoing support groups. Participants were reassessed 10 weeks after completion of the intervention phase (T2). Assessment batteries at T1 and T2 were similar; all measures were read aloud by a research assistant to ensure comprehension. Participants were compensated US$20 at T1 and US$30 at T2. Only women who completed both assessments were included in this study.

Measures

Demographic Data Form (DDF).

Sociodemographic information was collected using the self-report DDF. Designed for a series of studies conducted by this research group on IPV-exposed and recently suicidal African American women, it includes questions about age, relationship status, employment status, income, and housing status.

Beck Depression Inventory–II (BDI-II).

On the 21-item, self-report BDI-II (Beck, Steer, & Brown, 1996), participants rated depressive symptoms on a scale from 0 to 3; item responses were summed to create a total score, with higher ratings indicating greater symptom severity. The BDI-II has yielded high internal consistency estimates (Beck et al., 1996); comparable assessments of depressive symptoms across gender, race, and ethnic groups (Whisman, Judd, Whiteford, & Gelhorn, 2013); good internal reliability estimates; and evidence of convergent validity (Joe, Woolley, Brown, Ghahramanlou-Holloway, & Beck, 2008). In this study, the BDI-II yielded excellent and good internal consistency estimates at T1 (α = .94) and T2 (α = .85), respectively.

Beck Hopelessness Scale (BHS).

On the 20-item, self-report BHS (Beck & Steer, 1993), respondents rated items true (1) or false (0) of themselves. Item responses were summed; higher scores indicate more hopelessness. The BHS has demonstrated good internal consistency and high test–retest reliability (Dyce, 1996; Holden & Fekken, 1988), and prior research attests to its predictive and concurrent validity (Beck, Steer, Kovacs, & Garrison, 1985; Kocalevent et al., 2017), including in racially diverse samples (Beck et al., 1985). In this study, the BHS yielded excellent internal consistency estimates at T1 (α = .95) and T2 (α = .90).

Suicide Intent Scale (SIS).

The SIS assesses suicide intent, including suicide risk and attempt severity (Beck, Schuyler, & Herman, 1974). The measure assesses objective circumstances surrounding respondents’ suicide attempts (Items 1–8) and intentions and expectations regarding attempts (Items 9–15); five additional items assess other aspects of risk but are omitted when scoring the SIS (Beck, Morris, & Beck, 1974). A total score was calculated by summing the measure’s first 15 items; higher scores reflect greater suicidal intent. Studies attest to the SIS’s construct validity (Minkoff, Bergman, Beck, & Beck, 1973) and predictive utility (Stefansson, Nordstrom, & Jokinen, 2012). In this study, the SIS yielded borderline to adequate internal consistency estimates at T1 (α = .59) and T2 (α = .67).

Perceived Neighborhood Disorders Scale (PNDS).

The 13-item, self-report PNDS (Ross & Mirowsky, 1999) assesses perceptions of disorder in the respondent’s neighborhood, including neighborhood cleanliness (physical order), vandalism (physical disorder), perceived safety (social order), and crime (social disorder). Items are rated on a scale from 1 (strongly agree) to 4 (strongly disagree). Responses are summed; higher scores indicate greater perceived neighborhood disorder. Research has demonstrated evidence of the reliability and external validity of this measure (Ross & Mirowsky, 1999), including in predominantly African American, low-income samples (Gapen et al., 2011). In this study, the PNDS yielded an excellent internal consistency estimate at T1 (α = .93).

Social Support Behaviors Scale (SS-B).

The 45-item, self-report SS-B (Vaux, Riedel, & Stewart, 1987) assesses social support from family and friends using a 5-point Likert-type scale. Support domains assessed include emotional support, socializing, practical assistance, financial assistance, and advice or guidance. Separate scores for family and friends are obtained by summing the appropriate items; higher scores indicate greater social support. Studies indicate good internal consistency estimates for each scale (Wang, Wong, Tran, Nyutu, & Spears, 2013). In this study, the SS-B yielded excellent internal consistency estimates for the family and friends subscales at T1 (α = .99 for each).

Data Analytic Plan

Preliminary analyses were performed to identify univariate and multivariate outliers and nonnormal distributions. Square root transformations were applied as needed (Tabachnick & Fidell, 2013). These preliminary analyses were performed using SPSS 25.

