Factors Influencing Resource Use by African American and African Caribbean Women Disclosing Intimate Partner Violence

Abstract

Many victims of intimate partner violence (IPV) do not access services. Education and severity of physical violence have previously been shown to predict resource utilization, but whether these hold true specifically among women of African descent is unknown. This paper furthers our understanding of the relationship between IPV and resource use, considering socio-demographics and aspects of IPV by presenting results from a study conducted with African American and African Caribbean women in Baltimore, Maryland and the U.S. Virgin Islands. Of the 545 women included in this analysis, 95 (18%) reported emotional abuse only, 274 (50%) reported experiencing physical abuse only, and 176 (32%) had experienced both physical and sexual abuse by an intimate partner. Resource utilization was relatively low among these women, with only 57% seeking any help. Among those who did, 13% sought medical, 18% DV, 37% community and 41% criminal justice resources. Generalized linear model results indicated that older age, severe risk for lethality from IPV and PTSD were predictive of certain types of resource use, while education, insurance status, and depression had no influence. Perceived availability of police and shelter resources varied by site. Results suggest that systems that facilitate resource redress for all abused women are essential, particularly attending to younger clients who are less likely to seek help, while building awareness that women accessing resources may be at severe risk for lethality from the violence and may also be experiencing mental health complications. In addition, greater efforts should be made on the community level to raise awareness among women of available resources.

Intimate partner violence (IPV) remains a major public health concern in the United States, where at least a third of women have experienced IPV in their lifetime (Black et al., 2011; Moracco, Runyan, Bowling, & Earp, 2007). African American women are disproportionately affected by IPV in the U.S., with 43.7% of Black women reporting rape, physical violence or stalking by a partner, compared to 34.6% of White women (Black et al., 2011). In addition, a study conducted in 16 U.S. states and two territories found that 22.5% of women of African descent in the U.S. Virgin Islands (USVI) have been victims of IPV (Breiding, Black, & Ryan, 2008). Consequences of IPV for African American women include increased disparities in both physical (e.g., cardiovascular disease, HIV/STIs, heightened stress) and mental health conditions (e.g., depression, post-traumatic stress disorder) (D. W. Campbell, Sharps, Gary, Campbell, & Lopez, 2002). The most severe consequences are intimate partner homicide for women (also known as femicide) and near fatal femicide (J. C. Campbell, Webster, & Glass, 2009) African American women experience higher rates of femicide than White women, although the cumulative rates are similar.

Resource Use and the Influence of Type of Abuse Experienced

As part of the process of coping with the abuse and potentially leaving an abusive relationship, women may access a number of resources through different pathways. These can be formal services such as protective orders and domestic violence (DV) shelters or informal resources such as friends or family. Several population-based studies have shown that 58–80% of abused women opted to disclose information about the abuse and sought support at least once from any informal resource while up to 64% of women reported use of formal services (Ansara & Hindin, 2010; Coker, Derrick, Lumpkin, Aldrich, & Oldendick, 2000; Du Mont, Forte, Cohen, Hyman, & Romans, 2005; Fanslow & Robinson, 2010; Goodkind, Sullivan, & Bybee, 2004; Postmus, Severson, Berry, & Yoo, 2009). Social support from family, friends, and coworkers are common informal resources (Coker et al., 2000), while health professionals (47.2%) and the police (34.2%) are commonly used formal resources (Ansara & Hindin, 2010).

For women of color, use of informal resources including friends and faith-based communities are most often the first line of contact before reaching out to formal providers (T. B. Bent-Goodley, 2006; T. B. Bent-Goodley, 2007; El-Khoury et al., 2004; Paranjape, Tucker, Mckenzie-Mack, Thompson, & Kaslow, 2007)Typically African American women reach outside of their informal network to receive formal help when the violence has severely escalated, when they are afraid of hurting their partner or themselves, or when they are trying to stop an abusive incident from occurring (T. B. Bent-Goodley, 2007). If and when they seek formal help, resources that are used include the police (Flicker et al., 2011), () and outpatient psychiatric care (Paranjape, Heron, & Kaslow, 2006), which are consistent with the need for an immediate resolution to the IPV incident(s).

The type and severity of IPV experienced has been shown to affect women’s help-seeking behaviors. In a population-based sample of 1,509 abused women, those who experienced sexual IPV were more likely to seek medical and legal services and women experiencing physical IPV were three times more likely to seek legal services than women who experienced psychological IPV alone (Duterte et al., 2008). Other research has shown that higher levels of psychological abuse are significantly correlated with use of informal resources (Goodkind et al., 2004). Women who experienced severe physical abuse from an intimate with acts of coercive control and verbal abuse were more likely to have told someone about the violence when compared with women who experienced moderate physical violence or women who experienced sexual abuse alone (Ansara & Hindin, 2010; Fanslow & Robinson, 2010). Increased severity of physical violence is also associated with an increase in women’s utilization of medical and legal services while increased severity of sexual and psychological IPV are only associated with increasing rates of legal help seeking (Duterte et al., 2008). Furthermore, women who experienced severe psychological IPV or severe physical IPV were almost twice as likely to call the police as other abused women (Bonomi, Holt, Martin, & Thompson, 2006). Relatedly, fear for one’s own life was a strong predictor of abused women’s help-seeking behaviors (Barrett & Pierre, 2011; Hyman, Forte, Du Mont, Romans, & Cohen, 2009).

