Shared Risk Factors Among Women for Intimate Partner Violence in the United States: A Secondary Analysis

Abstract

Multilevel risk factors may increase the risk of experiencing intimate partner violence among women. The overall goal of this study was to provide a comprehensive view of factors that may be associated with three forms of intimate partner violence. The primary aim was to explore associations between understudied factors and women's experiences of physical and sexual violence and stalking by an intimate partner. Secondary analysis of existing health registry data was conducted. Our evidence-driven strategy was based on a multipronged analytical approach informed by existing literature and the social–ecological model. We created an evidence-based hierarchical list comprised of three tiers. Three separate multiple logistic regression analyses were performed. Several shared risk factors were retained across all three forms including low levels of formal education, past experiences of non- partner sexual violence, residential instability, presence of children, experiences of a traumatic event and panic attacks, status of receiving US government benefits, and barriers to healthcare access. Results contribute to future research on intimate partner violence prevention by providing preliminary evidence of emerging factors associated with experiencing three forms of intimate partner violence.

Introduction

Intimate partner violence (IPV) is a serious preventable public health issue that detrimentally impacts the health and wellbeing of women around the world. Nearly one in three women has reported experiences of either sexual or physical violence by an intimate partner and/or sexual violence by nonpartner during their lifetime (World Health Organization [WHO], 2021). In the United States (US), nearly one in four women has reported experiences of sexual and physical violence and/or stalking by an intimate partner in their life (Smith et al., 2018). Four main forms of IPV include physical violence, sexual violence, stalking, and psychological aggression (including coercive tactics) perpetrated by a current or former intimate partner (i.e., spouse, boyfriend/girlfriend, dating partner, or ongoing sexual partner) (Niolon et al., 2017). IPV-related adverse health outcomes include, but are not limited to, traumatic brain injury, multiple bone fractures, physical and mental impairments, depression, posttraumatic stress disorder, suicidal attempts, anxiety, migraines, and substance use disorders among women from diverse sociodemographic backgrounds (Chatav Schonbrun et al., 2013; Cheng & Lo, 2019; Cho et al., 2017; Karakurt et al., 2017; Ralston et al., 2019; Young-Wolff et al., 2013). As a result, women who experience IPV in their lifetime have higher utilization rates of and higher annual healthcare costs compared to women with no IPV experiences (Rivara et al., 2007). Overall, IPV-related healthcare and associated loss of productivity have cost approximately $5.8 billion USD annually (Lutgendorf, 2019).

Background

Scholars have applied a variety of theoretical theories or models to examine risk and protective factors that may be associated with IPV victimization or perpetration including ecological model, feminist theory, exchange theory, and resource-based power theory (Mengistu, 2019). The social–ecological model has been predominantly used to understand the associations between multilevel risk factors and IPV victimization in the development of IPV prevention programs (Center for Disease Control and Prevention [CDC], 2022; Krug et al., 2002). In fact, the social–ecological model explains how the combination of individual, relational, community, and societal-level factors associated with a greater risk of experiencing IPV (CDC, 2022; Kelly, 2011; Krug et al., 2002; Mengistu, 2019). In this model, the individual level has represented biological and personal factors. In the context of IPV, well-documented factors at this level have included younger age, lower socioeconomic status, drug and alcohol abuse, mental health problems, personality disorders, prior experiences of IPV, and experiences of child abuse and neglect (Cho et al., 2020; Krug et al., 2002; Li et al., 2019; Mengistu, 2019). The relational level has focused on risk factors associated with close relationships such as conflicts and dissatisfaction among partners, male dominance within a relationship, and financial issues (Krug et al., 2002; Lutgendorf, 2019; Mallory et al., 2016; WHO & Pan American Health Organization, 2012). The community level has examined the characteristic risk factors in different settings (i.e., neighborhoods) where social interactions occur. In the context of IPV, previously identified factors have included high residential instability, isolation, or lack of social support and social services for women, poverty, and armed conflict in society (Hardesty & Ogolsky, 2020; Krug et al., 2002; Mengistu, 2019). The final level, societal, has represented factors such as patriarchal sociocultural norms as well as economic, educational, and health policies that support or condone IPV against women within a society. Societal-level factors associated with IPV in current literature have included religious support of IPV, society that promotes male dominance, normalization of IPV, socioeconomic inequalities, and lack of civil rights and legal protections for women experiencing IPV (Guedes et al., 2016; Hardesty & Ogolsky, 2020; Lutgendorf, 2019; Mengistu, 2019; WHO & Pan American Health Organization, 2012). In the present study, the social–ecological model was used as a foundational framework to identify and categorize variables that were available in a women's health registry dataset as individual, relational, community, or societal. Furthermore, the social–ecological model helped lay the groundwork to inform the exploration of understudied societal-level factors (i.e., the health, economic, and government policies) and women's self-reported experiences of IPV—factors which are often overlooked in the development of effective and efficient IPV interventions.

Despite the well-documented factors discussed above, it is plausible that additional factors associated with increased risk for experiencing IPV or high risk for revictimization of IPV have not been substantiated in the current IPV literature. Such potential factors worthy of further exploration have included the woman's status of receiving US government benefits (e.g., welfare, financial assistance, and legal financial aid) and issues of access to healthcare services. Prior studies have shown perpetrators of IPV use a variety of assaultive and coercive behaviors to obstruct women from sustained employment, education, and professional training. This results in women feeling inadequate and financially dependent on their perpetrators and thus staying in the abusive relationship (Lyon, 2002; Postmus et al., 2012). For example, between 20% and 50% of women receiving US government benefits or welfare or legal financial aid called Temporary Assistance for Needy Families (TANF), a program that replaced Aid to Families with Dependent Children (AFDC), reported experiencing IPV in the past 12 months, and 50–60% reported IPV experiences in their lifetime (An & Choi 2019; Pompa, 2007; Steiner et al., 2019). If women in the US receiving government benefits are more likely to experience IPV, then these women may need additional financial assistance from government services to break the cycle of violence (Whitesell, 2019).