To test the three study hypotheses, six regression analyses were performed in Mplus Version 7.11 (Muthen & Muthen, 2012). For each analysis, one of the T2 mental health outcomes (i.e., depressive symptoms [BDI-II], hopelessness [BHS], suicidal intent [SIS]) was regressed on T1 neighborhood disorder and either T1 social support from family or T1 social support from friends. In these regression analyses, the T1 value of each dependent variable (i.e., mental health outcome) was included as a covariate, as was treatment condition. To test the first two hypotheses, only the main effects yielded by these regression analyses were examined. To test the third hypothesis, interaction terms were computed using standardized scores. For analyses that yielded significant interaction effects, simple slopes analysis was used to probe interactions at ±1 SD from the mean of the moderator (Aiken & West, 1991). If zero was not included in the confidence interval (CI), the analysis was considered significant.

Results

Main Effects

Multiple regression results can be found in Table 1. We found no significant main effects for neighborhood disorder, or for either type of social support, on any of the outcomes assessed. Thus, there was no support for the first two hypotheses.

Table 1.

Regression Analyses Examining Main and Interactive Effects of ND and SS on Mental Health Outcomes.

	ND × SS-Family				ND × SS-Friends
	b	SE b	95% CI	β	b	SE b	95% CI	β
Depressive symptoms
BDI (T1)	0.39	0.10	[0.19, 0.59]	.40	0.39	0.11	[0.19, 0.60]	.41
Condition	12.15	2.83	[6.60, 17.70]	.47	10.08	3.01	[4.19, 15.97]	.39
ND	2.04	4.86	[−7.49, 11.56]	.05	−2.34	4.90	[−11.95, 7.27]	−.05
SS	1.31	1.18	[−1.00, 3.63]	.12	−0.01	1.28	[−2.52, 2.51]	−.00
ND × SS	−3.65	1.44	[−6.47, −0.82]	−.27	−2.34	1.56	[−5.40, 0.72]	−.17
Hopelessness
BHS (T1)	0.23	0.09	[0.04, 0.41]	.31	0.29	0.09	[0.13, 0.46]	.40
Condition	0.08	0.21	[−0.34, 0.49]	.04	0.05	0.21	[−0.36, 0.46]	.03
ND	−0.40	0.35	[−1.08, 0.28]	−.14	−0.52	0.35	[−1.21, 0.16]	−.18
SS	−0.09	0.09	[−0.28, 0.09]	−.12	−0.05	0.09	[−0.22, 0.13]	−.06
ND × SS	−0.30	0.11	[−0.52, −0.07]	−.30	−0.29	0.10	[−0.48, −0.09]	−.32
Suicide intent
SIS (T1)	0.28	0.19	[−0.08, 0.65]	.22	0.30	0.21	[−0.18, 0.70]	.23
Condition	0.08	1.29	[−1.73, 3.33]	.09	0.40	1.35	[−2.15, 3.02]	.04
ND	0.32	2.19	[−3.98, 4.62]	.02	−0.06	2.38	[−4.33, 5.11]	−.04
SS	−0.16	0.58	[−1.29, 0.98]	−.04	0.02	0.74	[−1.33, 1.52]	.01
ND × SS	−0.29	0.70	[−11.66, 1.08]	−.06	−1.52	0.88	[−3.61, −0.05]	−.32

Note. Bolded values indicate a significant effect (95% CI does not contain zero). ND = neighborhood disorder; SS = social support; CI = confidence interval; BDI = Beck Depression Inventory; BHS = Beck Hopelessness Scale; SIS = Suicide Intent Scale.

Interaction Effects

Depressive symptoms.

A significant interaction effect was found for neighborhood disorder (T1) and social support-family (T1) on depressive symptoms (T2). Simple slopes analysis revealed that high levels of social support-family buffered against the harmful effects of neighborhood disorder on depressive symptoms (b = −13.34, 95% CI = [−26.58, −0.11]; Figure 1a). The association did not vary as a function of social support-family when such support was low (b = −4.74, 95% CI = [−14.38, 4.89]). Neighborhood disorder (T1) and social support-friends (T1) did not interact to predict depressive symptoms (T2). Thus, the third hypothesis was supported with regard to depressive symptoms only in the context of social support from family.

Figure 1.

Significant interaction effects for neighborhood disorder, social support, and mental health outcomes plotted using values ±1 SD from the mean of neighborhood disorder and social support.

Hopelessness.