Barriers to Resource Utilization

When examining barriers to resource utilization, Fanslow and Robinson (2009) identified 63.4% of abused women did not seek help from formal services due to their perception of the violence to be “normal or not serious”. This perception of “normality” has resulted in women enduring the violence without any help (Morrison, Luchok, Richter, & Parra-Medina, 2006). Other external barriers such as lack of money, insurance, or time have been cited as reasons for not approaching a DV agency or counselor (Fugate, Landis, Riordan, Naureckas, & Engel, 2005). Furthermore, women report a lack of knowledge about available resources as a barrier to help-seeking (Postmus et al., 2009).

Women of color experience barriers to resource utilization at the individual, institutional and structural levels. In addition to relying on informal sources until there is an escalation of violence, at which time formal resources are sought, women of color may feel personal shame and embarrassment with regards to experiencing IPV and therefore maintain family secrecy about their experiences of IPV. As a means of securing the family’s image and stability, it has been commonplace to avoid disclosing family matters to individuals outside of the family (T. B. Bent-Goodley, 2007). Other individual level barriers include the failure to recognize events as IPV or that IPV is wrong; self-doubt and low self-esteem; fear of losses; fear of perpetrator; or the desire to protect the perpetrator (Petersen, Moracco, Goldstein, & Andersen Clark, 2005).

At the institutional level, barriers to resource utilization for women of color include lack of cultural competence, language barriers, and stereotyping and labeling (Ahmed & McCaw, 2010; T. B. Bent-Goodley, 2007) Many women question the cultural sensitivity of service providers. Latta and Goodman (2005) found that Haitian women regarded the lack of culturally sensitive shelters and the scarcity of shelters serving the Haitian community, in particular, as barriers to their seeking help for IPV(R. E. Latta & Goodman, 2005b). These women feared being misunderstood and marginalized by formal agencies that seemed to exclude both help-providers and help-seekers from the Haitian community. Culturally appropriate services can improve the uptake of IPV-related services for women of color (Gillum, 2008; Gillum, 2009). As cultural considerations are important to consider in the provision of services, the recognition that there is diversity within the Black population (i.e., African Americans, Caribbean Americans, Africans, other persons of Black race who may not self-identify as African American) is also extremely important. In the US and its territories, it is crucial to disentangle the ethnic heterogeneity among Blacks (Griffith, Johnson, Zhang, Neighbors, & Jackson, 2011), various dialects, and cultural experiences in efforts to deliver culturally competent services. Finally, women of color experience several stereotypes and labels that discourage them from receiving assistance to deal with IPV (Bent-Goodley, 2007). African American women can be stereotyped as being strong enough to respond to violence which can lead to being denied shelter services (T. B. Bent-Goodley, 2007). Such events create a perception that the formal resources are unresponsive and disinterested in helping them (T. B. Bent-Goodley, 2007).

At the structural level, the preponderance of literature has focused on African American women’s experiences. Their responses to violent and abusive behavior may be influenced by the chronic experiences of racism and discrimination, and the social contexts in which they live which may include poverty and less access to care and services (T. B. Bent-Goodley, 2007; D. W. Campbell et al., 2002). For example, African Americans are more likely to be prosecuted as a result of IPV, a form of discriminatory treatment that leads to African American women being unwilling to reach out to the criminal justice system, despite the need for assistance (T. B. Bent-Goodley, 2007). African American women have also experienced a higher degree of child removals when IPV is involved compared to similar cases across other ethnic groups. As a result, many African American women often choose not to disclose experiences of IPV to child welfare workers because they know they would be at greater risk of losing their children (T. Bent-Goodley & Brade, 2006). Women living in poverty with fewer resources available may be less free to conceptualize IPV as intolerable because of the reduced likelihood that the problem will be solved.

The Current Study

We sought to build upon previous studies that have addressed resource utilization among abused women of color. The purpose of our study was to examine resource utilization among abused women of African descent in two distinct settings (Baltimore, MD and the U.S. Virgin Islands (USVI)) and understand the factors that may influence this utilization. We also examined the perceived availability of resources among women who did not access them and whether site location was an independent predictor of this perceived availability. Attending to the potential varied experiences of Black women with diverse backgrounds, our sample included African American and African Caribbean women, and we assessed for differences in racial background as well as site location. The results have implications for resource professionals and can help to inform the development of violence detection and violence prevention initiatives.

Method

Study population

The study population was composed of African American and African Caribbean women in Baltimore, MD and St. Croix and St. Thomas, U.S. Virgin Islands (USVI). The USVI is an unincorporated US territory, consisting of the four major islands of St. Croix, St. Thomas, St. John and Water Island, and approximately 50 small mostly uninhabited islands and cays. According to the 2010 US Census, the population of the islands is 106,405(U.S. Census Bureau, 2011). Approximately 80% are black, African American or of African ancestry, 15 % white and 14% Hispanic. Approximately 66.8% of the population have been born in the USVI and have U.S. citizenship.

