Health Care Professionals,Rurality,and Intimate Femicide

Abstract

Research demonstrates that disparities exist in access to quality rural health care. With studies showing that intimate partner violence is more severe and homicide is more prevalent in rural areas, scholars have begun to turn to the inaccessibility of health care in these areas as an explanation. The current study sets out to further this limited body of literature by examining the importance of rurality on the relationship between the availability of health care professionals and intimate femicide at the county level. Results indicate that rurality moderates the relationship between the availability of health care professionals and intimate femicide; however, results are not as predicted.

Keywords

place domestic violence health care access barriers to services NVDRS

Introduction

While largely understudied, research has begun to establish the importance of accounting for place in the study of intimate partner violence (IPV; DeKeseredy & Schwartz, 2009; Peek-Asa et al., 2011; Weisheit, Falcone, & Wells, 2006), intimate partner homicide (IPH; Gallup-Black, 2005; Jennings & Piquero, 2008), and femicide in particular (Beyer, Wallis, & Hamberger, 2015; Gillespie & Reckdenwald, 2017; Sinauer, Michael Bowling, Moracco, Runyan, & Butts, 1999). This small body of research has shown that rural areas have an increased prevalence or risk of IPV (DeKeseredy & Schwartz, 2009; Peek-Asa et al., 2011; Websdale, 1998; Weisheit et al., 2006) and higher rates of IPH (Gallup-Black, 2005; Jennings & Piquero, 2008; Sinauer et al., 1999). Literature has also shown that rural violence differs in comparison with urban violence (Bouffard & Muftic, 2006), with the differences for rural women lying in type and severity of the abuse (Peek-Asa et al., 2011; Websdale & Johnson, 1998), and impacted by characteristics of rural areas and their residents (Websdale, 1998; Weisheit & Donnermeyer, 2000).

Another body of literature has examined barriers to accessing IPV services in rural areas (Edwards, 2015; National Rural Health Association, 2002; Peek-Asa et al., 2011). While this literature highlights the impact of geographic isolation, lack of transportation, service availability, and inadequate training for medical professionals to recognize and screen for IPV, there is scarce knowledge on the availability of health care professionals in rural areas in the context of IPV or IPH. One study interviewed medical staff in rural areas and found that rural providers felt there was a shortage of staff, funding, and training to effectively combat IPV (Eastman & Bunch, 2007). In a related line of research, scholars have emphasized the importance of gender inequality in explaining gendered violence (Whaley & Messner, 2002). It is suggested that high levels of gender inequality is associated with high levels of gendered violence when men use violence to control women who are viewed as inferior. Under this view, increasing gender equality should reduce gendered violence and provide women with opportunities and resources to leave violent relationships. In contrast, it has also been argued that increasing gender equality may have a detrimental effect by increasing violence against women, as men feel threatened by a loss of control (Whaley & Messner, 2002).

The current study intends to bridge the literatures on gendered IPH and access to health care professionals while considering the impact of rurality. The importance of structural gender inequality and economic deprivation will be examined as well. Specifically, influences of the availability of health care professionals (i.e., physicians, nurses, and dentists), economic deprivation (i.e., poverty, unemployment), gender inequality (i.e., income, education, employment), and rurality on females killed by male intimate partners (i.e., intimate femicide) are examined. The purpose of this study is to determine the role that rurality plays in regard to health care availability and intimate femicide.

Literature Review

Place, IPV, and IPH

Battered women in rural areas are often overlooked or ignored in research (Websdale, 1995; Websdale & Johnson, 1998, for review of studies comparing IPV in rural and urban areas see Edwards, 2015) despite victimization experiences in rural settings being very different and affected by several key issues such as geography, isolation, subcultural attitudes surrounding gender, enforcement officers, economic disadvantages, education level, and availability of services (Dudgeon & Evanson, 2014; Neill & Hammatt, 2015). IPV research tends to focus on the urban setting while ignoring “space and place,” which implies little to no difference between rural and urban locales (Pruitt, 2008). Yet, violence studies have more recently demonstrated the importance of nonurban place and have shown that rural violence differs in comparison with urban violence (Bouffard & Muftic, 2006) in terms of the type and severity of the abuse (Websdale & Johnson, 1998) as well as the rural context associated with lifestyle (Websdale & Johnson, 1998; Weisheit & Donnermeyer, 2000).

Research has indicated the severity of physical abuse is higher for rural women compared with nonrural women (Logan, Walker, Cole, Ratliff, & Leukefeld, 2003; Peek-Asa et al., 2011; Websdale & Johnson, 1998). Moreover, frequency and severity of female-victim IPV increases as rural locales become increasing more rural, with isolated areas being worse off (Peek-Asa et al., 2011). Peek-Asa et al.’s (2011) clinic-based survey research of 1,478 women showed that rural and isolated women reported having the highest prevalence of IPV. Beyer et al. (2015) stated that the isolation rural women face may increase the risk of violence due to lack of intervention abilities and because the isolation of the rural locale may shield violence from view. Rural subculture may also play a role, including a subculture of acceptance toward violence and the use of guns (Websdale, 1995). This subcultural attitude also entails rural individuals favoring less governmental interference compared with urban areas concerning IPV specifically (Edwards, 2015).

Rurality has also been shown to be associated with higher rates of IPH (Gallup-Black, 2005; Jennings & Piquero, 2008) and femicide particularly (Sinauer et al., 1999). In the examination of domestic homicides trends, Gallup-Black (2005) found place to be significant with rates of IPH highest in rural areas. In addition, during the time period examined, IPH declined for all locations with the exception of rural areas, where IPH increased. This same study found that the more isolated an area, the greater chance the homicide perpetrator was an intimate partner or close family member. Likewise, Sinauer et al. (1999) examined female homicide rates in North Carolina between 1988 and 1993 and found the percentage committed by intimate partners to be higher in rural counties compared with urban.

Economic Deprivation and Gender Inequality

Female victims of IPV in rural areas may struggle to find appropriate help for financial reasons. Research has suggested that rural areas may be more socioeconomically disadvantaged than urban areas, which affects IPV rates, responses, and adequate care (Sandberg, 2013). Websdale (1998) provided support for this in which shelter workers in rural areas disclosed that many rural women have more difficulties obtaining help due to lack of financial resources. Research indicates that lack of income is also a factor that makes it more challenging for women to leave a violent relationship (Averill, Padilla, & Clements, 2007; Eastman & Bunch, 2007; Neill & Hammatt, 2015). Expanding on this, Peek-Asa et al. (2011) found that a third of women in rural areas did not have any type of insurance coverage and Laditka, Laditka, Olatosi, and Elder (2007) found that rural residents are more likely to be uninsured than urban residents, making it even more difficult to afford proper health care and social services.

The literature highlights the importance of considering economic deprivation in the study of gender-specific IPH at the macro-level as well (Dugan, Nagin, & Rosenfeld, 1999; Gillespie & Reckdenwald, 2017; Reckdenwald & Parker, 2012). For instance, in a county-level examination in North Carolina, Madkour, Martin, Halpern, and Schoenback (2010) found that county disadvantage, in terms of poverty, female-headed households, public assistance payments, unemployment rate, and education, affects the risk of female-victim IPH. Disadvantage was attributed to the variation in response by law enforcement, the availability of accessible services, and funding for IPV services.

This line of research also emphasizes the role of gender inequality in gendered violence with two competing perspectives—the amelioration hypothesis and the backlash perspective. In the amelioration hypothesis, violence may be used as a method of control when men hold more status and power over women and women are viewed as inferior (Whaley & Messner, 2002). In this sense, increased levels of gender inequality are associated with higher levels of gendered violence. If this is the case, reducing gender inequality (i.e., increasing occupational, economic, and educational equality for women relative to men) should create an ameliorative process thereby reducing female victimization. In support of amelioration, Gillespie and Reckdenwald (2017) showed the importance of gender equality as a protective factor against IPH in rural counties in North Carolina. Alternatively, the backlash hypothesis, originally discussed in the early 1970s, states that increasing gender equality may increase violence against women, as men feel threatened by a loss of control. Giving support to this hypothesis, Whaley and Messner (2002) found that male perpetrated homicides against women in the South were significantly and positively associated with gender equality.

Barriers to Accessing Services in Rural Areas

Research has found that there are various obstacles to accessing social, medical, and IPV services in rural areas (Edwards, 2015; National Rural Health Association, 2002). One of these obstacles is the geographical isolation and remote living situations common in rural areas. This may make women more vulnerable to abuse, as it augments an abusers control, keeps women away from other men and the larger community, and furthers other forms of isolation (Sandberg, 2013; Websdale, 1995; Websdale & Johnson, 1998). Peek-Asa et al. (2011) found that rural women had over 3 times the distance to travel to arrive at the nearest IPV social service when comparing the distance with urban women. In the same study, it was found that 25% of rural women lived 40 miles or further from the closest IPV program in comparison with less than 1% of urban women.