Another factor worthy of further exploration is access to healthcare. Women who experience IPV report poorer health status and less access to healthcare services compared to women with no experience of IPV (Massetti et al., 2018; Thomas et al., 2020). Adverse health outcomes of IPV among women who have experienced IPV have been well documented in the current literature; however, to our knowledge, there remains limited information about barriers to healthcare access among women who experience IPV in the US and worldwide. Thus, there is an urgent need for studies to explore and identify potential associations between receipt of US government benefits, barriers to healthcare access, and women's experiences of IPV. Generating evidence about these understudied factors and IPV is important to the development of future IPV interventions and proposing changes to current health, economic, and government policies.

Purpose

The purposes of this secondary analysis study were twofold. The primary aim was to explore the wide variety of variables present in an existing women's health registry dataset to identify if understudied factors (i.e., government benefits or welfare or financial assistance or legal financial aid and issues of access to healthcare services) were associated with women's experiences of physical IPV, sexual IPV, and stalking by an intimate partner. The secondary aim was to identify which factors were associated with all three IPV experiences. While the overall goal of this study was to explore associations between understudied factors and three forms of IPV experiences, all factors available within the registry were included to provide all the overarching factors that may be associated with different forms of IPV experiences.

Methods

We conducted a secondary analysis of existing women's health registry data. For the purposes of this study, the research team used the most recent self-reported responses from a registry participant to identify experiences of any three forms of IPV.

Data Source

The data source was the Kentucky Women's Health Registry (KWHR) currently known as Wellness, Health, and You dataset. The KWHR is a longitudinal study aimed at understanding how risk and protective behavioral factors might influence women's health in the state of Kentucky, in the US. Data used in this analysis were gathered through annual health surveys developed by the University of Kentucky (UK) research team (Sprang et al., 2020). The resultant dataset included data gathered from 16,645 women via convenience sampling from 2006 to 2014. The UK research team established the following inclusion criteria for participating in the KWHR: (a) self-identify as a woman, (b) age of 18 years and older, and (c) residence in the state of Kentucky (US). In the current study, a total of 8886 women who self-reported experiences of any of three forms of IPV (i.e., physical, sexual, and stalking) were included.

Ethical Considerations

Analysis of the KWHR dataset began once approval from the University of Cincinnati Institutional Review Board was received. The research team requested data for this project through UK's Center for Clinical and Translational Science and the UK Clinical Research Support Office. Data received from the KWHR was stored using password-protected digital files on the research-designated server of the authors’ computer in University of Cincinnati institution. Personal identifiers had been removed by the UK research team to maintain the confidentiality of participants in the dataset. For this project, only members of the University of Cincinnati research team had access to the deidentified dataset. This project was conducted between October 2020 and September 2021.

Study Approach

Due to a large number of variables contained in the dataset, the research team developed a specific strategy to explore understudied risk factors that might be associated with women's experiences of IPV in the state of Kentucky and examine which factors were associated with all forms of IPV experience. Our evidence-driven strategy was based on a multipronged analytical approach informed by existing literature and the social–ecological model about factors that may be associated with increased risk for experiencing IPV.

First, the research team conducted a comprehensive literature search on IPV against women to create a hierarchy of variables. As depicted in Table 1, using current IPV literature, an evidence-based list comprised of three tiers was generated to guide the exploration of variables associated with women's experiences of three forms of IPV. The first tier included variables that are well established in the IPV literature as being associated with women's experiences of IPV. The second tier contained variables with evidence in the literature as being potentially associated with or a covariate of women's experiences of IPV. The third tier contained variables that have limited or questionable evidence about their association with experiences of IPV. Once an evidence-based tier list was created, the research team went back to the registry dataset with the tier list to determine which available variables might be associated with women's experiences of IPV. Lastly, the identified variables were labeled as individual, relational, community, and societal as informed by the social–ecological model. Finally, the research team performed three separate multiple logistic regression analyses (one for each IPV experience) to explore factors in a tiered fashion that were associated with women's self-reported experiences of three forms of IPV.

Table 1.

A Summary of Variables Included in Tier 1, 2, and 3 Lists.

Tier 1 variables (Established in literature as associated with experiences of IPV)	Tier 2 variables (Documented in literature as potentially associated with or covariate of women’s experiences of IPV)	Tier 3 variables (Limited or questionable evidence about association with experiences of IPV)
Sociodemographic factors Race Ethnicity Age Level of education Employment status Occupation Income Marital status Past experience of VAW Non-Partner Sexual Violence Childhood experience of violence Physical abuse Sexual abuse Substance use disorders Drug Alcohol	Health status General health status Physical health issues Mental health issues Stress Social support Need friend’s support Family care for children or adults Housing House type House payment Living with someone (partner, children, parents, others) or alone Use of health services Emergency room visit in the past 12 months Admitted to the hospital in the past 12 months Satisfaction with healthcare services	Government benefits Insurance type Barriers to healthcare access Get to doctor appointments Could not get an appointment Type of medical care provider No money to pay the co-pay/cost of a visit Transportation problems No time off work Forgot my insurance/Medicaid card Childcare problem Too busy Did not want anyone to know I had a problem Too sick/too tired to go

Variables

Dependent variables. In the registry, each question about the woman's experience with IPV was captured using the following answer options: “yes, in the past 12 months,” “yes, in my lifetime,” “no,” and “no response.” The question about physical IPV was: Has an intimate partner hit, kicked, punched, or otherwise hurt you? The sexual IPV question asked: Has an intimate partner used force (like hitting, holding down, or using a weapon) to make you have sex? Finally, the question about stalking by an intimate partner read: Has an intimate partner ever repeatedly followed you, spied on you, made unsolicited phone calls to your place of work or at home, damaged your property, or stalked you in any way? The registry did not include questions asking women about past experiences of emotional, psychological, or financial forms of IPV. For this study, participant responses of “yes, in the past 12 months” and “yes, in my lifetime” were combined to represent an affirmative response to women's experiences of IPV. Therefore, each of the three IPV outcomes was coded as dichotomous variables with category 0 for no past IPV experience reported and category 1 for past IPV experience reported.