Participants in the multi-site study were recruited from primary care, prenatal or family planning clinics. Participation was restricted to women aged 18–55 with an intimate relationship within the past 2 years and who identified as being of racial or ethnic heritage that included African descent. After the informed consent process, eligible women completed a survey composed of field-tested instruments via an audio computer-assisted self-interview (ACASI). Women took the survey once in private offices in the clinics. Most participants completed it during the wait time for their appointments, as the average duration for completion was 30 minutes. If they were called to their appointment, they returned immediately afterwards to continue the survey. Remuneration of a $20 gift card per participant was provided for their time.

All women reporting a history of IPV were enrolled as cases. Women who only reported history of child abuse or abuse by someone other than an intimate partner were deemed ineligible for participation. Because the prevalence of non-abuse is higher in the population than abuse and we wanted comparable sizes of groups, never-abused women were randomly selected as controls. The random selection was done via the ACASI program, and the never-abused women who were eligible but not enrolled were considered “non-selected controls.” Of the 1315 eligible women, 403 were non-selected controls, 11 eligible women did not finish the survey, and 901 women participated in the study. This analysis focuses on the 543 women reporting abuse by an intimate. IRB approval was obtained from the Johns Hopkins University and the University of the Virgin Islands.

Dependent variables

Outcomes of interest pertained to women’s use of resources to help with their abusive situation. The measure of resource use asked about the use of community, criminal justice, DV, and health care resources. Under community resources, women were asked if they had talked to a co-worker, a boss, a counselor, and/or someone at church for help with their abuser. Criminal justice resources pertained to contacting the police and/or obtaining a restraining order. DV resource use included talking with a DV advocate and/or utilizing a DV shelter. Talking with a health care professional in an emergency department (ED) and/or primary care setting comprised health care resources. Responses were initially tallied to assess frequency of accessing more than one form of resource in a given category. We then dichotomized those responding yes to any resource in a given category classified as having used that type of resource.

Correlates of Resource Use

We assessed the role of socio-demographics on resource use, including age, marital status, education level, employment status, having children less than 18 living at home, and having health insurance.

Additional correlates of interest pertained to mental health and abuse. Mental health correlates included self-reported depressive symptoms, measured by the Center for Epidemiologic Studies Depression Scale (CESD-10 (Andresen, Malmgren, Carter, & Patrick, 1994). Participants’ scores were summed and then categorized into either not suggestive of depression (score <10) or having symptoms suggestive of depression (score 10+), a cutpoint previously established in the literature (Andresen et al., 1994). The presence of post-traumatic stress disorder (PTSD) symptoms was assessed using the 4-item Primary Care Post Traumatic Stress Disorder Screening (range 0–4) (PC-PTSD; (Prins et al., 2003). We used the recommended cut-off (3+) to classify participants as positive for PTSD.

Although both abused (physical, sexual, and/or emotional abuse by an intimate partner) and non-abused women were recruited, only women reporting abuse (n=543) were included in this analysis. Abuse-related correlates included type of intimate partner abuse (IPA), risk for lethality from the abuse, history of previously ending the relationship with the abuser, and perceived community attitudes towards partner abuse. We classified women into five types of IPA experienced, based on their responses to the Women’s Experience of Battering (WEB), the Abuse Assessment Screen (AAS), and the Severity of Violence Against Women instrument ((SVAWS (Marshall, 1992)). These groupings included (1) psychological only, (2) physical OR sexual only, (3) physical AND sexual without psychological, (4) physical OR sexual with psychological, (5) all three forms. Risk for lethality from abuse was determined through the Danger Assessment ((DA) (J. C. Campbell et al., 2009)). Weights were assigned and summed scores (range 0–36) were categorized into three categories: limited (0), variable/increased (1–13), and severe/extreme danger (14–36). Women were also asked whether or not they had left the abuser previously. To ascertain perceived norms surrounding IPV, women were asked whether, in their community, it is “considered OK to hit your wife or girlfriend under certain circumstances?” Those responding positively were classified as being in communities that held more permissive norms of IPV.

Awareness of resources

Among women who did not access resources, we assessed their awareness of resources. Responses to questions regarding whether a participant thought a given resource was available were dichotomized into being “not aware” or “aware” of that resource.

Data Analysis

We assessed internal consistency of the CESD-10 and PC-PTSD scales (Cronbach’s alpha 0.84 and 0.79, respectively). Chi-square tests were used to describe the associations between variables of socio-demographics, type of IPV, depressive symptoms, and PTSD and the outcome variables of resource utilization as well as to assess the associations between sites and perceptions of resource availability. A series of generalized linear models were conducted and included variables that had p-values of less than 0.15 in bivariate analyses or theoretically relevant. Adjusted odds ratios (AORs) and 95% confidence intervals (95% CI) were obtained from the logistic regressions. Final models were selected for best fit based on their Akaike Information Criterion values. All statistical analyses were performed using Stata 11.