This geographic isolation to services is made even more difficult by the lack of public transportation in rural areas (Bledsoe, Yankeelov, Barbee, & Antle, 2004; Neill & Hammatt, 2015). Stommes and Brown (2002) found that public transit was only available in half of rural counties nationwide, making it necessary for inhabitants to drive long distances to access social or medical services (National Rural Health Association, 2002; Van Hightower & Gorton, 2002). In combination with geographic isolation, lack of reliable public transportation makes it difficult to seek help and limits options for leaving an abusive relationship (Jennings & Piquero, 2008; Lewis, 2003; Peek-Asa et al., 2011). Research has shown that transportation difficulties prove to be more prevalent to accessing IPV services in rural locations as opposed to urban and suburban ones (Peek-Asa et al., 2011; Zielewski & Macomber, 2007). Making access even more difficult, many rural households do not own a cell phone or have a computer in the home and, when they do, phone or Internet service is lacking in these remote areas (O’Hare & Johnson, 2004; Stommes & Brown, 2002).

In addition, research has consistently shown that services in rural areas are “few and far between” in comparison with urban locations (see Edwards, 2015). More specifically, lack of services extends to social services, shelters, police, and courts, which are essential for IPV victims in seeking help (Sandberg, 2013). In regard to IPV services specifically, Peek-Asa et al. (2011) found that IPV resources were further away, serviced more counties, and had fewer on-site services than those in urban locations. This is true of health care services as well, in which Sandberg (2013) stated that the lack of health services in rural locales may have dire health consequences for IPV victims, even suggesting that it may increase the risk of IPH.

With the lack of services also comes lack of adequate IPV services and training in both the social and health care arenas. The same study by Peek-Asa et al. (2011) found that programs in rural areas for IPV had fewer resources in general and had to serve whole counties. In a recent study, Roush and Kurth (2016) surveyed health care professionals in rural areas and found 68% were unsure if there were IPV detection and management guidelines. Also, health care professionals lacked knowledge when it came to important IPV information such as prevalence, the increased risk of injury faced when leaving an abusive relationship, and what course of action to take when IPV is disclosed. These findings emphasize the lack of proper resources and training available for the few services that are offered in rural areas, making access and treatment that much more difficult to acquire.

Training and resources for detecting and handling IPV also extends to police and criminal justice agencies. Studies show a rural/urban divide in this sense in which law enforcement may be more dispersed and less available in rural locales (Weisheit et al., 2006). Reaves (2015) found that rural police forces have fewer officers who are spread out over vast areas, which may lead to extended response times when women call the police for domestic violence reasons (Eastman et al., 2007; Riddell, Ford-Gilboe, & Leipert, 2009). Regarding responses to IPV calls, research has found that rural police are less likely to take action in an IPV complaint and also less likely to remove the perpetrator from the home when compared with urban police (Bachman & Coker, 1995; Bell, 1986; Websdale & Johnson, 1997). A study by Schafer and Giblin (2010) found that fewer than half of all police and sheriff departments had formalized agreements to facilitate coordinated responses between social service providers when dealing with IPV. This is consistent with other studies, which found that suburban and urban police are more likely to refer victims to services when arrests are not made than rural police (Websdale & Johnson, 1997), and that rural IPV victims are more likely than urban victims to self-refer or be referred by family, friends, or a legal service provider (Grossman, Hinkley, Kawalski, & Margrave, 2005).

In addition, criminal justice interventions may be weaker in rural areas (Logan, Walker, Hoyt, & Faragher, 2009). Studies have found that those in rural areas find it more difficult to efficiently obtain a protective order and have it properly enforced (Logan, Stevenson, Evans, & Leukefeld, 2004; Logan, Shannon, & Walker, 2005). In the efforts to launch, investigate, and prosecute a domestic violence case, research has identified the importance of the relationship between health care providers and criminal justice agencies. Having accurate medical reports, such as photographs of injuries and exact quoted phrases from the victim can be crucial in the decision to launch a criminal case (Isaac & Enos, 2001). This is important to note because, as discussed earlier, medical practitioners may not be aware of the appropriate actions and guidelines to follow when a woman discloses abuse—a factor that could prove fundamental in criminal justice interventions and coordinated responses to IPV.

Availability of Health Care Professionals

It is well documented that there is a skewed distribution of health care professionals, where health care professionals are more common in urban and wealthy areas as opposed to rural areas (Dessault & Franceschini, 2006; Lehmann, Dieleman, & Martineaeu, 2008; Vanasse et al., 2007). In addition, the availability of health care professionals declines as communities become more rural (Institute of Medicine, 2005). The Bureau of Health Professionals (1992) documented that roughly 20% of the U.S. population lives in rural communities, but only 9% of the nation’s physicians practice in these areas. More recently, the Health Resources and Services Administration (HRSA; 2017) reported that 60% of Primary Medical Health Professional Shortage areas are located in rural areas.

There are a variety of occupations that fall under the category of health care professional, such as physicians, nurses, nurse practitioners, and dentists. More specifically within the field of physicians, rural areas have consistently had a disproportionately smaller amount. In 2005, the ratio of physicians to the population size in urban counties was 136% greater than the ratio in rural counties (Fordyce, Chen, Doescher, & Hart, 2007; Ricketts, 2005). The same can be said of registered nurses, where the ratio of registered nurses to the population size in urban counties was 130% larger than the ratio in rural counties (Ricketts, 2005). A study by Knapp and Hardwick (2000) used computer based mapping to identify health care providers (e.g., primary care physicians, physician assistants, nurse practitioners, nurse midwives, dentists) in rural areas. Results found that all providers in rural areas were at levels noticeably lower than national averages. The research also showed that providers were not available for 5.8 million rural residents in their zip code.

Barriers have been uncovered that prevent health care professionals from effectively identifying IPV, including lack of time, training/education, and interventions to help victims (Waalen, Goodwin, Spitz, Petersen, & Saltzman, 2000). Shortage of health care professionals in rural areas exacerbates problems in identifying IPV. One study by Eastman and Bunch (2007) found that rural providers felt the demand for health care services outweighed the availability of resources. This was coupled with the finding that many providers also felt there was a shortage of staff, funding, and training in rural areas when comparing them with urban locations. Practitioners in rural areas also see more patients, which may in turn reduce quality of care (Bronstein, Johnson, & Fargason, 1997). The shortage of health care professionals, coupled with the lack of resources and adequate training to identify IPV in rural areas, may have unanticipated consequences for inhabitants in terms of health, injury, and even death.

Current Study

The current study intends to connect the rural IPV and homicide research with the limited research on health care availability. With this small body of research showing that rural disparities exist in access to health care (Grossman et al., 2005; Peek-Asa et al., 2011) as well as research indicating that rural areas have an increased prevalence of IPV (e.g., DeKeseredy & Schwartz, 2009) and higher rates of IPH (Gallup-Black, 2005; Jennings & Piquero, 2008; Sinauer et al., 1999), the aim of the current study is to explore the influence of rurality and, specifically, the impact of the availability of health care professionals on female-victim IPHs (i.e., intimate femicides). Interactions with health care professionals offer an opportunity to identify IPV, intervene by providing victims with resources and support, and treat assaults/wounding associated with IPV, which in turn may prevent or put a stop to future IPH, and in particular intimate femicide. Our goal is to determine whether intimate femicide is greater in rural areas, the impact of health care professionals on intimate femicide rates, and whether rurality has a moderating role on the relationship between health care professionals and intimate femicide. Based on the existing literature, we predict that intimate femicide in rural counties will be a function of lack of health care professionals, while considering gender inequality and economic deprivation.

Data and Method

Data

The current study utilized county-level data from multiple data sources. The dependent variable is based on data from the restricted use National Violent Death Reporting System (NVDRS),¹ an active incident-based violent death surveillance system, for the years 2005 to 2014.² Information is based on 961 counties, which were all the counties in 16 select states (Alaska, Colorado, Georgia, Kentucky, Maryland, Massachusetts, New Jersey, New Mexico, North Carolina, Oklahoma, Oregon, Rhode Island, South Carolina, Utah, Virginia, and Wisconsin) that were participating in 2005. The NVDRS expanded to include two more states (Ohio and Michigan) in 2010; however, these states were not included in the analyses as to keep the states consistent throughout the study period. The NVDRS includes all types of violent deaths (i.e., homicide, suicide, unintentional firearm deaths, legal intervention deaths, and deaths with an undetermined intent) and corresponding data from death certificates, state and local medical examiners, coroners, law enforcement, and toxicology reports. As information is cross-referenced across multiple data sources, the NVDRS is not subject to the incomplete or nonreporting issues by law enforcement that plagues Supplemental Homicide Reports (SHR; Pridemore, 2005). Information includes demographic information for the victim and suspect, manner of death, relationship and life stressors, other circumstances of the violent death, injury and death location, mental health and substance abuse issues, criminal involvement, and weapon information. However, for purposes of the current study, only county-level total counts of homicide over the study period were utilized.