Independent variables. Tier 1 list variables included factors established in IPV literature and primarily consisted of individual-level factors such as sociodemographic variables (i.e., race, ethnicity, age, level of formal education, employment status, occupation, income, and marital status, sexual violence by other than an intimate partner, childhood experiences of physical and sexual violence, and substance use disorders). Tier 2 list variables contained factors from existing literature with some evidence of being associated with or a covariate of women's experiences of IPV. The tier included a mixture of individual, relational, community, and societal levels of factors (i.e., general health status, physical and mental health issues, level of stress, social support, religious or spiritual beliefs, housing issues, and use of healthcare services). Finally, Tier 3 list variables included factors for which limited or questionable association with women's experiences of IPV is documented in the current literature. This tier contained a large societal level of risk factors (i.e., government benefits, health insurance type, and issues of access to healthcare services) presented in Table 1.

Data Analysis

After creating the hierarchal tier lists, the research team performed multiple logistic regression analyses to quantitatively explore factors in each tier list that may be associated with women's past experiences of three forms of IPV. For each outcome, the research team followed a model-building strategy that respected the established tiers along with a common regression model-building method. The multiple logistic regression analyses were performed with the data from the 8886 women in the registry who self-reported experiencing at least one form of IPV.

Model-Building Strategy

For each of the three outcomes (i.e., physical IPV, sexual IPV, and stalking by an intimate partner), the research team built a multiple logistic regression model via the following strategy. Each model first performed forward selection (p-values entered equal to 0.05, and p-values removed equal to 0.10) with the tier 1 variables. After tier 1 variables were entered, they were then fixed in the model and could not be removed. Tier 2 variables were considered next, and forward selection (p-values entered equal to 0.05, and p-values removed equal to 0.10) was used on these variables while adjusting for the fixed tier 1 variables from the previous forward selection. Lastly, tier 3 variables were considered, and forward selection (p-values entered equal to 0.05, and p-values removed equal to 0.10) was used on these variables while adjusting for the fixed tier 1 and 2 variables from the previous forward selection procedures. The resultant model respects the importance of the large nature of the data and the hierarchy of the variables’ importance as found in the existing literature. After the three regression models were built using the model-building strategy, a Venn diagram was created to visually display the associations and the shared risk factors found in one, two, or across all three of the outcomes (i.e., physical IPV, sexual IPV, and stalking by an intimate partner) under investigation (see Figure 1). Individual adjusted odds ratios (OR) and their 95% confidence intervals (CIs) are also examined for general trend direction and range.

Figure 1.

Venn Diagram.

Based on our model-building strategy outlined above, the model appreciates the hierarchy of the variables according to the previous literature first, followed by the statistical evidence we found about those variables within the existing data. Due to this strategy, some variables in higher tiers may later become nonstatistically significant after lower-tier variables are entered. Rather than removing those variables at this step, we wanted to appreciate the previously established hierarchy and keep the variable in the model rather than removing it. For this reason, the final model for each IPV outcome may contain some variables that have confidence intervals (CIs) that contain 1; however, those variables were significant (p-value < 0.05) at the time they were entered into the model.

Applying this model-building strategy guided our examination of the associations and connections between factors and three forms of IPV experience within the health registry dataset. While the research team applied a rigorous data analysis approach, there may be some missing data. In this study, the missing data may be explained for several reasons. First, the health registry questions were revised over time, so the changes in the survey questions or items may explain why some of the data may be missing. Second, some of the survey responses (i.e., do not know, choose not to answer, or skipped question) were coded as blank. Finally, the sensitivity of questions related to IPV experience may cause the missing responses.

Results

Sociodemographic Characteristics

The overall characteristic of participants in the registry were self-identified women who were between 18 and 110 years old, were predominately (95%) White or Caucasian, had a higher level of formal education (i.e., 22% bachelor's degree and 33% postgraduate training), and had higher annual house income (i.e., $75,000 or more) than national average. Participants who self-reported experiences of physical IPV were on average 48.6 years old. Participants who self-reported experiences of sexual IPV were on average 47.9 years old. Participants who self-reported experiences of stalking by an intimate partner were on average 45.7 years old. Participants who self-reported experiences of any form of IPV were White or Caucasian and had lower education and an annual income level compared to participants with no self-reported experiences of IPV. Table 2 shows women's self-reported experiences of three forms of IPV, and Table 3 demonstrates the sociodemographic characteristics of participants with and without self-reported experiences of physical IPV, sexual IPV, and stalking by an intimate partner.

Table 2.

Women’s Self-Reported Experiences of Physical IPV, Sexual IPV, and Stalking by an Intimate Partner Anytime in the Past From Registry Dataset.

Self-reported experiences of IPV	N	% of total sample	% of IPV reporters
Physical IPV only	1494	9.0%	28.3%
Sexual IPV only	178	1.1%	3.4%
Stalking IPV only	968	5.8%	18.4%
Physical and sexual IPV only	444	2.7%	8.4%
Sexual and stalking IPV only	104	0.6%	2.0%
Stalking and physical IPV only	1099	6.6%	20.8%
Physical, sexual, & stalking IPV	984	5.9%	18.7%
Total IPV reporters	5271	31.7%
No IPV reported	11,374	68.3%	N/A
Total sample	16,645

Table 3.