Results

Characteristics of Participants

Of the 543 abused women included in this analysis, 89 (16%) reported lifetime psychological partner abuse only, 77 (14%) reported experiencing physical OR sexual IPV only, 26 (5%) reported physical AND sexual only, 188 (35%) experienced either physical OR sexual in combination with psychological, and 163 (30%) reported all three forms of IPA. One-hundred and fifty nine (29%) women were from Baltimore, MD, 207 (38%) were from St. Thomas, USVI and 177 (33%) were from St. Croix, USVI (Table 1). The majority of the women (495, 91%) identified as being black African American or African Caribbean, while 38 (7%) identified as being Spanish or Hispanic African American and another 10 (2%) considered themselves other mixed race with African descent. The average age was 29.3 years (8.92 SD). One-fourth (24.5%) were pregnant, over three-fourths (76%) had at least one child under 18 living at home, and 44% were either partnered or married. Approximately 80% of the participants had at least a high school diploma or equivalent, nearly half of the sample (49%) was employed, and 65% had some form of health insurance including 40% on Medicaid or government subsidized insurance programs. The vast majority (93%) reported an individual annual income of less than 24,000USD.

Table 1.

Bivariate Associations Between Resource Use and Socio-Demographic Variables Among a Sample of African American and African Caribbean Women Reporting Experiences of IPV

		Type of Resource
		Community		Criminal Justice		Domestic Violence		Health Care
Socio-demographic Variables	Sample (n=543)	Not used (n=341)	Used (n=202)	Not used (n=322)	Used (n=221)	Not used (n=445)	Used (n=98)	Not used (n=473)	Used (n=70)
Site, n (%)
Baltimore	159 (29)	93 (28)	66 (33)	78 (24)	81 (37)	126 (28)	33 (34)	139 (30)	20 (29)
St. Thomas	207 (38)	141 (41)	66 (33)	150 (47)	57 (26)	180 (41)	27 (27)	181 (38)	26 (37)
St. Croix	177 (33)	107 (31)	70 (34)	94 (29)	83 (37)	139 (31)	38 (39)	153 (32)	24 (34)
Race, n (%)
Black AA/AC	495 (91)	317 (93)	178 (88)	296 (92)	199 (90)	412 (93)	83 (85)	433 (91)	62 (89)
Hispanic AA/AC	38 (7)	20 (6)	18 (9)	23 (7)	15 (7)	25 (5)	13 (13)	31 (7)	7 (10)
Other mixed race	10 (2)	4 (1)	6 (3)	3 (1)	7 (3)	8 (2)	2 (2)	9 (2)	1 (1)
Age, n (%)
18–24	204 (38)	135 (40)	69 (34)	135 (42)	69 (31)	177 (40)	27 (28)	185 (39)	19 (28)
25–34	201 (37)	130 (38)	71 (35)	112 (35)	90 (40)	162 (36)	39 (40)	184 (39)	17 (24)
35–44	92 (17)	52 (15)	40 (20)	52 (16)	40 (18)	76 (17)	16 (16)	75 (16)	17 (24)
45+	46 (8)	24 (7)	22 (11)	23 (7)	23 (11)	30 (7)	16 (16)	29 (6)	17 (24)
Marital Status, n (%)
Single	264 (49)	167 (49)	97 (48)	149 (46)	115 (52)	221 (49)	43 (44)	228 (48)	36 (51)
Partnered	159 (29)	112 (33)	47 (23)	107 (33)	52 (24)	136 (31)	23 (23)	147 (31)	12 (17)
Married	80 (15)	44 (13)	36 (18)	48 (15)	32 (14)	61 (14)	19 (19)	64 (14)	16 (23)
Divorced/Separated	40 (7)	18 (5)	22 (11)	18 (6)	22 (10)	27 (6)	13 (13)	34 (7)	6 (9)
Employed, n (%)	266 (49)	176 (52)	90 (45)	167 (52)	99 (45)	226 (51)	40 (41)	244 (52)	22 (31)
Children <18, n (%)	415 (76)	265 (78)	150 (75)	236 (73)	179 (81)	336 (76)	79 (81)	363 (77)	52 (74)

p <0.05 denoted in bold type

Associations Between Correlates of Interest and Resource Use

Overall prevalence of any resource use was 57%. Of the specific types of resources, 37% of women accessed informal community resources, 54% of whom utilized more than one form. Less than half (41%) of women accessed criminal justice resources, although 58% of these women used both police and protective orders. Less than one in five women (18%) accessed DV specific resources, and 75% of these women accessed either a DV hotline or a DV shelter, but not both. Only 13% of women used health care resources, and 70% of these women spoke with either their primary care provider or an ED provider but not both.

In bivariate associations, education level, health insurance status, and being pregnant were not associated with any of the different resources, while other variables (see Table 1) were significantly associated with some but not all resource types. Of the mental health and abuse correlates, depression was not correlated with any kind of resource use, but all other variables had significant associations with at least one type of resource (see Table 2).