In separate analyses predicting hopelessness (T2), significant interaction effects were found for neighborhood disorder (T1) and social support-family (T1), and for neighborhood disorder (T1) and social support-friends (T1). Simple slopes analysis revealed a similar pattern of results for each interaction effect: High levels of social support-family (b = −0.30, 95% CI = [−0.51, −0.08]) and social support-friends (b = −1.68, 95% CI = [−2.72, −0.63]) mitigated risk for hopelessness associated with neighborhood disorder (Figure 1b and 1c). The association between neighborhood disorder and hopelessness did not vary as a function of social support-family (b = −0.14, 95% CI = [−0.36, 0.09]) when such support was low. At 1 SD below the mean of social support-friends, such support still mitigated risk for hopelessness associated with neighborhood disorder (b = −1.02, 95% CI = [−1.78, −0.26]); it was only at very low levels of social support-friends (e.g., 2 SDs below the mean) that social support-friends did not influence the association between neighborhood disorder and hopelessness (b = −0.18, 95% CI = [−0.41, 0.05]). Thus, the third hypothesis was fully supported with respect to hopelessness.

Suicidal intent.

Neighborhood disorder (T1) significantly interacted with social support-friends (T1) to predict suicidal intent (T2). Simple slopes analysis revealed that high levels of social support-friends buffered against the dangerous effects of neighborhood disorder on suicidal intent (b = −6.19, 95% CI = [−16.38, −0.13]; Figure 1d). The association between neighborhood disorder and suicide intent did not vary as a function of social support-friends when support was low (b = −2.68, 95% CI = [−8.67, 2.36]). Neighborhood disorder (T1) and social support-family (T1) did not interact to predict suicidal intent (T2). Thus, the third hypothesis was supported with regard to suicidal intent only in the context of social support from friends.

Discussion

Findings from this study advance our understanding of the impact of ecological systems on IPV survivors’ mental health outcomes, highlighting the causal, interactive roles of understudied exosystem and microsystem variables on depressive symptoms, hopelessness, and suicidality. We found evidence of the powerful role that social support plays in mitigating risk for low-income, African American women residing in highly disordered neighborhoods. Support from family and friends in interaction with neighborhood disorder impacts mental health outcomes, but the effects are not uniform. Both sources of support serve as a protective role against the impact of neighborhood disorder on hopelessness. However, social support from family buffers against the longitudinal effects of neighborhood disorder on depressive symptoms, whereas social support from friends buffers against heightened suicide intent.

Our results are consistent with ecological models of IPV and gender-based violence (Heise, 1998; Teaster et al., 2006) and build upon a small literature applying these models to IPV survivors’ mental health. They converge with research showing aspects of women’s proximal environments to be associated with both IPV severity and mental health symptomatology (DePrince, Buckingham, & Belknap, 2013). Building upon that literature, our findings support a central tenet of the ecological systems theory, namely that behavior is best understood by attending to the effects of conditions at different system levels on each other and on individuals. Thus, it is consistent with both theory and some (Beeble et al., 2011), but not all (e.g., Coker et al., 2002), extant research that the data yielded no prospective main effects of neighborhood disorder or social support on any outcome variable. Given the small and homogeneous nature of the sample, it is premature to conclude that neither neighborhood disorder nor social support directly affects mental health outcomes. However, our data appear to show that specifically among IPV-exposed African American women, social support, at the level of the microsystem, may operate differently based on conditions at the exosystem (in our study, level of neighborhood disorder). In this sense, our findings align with studies demonstrating interaction effects of neighborhood and social factors on fear and mistrust (Ross & Jang, 2000), depression (Kim & Ross, 2009), antisocial behavior in African American children (Schofield et al., 2012), and psychological distress among African American women (Cutrona, Russell, Hessling, Brown, & Murry, 2000). Our data additionally suggest that support from family members and friends operate in unique ways among women in highly disordered neighborhoods, resulting in the differential prediction of depressive symptoms and suicide intent.