Sociodemographic Characteristics of Participants With and Without Self-Reported Experiences of Physical IPV, Sexual IPV, and Stalking by an Intimate Partner.

	Physical IPV				Sexual IPV				Stalking IPV
	No		Yes		No		Yes		No		Yes
	Row %	N	Row %	N	Row %	N	Row %	N	Row %	N	Row %	N
Race/ethnicity
American Indian or Alaska Native	55.41%	128	44.59%	103	74.67%	171	25.33%	58	64.94%	150	35.06%	81
Asian	83.84%	83	16.16%	16	92.00%	92	8.00%	8	83.84%	83	16.16%	16
Native Hawaiian or other Pacific Islander	71.43%	15	28.57%	6	81.82%	18	18.18%	4	77.27%	17	22.73%	5
White or Caucasian	76.08%	11,911	23.92%	3744	90.00%	14,079	10.00%	1565	81.14%	12,709	18.86%	2955
Race other	61.19%	41	38.81%	26	81.82%	54	18.18%	12	62.69%	42	37.31%	25
Hispanic or Latino	73.53%	75	26.47%	27	86.27%	88	13.73%	14	77.45%	79	22.55%	23
Highest level of education
Less than high school	50.00%	23	50.00%	23	77.78%	35	22.22%	10	65.91%	29	34.09%	15
Some high school	50.70%	72	49.30%	70	75.35%	107	24.65%	35	63.19%	91	36.81%	53
High school diploma	75.81%	962	24.19%	307	90.94%	1155	9.06%	115	85.15%	1078	14.85%	188
GED	49.65%	142	50.35%	144	75.89%	214	24.11%	68	68.18%	195	31.82%	91
Some college	68.62%	1953	31.38%	893	86.14%	2455	13.86%	395	75.81%	2163	24.19%	690
Vocational or technical certificate or degree—2007	67.97%	471	32.03%	222	84.93%	586	15.07%	104	76.09%	525	23.91%	165
Associate degree	68.41%	1072	31.59%	495	86.65%	1356	13.35%	209	75.80%	1184	24.20%	378
Baccalaureate degree	78.87%	2967	21.13%	795	91.47%	3440	8.53%	321	81.85%	3081	18.15%	683
Postgraduate training	82.21%	4556	17.79%	986	92.65%	5129	7.35%	407	85.27%	4734	14.73%	818
Education other	75.99%	231	24.01%	73	87.87%	268	12.13%	37	77.56%	235	22.44%	68
Current legal marital status
Married	80.27%	8416	19.73%	2069	92.04%	9637	7.96%	834	85.05%	8918	14.95%	1567
Separated	45.02%	113	54.98%	138	71.08%	177	28.92%	72	52.02%	129	47.98%	119
Divorced	54.75%	1343	45.25%	1110	79.64%	1952	20.36%	499	65.32%	1603	34.68%	851
Widowed	79.10%	753	20.90%	199	91.40%	872	8.60%	82	87.42%	834	12.58%	120
Never married	78.80%	1743	21.20%	469	90.66%	2010	9.34%	207	79.09%	1755	20.91%	464
Primary job/occupation
Professional	78.03%	8251	21.97%	2323	91.02%	9614	8.98%	949	82.26%	8702	17.74%	1877
Clerical	72.18%	3412	27.82%	1315	87.74%	4145	12.26%	579	78.75%	3724	21.25%	1005
Retail	65.82%	1483	34.18%	770	83.76%	1888	16.24%	366	72.19%	1633	27.81%	629
Factory work	54.15%	437	45.85%	370	75.96%	613	24.04%	194	65.76%	532	34.24%	277
Farming	64.29%	369	35.71%	205	84.52%	486	15.48%	89	70.12%	399	29.88%	170
Mining	9.09%	1	90.91%	10	72.73%	8	27.27%	3	72.73%	8	27.27%	3
Job other	70.55%	1224	29.45%	511	85.77%	1483	14.23%	246	77.10%	1340	22.90%	398
Job status
Working for pay at one job	76.08%	7071	23.92%	2223	90.39%	8386	9.61%	892	80.46%	7476	19.54%	1816
Working for pay at more than one job	70.74%	979	29.26%	405	86.79%	1202	13.21%	183	76.90%	1062	23.10%	319
Working without pay in a family business	75.52%	108	24.48%	35	88.65%	125	11.35%	16	88.03%	125	11.97%	17
Currently seeking work	60.10%	119	39.90%	79	80.30%	159	19.70%	39	62.24%	122	37.76%	74
On layoff from a job	65.19%	88	34.81%	47	86.03%	117	13.97%	19	71.32%	97	28.68%	39
Homemaker	83.27%	672	16.73%	135	93.05%	750	6.95%	56	87.14%	705	12.86%	104
Student	78.37%	471	21.63%	130	88.43%	535	11.57%	70	75.87%	459	24.13%	146
Retired	84.06%	2415	15.94%	458	94.47%	2718	5.53%	159	91.92%	2649	8.08%	233
Disabled/unable to work	45.20%	339	54.80%	411	69.71%	520	30.29%	226	55.73%	418	44.27%	332
Job other	69.30%	158	30.70%	70	85.22%	196	14.78%	34	75.88%	173	24.12%	55
Annual household income—2011
Less than $10,000	40.71%	103	59.29%	150	66.27%	167	33.73%	85	51.00%	128	49.00%	123
Less than $15,000	54.40%	99	45.60%	83	78.02%	142	21.98%	40	68.31%	125	31.69%	58
Less than $20,000	65.02%	132	34.98%	71	86.83%	178	13.17%	27	71.92%	146	28.08%	57
Less than $25,000	66.10%	234	33.90%	120	83.76%	294	16.24%	57	77.30%	269	22.70%	79
Less than $35,000	69.93%	535	30.07%	230	86.09%	656	13.91%	106	78.04%	597	21.96%	168
Less than $50,000	76.60%	1080	23.40%	330	90.12%	1268	9.88%	139	82.44%	1160	17.56%	247
Less than $75,000	78.53%	1496	21.47%	409	91.71%	1748	8.29%	158	83.83%	1597	16.17%	308
$75,000 or more	82.31%	3485	17.69%	749	92.71%	3916	7.29%	308	86.38%	3660	13.62%	577
Don’t know/not sure	68.24%	116	31.76%	54	87.72%	150	12.28%	21	77.06%	131	22.94%	39
House type
House	77.90%	10,749	22.10%	3049	90.98%	12,548	9.02%	1244	82.79%	11,427	17.21%	2375
Apartment	68.81%	867	31.19%	393	85.32%	1075	14.68%	185	73.87%	933	26.13%	330
Condominium	75.32%	412	24.68%	135	89.23%	489	10.77%	59	81.97%	450	18.03%	99
Trailer/mobile home	58.01%	315	41.99%	228	80.11%	435	19.89%	108	67.95%	371	32.05%	175
Shelter/group home	23.68%	54	76.32%	174	58.04%	130	41.96%	94	33.93%	76	66.07%	148
No permanent home	92.31%	24	7.69%	2	96.15%	25	3.85%	1	100.00%	26	0.00%	0
Choose not to answer	42.86%	18	57.14%	24	69.05%	29	30.95%	13	59.52%	25	40.48%	17
Government benefits/welfare/financial aid
No	81.06%	9123	18.94%	2131	92.61%	10,424	7.39%	832	83.92%	9449	16.08%	1810
Yes	63.28%	3171	36.72%	1840	82.73%	4135	17.27%	863	73.92%	3711	26.08%	1309