Table 2.

Bivariate Associations Between Resource Use and Mental Health and Violence-Related Variables Among a Sample of African American and African Caribbean Women Reporting Experiences of IPV

		Type of Resource
		Community		Criminal Justice		Domestic Violence		Health Care
Mental Health and IPV-related Variables	Sample (n=543)	Not used (n=341)	Used (n=202)	Not used (n=322)	Used (n=221)	Not used (n=445)	Used (n=98)	Not used (n=473)	Used (n=70)
Depressive Symptoms, n (%)	268 (49)	165 (48)	103 (51)	154 (48)	114 (52)	217 (49)	51 (52)	232 (49)	36 (51)
PTSD, n (%)	117 (21)	55 (16)	62 (31)	55 (17)	62 (28)	87 (19)	30 (31)	96 (20)	21 (30)
Type of IPV, n (%)
Psych only	89 (16)	68 (20)	21 (10)	60 (19)	29 (13)	82 (18)	7 (7)	78 (16)	11 (16)
Phys OR Sex only	77 (14)	64 (19)	13 (6)	59 (18)	18 (8)	72 (16)	5 (5)	74 (16)	3 (4)
Phys & Sex only	26 (5)	15 (4)	11 (6)	14 (4)	12 (5)	20 (4)	6 (6)	22 (5)	4 (6)
Phys OR Sex + Psych	188 (35)	118 (35)	70 (35)	112 (35)	76 (34)	151 (34)	37 (38)	163 (34)	25 (36)
All Three Forms	163 (30)	76 (22)	87 (43)	77 (24)	86 (39)	120 (27)	43 (44)	136 (29)	27 (39)
Community Permissive of Violence, n (%)	123 (23)	75 (22)	48 (24)	83 (27)	40 (18)	105 (24)	18 (18)	100 (22)	23 (33)
Lethality of IPV, n (%)
Variable danger	187 (35)	138 (40)	49 (25)	138 (43)	49 (23)	165 (37)	22 (23)	167 (35)	20 (29)
Increased danger	224 (42)	144 (42)	80 (40)	147 (46)	77 (35)	190 (43)	34 (35)	192 (41)	32 (46)
Severe/Extreme danger	127 (24)	58 (17)	69 (35)	35 (11)	92 (42)	86 (20)	41 (42)	110 (23)	17 (25)
Previous break-up with abuser (n=520)	427 (83)	255 (80)	172 (86)	231 (78)	196 (89)	342 (82)	85 (87)	371 (83)	56 (81)

p <0.05 denoted in bold type

Socio-demographic Correlates

Certain socio-demographic correlates were significant predictors of resource utilization (see Table 3). Site location was a significant independent correlate for criminal justice resource use, with women in St. Thomas being 64% less likely to access these (AOR 0.36, 95% 95% CI, 0.21–0.62) than women in Baltimore. Age was an independent correlate for several resources. When compared to the youngest women (18–24), women over 45 years of age were independently more likely to utilize all but community resources, while women 35–44 years old were almost three times as likely as 18–24 year olds to access medical resources (AOR 2.65, 95% CI 1.19, 5.89). Currently employed women were 70% (AOR 0.30, 95% CI 0.16–0.56) less likely than unemployed women to report utilizing health care resources. Race, marital status and having children under 18 years old at home were not independent correlates of any type of resource use.

Table 3.

Odds Ratios (ORs) and Adjusted Odds Ratios (AORs) for Correlates of Resource Use Among African American and African Caribbean Women