The Census’ American Community Survey (ACS) 2013 (5-year estimates) was used for information on employment, income, education, poverty, county population estimates for rate calculations, and two of our control variables. The Area Health Resources Files data (AHRF) 2010, a Health Resources and Services Administration (HRSA) database, was utilized for information on the availability of health care professionals and hospitals across counties. The AHRF provides data at the county, state, and national level from the HRSA Bureau of Health Workforce. It includes information about the population, economy, and environment, health care professions, health facilities, health professionals training, and hospital utilization and expenditures. The Department of Agriculture’s Rural-Urban Continuum codes (Economic Research Service, 2012) is the source of information for our measure of county rurality. The Domestic Violence Service Directory collected by the National Coalition Against Domestic Violence (2008) is the source of data for domestic violence service availability. Our measure for mandatory arrest laws by state comes from the National Criminal Justice Reference Service (2008). The measure of the police officer rate was gathered from the FBI’s Law Enforcement Officers Killed and Assaulted (LEOKA) in 2012.³

Variables

Dependent measure

Intimate femicide rate

The dependent variable is the county female-victim IPH rate, which we refer to as the intimate femicide rate. Intimate femicide included homicides committed by husbands, former husbands, boyfriends, and former boyfriends. Ten years of data are used for the current study (years 2005-2014) due to the rare nature of intimate femicide. During this period of time, 3,543 female victims were killed by their male intimate partner. Of the counties, 679 experienced at least one intimate femicide, making 282, or roughly 29% of counties characterized with no intimate femicide incidents over the 10-year period. With the difficultly of assessing rare events at small levels of geography, common practice in the literature is to pool data across multiple years (e.g., Gallup-Black, 2005; Gillespie & Reckdenwald, 2017; Madkour et al., 2010; Osgood & Chambers, 2000). Following this practice and previous research, the current study pooled data over 10 years to increase the variation of intimate femicide across counties (Gillespie & Reckdenwald, 2017).⁴ Rates per 100,000 females were calculated by summing the counts of intimate femicide in each county across the 10 years of data and dividing this by the at-risk population (i.e., female population 15 years and older as recorded in the ACS 2013, multiplied by 10 to coincide with the summed intimate femicide counts).

Independent measures

Rurality

With research showing that the context of rural life may shape domestic abuse (Websdale, 1995, 1998; Websdale & Johnson, 1998), whether a county was considered to be in a rural geographical area was included in the analyses. The Department of Agriculture’s Rural-Urban Continuum codes (Economic Research Service, 2013) was the source of data for our measure of county rurality. This rural-urban classification is commonly used in studies examining crime differences across place (e.g., Gallup-Black, 2005; Gillespie & Reckdenwald, 2017; Kaylen & Pridemore, 2011; Lee & Ousey, 2001; Osgood & Chambers, 2000; Wells & Weisheit, 2004). Rural-Urban Continuum codes classify metropolitan (metro) counties by the population and nonmetropolitan counties by the amount of urbanization and adjacency to metro areas. Continuum codes divide counties into three metro groupings and six nonmetro groupings.⁵ Due to small sample issues and the desire to compare rural counties with other types of counties, the current study collapsed the Rural-Urban code categories into two groupings: nonrural (urban metropolitan and nonmetropolitan counties) and rural (Gillespie & Reckdenwald, 2017). Rural intimate femicides occurred in counties that were coded as having less than 2,500 urban population, either adjacent to a metro area or not adjacent to a metro area (n = 176). Nonrural intimate femicides occurred in counties that were coded as metro counties or those counties classified as nonmetro, but had an urban population (n = 785).

Health care professional availability

To measure the availability of health care professionals, four measures were included from 2010 as this year represented the pooled years most closely and had the most complete data available. The measures are number of physicians in county per 100,000 population, number of registered nurses in county per 100,000 population, number of nurse practitioners in county per 100,000 population, and number of dentists in county per 100,000 population. These four health care professionals were selected as they represent key agents that women have contact with as they visit outpatient medical offices for health-related appointments as well as emergency departments. Physicians and nurses are commonly examined in studies related to identification of IPV in health care settings (e.g., Gutmanis, Beynon, Wathen, & MacMillan, 2007). More recently, awareness of dentists’ role in identification of IPV is being recognized (Halpern, 2010; Senn, McDowell, & Alder, 2001). Finally, to account for the availability of health care professionals in hospitals, a measure for whether a hospital was located in a county (yes or no) was included in the analyses.

Gender inequality

Four ratio measures of females’ status relative to males were used to measure gender inequality in regard to employment, education, and income in a county (Whaley & Messner, 2002): the ratio of females to males in managerial or professional jobs (F:M managerial and professional employment ratio), the ratio of females to males 16 years or older employed in the labor force (F:M employment ratio), the ratio of females to males 25 years or older with 4 or more years of college education (F:M with four or more years of college education ratio), and the ratio of female to male median income (F:M median income ratio). Scores of 1.00 on these measures indicate equality between females and males, scores less than 1.00 indicate females are at a disadvantage to males, and scores greater than 1.00 indicate males are at a disadvantage to females.

Female economic deprivation

Three measures were used to capture female economic deprivation in the county: the proportion of female-headed households, the proportion of females below poverty, and the proportion of females unemployed.

Control measures

Two control measures were included that represent predictors commonly used in homicide research. Region (i.e., whether the county was located in the south) was included because research consistently indicates high rates of lethal violence in the south (Huff-Corzine, Corzine, & Moore, 1986). Also, in line with previous research (Bouffard & Muftic, 2006; D’Alessio & Stolzenberg, 2010), the county female to male sex ratio (equal numbers of males and females in the county is represented by a 1.0) was included in the analyses. In addition, three measures were included that are important when examining IPV and IPH. Whether a shelter or safe house was present in the county (yes or no)⁶ was included to account for domestic violence service availability. Next, we also accounted for state-level law enforcement domestic violence policy, by measuring whether a county was in a state that had a formal mandatory arrest policy regarding domestic violence situations (yes or no).⁷ Finally, as police officers are often the first line of defense for domestic abuse, the law enforcement officer rate per 1,000 population in the county was included.

Methodological Considerations

Collinearity among structural predictors is a common issue when doing macro-level research. To determine if collinearity was an issue, we first examined bivariate correlations between variables and then examined variance inflation factors (VIFs) for all examined predictors. Examination of the bivariate correlation matrices indicated collinearity among our measures. Furthermore, multicollinearity issues were apparent when examining some of our measures for the availability of health care professionals (O’Brien, 2007). Though multicollinearity issues were not present for our variables measuring gender inequality or female economic deprivation, based on prior research data reduction using principal components analysis with varimax rotation was conducted on all of our main predictors and three indices were created. The first index combined the four ratio measures of F:M employed in managerial or professional jobs, F:M employment, F:M with 4 or more years of college education, and F:M median income. This index is referred to as the gender inequality index (eigenvalue = 1.755) and explains 43.9% of the variance in the original four measures (α = .544). Though reliability is somewhat low, prior use of these four variables to measure the underlying theoretical construct of gender inequality is used to justify the use of this index (DeJong, Pizarro, & McGarrell, 2011; Parker & Reckdenwald, 2008; Reckdenwald & Parker, 2008; Whaley, Messner, & Veysey, 2011). Higher scores on this index signify less inequality of females compared with males (i.e., increasing equality). The second index combined the three measures of the proportion of female-headed households, the proportion of females unemployed, and the proportion of females in poverty. The index is referred to as the female disadvantage index (eigenvalue = 2.083) and explains 69.42% of the variance in the original three measures (α = .744). The third index combined the four measures of the physician rate, the registered nurse rate, the nurse practitioner rate, and the dentist rate. The index is referred to as the health care professionals index (eigenvalue = 2.806) and explains 70.14% of the variance in the original four measures (α = .845).

Analytic Plan

After examination of our dependent variable, it was clear that ordinary least squares as an estimator is inappropriate as the dependent variable is based on discrete counts of rare events and has a skewed distribution (Osgood, 2000; Osgood & Chambers, 2000). Furthermore, there is evidence of overdispersion in the distribution of the dependent variable (i.e., the variance exceeds the mean) and an excessive number of zeros (i.e., many counties had zero intimate femicides during the 10-year time period examined). Due to this, zero-inflated negative binomial regression was deemed most appropriate to use for the multivariate analyses. To convert the count models to the equivalent of a rate, the natural logarithm of the female population at risk of an intimate femicide is included in the multivariate analyses (i.e., females aged 15 years and older multiplied by 10 years of data) as an offset term in STATA.