Note. N = 4021. Participants who self-reported experience of physical IPV were on average of 48.6 years old. Participants who did not self-report experience of physical IPV were on average of 50.7 years old.

N = 1710. Participants who self-reported experience of sexual IPV were on average of 47.9 years old. Participants who did not self-report experience of sexual IPV were on average of 50.4 years old.

N = 3155. Participants who self-reported experience of stalking by an intimate partner were on average of 45.7 years old. Participants who did not self-report experience of stalking by an intimate partner were on average of 51.2 years old.

Recurrent Factors Retain Across Physical IPV, Sexual IPV, and Stalking

The following section represents the recurrent factors that are retained across physical IPV, sexual IPV, and stalking by an intimate partner depicted in Figure 1 (see Appendix A for reported adjusted odds ratio [OR] and individual 95% confidence intervals [CIs] for each IPV outcome).

Tier 1 Variables and IPV Experiences

Physical IPV. Women who reported more formal education (i.e., postgraduate training, baccalaureate degree, vocational or technical certificate or degree, or other education) were estimated to have lower adjusted odds of physical IPV (adjusted OR  =  0.744, 95% CI [0.613, 0.903]) compared to women who reported less formal education (i.e., some high school, GED, high school diploma, some college, or associate degree). Women who had not been exposed to nonpartner sexual violence were estimated to have lower adjusted odds of physical IPV (adjusted OR  =  0.379, 95% CI [0.287, 0.5]) compared to women who had been exposed to nonpartner sexual violence in the past 12 months and in their lifetime.

Sexual IPV. Women who reported their highest level of education (i.e., postgraduate training, associate degree or some college, or other education) were estimated to have higher adjusted odds of sexual IPV (adjusted OR  =  1.073, 95% CI [0.823, 1.398]) compared to women with less formal education (i.e., a vocational or technical certificate, baccalaureate degree, general educational development, high school diploma, or some high school). Women who had an annual household income less than $20,000, less than $35,000, or $75,000 or more were estimated to have lower adjusted odds of sexual IPV (adjusted OR  =  0.555, 95% CI [0.390, 0.791]) compared to women who reported their income in the categories of less than $10,000 and less than $25,000. Women who had not been exposed to nonpartner sexual violence were estimated to have lower adjusted odds of sexual IPV (adjusted OR  =  0.218, 95% CI [0.162, 0.292]) compared to women who had been exposed to nonpartner sexual violence in the past 12 months and in their lifetime.

Stalking by an Intimate Partner. Women who reported more formal education (i.e., postgraduate training, baccalaureate degree, vocational or technical certificate or degree, or other education) were estimated to have lower adjusted odds of stalking (adjusted OR  =  0.876, 95% CI [0.578, 1.327]) compared to women who reported less formal education (i.e., some high school, GED, high school diploma, some college, or associate degree). Women who had an annual income of more than $35,000 were estimated to have lower adjusted odds of stalking (adjusted OR  =  0.781, 95% CI [0.569, 1.073]) compared to women who had an annual income less than $25,000 or less than $10,000. Women who had not been exposed to nonpartner sexual violence were estimated to have lower adjusted odds of stalking (adjusted OR  =  0.384, 95% CI [0.29, 0.507]) compared to women who had been exposed to nonpartner sexual violence in the past 12 months and in their lifetime.