	Type of Resource
	Community		Criminal Justice		Domestic Violence		Health Care
Correlates	OR (95% CI)	AOR (95% CI)	OR (95% CI)	AOR (95% CI)	OR (95% CI)	AOR (95% CI)	OR (95% CI)	AOR (95% CI)
Site (Baltimore, ref)
St. Thomas	0.66 (0.43–1.01)	0.62 (0.37–1.03)	0.36 (0.24–0.57)	0.36 (0.21–0.62)	0.57 (0.33–0.99)	0.58 (0.30–1.12)	0.99 (0.54–1.86)	--
St. Croix	0.92 (0.60–1.42)	0.89 (0.55–1.46)	0.85 (0.55–1.31)	0.88 (0.53–1.47)	1.04 (0.62–1.76)	1.02 (0.55–1.90)	1.09 (0.58–2.06)	--
Race (AA/AC, ref)
African-Hispanic	1.60 (0.83–3.11)	--	0.97 (0.49–1.90)	--	2.58 (1.27–5.25)	2.21 (0.93–5.24)	1.58 (0.67–3.74)	--
Mixed Other	2.67 (0.74–9.59)	--	3.47 (0.89–13.58)	--	1.24 (0.26–5.95)	0.92 (0.13–6.39))	0.78 (0.10–6.23)	--
Age (18–24, ref)
25–34	1.07 (0.71–1.61)	--	1.55 (1.04–2.32)	1.25 (0.76–2.05)	1.58 (0.92–2.69)	1.45 (0.80–2.63)	0.90 (0.45–1.79)	0.98 (0.48–2.01)
35–44	1.51 (0.91–2.49)	--	1.51 (0.91–2.49)	1.30 (0.69–2.43)	1.38 (0.70–2.71)	1.18 (0.54–2.60)	2.21 (1.09–4.48)	2.65 (1.19–5.89)
45+	1.79 (0.94–3.43)	--	1.96 (1.02–3.74)	2.62 (1.20–5.71)	3.50 (1.69–7.25)	5.17 (2.11–12.67)	5.71 (2.66–12.23)	8.70 (3.61–20.99)
Marital Status (single, ref)
Partnered	0.72 (0.47–1.10)	0.81 (0.51–1.30)	0.63 (0.42–0.95)	0.93 (0.57–1.51)	0.87 (0.50–1.51)	1.22 (0.65–2.28)	0.52 (0.26–1.03)	--
Married	1.41 (0.85–2.34)	1.63 (0.92–2.86)	0.86 (0.52–1.44)	0.98 (0.53–1.82)	1.60 (0.87–2.95)	1.53 (0.74–3.19)	1.58 (0.83–3.04)	--
Divorced/Separated	2.10 (1.08–4.12)	1.74 (0.83–3.69)	1.58 (0.81–3.09)	1.45 (0.63–3.34)	2.47 (1.18–5.18)	1.87 (0.77–4.54)	1.12 (0.44–2.85)	--
Employed	0.75 (0.53–1.07)	--	0.75 (0.53–1.06)	--	0.67 (0.43–1.04)	--	0.43 (0.25–0.74)	0.30 (0.16–0.56)
Children <18 at home	0.83 (0.56–1.24)	--	1.55 (1.02–2.36)	1.50 (0.90–2.50)	1.35 (0.78–2.32)	--	0.88 (0.49–1.56)	--
Depressive Symptoms	1.11 (0.78–1.57)	0.70 (0.47–1.06)	1.16 (0.83–1.64)	0.69 (0.45–1.07)	1.14 (0.74–1.77)	0.86 (0.51–1.44)	1.10 (0.67–1.82)	1.03 (0.58–1.84)
PTSD	2.30 (1.52–3.49)	1.73 (1.06–2.83)	1.89 (1.25–2.86)	1.64 (0.97–2.79)	1.82 (1.11–2.96)	1.16 (0.64–2.10)	1.69 (0.96–2.94)	1.59 (0.84–3.04)
Type of IPV (emotional only, ref)
Phys OR Sex Only	0.66 (0.30–1.42)	1.02 (0.47–1.06)	0.63 (0.32–1.26)	1.12 (0.97–2.78)	0.81 (0.25–2.67)	1.58 (0.45–5.56)	0.29 (0.07–1.07)	0.67 (0.16–2.71)
Phys & Sex Only	2.37 (0.94–5.95)	2.02 (0.72–5.60)	1.77 (0.72–4.31)	2.34 (0.80–6.89)	3.51 (1.06–11.61)	5.91 (1.58–22.08)	1.28 (0.73–4.45)	2.16 (0.55–8.45)
Phys OR Sex + Psych	1.92 (1.08–3.40)	1.97 (1.05–3.68)	1.40 (0.83–2.39)	1.79 (0.95–3.39)	2.87 (1.23–6.73)	3.86 (1.51–9.83)	1.09 (0.51–2.32)	1.65 (0.68–3.97)
All Three Forms	3.71 (2.08–6.60)	2.97 (1.55–5.69)	2.31 (1.35–3.96)	1.70 (0.87–3.30)	4.20 (1.80–9.79)	4.58 (1.77–11.85)	1.40 (0.66–2.99)	1.67 (0.69–4.08)
Community Permissive of IPV	1.09 (0.72, 1.65)	0.94 (0.59–1.49)	0.61 (0.40–0.94)	0.52 (0.31–0.87)	0.72 (0.41–1.26)	0.53 (0.29–0.99)	1.78 (1.03–3.07)	1.75 (0.95–3.21)
Lethality of IPV (Variable, ref)
Increased Risk	1.56 (1.02, 2.39)	1.20 (0.75–1.93)	1.47 (0.96–2.26)	0.87 (0.53–1.41)	1.34 (0.75–2.38)	0.96 (0.51–1.82)	1.39 (0.77–2.52)	1.08 (0.55–2.10)
Severe/Extreme Risk	3.35 (2.08, 5.40)	1.91 (1.09–3.34)	7.40 (4.46–12.30)	4.55 (2.50–8.30)	3.58 (2.00–6.39)	2.46 (1.24–4.89)	1.29 (0.65–2.57)	1.10 (0.49–2.44)
Previous break-up with abuser	1.85 (1.10, 3.09)	1.28 (0.73–2.23)	2.47 (1.47–4.15)	1.84 (1.02–3.35)	1.53 (0.80–2.95)	1.16 (0.55–2.45)	0.93 (0.48–1.82)	0.86 (0.41–1.79)

p <0.05 denoted in bold type

-- Not included in multivariate model due to lack of significance in bivariate model and no contribution to model fit.