Results

Descriptive Statistics

The descriptive statistics for all the variables can be viewed in Table 1. Examination of the dependent variable shows that across all 961 counties, the mean intimate femicide rate over the 10 years of pooled data is 1.29 per 100,000 females at risk (SD = 1.80). Examination of our variables representing the availability of health care professionals indicate that, across all counties, there are on average roughly 60 physicians, 40 registered nurses, 30 nurse practioners, and 38 dentists per 100,000 persons. Also, a hospital was located in the majority of counties (75.7%). Our measures of gender inequality show that women benefit compared with men in terms of mangagerial and professional employment (M = 1.46, SD = 0.39), overall employment (M = 1.02, SD = 0.04), and education (M = 1.10, SD = 0.33). It appears that on average, across all counties, there are 146 women in managerial and professional jobs, 102 women employed, and 110 women with 4 or more years of college education for every 100 men. Although women on average are better employed and educated, men still benefit in terms of income, with women’s income only 64% of men’s income. In regard to economic deprivation across counties, on average 18% of household are headed by a female, 10% of females are living below poverty, and 9% of females are unemployed. In addition, looking at our control variables, 84.6% of counties have a shelter or safe house, 45.2% have mandatory arrest policies in place for domestic violence, and 68.7% are considered to be in the southern region. Furthermore, on average there are close to 18 police officers per 1,000 population and equal numbers of males and females across counties.

Table 1.

Descriptive Statistics of All Variables (N = 961).

	n	M	SD	Range
Dependent measure
Intimate femicide rate		1.29	1.80	0-30.21
Place
Rural county	176
Nonrural county	785
Health care professional availability
Physician rate		60.23	47.30	0-508.34
Registered nurse rate		40.47	39.30	0-476.14
Nurse practitioner rate		30.07	30.07	0-409.43
Dentist rate		38.42	29.36	0-341.24
Hospital (1 = yes)	727
Gender inequality
F:M managerial and professional employment ratio		1.46	0.39	0.71-4.45
F:M employed in the labor force ratio		1.02	0.04	0.89-1.23
F:M with 4 or more years of college education ratio		1.10	0.33	0.32-3.98
F:M median income ratio		0.64	0.13	0.30-2.87
Female economic deprivation
Proportion of female-headed households		0.18	0.06	0.01-0.45
Proportion of females below poverty		0.10	0.04	0.02-0.23
Proportion of females unemployed		0.09	0.04	0-0.25
Controls
Shelter or safe house (1 = yes)	813
Mandatory arrest law (1 = yes)	434
Officer rate		17.91	13.11	0-154.00
South (1 = yes)	660
F:M sex ratio		1.01	0.09	0.36-1.45

Note. F:M = female: male.

Bivariate Analyses

Comparison between rural and nonrural counties

Bivariate analyses were conducted to differentiate between rural and nonrural intimate femicides. The differences in proportions for variables measuring the availability of health care professionals, female inequality, and female economic deprivation between rural and nonrural intimate femicides are shown in Table 2. First, results show that the intimate femicide rate differs significantly between rural and nonrural counties. On average, for rural counties, there were 1.55 intimate femicides per 100,000 at-risk females compared with 1.24 intimate femicides for nonrural counties. Furthermore, results indicate that differences in the availability of health care professionals exist in the counties in which these intimate femicides occurred. As expected, the physician rate, the registered nurse rate, and the dentist rate are lower in rural counties compared with nonrural counties (t = 5.77, p ≤ .001; t = 2.37, p ≤ .05; t = 8.39, p ≤ .001, respectively). In addition, the percentage of rural counties that have a hospital is considerably lower than for nonrural counties (40.3 vs. 83.6; t = 145.83, p ≤ .001). Examination of our measures of female inequality indicate that rural females are better positioned compared with males in managerial and professional jobs (t = −8.04, p ≤ .001), employment (t = −5.85, p ≤ .001), and education (t = −6.62, p ≤ .001) than nonrural females. As far as economic deprivation is concerned, rural counties have a lower proportion of female-headed households (0.16 vs. 0.19, t = 4.77, p ≤ .001) and females who are unemployed (0.08 vs. 0.09, t = 2.61, p ≤ .01) compared with nonrural counties; however, rural counties are characterized by a slightly higher proportion of females living in poverty than nonrural counties (0.11 vs. 0.10, t = −3.02, p ≤ .01).

Table 2.

Mean Differences Between Rural and Nonrural Counties (N = 961).

	Rural (n = 176)	Nonrural (n = 785)	t/χ²
	M (SD)/Freq (%)	M (SD)/Freq (%)	t/χ²
Dependent measure
Intimate femicide rate	1.55 (3.29)	1.24 (1.25)	−2.05*
Health care professional availability
Physician rate	41.92 (43.74)	64.34 (47.12)	5.77***
Registered nurse rate	34.14 (34.55)	41.89 (40.17)	2.37*
Nurse practitioner rate	30.89 (33.27)	29.89 (29.34)	−0.40
Dentist rate	21.99 (21.69)	42.02 (29.60)	8.39***
Hospital (yes = 1)	71 (40.30)	656 (83.60)	145.83***
Gender inequality
F:M managerial and professional employment ratio	1.67 (0.59)	1.41 (0.31)	−8.04***
F:M employed in the labor force ratio	1.03 (0.06)	1.01 (0.03)	−5.85***
F:M with 4 or more years of college education ratio	1.24 (0.47)	1.07 (0.28)	−6.62***
F:M median income ratio	0.65 (0.15)	0.64 (0.13)	−0.84
Female economic deprivation
Proportion of female-headed households	0.16 (0.07)	0.19 (0.06)	4.77***
Proportion of females below poverty	0.11 (0.04)	0.10 (0.04)	−3.02**
Proportion of females unemployed	0.08 (0.04)	0.09 (0.03)	2.61**
Controls
Shelter or safe house available (yes = 1)	148 (84.10)	665 (84.70)	0.043
Mandatory arrest law (yes = 1)	80 (45.50)	354 (45.10)	0.01
Officer rate	19.54 (17.09)	17.54 (12.02)	−1.84
South (yes = 1)	112 (63.60)	548 (69.80)	2.55
F:M sex ratio	0.96 (0.13)	1.02 (0.08)	8.00***

Note. F:M = female: male.

p ≤ .05. **p ≤ .01. ***p ≤ .001.

Comparison between counties with zero, midlevel, and high-level intimate femicide rates

A second set of bivariate analyses were conducted to examine county characteristics across three levels of intimate femicide rates. Specifically, we were interested in determining how counties with zero intimate femicides over the 10-year period differed from counties with a “mid-level” intimate femicide rate and counties with a “high-level” intimate femicide rate.⁸ If in fact health care professionals mitigate intimate femicide across place, we would expect to see counties with no intimate femicides over the study period generally characterized as being nonrural and having high health care professional availability and counties with “high-level” intimate femicide characterized as being rural and having low health care professional availability.

Results in Table 3 do indeed show significant differences across place and health care professional availability as well as gender inequality, female economic deprivation, and many of our control variables. Descriptive statistics show that the majority of “zero” intimate femicide counties are in fact nonrural counties; however, inspection of the adjusted residuals (Delucchi, 1993) indicate an association between place and occurrence such that “zero” intimate femicide counties are more likely than would be expected to be rural. Post hoc analysis⁹ shows “zero” and “high-level” intimate femicide counties are similar in regard to the physician rate, registered nurse rate, nurse practitioner rate, and the dentist rate. It appears that the differences arise in the “mid-level” intimate femicide counties. For instance, the physician rate and the dentist rate in “mid-level” counties are significantly higher than “zero” and “high-level” intimate femicide counties. Also, the registered nurse rate is higher in “mid-level” counties compared with “zero” intimate femicide counties. In addition, adjusted residuals indicate a hospital present in the county is less likely than would be expected in counties with “zero” and “high-level” intimate femicide, but higher than expected in “mid-level” intimate femicide counties. In regard to gender inequality, though there are variations between the three groupings, overall, females are better positioned than men in terms of employment and education across all counties regardless of the level of intimate femicide. Regarding economic deprivation, generally speaking, “zero” intimate femicide counties have a significantly lower level of female-headed households, females living below poverty, and females unemployed, while “high-level” intimate femicide counties have higher levels of these disadvantages.

Table 3.

Mean Differences Between Counties with Zero, Midlevel, and High-level Intimate Femicide Rates.