Tier 2 Variables and IPV Experiences

Physical IPV. Women who reported currently living in a house, condominium, or apartment were estimated to have lower adjusted odds of physical IPV (adjusted OR  =  0.649, 95% CI [0.438, 0.96]) compared to women who reported living in a trailer, mobile home, shelter, or group home. Women who lived with children were estimated to have higher adjusted odds of physical IPV (adjusted OR  =  1.119, 95% CI [0.891, 1.404]) compared to women who did not live with children. Women who reported not feeling the need for a friend's support were estimated to have higher adjusted odds of physical IPV (adjusted OR  =  1.208, 95% CI [0.909, 1.607]).

Women who reported experiencing a traumatic event (i.e., women having experienced or witnessed actual or threats of death or serious injury that involved intense fear or horror in the past) were estimated to have higher adjusted odds of physical IPV (adjusted OR  =  2.052, 95% CI [1.669, 2.523]). Women who had not experienced panic attacks within two straight weeks in their lifetime and the past 12 months were estimated to have lower adjusted odds of physical IPV (adjusted OR  =  0.733, 95% CI [0.596, 0.901]) compared to women who had panic attacks within two straight weeks in their lifetime and the past 12 months. Women who reported no trouble concentrating on things at least two straight weeks in their lifetime were estimated to have lower adjusted odds of physical IPV (adjusted OR  =  0.713, 95% CI [0.571, 0.89]) compared to women who had trouble concentrating on things at least two straight weeks in the past 12 months.

Sexual IPV. Women who reported currently living in a house, condominium, or apartment were estimated to have lower adjusted odds of sexual IPV (adjusted OR  =  0.758, 95% CI [0.489, 1.176]) compared to women who lived in a trailer, mobile home, shelter, or group home or had no permanent home. Women who lived with children were estimated to have higher adjusted odds of sexual IPV (adjusted OR  =  1.400, 95% CI [1.059, 1.851]) compared to women who did not live with children. Women who did not live with their husbands were estimated to have higher adjusted odds of sexual IPV (adjusted OR  =  1.141, 95% CI [0.873, 1.492]). Women who reported not feeling the need for a friend's support were estimated to have higher adjusted odds of sexual IPV (adjusted OR  =  1.214, 95% CI [0.274, 0.857]).

Women who reported experiencing a traumatic event (i.e., women having experienced or witnessed actual or threats of death or serious injury that involved intense fear or horror in the past) were estimated to have higher adjusted odds of sexual IPV (adjusted OR  =  2.166, 95% CI [1.653, 2.838]). Women who had not experienced panic attacks within two straight weeks in their lifetime and the past 12 months were estimated to have lower adjusted odds of sexual IPV (adjusted OR  =  0.785, 95% CI [0.584, 1.055]) compared to women who had panic attacks within two straight weeks in their lifetime and the past 12 months. Women who reported that sleep does not refresh or restore their energy were estimated to have higher adjusted odds of sexual IPV (adjusted OR  =  1.337, 95% CI [1.006, 1.777]).

Stalking by an Intimate Partner. Women who reported currently living in a house, condominium, or apartment were estimated to have lower adjusted odds of stalking (adjusted OR  =  0.831, 95% CI [0.561, 1.23]) compared to women who reported living in a trailer, mobile home, shelter, or group home. Women who lived with children were estimated to have higher adjusted odds of stalking (adjusted OR  =  1.264, 95% CI [0.997, 1.602]) compared to women who did not live with children. Women who did not live with their husbands were estimated to have higher adjusted odds of stalking (adjusted OR  =  1.636, 95% CI [1.289, 2.077]).

Women who reported experiencing a traumatic event (i.e., women having experienced or witnessed actual or threats of death or serious injury that involved intense fear or horror in the past) were estimated to have higher adjusted odds of stalking (adjusted OR  =  1.869, CI [1.491, 2.342]). Women who had not experienced panic attacks within two straight weeks in their lifetime and the past 12 months were estimated to have lower adjusted odds of stalking (adjusted OR  =  0.751, 95% CI [0.591, 0.954]) compared to women who had panic attacks within two straight weeks in their lifetime and the past 12 months. Women who reported that sleep does not refresh or restore their energy were estimated to have higher adjusted odds of stalking (adjusted OR  =  1.173, 95% CI [ 0.925, 1.488]). Women who reported no trouble concentrating on things at least two straight weeks in their lifetime were estimated to have lower adjusted odds of stalking (adjusted OR  =  0.718, 95% CI [0.558–0.925]) compared to women who had trouble concentrating on things at least two straight weeks in the past 12 months.

Physical IPV	Sexual IPV	Stalking	Recurrent
1 = Marital status  2 = Problem quit drinking 3 = Pain pills illegally  4 = Live in religious order 5 = General health status 6 = WIC - Did not receive  7 = Food Stamps - Did not receive 8 = Back chronically painful 9 = Other body areas chronically painful 10 = Religious/Spiritual Beliefs keep me from using alcohol, caffeine, or tobacco products	12 = American Indian or Alaska Native 13 = Retail 14 = Too busy 15 = Days poor physical health 16 = Little interest/pleasure in doing things at least two straight days 17 = Routine check-up 18 = Satisfied with the amount of time you had to wait 19 = Take medications as directed	24 = Medicare 25 = Government benefits 26 = Get to doctor appointments 27 = Couldn't get an appointment 28 = Childcare problem 29 = Takes OTC pain medicine 30 = House payment 31 = Live partner 2007- 32 = Currently the amount of stress 33 = Takes OTC sleeping medicine 34 = Takes OTC stimulants 35 = Methamphetamine	11= Needs friend’s support 20 = Annual Household income 21= Too sick/too tired to go 22 = Lives with Husband 23 = Sleep does not refresh or restore my energy 36 = Trouble concentrating on things at least two straight weeks 37 = Highest levels of Education 38 = Non-partner sexual violence (anyone other than an intimate partner) 39 = House type 40 = Lives with children 41 = AFDC received more than 1 year ago 42 = Traumatic event 43 = Panic or episodes of panic attacks at least two straight weeks

Tier 3 Variables and IPV Experiences

Physical IPV. Women who had received US government financial assistance more than 1 year ago from the Temporary Assistance for Needy Families (TANF), a program that replaced Aid to Families with Dependent Children (AFDC), were estimated to have higher adjusted odds of physical IPV (adjusted OR  =  2.02, 95% CI [1.430, 2.853]).