The link between specific sources of support and outcomes, in the context of high disorder, may be due to differences in the type and effectiveness of support from survivors’ family versus friends. Regarding type of support, a qualitative study of Afro-Trinidadian women found that family members often provided IPV-exposed women with material resources (e.g., money, shelter) and confronted their abusive partners, whereas friends gave advice and safe havens (Hadeed & El-Bassel, 2006). Regarding support effectiveness, a mixed-methods study of Jordanian women found that family support may serve a protective role for IPV-exposed women, yet a family’s likelihood of providing effective support varied based on history of intrafamilial violence and whether a woman had children with her abusive partner (Clark, Silverman, Shahrouri, Everson-Rose, & Groce, 2009). Unfortunately, much of the limited data on source-specific support is from other countries; accordingly, there may be differences in cultural norms and values that impact those data, and it is challenging to extrapolate them to low-income, African American, IPV-exposed women. In the United States, one study found an association between support from friends and more frequent IPV, although this relation was tempered for women in more disordered neighborhoods (Wright, 2015). The effect could be attributable to IPV survivors’ tendencies to form relationships with women with similar interpersonal trauma histories, but additional research is needed to confirm this possibility and identify implications for mental health outcomes among IPV survivors.

Our study is limited by several factors. The small sample only included individuals from a specific sociodemographic group; accordingly, caution should be used when drawing conclusions from our findings. More specifically, the relevance of the findings to IPV survivors at large, or to women from other ethnoracial and social class groups, is unknown and cannot be extrapolated from our results. Nevertheless, IPV-exposed, low-income, African American women face many barriers to study participation and retention; therefore, this group has traditionally been understudied and under-served. Because they represent a population at elevated risk, we believe that concerns over statistical power must be balanced with the clinical relevance of our findings and the importance of their dissemination. Conducting research with individuals from a particular sociodemographic group also provides an in-depth and culturally relevant understanding of the applicability of an ecological model. The second limitation relates to measurement issues. Neighborhood disorder, social support, and mental health outcomes were only assessed by self-report measures, and internal consistency estimates for the SIS were low. Future research could build upon our findings by using objective indicators of disorder and support, and clinician-administered assessments of psychological symptoms. However, research shows that perceptions of neighborhood disorder and social support may be the most proximal predictors of mental health outcomes; for instance, perceived neighborhood disorder has been shown to mediate the relationship between neighborhood disadvantage and depression (Kim, 2010). Accordingly, it is valuable to understand the associations between perceptions and outcomes. Finally, only one microsystem-level factor, social support, and only one exosystem-level factor, neighborhood disorder, were considered. In the future, more comprehensive inclusion of other potentially relevant microsystem and exosystem variables could yield a richer conceptualization of ecological risk for poor mental health outcomes in IPV-exposed women.

Recognizing these limitations, our findings suggest meaningful avenues for future research. Studies should aim to delineate the differential mechanisms by which social support influences the association between neighborhood disorder and poor mental health outcomes following IPV. A number of causal frameworks linking neighborhood disorder and depression have been proposed (Blair et al., 2014). Of most relevance, women who perceive greater disorder in their neighborhoods may become fearful and reclusive and, as a result, forfeit opportunities to form protective formal and informal social networks; in the absence of existing support from family, this forfeiture may render them more vulnerable to depressive symptoms (Blair et al., 2014; Ross & Jang, 2000). On the contrary, existing, supportive bonds with friends may buffer against risk factors for women’s engagement in suicidal behavior engendered in highly disordered neighborhoods. For instance, among IPV-exposed women from disordered neighborhoods, social support may protect against thwarted belongingness and hopelessness, two factors associated with suicidal ideation and attempts (Klonsky & May, 2015; Latkin & Curry, 2003). Alternatively, social support may have a reparative effect following exposure to the type of fear-inducing experiences that contribute to acquired capability for death by suicide and suicide attempts (Joiner, 2005; Van Orden et al., 2010). Given our differential findings for support from family versus friends, it is important that future research clarifies the unique effects of different sources of social support as they impact these pathways.

It is important that providers serving IPV-exposed women are mindful of the individual, relational, social, environmental, and cultural factors that influence survivors’ disclosure, social engagement, help-seeking, and treatment attendance and response (Ilardi & Kaslow, 2009). At the individual level, providers should address negative self-referential and social cognitions (Beck et al., 2011) that impede support-seeking among IPV survivors (Enander, 2010). Group-based interventions could target microsystem-level factors and provide an atmosphere for bonding; interventions could help survivors enhance self-efficacy and develop social relationships (Davis et al., 2009). Individual and group interventions should help women navigate exosystem-related barriers to forming relationships (e.g., high crime) that are uniquely relevant to the IPV survivors’ common mental health issues (e.g., hyperarousal, avoidance). Community-level interventions are also needed to increase social cohesion and belongingness, strengthen social ties between neighbors, and enhance the collective efficacy of women in disordered neighborhoods with high rates of IPV (Cutrona et al., 2006; Jones, Heim, Hunter, & Ellaway, 2014; Wright, Pinchevsky, Benson, & Radatz, 2015). Outreach programs to community institutions that influence how women and their social networks think about survivors of IPV are equally needed. For instance, partnerships between survivors, their advocates, and community networks might be formed to educate community groups on the challenges faced by IPV survivors so that marginalized women experience less ostracism and receive greater social support (Nash, 2005; Neietlisbach & Maercker, 2009). Such programs could be linked to relatively new and innovative efforts to address neighborhood disorder at the community level, such as through community empowerment (Aiyer, Zimmerman, Morrel-Samuels, & Reischl, 2015) and cross-sector public–private partnerships aimed at improving the health and well-being of community members (Mattessich & Rausch, 2014).