	Zero (n = 282)Freq (%)/M (SD)	Midlevel (n = 582)Freq (%)/M (SD)	High-level (n = 97)Freq (%)/M (SD)	F/χ²
Place
Rural county (yes = 1)	109 (38.70)^a	36 (6.20)^b	31 (32.00)^a	117.54***
Health care professional availability
Physician rate	53.77 (56.28)	65.76 (40.67)^c	45.85 (50.81)^d	11.33***
Registered nurse rate	34.86 (42.53)	43.49 (35.77)^c	38.72 (47.50)	4.72**
Nurse practitioner rate	27.75 (33.48)	31.03 (27.33)	31.06 (34.96)	1.12
Dentist rate	31.67 (36.35)	43.09 (25.33)^c	28.32 (23.17)^d	21.32***
Hospital (yes = 1)	153 (54.30)^b	514 (88.30)^a	60 (61.90)^b	130.78***
Gender inequality
F:M managerial and professional employment ratio	1.52 (0.52)	1.41 (0.30)^c	1.62 (0.39)^c,d	17.46***
F:M employed in the labor force ratio	1.02 (0.05)	1.02 (0.03)	1.02 (0.04)^d	3.66*
F:M with 4 or more years of college education ratio	1.18 (0.43)	1.04 (0.23)^c	1.18 (0.41)^d	21.81***
F:M median income ratio	0.65 (0.20)	0.64 (0.08)	0.66 (0.11)	2.17
Female economic deprivation
Proportion of female-headed households	0.17 (0.07)	0.19 (0.06)^c	0.21 (0.07)^c,d	14.26***
Proportion of females below poverty	0.10 (0.04)	0.09 (0.04)^c	0.12 (0.04)^c,d	21.86***
Proportion of females unemployed	0.09 (0.04)	0.09 (0.03)^c	0.10 (0.04)^c	5.06**
Controls
Shelter or safe house available (yes = 1)	220 (78.00)^b	513 (88.10)^a	80 (82.50)	19.35***
Mandatory arrest law (yes = 1)	141 (50.00)	259 (44.50)	34 (35.10)^b	9.76**
Officer rate	19.68 (16.08)	17.11 (11.61)^c	17.55 (11.45)	3.70*
South (yes = 1)	170 (60.30)^b	413 (71.00)	77 (79.40)^a	25.13***
F:M sex ratio	0.97 (0.12)	1.03 (0.06)^c	1.00 (0.12)^c,d	35.31***

Note. Fisher’s Least Significant Difference (LSD) post hoc test was conducted for the continuous level variables. F:M = female: male.

Adjusted residuals exceed ± 2 and are higher than expected if two variables are independent.

Adjusted residuals exceed ± 2 and are lower than expected if two variables are independent.

Significantly differs from “Zero.”

Significantly differs from “Mid-level.”

p ≤ .05. **p ≤ .01. ***p ≤ .001.

Zero-Inflated Negative Binomial Regression Analyses

To determine the main effects of the availability of health care professionals and rurality on intimate femicide rates, we ran four regression models (See Table 4). Model 1 presents the influence of the health care professionals index and the measure of hospital presence, Model 2 presents the influence of rurality, Model 3 presents the influence of the health care professionals index and hospital presence while controlling for rurality, and Model 4 presents the influences of all the measures included in Model 3 with the addition of gender inequality and female economic deprivation indices. Model 1 indicates that the health care professionals index is not significantly associated with intimate femicide rates. However, whether a county has a hospital or not is significantly related such that a hospital present in a county decreases the expected mean rate of intimate femicide by 88%. Model 2 shows that rurality is significantly associated with intimate femicide rates. Results suggest that rural counties, compared with nonrural counties, increase the expected mean rate of intimate femicide by a factor of 17.38. Results shown in Model 3 indicate that when rurality is controlled for, the health care professionals index obtains significance, such that a one-unit increase in the health care professionals index coincides with a 7% decrease in the expected mean intimate femicide rate. Also, whether a hospital is located in the county and rurality remain significant predictors of intimate femicide. Once gender inequality and female economic disadvantage are controlled for in Model 4, our measures of health care availability and rurality remain important predictors of intimate femicide in the predicted direction. Also, many of our control variables are significantly associated with intimate femicide. Having a shelter or safe house in a county decreases the expected mean rate of intimate femicide (by 43%) as does an increase in the female to male sex ratio. Finally, an increase in the officer rate is associated with a 3% increase in the expected mean intimate femicide rate.

Table 4.

Main Effects Models of Availability of Health Care Professionals and Rurality Controlling for Gender Inequality and Economic Deprivation on Intimate Femicide Rates (N = 961).

	Model 1		Model 2		Model 3		Model 4
	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR
Health care professionals index	−0.03 (0.02)	0.97	—	—	−0.07 (0.02)	0.93**	−0.06 (0.02)	0.94**
Hospital	−2.14 (0.16)	0.12***	—	—	−1.13 (0.16)	0.32***	−1.18 (0.15)	0.31***
Rural	—	—	2.86 (0.18)	17.38***	2.32 (0.18)	10.20***	2.37 (0.18)	10.69***
Gender inequality index	—	—	—	—	—	—	0.08 (0.03)	1.08**
Female disadvantage index	—	—	—	—	—	—	0.10 (0.03)	1.10***
Shelter or safe house	−0.55 (0.19)	0.58**	−0.72 (0.18)	0.49***	−0.73 (0.17)	0.48***	−0.56 (0.17)	0.57**
Mandatory arrest law	−0.57 (0.15)	0.56***	−0.28 (0.14)	0.76***	−0.33 (0.113)	0.72*	−0.22 (0.13)	0.81
Officer rate	0.04 (0.01)	1.04***	0.03 (0.01)	1.04***	0.03 (0.01)	1.03***	0.03 (0.01)	1.03***
South	0.258 (0.17)	1.32	0.57 (0.16)	1.77***	0.31 (0.16)	1.37*	0.09 (0.16)	1.09
F:M sex ratio	−4.61 (0.80)	0.01***	−4.82 (0.69)	0.01***	−3.84 (0.70)	0.02***	−4.02 (0.76)	0.02***
Constant	–3.78 (0.80)***		–6.86 (0.71)***		–6.69 (0.72)***		–6.52 (0.79)***
Log likelihood	–1628.32		–1577.36		–1539.84		–1521.80
Likelihood ratio χ²(df)	298.58 (7)***		400.50 (6)***		475.55 (8)***		511.62 (10)***

Note. IRR = incidence rate ratio; F:M = female: male.

p ≤ .05. **p ≤ .01. ***p ≤ .001.

Table 5 presents the interaction models between the availability of health care professionals and rurality and intimate femicide. Model 1 shows a significant relationship for our interaction between the availability of health care professionals and rurality on intimate femicide rates, such that a one-unit increase in the availability of health care professionals, in rural counties, corresponds with a 28% increase in the expected mean intimate femicide rate. Model 2 shows our interaction effect while controlling for gender inequality and female economic deprivation. Results indicate that, while controlling for gender inequality and female economic deprivation, the interaction between the availability of health care professionals and rurality continues to be a significant predictor of intimate femicide rates. Specifically, a one-unit increase in the availability of health care professionals, in rural counties, is associated with a 37% increase in the expected mean intimate femicide rate. Also, both indices measuring gender inequality and female economic deprivation are significantly related to intimate femicide rates. It appears that a one-unit change in the gender inequality index (i.e., females gaining equality) and the female disadvantage index (i.e., females are more disadvantaged) coincides with a 10% increase in the expected mean intimate femicide rate.

Table 5.

Interaction Effects Models of Availability of Health Care Professionals and Rurality Controlling for Gender Inequality and Economic Deprivation on Intimate Femicide Rates (n = 961).

	Model 1		Model 2
	b (SE)	IRR	b (SE)	IRR
Rural × Health care professionals index	0.24 (0.07)	1.28**	0.32 (0.07)	1.37***
Rural	2.45 (0.18)	11.54***	2.51 (0.18)	12.34***
Health care professionals index	−0.01 (0.02)	0.91***	−0.09 (0.02)	0.91***
Hospital	−1.09 (0.15)	0.34***	−1.11 (0.15)	0.33***
Gender inequality index	—	—	0.10 (0.03)	1.10**
Female disadvantage index	—	—	0.10 (0.03)	1.10***
Shelter or safe house	−0.74 (0.16)	0.47***	−0.55 (0.16)	0.58**
Mandatory arrest law	−0.31 (0.13)	0.74*	−0.17 (0.13)	0.84
Officer rate	0,03 (0.01)	1.03***	0.03 (0.01)	1.03***
South	0.35 (0.16)	1.42*	0.14 (0.16)	1.15
F:M sex ratio	−3.70 (0.70)	0.02***	−3.81 (0.75)	0.02***
Constant	–6.87 (0.71)***		–6.83 (0.77)***
Log likelihood	–1534.25		–1512.81
Likelihood ratio χ²(df)	486.73 (9)***		529.60 (11)***

Note. IRR = incidence rate ratio; F:M = female: male.

p ≤ .05. **p ≤ .01. ***p ≤ .001.