Mental Health and Abuse-Related Correlates

Depression was not significantly related to any kind of resource use. When controlling for other correlates, women with PTSD symptoms remained nearly twice as likely to access community resources (AOR 1.73, 95% CI 1.06, 2.83) compared to their counterparts.

Type of IPV was an independent correlate of community and domestic violence resource use. Women reporting a combination of physical and sexual IPV (without psychological) were independently two and six times more likely to report utilizing community and DV specific resources, respectively, than those reporting emotional abuse only (see Table 3). Women reporting either physical or sexual (with psychological) were also nearly two and four times more likely to access community and domestic violence resources, respectively. The same held true for women who experienced all three forms of IPA, who were three and four times more likely to seek community and domestic violence resources, respectively (see Table 3). Women at severe or extreme risk for lethality from IPV were independently more likely to access community, criminal justice and DV specific resources than women at variable risk.

A previous break-up with the abuser independently increased a woman’s odds of accessing criminal justice resources (see Table 3). Women who felt their communities were more permissive of IPV were 48% and 47% less likely to access criminal justice resources (AOR 0.52, 95% CI 0.31–0.87) and domestic violence resources (AOR 0.53, 95% CI 0.29–0.99) than women in communities not permissive of IPV.

Awareness of Resources Available

Many women who did not access resources believed the resources in question were not available to them. Sixty percent of the women who did not access community resources were not aware of any. No significant site-to-site differences existed regarding this belief. Half (50%) of women who did not access criminal justice resources believed that neither contacting police nor obtaining a restraining order was available to them. Although there were no differences by site in knowledge of restraining order availability, women in St. Thomas were 59% less likely (AOR 0.41, 95% CI 0.22–0.77) than women in Baltimore to consider police an available resource. Of women not seeking DV resources, 65% felt that neither shelters nor hotlines were available. Site was not an independent predictor of knowing about DV resources in the communities. Regarding medical services, 66% were unaware that primary providers and ER personnel were available to help. Women in the three sites did not differ significantly in this knowledge when controlling for socio-demographics. However, women who were college graduates were nearly four times as likely as those with less than a high school education to know that primary care resources were available (AOR 3.83, 95% CI 1.37–10.71), but the relationship did not hold with ED services.

Discussion

Overall resource utilization was relatively low among these women, with health care and DV specific use substantially lower than community and criminal justice resources. Our results are not entirely surprising, however, and somewhat mirror results of previous studies conducted in the U.S. and Canada that included minority groups (Ansara & Hindin, 2010). Contrary to findings in Canada where health professionals were the most frequent resource used (Ansara & Hindin, 2010), health professionals were least utilized in our sample of women.

Economic independence, proxied here by employment status and educational level, yielded unexpected findings. We found no relationship between education level and resource use and an inverse relationship pertaining to employment status, insomuch as those who were employed were less likely to utilize health care resources. Although not significant, the inverse relationship existed for other resource types as well, which contrasts with previous research which indicated household economics and dependency on the abuser were important considerations in a woman’s ability to seek help and eventually to leave (Henning & Klesges, 2008; Kim & Gray, 2008; Strube & Barbour, 1984). Further research with more specific measures of economic independence is warranted.

The site differences between St. Thomas, St. Croix and Baltimore also require further examination. These differences, present for resource utilization as well as perceived availability, may be related to the contained nature of the island society and reticence to come forth for fear of potential breaches of confidentiality. However, because the differences were not present between St. Croix and Baltimore, other elements may be influencing women’s resource use in St. Thomas. Contextually and culturally, the two islands differ, and stigma, fear, potential relationships with responders, perceived lack of support and community tolerance for abuse may inadvertently act as barriers to resource utilization on St Thomas. Judicial, police and health care professionals may vary in responsiveness to partner violence across islands (personal communication, Director, Domestic Violence and Sexual Assault Council, St Croix, 2012).

While no prior studies have examined political and cultural differences in the Virgin Islands, several studies have examined community response and needs of abused African American women including reports of dissatisfaction with services received due to lack of cultural competence (Gillum, 2008), and limited use of available resources, especially mental health treatment, with African American women less likely to use mental health treatment and criminal justice interventions (Wilson, Silberberg, Brown, & Yaggy, 2007). Morrison (2011) reported African American women relying on their informal networks for instrumental support but not finding these informal networks emotionally supportive. Participants in Morrison’s study noted that the African American community at-large believes victims of violence to be “stupid” for remaining in violent relationships (Morrison et al., 2006).

An identified need exists in the USVI for training first responders in sensitivity to and appropriate management of domestic violence calls. In the small USVI population, relationships are close and may sometimes influence the level of action taken when a complaint is made (personal communication, Executive Director, St. Thomas VI Family Resource Center). First responders will need assistance to resist subjectivity and to approach each situation without prejudice. Additional research is needed to investigate variations in resource utilization by abused women in the Virgin Islands and other Caribbean islands.