In conclusion, it is imperative that research on determinants of mental health in underserved and marginalized survivors of IPV go beyond addressing individual or situational risk factors for poor outcomes to also address system-level variables. Without attention to these factors, and the ways in which they interact with other relevant variables (e.g., history of interpersonal trauma), services for IPV survivors may be underutilized and ineffective. For low-income, African American women who live in disordered neighborhoods, we recommend that interventions include a significant focus on women’s fostering of relationships with support figures, including family and friends. This may be critical for protecting against severe psychological symptomatology, such as suicidal intent, and predictors of suicidal behavior, such as hopelessness. Providers need to reduce logistical barriers to accessing services among women in highly disorder neighborhoods. To reduce mental health disparities in this high-risk group, interventions must consider IPV survivors in their multiple, intersecting contexts.

Author Biographies

Alison M. Pickover is a PhD candidate in clinical psychology at the University of Memphis. Currently, she is completing her clinical internship at Emory University School of Medicine/Grady Health System. Her research examines individual- and system-level factors that influence outcomes in survivors of interpersonal trauma. She is also interested in using Geographic Information Systems (GIS) technology to identify geospatial barriers to care among underserved groups. Through her work, she aims to improve interventions and inform policy to benefit the most vulnerable communities.

Jabeene Bhimji is a PhD candidate in clinical psychology at Idaho State University. She is completing her clinical internship at Emory University School of Medicine/Grady Health System. Her research interests surround adult attachment and interpersonal relationships, as well as the impact that technology has on these areas.

Shufang Sun is a PhD candidate of counseling psychology at the University of Wisconsin–Madison. Her scholarly interests include health promotion with cultural and sexual minorities from a minority stress perspective, intersectionality, and development of culturally adapted treatment. Her work has appeared in several journals including Journal of Counseling Psychology and Professional Psychology: Research and Practice. Currently, she is completing clinical internship at Emory University School of Medicine/Grady Health System.

Anna Evans received a bachelor of science degree in biological sciences from the University of Georgia in 2012. Presently, she is a graduate student in the Applied Adolescent and Child Psychology master’s degree program at the University of Washington. Her current research interests are in pediatric psychology and health psychology.

Lucy J. Allbaugh, PhD, is a postdoctoral resident in health service psychology at the Emory University School of Medicine at Grady Health System. She received her PhD in clinical psychology from Miami University in 2017 after completing the predoctoral internship with the Emory School of Medicine. Her overarching research goals are to understand what promotes women’s health, resilience, and well-being, and prior work has explored the risk posed by trauma exposure, the critical role of supportive relationships and community ties in fostering resilience and well-being, and factors associated with women’s positive relational development.

Sarah E. Dunn, PhD, is an Emory University School of Medicine Department of Psychiatry and Behavioral Sciences faculty member based at Grady Health System. Board certified in clinical psychology through the American Board of Professional Psychology (ABPP), she is a board certification examiner and an applicant mentor. She has a number of areas of expertise including crisis management, risk assessment, and emergency room psychology; the assessment and treatment of suicidal persons, individuals with serious and persistent mental illness and personality disorders, and individuals exposed to family violence; forensic evaluations; psychological assessment; and clinical supervision.

Nadine J. Kaslow, PhD is a professor and vice chair for faculty development in the Department of Psychiatry and Behavioral Sciences at Emory University School of Medicine, chief psychologist at Grady Memorial Hospital, director of the Psychology Postdoctoral Fellowship Program in professional psychology, and chair of the Emory Medical Care Foundation Research Committee. Her major interests are in suicide and family violence, couple and family therapy, severe mental illness and personality disorders, integrationist approaches to psychotherapy, and competency-based approaches to education/training and supervision.