Discussion

Medical associations, including the American Medical Association and the American College of Obstetricians and Gynecologists (ACOG), recommend screening for IPV by health care professionals. Though barriers have been acknowledged and the best approach to identify IPV in the health care setting remains unclear (for a review see Waalen et al., 2000), we believe health care professions are important resources for both intervention and prevention of IPV and IPH. Research documenting that health care professionals are more common in urban and wealthy areas as opposed to rural areas (Dessault & Franceschini, 2006; Lehmann et al., 2008; Vanasse et al., 2007) points to the importance of determining what role, if any, place has in the relationship between the availability of health care professionals and intimate femicide.

Bivariate analyses support previous research indicating that rates of IPH are higher in rural areas (Gallup-Black, 2005; Jennings & Piquero, 2008; Sinauer et al., 1999) and femicide, in particular, is higher in rural areas as well (Beyer et al., 2015; Gillespie & Reckdenwald, 2017; Sinauer et al., 1999). In addition, support is found for the lack of health care professionals in rural areas (Dessault & Franceschini, 2006; Lehmann et al., 2008; Vanasse et al., 2007). As expected, the physician rate, the registered nurse rate, and the dentist rate are significantly lower in rural counties compared with nonrural counties. Likewise, rural residents are significantly less likley to have access to a hospital in the county. Interestingly, it also appears that compared with nonrural females, rural females are better positioned compared with males in employment and education and are less disadvantaged in terms of unemployment and being the sole provider of the household; though, rural females are more likely to live in poverty. Despite being more educated and employed (professionally and otherwise), rural poverty may result in economic dependence on an abusive partner and affect females’ ability to leave abusive relationships.

Examination of intimate femicide counts across counties during this 10-year period shows that almost one third of counties did not have an intimate femicide. In comparison, 10% of counties had 10 or more intimate femicides. The makeup of these counties is important to the overall questions regarding the impact of place on the role of health care professionals of the present study. The majority of “zero” and “high-level” intimate femicide counties are nonrural, though the number of rural counties falling in both of these levels is higher than expected, suggesting place to be a factor, albeit complex, in intimate femicide occurrence. Adding to the complexity of this relationship, bivariate analyses show “zero” and “high-level” intimate femicide counties to be similarly situated in terms of the availability of health care professionals. Notably, variation is evident for “mid-level” intimate femicide counties and is especially apparent in higher rates of physicians and dentists compared with both “zero” and “high-level” intimate femicide counties.

Multivariate analyses further indicate that rurality is an important predictor of intimate femicide, with a rural county increasing the expected mean rate of intimate femicide by a factor of over 10 in each of the multivariate models. The availability of health care professionals in a county has a negative relationship with intimate femicide in the models; however, interaction models suggest a more complex relationship between our main predictors. It appears that rurality does indeed moderate the relationship between the availability of health care professionals and the intimate femicide rate in counties, but results are contrary to our predictions. Results suggest the availability of health care professionals, in rural counties, is associated with an increase in the expected mean intimate femicide rate, not a decrease as we predicted. Also, there is a positive relationship between female economic deprivation and intimate femicide, with increases in economic deprivation among females resulting in increases in county-level intimate femicide. As posited from the bivariate results, it seems that as women become more economically dependent on men, it makes it difficult to leave an abusive relationship. Also, it appears there is a backlash effect as females gain equality relative to males across counties.

Explanations behind these findings may lie in factors unaccounted for in the current study. Previous research has discussed many barriers that exist in obtaining services by health care professionals (Edwards, 2015; National Rural Health Association, 2002; Peek-Asa et al., 2011). For instance, research has shown that isolation and remote living situations is common for many rural residents (Peek-Asa et al., 2011). Coupled with economic deprivation, geographic isolation in rural areas would make it more difficult to see doctors, nurses, and dentists who may be able to recognize IPV before it becomes deadly, even if these health care professions were available in a county. However, we were unable to account for access of these providers in the current study. In addition, we were unable to account for the qualifications or training of health care professionals in these counties. Research has shown that health care professionals in rural areas lack knowledge on IPV (Roush et al., 2016). Other research indicates that screening for IPV in general is low among health care professionals and in rural areas in particular (Choo, Newgard, Lowe, Hall, & McConnell, 2011; Sharps et al., 2001). Lack of knowledge may affect identification of risk factors for IPH by health care professionals, which in turn limits their ability to provide education and support, and effective referrals.

Furthermore, we were unable to measure whether or not abused women who need to see health care professionals actually use these services. Research shows that many rural women are afraid to use services because of the close network where they know their doctor (Websdale & Johnson, 1998). In addition, the literature has indicated many strategies that abusive partners use to keep their partner isolated (Sandberg, 2013), which may prevent abused women from visiting a doctor. Other reasons may include lack of money to pay for the services or the reluctance to disclose abuse to health care professionals. We were also not able to account for whether resources would be available to women if they did choose to leave a violent relationship. Although, more professionals may be able to screen women for abuse, educate them on their risk of IPH, help them develop a safety plan to reduce their risk of death, and direct them to resources, if these resources are lacking or unavailable to help these women transition out of the violent relationship and keep them safe, an increase in the number of health care professionals in these rural counties may not have the intended impact. Finally, we were not able to account for factors over time that may have affected the intimate femicides rate, the Affordable Care Act that was signed into law in 2010 being one of them.

There are limitations specifically related to the sources of data that should be mentioned as well. The NVDRS is limited by 16 states and therefore results may not be generalizable across the United States. For a detailed discussion of challenges associated with violent death surveillance, see Paulozzi, Mercy, Frazier, and Annest (2004) and Weiss, Gutierrez, Harrison, and Matzopoulos (2006). The AHRF includes information from over 50 databases (e.g., decennial census, American Medical Association, National Sample Survey of Nurses), which is a strength in terms of breadth of information but can also be a limitation as one has to consider all of the limitations associated with each included data source. Furthermore, the years available differ across variables, making longitudinal analyses not only difficult but also restricting which variables are appropriate for the study frame. Moreover, numbers of physicians are based on figures from the American Medical Association, which have been shown to have some discrepancies with licensing boards in states. Finally, the number of physicians in each county is calculated based on preferred mailing addresses of physicians. In some cases, their preferred mailing address may be in one county and their practice address in another county.

Conclusion

Our study further contributes to the understanding of the impact of rurality on intimate femicide; however, questions still remain regarding health care availability in these areas. Future research is needed to further understand the impact of health care professionals on intimate femicide and ascertain the validity of these possible explanations. In addition, it is important to determine better measures to accurately account for “availability” and “access” while considering supply of services, opportunity to obtain services (e.g., services are physically accessible and affordable), and barriers to obtaining services (see Gulliford et al., 2002). Also, future research should consider the impact of the current and former relationship status on the risk of death for intimates across rurality as there is an increased risk of homicide when an abusive relationship is ending (Block & Christakos, 1995; Hotton, 2001; Johnson & Hotton, 2003). Finally, though it is important to understand counties that do not experience intimate femicide and counties that experience a high-level of intimate femicide, “mid-level” counties are important as well. After all, this is where we noticed many differences in the availability of health care professionals. Addressing IPV and IPH in rural areas is a multifaceted process and it is important that intervention and prevention efforts, whatever they may be, be tailored specifically to each community.

Footnotes

Acknowledgements

Contributors to this report included participating Violent Death Reporting System states; participating state agencies, including state health departments, vital registrars’ offices, coroners’ and medical examiners’ offices, crime laboratories, and local and state law enforcement agencies; partner organizations, including the Safe States Alliance, National Violence Prevention Network, National Association of Medical Examiners, National Association for Public Health Statistics and Information Systems (NAPHSIS), Council of State and Territorial Epidemiologists (CSTE), and Association of State and Territorial Health Officials; federal agencies, including the Department of Justice (Bureau of Justice Statistics and the Federal Bureau of Investigation), the Department of the Treasury (Bureau of Alcohol, Tobacco, and Firearms), and the International Association of Chiefs of Police; and other stakeholders, researchers, and foundations, including The Joyce Foundation, the National Institute for Occupational Safety and Health, and the National Center for Health Statistics, Centers for Disease Control and Prevention (CDC).