Sexual IPV. Women who had received US government financial assistance more than 1 year ago from the Temporary Assistance for Needy Families (TANF), a program that replaced Aid to Families with Dependent Children (AFDC), were estimated to have higher adjusted odds of sexual IPV (adjusted OR  =  2.155, 95% CI [1.489, 3.120]). Women who had some issues related to access to healthcare services (i.e., feeling too sick and/or tired to go to a doctor) were estimated to have higher adjusted odds of sexual IPV (adjusted OR  =  1.379, 95% CI [0.959, 1.983]).

Stalking by an Intimate Partner. Women who had received US government financial assistance more than 1 year ago from the Temporary Assistance for Needy Families (TANF), a program that replaced Aid to Families with Dependent Children (AFDC), were estimated to have higher adjusted odds of stalking (adjusted OR  =  1.927, 95% CI [1.379, 2.692]). Women who had some issues related to access to healthcare services (i.e., feeling too sick and/or tired to go to a doctor) were estimated to have higher adjusted odds of stalking (adjusted OR  =  1.469, 95% CI [1.072–2.013]).

Discussion

This study explores factors that might be associated with increased risk for experiencing IPV among women within the existing health registry dataset. Results of three separate multiple logistic regression analyses (one for each type of IPV experience) found several recurrent factors retained across all three IPV experiences (i.e., physical IPV, sexual IPV, and stalking by an intimate partner). These factors include low levels of formal education, past experiences of nonpartner sexual violence, residential instability, presence of children, experiences of a traumatic event and panic attacks, status of receiving US government benefits (i.e., AFDC or TANF) or welfare or financial assistance or legal financial aid, and barriers to healthcare access. In addition, several tier 3 factors appear to be associated with at least one form of IPV experience. Figure 1 displays a Venn diagram of the retained factors found to be associated with one form of IPV or shared between two or all three forms of IPV.

Study results extend prior evidence of well-documented risk factors for experiencing different forms of IPV among women. For instance, according to the hierarchical evidence-based tier list and social–ecological model, a considerable number of factors (i.e., individual, relational, community, and societal levels) demonstrate significance in this registry as well as overlap in tiers 1, 2, and 3 across three forms of IPV experience (Mengistu, 2019; WHO, 2012). These factors include lower socioeconomic status, marital status, housing issues, substance use disorders, childcare issues, living with children, social or friends support, poor health status, experiences of a traumatic event, and nonpartner sexual violence (Cho et al., 2020; Li et al., 2019; Lutgendorf, 2019; Niolon et al., 2017). The overall goal of this study was to provide a comprehensive view of factors that may be associated with three forms of IPV experience within the registry; the primary aim was to explore for potential associations between tier 3 factors (i.e., US government benefits, health insurance characteristics, and issues of access to healthcare) and women's experiences of IPV because these factors have either been understudied or have limited support in existing literature. Indeed, study results show that US government benefits or welfare or financial assistance (i.e., food stamps, AFDC, and WIC) or legal financial aid, issues of access to healthcare, and health insurance type were all factors in the tier 3 list associated with three forms of IPV experience. In addition, the results of this study underline the urgent need for more research examining large societal-level risk factors or policies associated with women's economic and education level, structural barriers to accessing healthcare services, and experiences of IPV in the US and across the globe (CDC, 2022; Krug et al., 2002). In the context of US, Whitesell (2019) found that women who receive government benefits have low socioeconomic status and are more likely to experience IPV. Likewise, women receiving government benefits have experienced a higher prevalence of IPV and IPV-related health issues (An & Choi, 2019; Yoshihama et al., 2006). Experiencing IPV may limit a woman's ability to comply with minimum work requirements mandated by the US federal government to receive the TANF program benefits (Booshehri et al., 2018; Gallagher, 2011; Spencer et al., 2020; Wahler et al., 2015). Hsu (2017) examined the association between the timing of TANF transfer payments and reporting IPV to the police across 21 states in the US. In the same study, the researcher noted “a causal relationship between timing close to welfare payments and IPV against women” (pp. 1029–1030). Moreover, the current study results show several factors retained that are associated with accessing healthcare services in women experiencing IPV that warrant further exploration. These factors include barriers to scheduling an appointment, poor mental and physical health status, low satisfaction with received healthcare services, health insurance type, transportation to healthcare services, and childcare issues (Makaroun et al., 2020; Saletti-Cuesta et al., 2018; The National Consensus Guidelines, 2014).

Research Implications

The primary focus of this study was to lay the groundwork for future research exploring potential causal relationships between factors in the evidence-based tier list (i.e., tiers 1, 2, and 3) and three forms of IPV experience. Results could inform the development of IPV prevention programs and interventions. Indeed, study results contribute to future research on IPV prevention by providing preliminary evidence of emerging factors associated with women's experiences of three forms of IPV. While the research team urges readers to interpret the estimates with caution due to limitations with using secondary data from a health registry, it is important that future research explores possible associations between three forms of IPV experience and factors in the tier 3 list. Furthermore, future research is needed to extend newly identified large societal-level risk factors in the tier 3 list as well as explore the associations between experiences of emotional, psychological, or financial IPV among women from diverse sociodemographic and cultural backgrounds. To our knowledge, there is limited published research examining the associations between US government benefits or welfare or financial assistance or legal financial aid and three forms of IPV (i.e., sexual IPV, physical IPV, and stalking by an intimate partner) as well as psychological and financial IPV experience. Future research needs to explore possible causal relationships between different forms of IPV experience and receipt of US government benefits. It is well documented that IPV is associated with poor mental and physical health, which in turn, detrimentally affects women's ability to work and follow the requirements associated with receiving government benefits in the US. Thus, future community-based IPV interventions may partner with government assistance departments in the development of effective and efficient IPV prevention programs among women from different sociodemographic backgrounds.