The increased likelihood of those experiencing physical and sexual IPV compared to women experiencing emotional abuse only to access resources is not entirely surprising. Emotional abuse only can be much harder to detect, even for the women experiencing it. If one does not identify her abuse as such, she is not likely to utilize resources. Knowledge, attitudes, and beliefs about abuse develop within sociocultural contexts and influence how women define and respond to experiences (D. W. Campbell et al., 2002; Petersen et al., 2005).

Nearly half (49%) of the abused women in our study reported symptoms indicative of depression, which is not uncommon as research has demonstrated a strong correlation between IPV and depression ((J. C. Campbell, 2002; Dienemann et al., 2000; Hankin, Smith, Daugherty, & Houry, 2010; Humphreys & Lee, 2009). Our findings that the presence of depressive symptomology was not significantly related to using or not using resources are somewhat surprising on a pragmatic level, considering that there might be a presumption that depression would serve as a barrier to help-seeking for or leaving an abusive relationship. However, little extant research explores depression as a direct correlate of IPV resource utilization. Rather, research has examined the influence of IPV on depressed women utilizing mental health resources, in which they found little difference between women experiencing abuse and other women in terms of their help-seeking behaviors for depression (Van Hook, 1999). Clearly, further research is warranted to examine this relationship and explain this rather surprising finding.

Consistent with prior research, severe risk for lethality from IPV and the elevated PTSD that is in part a function of fear was a strong independent predictor of all resource utilization except health care (Duterte et al., 2008; Fanslow & Robinson, 2010; Wright & Johnson, 2009). Women experiencing severe violence may view certain resources, particularly finding shelter from formal resources or from the community or gaining protection from the police, as more immediately valuable if they are acting on their elevated sense of self-preservation from a severely abusive partner. Medical resources may not seem as immediately germane as getting physically separated from this abuser. However, further research into the perceptions is warranted, particularly using qualitative methods. Regardless, this indicates a need for resource systems to recognize this elevated danger and take appropriate actions. Practitioners and advocates should assess and document specifically for both IPV and PTSD, taking both into account when developing safety plans with women and making appropriate referrals for mental health care. In addition, providers and advocates must be aware that lesser forms of violence may not prompt women to seek help.

Overall, many women in our study who did not access resources reported they were unaware of them. We were unable to ascertain in our quantitative data between true unawareness or beliefs that responses might not be helpful or sensitive to their situation or needs. Further research is required that focuses specifically on this issue. Either way, efforts by the community to increase awareness of existing resources and build additional avenues would be beneficial to provide assistance in domestic violence situations. Practitioners in systems providing legal protection, health care and community support must also be fully educated in their roles and responsibilities in providing competent services. This recommendation is supported by Rodriquez, et al (2009), who conducted an extensive review of identified barriers to help seeking and use of resources by women who have experienced IPV. These barriers included a variety of patient, provider, and health system/community factors. Attention to the barriers to mental health care for ethnically diverse survivors of IPV can direct the development of more effective strategies for health care practice and policy (Rodriguez, Valentine, Son, & Muhammad, 2009).

Limitations of this study include its cross-sectional nature, lack of a validated scale to effectively characterize economic independence, abbreviated mental health measures, and limited generalizability of study findings. The cross-sectional design of the study prohibits drawing conclusions about the specific causal relationships. The study would have been strengthened by the use of a validated scale for economic independence, rather than proxy markers. Mental health indicators assessed in our study included depressive symptoms and PTSD, both of which were measured by validated but abbreviated scales. As such, we could not evaluate PTSD symptom clusters, which may further elucidate the relationship between mental health and resource utilization. In addition, this study utilized a convenience sample of women of African descent in three locations. The results cannot be generalized to different African American and African Caribbean communities in the US, US territories, or Caribbean countries.

Despite these limitations, the study findings substantially add to the limited data on correlates of resource utilization among women of African descent. Our findings build on explaining factors associated with low resource utilization. Such knowledge is critical to advocate for the integration of culturally tailored services to effectively reach out to and serve women of color. Specifically, those working within the systems noted here need to increase and sustain efforts to advertise resources to African American and African Caribbean communities and make sure the communities realize those resources are nonbiased and culturally appropriate. Moreover, since women of color often do not seek help until violence escalates, it is crucial for services to intervene to prevent potentially lethal outcomes (i.e., intimate partner homicide) and improve overall safety. There is also a need to facilitate resource redress for all abused women, attending specifically to younger women and those experiencing less lethal forms of abuse (i.e. emotional), as they are less likely to utilize the different avenues for help. In sum, domestic violence services are under-utilized by women of color due to barriers faced at the individual, institutional, and structural levels (T. B. Bent-Goodley, 2007; Griffith et al., 2011; R. E. Latta & Goodman, 2005a; Petersen et al., 2005). Although domestic violence services cannot alter individual-level barriers, they have a duty to address the institutional and structural-level barriers (e.g., culturally appropriate services, addressing racism and discrimination within domestic violence services) that may increase uptake of services for abused women of color.