Authors’ Note

The analyses, results, and conclusions presented here represent those of the authors and not necessarily reflect those of the Centers for Disease Control and Prevention (CDC). This research uses data from National Violent Death Reporting System (NVDRS), a surveillance system designed by the CDC’s National Center for Injury Prevention and Control. The findings are based, in part, on the contributions of the 32 funded states that collected violent death data and the contributions of the states’ partners, including personnel from law enforcement, vital records, medical examiners/coroners, and crime laboratories. Persons interested in obtaining data files from NVDRS should contact CDC’s National Center for Injury Prevention and Control, 4770 Buford Hwy, NE, MS F-64, Atlanta, GA 30341-3717, (800) CDC-INFO (232-4636).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Amy Reckdenwald, PhD, is an associate professor in the Department of Sociology at the University of Central Florida. Her research interests include violent victimization and offending; particularly as it relates to intimate partner homicide, domestic violence, and sexual violence. Her work has appeared in journals such as Criminology, Trauma, Violence and Abuse, Journal of Criminal Justice, Feminist Criminology, and Violence and Victims.

Alexis Yohros, MA, is a doctoral student at Northeastern University in the School of Criminology and Criminal Justice. Her research lies in the causes and consequences of violent crime, more specifically in the areas of homicide and domestic violence.

Alec Szalewski, MA, is a doctoral student in the Department of Sociology at the University of Central Florida. His teaching and research interests include interpersonal violence and homicide, mental health, and drugs.

References

Averill

J. B.

Padilla

A. O.

Clements

P. T.

(2007). Frightened in isolation: Unique considerations for research of sexual assault and interpersonal violence in rural areas. Journal of Forensic Nursing, 3, 42-46. doi:10.1111/j.1939-3938.2007.tb00091.x

Bachman

Coker

A. L.

(1995). Police involvement in domestic violence: The interactive effects of victim injury, offender’s history of violence, and race. Violence and Victims, 10, 91-106.

Bell

D. J.

(1986). Domestic violence in small cities and towns: A pilot study. Journal of Crime & Justice, 9, 163-181.

Beyer

Wallis

A. B.

Hamberger

K. L.

(2015). Neighborhood environment and intimate partner violence: A systematic review. Trauma, Violence, & Abuse, 16, 16-47. doi:10.1177/1524838013515758

Bledsoe

L. K.

Yankeelov

P. A.

Barbee

A. P.

Antle

B. F.

(2004). Understanding the impact of intimate partner violence mandatory reporting law. Violence against Women, 10, 534-560. doi:10.1177/1077801204264354

Block

C. R.

Christakos

(1995). Intimate partner homicide in Chicago over 29 years. Crime & Delinquency, 41, 496-526.

Bouffard

L. A.

Muftic

L. R.

(2006). The “rural mystique”: Social disorganization and violence beyond urban communities. Western Criminology Review, 7(3), 56-66.

Bronstein

J. M.

Johnson

V. A.

Fargason

C. A.

Jr. (1997). How rural physicians compare on cost and quality measures for Medicaid ambulatory care episodes. Journal of Rural Health, 13, 126-135.

Bureau of Health Professions. (1992). Rural health professions facts: Supply and distribution of health professions in rural America. Rockville, MD: Health Resources and Services Administration.

10.

Bureau of Health Workforce Health Resources and Services Administration (HRSA) U.S. Department of Health & Human Services. (2017). Designated Health Professional Shortage Areas Statistics. Retrieved from https://datawarehouse.hrsa.gov/tools/hdwreports/reports.aspx

11.

Choo

E. K.

Newgard

C. D.

Lowe

R. A.

Hall

M. K.

McConnell

K. J.

(2011). Rural-urban disparities in emergency department intimate partner violence resources. Western Journal of Emergency Medicine, 12, 178-183.

12.

D’Alessio

S. J.

Stolzenberg

(2010). The sex ratio and male-on-female intimate partner violence. Journal of Criminal Justice, 38, 555-561.

13.

DeJong

Pizarro

J. M.

McGarrell

E. F.

(2011). Can situational and structural factors differentiate between intimate partner and “other” homicide? Journal of Family Violence, 26, 365-376.

14.

DeKeseredy

W. S.

Schwartz

M. D.

(2009). Dangerous exits: Escaping abusive relationships in rural America. New Brunswick, NJ: Rutgers University Press.

15.

Delucchi

K. L.

(1993). On the use and misuse of chi-square. In Keren

Lewis

(Eds.), A handbook for data analysis in the behavioral sciences (pp. 294-319). Hillsdale, NJ: Lawrence Erlbaum.

16.

Dessault

Franceschini

M. C.

(2006). Not enough there, too many here: Understanding geographical imbalances in the distribution of the health workforce. Human Resources and Health, 4, 2.

17.

Dudgeon

Evanson

T. A.

(2014). Intimate partner violence in rural U.S. areas: What every nurse should know. American Journal of Nursing, 114(5), 26-35.

18.

Dugan

Nagin

D. S.

Rosenfeld

(1999). Explaining the decline in intimate partner homicide: The effects of changing domesticity, women’s status, and domestic violence resources. Homicide Studies, 3, 187-214.

19.

Eastman

B. J.

Bunch

S. G.

(2007). Providing services to survivors of domestic violence: A comparison of rural and urban service provider perceptions. Journal of Interpersonal Violence, 22, 465-473. doi:10.1177/0886260506296989

20.

Eastman

Bunch

Hamilton Williams

Carawan

. (2007). Exploring the Perceptions of Domestic Violence Service Providers in Rural Localities. Violence Against Women, 13, 700-716. doi: 10.1177/1077801207302047

21.

Economic Research Service. (2013). Rural-urban continuum codes: Documentation. Retrieved from http://www.ers.usda.gov/data-products/rural-urban-continuum-codes/documentation.aspx

22.

Edwards

(2015). Intimate partner violence and the rural-urban-suburban divide: Myth or reality? A critical review of the literature. Trauma, Violence, & Abuse, 16, 359-373. doi:10.1177/1524838014557289

23.

Fordyce

M. A.

Chen

F. M.

Doescher

M. P.

Hart

L. G.

(2007). 2005 Physician supply and distribution in rural areas of the United States (WWAMI: The Rural Health Working Papers Series 116). Retrieved from https://depts.washington.edu/uwrhrc/uploads/RHRC%20FR116%20Fordyce.pdf.

24.

Gallup-Black

(2005). Twenty years of rural and urban trends in family and intimate partner homicide. Homicide Studies, 9, 149-173. doi:10.1177/1088767904274158

25.

Gillespie

L. G.

Reckdenwald

(2017). Gender equality, place, and female-victim intimate partner homicide: A county-level analysis in North Carolina. Feminist Criminology, 12, 171-191. doi:10.1177/1557085115620479

26.

Grossman

S. F.

Hinkley

Kawalski

Margrave

(2005). Rural versus urban victims of violence: The interplay of race and region. Journal of Family Violence, 20, 71-81. doi:10.1007/s10896-005-3170-y

27.

Gulliford

Figueroa-Munoz

Morgan

Hughes

Gibson

Beech

Hudson

(2002). What does “access to health care” mean? Journal of Health Services Research & Policy, 7, 3186-3188.

28.

Gutmanis

Beynon

Wathen

C. N.

MacMillan

H. L.

(2007). Factors influencing identification of and response to intimate partner violence: A survey of physicians and nurses. BMC Public Health, 7, 12-22.

29.

Halpern

L. R.

(2010). Orofacial injuries as markers for intimate partner violence. Oral and Maxillofacial Surgery Clinics, 22, 239-246.

30.

Hotton

(2001). Spousal violence after separation. Juristat, 21, 1-19.

31.

Huff-Corzine

Corzine

Moore

D. C.

(1986). Southern exposure: Deciphering the south’s influence on homicide rates. Social Forces, 64, 906-924.

32.

Institute of Medicine. (2005). Quality through collaboration: The future of rural health. Washington, DC: National Academies Press.

33.

Isaac

N. E.

Enos

V. P.

(2001). Documenting domestic violence: How health care providers can help victims—Research in brief. Washington, DC: National Institute of Justice, U.S. Department of Justice.

34.

Jennings

W. G.

Piquero

A. R.

(2008). Trajectories of non-intimate partner and intimate partner homicides, 1980-1999: The importance of rurality. Journal of Criminal Justice, 36, 435-443. doi:10.1016/j.jcrimjus.2008.07.002

35.

Johnson

Hotton

(2003). Losing control: Homicide risk in estranged and intact intimate relationships. Homicide Studies, 7, 58-84.

36.

Kaylen

M. T.

Pridemore

W. A.

(2011). A reassessment of the association between social disorganization and youth violence in rural areas. Social Science Quarterly, 92, 978-1001.

37.