Policy Changes in Healthcare Settings

Nurses play a vital role in the development of and changes to policies in healthcare settings. Currently, consistent use of IPV assessment protocols contributes to the prevention of IPV, timely identification of patients who experience IPV, and assessment of IPV-related health issues. The routine assessment protocols need to be comprehensive and include individual-, relational-, community-, and societal-level risk factors for women's experiences of various forms of IPV. To develop effective and efficient IPV prevention programs, it is important to understand the complexity of multiple levels of factors and women's various experiences of IPV and integration strategies to eliminate structural barriers in accessing healthcare services among women experiencing IPV from different demographic and cultural backgrounds.

Policy Changes at the Societal Level

Study results underscore an urgent need for societal policy changes to improve the health, educational, and economic statuses as well as increase the availability and accessibility of resources for women experiencing or survivors of IPV. Financial empowerment (i.e., not being financially dependent on a partner) is an important protective factor in preventing IPV against women. In other words, not having financial empowerment could result in not leaving the perpetrator and increased risk for revictimization of IPV. Financial empowerment policy changes need to focus on increasing women's education and potential for income, employment, and housing opportunities so that women experiencing IPV or survivor of IPV could be financially independent. However, in the US, women who seek financial assistance from the government encounter policy issues such as federal work requirements. It is important to understand women receiving financial assistance are more likely to experience IPV and have fewer financial and childcare resources. In addition, perpetrators physically and socially isolate women from achieving sustainable employment and professional job training, which in turn causes women's inability to meet the work requirements stated by the government. Women also experience IPV-related health issues, transportation, and childcare barriers, which in turn obstructs women from meeting the work requirements. As a result of these policy issues, women experiencing IPV are unable to follow the policies or regulations for receiving financial assistance that forces women to stay with their perpetrators. Moreover, women experiencing IPV encounter a lack of safe and affordable housing, which in turn makes women's decision of leaving the perpetrator exceedingly difficult. Therefore, there is a critical need to revise the existing regulations for receiving financial assistance to women who experience IPV or survivors of IPV and to increase the availability and accessibility of financial assistance programs, affordable housing and childcare services, and educational and professional job training opportunities so that women feel financially empowered which is one the most important protective factors in preventing IPV against women.

Limitations

Despite the important future research and policy implications of the study, it has some major limitations. First, the current research team was unable to control what data was contained in the dataset—health registry. Since the KWHR study did not employ an experimental design, direct causal relationships between variables cannot be inferred. Due to the original purpose of the health registry to collect information about women's health in Kentucky, the survey questions were developed by the registry developers. The research team also was unable to include relational-level risk factors and psychological aggression as a form of IPV. In addition, imputation methods were unable to be used in this study for several reasons: (a) the nature of data, (b) history of data that has been collected, (c) it is a health registry and gathered via the convenience sampling approach, (d) sensitivity of IPV questions, and (e) no clear idea about underlying mechanisms of some missing data. However, in this study, the research team developed a well-structured analysis approach guided by existing literature to explore possible factors that contribute to three forms of IPV experience. Finally, the purpose of this study was to examine the associations and connections between factors and women's experiences of three forms of IPV with the goal of identifying future research implications and policy changes for healthcare settings.

Moreover, results were generated from a sample that is different from the sociodemographic characteristics of the entire US. Participation of women who are racially and ethnically underprivileged minorities in the registry was low. This was likely due to the lack of racial and ethnic heterogeneity in Kentucky's population (87.5% Caucasian only and 96.9% non-Hispanic or Latino) during the period registry data were collected (US Census Bureau, n. d.). Therefore, results cannot be construed to represent the entire population of women who experience IPV in the US, as the study population is less racial and ethnically and culturally diverse. Although there are some major limitations to using the secondary data, it was valuable to use the health registry dataset. Important advantages of using the KWHR data included (a) a comprehensive dataset which is a health registry of women's health in a southern state in the US and (b) inclusion of many variables that may be associated with women's experience to three forms of IPV.

Conclusion

The overall purpose of this secondary analysis study was to explore factors in the existing registry that may be associated with women's experiences of three forms of IPV. In fact, the overarching goal was not to estimate women's experiences of three forms of IPV and factors; rather, it was to explore potential associations between these variables and three forms of IPV and to provide implications for further research. In addition, this study was aimed at exploring potential associations between factors in the tier 3 list that have been understudied in the current literature. IPV is a complex public health issue detrimentally affecting the health of millions of women regardless of race, ethnicity, education, and economic status around the world. Results of the study showed that there are important connections or shared risk factors among several factors and three forms of IPV experience. Future research is needed to expand our understanding of the shared and potential societal-level risk factors and three forms of IPV (i.e., sexual IPV, physical IPV, and stalking by an intimate partner) as well as psychological and financial IPV experience so effective IPV prevention programs can be developed. There is a need for effective routine assessment protocols for IPV in healthcare settings. In the context of US, there is an urgent need for policy changes to revise the regulations of receiving government benefits or welfare or financial assistance or legal financial aid to women who experience IPV or survivors of IPV. Future research may explore the possible causal relationships between receiving government benefits and barriers to access to healthcare and three forms of IPV experience.