Knapp

Hardwick

(2000). The availability and distribution of dentists in rural ZIP codes and primary care health professional shortage areas (PC-HPSA) ZIP codes: Comparison with primary care providers. Journal of Public Health Dentistry, 60, 43-48.

38.

Laditka

J. N.

Laditka

S. B.

Olatosi

Elder

K. T.

(2007). The health trade-off of rural residence for impaired older adults: Longer life, more impairment. Journal of Rural Health, 23, 124-132.

39.

Lee

M. R.

Ousey

G. C.

(2001). Size matters: Examining the link between small manufacturing, socioeconomic deprivation, and crime rates in nonmetropolitan communities. The Sociological Quarterly, 42, 581-602.

40.

Lehmann

Dieleman

Martineaeu

(2008). Staffing remote rural areas in middle- and low-income countries: A literature review of attraction and retention. BMC Health Services Research, 8, 19-28. doi:10.1186/1472-6963-8-19

41.

Lewis

S. H.

(2003). Unspoken crimes: Sexual assault in rural America. Enola, PA: National Sexual Violence Resource Center.

42.

Logan

T. K.

Shannon

Walker

(2005). Protective orders in rural and urban areas: A multiple perspective analysis. Violence Against Women, 11, 896-911.

43.

Logan

Stevenson

Evans

Leukefeld

(2004). Rural and urban women’s perceptions of barriers to health, mental health, and criminal justice services: Implications for victim services. Violence and Victims, 19(1), 37-62.

44.

Logan

T. K.

Walker

Cole

Ratliff

Leukefeld

(2003). Qualitative differences among rural and urban intimate violence victimization experiences and consequences: A pilot study. Journal of Family Violence, 18, 83-92.

45.

Logan

T. K.

Walker

Hoyt

Faragher

(2009). The Kentucky civil protective order study: A rural and urban multiple perspective study of protective order violation consequences, responses, and costs. Washington, DC: National Institute of Justice.

46.

Madkour

A. S.

Martin

S. L.

Halpern

C. T.

Schoenback

V. J.

(2010). Area disadvantage and intimate partner homicide: An ecological analysis of North Carolina counties, 2004-2006. Violence Victims, 25, 363-377.

47.

National Rural Health Association. (2002). What’s different about rural health care? Available from http://www.nrharural.org

48.

Neill

K. S.

Hammatt

(2015). Beyond urban places: Responding to intimate partner violence in rural and remote areas. Journal of Forensic Nursing, 11, 93-100. doi:10.1097/JFN.0000000000000070

49.

O’Brien

R. M.

(2007). A caution regarding rules of thumb for variance inflation factors. Quality & Quantity, 41, 673-690.

50.

O’Hare

W. P.

Johnson

K. M.

(2004). Child poverty in rural America. Population Reference Bureau Reports on America, 4, 1-21.

51.

Osgood

D. W.

(2000). Poisson-based regression analysis of aggregate crime rates. Journal of Quantitative Criminology, 16, 21-43.

52.

Osgood

D. W.

Chambers

J. M.

(2000). Social disorganization outside the metropolis: An analysis of rural youth violence. Criminology, 38, 81-116.

53.

Parker

K. F.

Reckdenwald

(2008). Women and crime in context: Examining the linkages between patriarchy and female offending across space. Feminist Criminology, 3, 5-24.

54.

Paulozzi

L. J.

Mercy

Frazier

Jr. Annest

J. L.

(2004). CDC’s national violent death reporting system: Background and methodology. Injury Prevention, 10, 47-52.

55.

Peek-Asa

Wallis

Harland

Beyer

Dickey

Saftlas

(2011). Rural disparity in domestic violence prevalence and access to resources. Journal of Women’s Health, 20, 1743-1749. doi:10.1089/jwh.2011.2891

56.

Pridemore

W. A.

(2005). A cautionary note on using county-level crime and homicide data. Homicide Studies, 9, 256-268.

57.

Pruitt

L. R.

(2008). Place matters: Domestic violence and rural difference. Wisconsin Journal of Law, Gender & Society, 23, 347-416.

58.

Reaves

B. A.

(2015). Local police departments, 2013: Personnel, policies, and practices. Washington, DC: Bureau of Justice Statistics.

59.

Reckdenwald

Parker

K. F.

(2008). The influence of gender inequality and marginalization on types of female offending. Homicide Studies, 12, 208-226.

60.

Reckdenwald

Parker

K. F.

(2012). Understanding the change in male and female intimate partner homicide over time: A policy- and theory-relevant investigation. Feminist Criminology, 7, 167-195. doi:10.1177/1557085111428445

61.

Ricketts

T. C.

(2005). Workforce issues in rural areas: A focus on policy equity. American Journal of Public Health, 95, 42-48.

62.

Riddell

Ford-Gilboe

Leipert

(2009). Strategies used by rural women to stop, avoid, or escape from intimate partner violence. Health Care for Women International, 30, 134-159.

63.

Roush

Kurth

(2016). Intimate partner violence: The knowledge, attitudes, beliefs, and behaviors of rural health care providers. American Journal of Nursing, 116, 24-47.

64.

Sandberg

(2013). Backward, dumb, and violent hillbillies? Rural geographies and intersectional studies on intimate partner violence. Affilia: Journal of Women and Social Work, 28, 350-365.

65.

Schafer

J. A.

Giblin

M. J.

(2010). Policing intimate partner violence in rural areas and small towns: Policies, practices, and perceptions. Women & Criminal Justice, 20, 283-301.

66.

Senn

D. R.

McDowell

J. D.

Alder

M. E.

(2001). Dentistry’s role in the recognition and reporting of domestic violence, abuse and neglect. Dental Clinics of North America, 45, 343-363.

67.

Sharps

P. W.

Koziol-McLain

Campbell

McFarlane

Sachs

(2001). health care providers’ missed opportunities for preventing femicide. Preventive Medicine, 33, 373-380.

68.

Sinauer

Michael Bowling

Moracco

K. E.

Runyan

C. W.

Butts

J. D.

(1999). Comparisons among female homicides occurring in rural, intermediate, and urban counties in North Carolina. Homicide Studies, 3, 107-128.

69.

Stommes

E. S.

Brown

D. M.

(2002). Transportation in rural America: Issues for the 21st century. Rural America, 16, 2-10.

70.

Vanasse

Ricketts

T. C.

Courteau

Orzanco

M. G.

Randolph

Asghari

(2007). Long term regional migration patterns of physicians over the course of their active practice careers. Rural Remote Health, 77, 812.

71.

Van Hightower

N. R.

Gorton

(2002). A case study of community-based responses to rural woman battering. Violence against Women, 8, 845-872.

72.

Waalen

Goodwin

M. M.

Spitz

A. M.

Petersen

Saltzman

L. E.

(2000). Screening of intimate partner violence by health care providers. Barriers and interventions. American Journal of Preventative Medicine, 19, 230-237.

73.

Websdale

(1995). Rural woman abuse: The voices of Kentucky women. Violence Against Women, 1, 309-338.

74.

Websdale

(1998). Rural women battering and the justice system—An ethnography. Thousand Oaks, CA: Sage.

75.

Websdale

Johnson

B. R.

(1997). The policing of domestic violence in rural and urban areas: The voices of battered women in Kentucky. Policing and Society, 6, 297–317. doi:10.1080/10439463.1997.9964759

76.

Websdale

Johnson

(1998). An ethnostatistical comparison of the forms and levels of woman battering in urban and rural areas of Kentucky. Criminal Justice Review, 23, 161-196.

77.

Weisheit

R. A.

Donnermeyer

J. F.

(2000). Change and continuity in crime in rural America. The Nature of Crime: Continuity and Change, 1, 309-357.

78.

Weisheit

R. A.

Falcone

D. N.

Wells

L. E.

(2006). Crime and policing in rural and small-town America (3rd ed.). Long Grove, IL: Waveland Press.

79.

Weiss

Gutierrez

M. I.

Harrison

Matzopoulos

(2006). The US National Violent Death Reporting System: Domestic and international lessons for violence injury surveillance. Injury Prevention, 12, ii58-ii62.

80.

Wells

L. E.

Weisheit

R. A.

(2004). Patterns of rural and urban crime: A county-level comparison. Criminal Justice Review, 29, 1-22.

81.

Whaley

R. B.

Messner

S. F.

(2002). Gender equality and gendered homicides. Homicide Studies, 6, 188-210.

82.

Whaley

R. B.

Messner

S. F.

Veysey

B. M.

(2011). The relationship between gender equality and rates of inter- and intra-sexual lethal violence: An exploration of functional form. Justice Quarterly, 30, 732-754.

83.

Zielewski

E. H.

Macomber

J. E.

(2007). Rural families’ connections to services in an alternative response system. Journal of Public Child Welfare, 1, 21-41. doi:10.1080/15548730802118215