Repetitive Head Injury and Cognitive, Physical, and Emotional Symptoms in Women Survivors of Intimate Partner Violence

Abstract

Many women survivors of intimate partner violence (IPV) experience repetitive head injuries in their lifetime, but limited research has examined the cumulative effects of multiple head injuries on post-concussion symptom severity in this population. This study examined how number of lifetime head injuries and episodes of loss of consciousness (LOC) due to head injuries were related to current cognitive, physical, and emotional symptoms among women survivors of IPV. Cisgender women from Kentucky were recruited following a protective order against an intimate partner, including 268 women with no reported lifetime head injuries and 250 women with one or more IPV-related head injuries (mean [M] = 17.2 head injuries, standard deviation [SD] = 50.5, median [Mdn] = 4, range = 1-515; M = 1.8 LOC episodes, SD = 4.3, Mdn = 1, range = 0-35, respectively). Participants underwent in-person interviews about lifetime physical and sexual IPV history, head injury history, and current cognitive, physical, and emotional symptoms. Sociodemographic characteristics, physical and sexual IPV severity, and current symptom severity were examined in relation to number of head injuries and LOC episodes. A higher number of head injuries was associated with greater age, White race, less than high school education, unemployment, and rural residence. No sociodemographic variables differed based on number of LOC episodes. Greater number of lifetime head injuries and LOC episodes correlated significantly with physical IPV severity (rho = 0.35, p < 0.001; rho = 0.33, p < 0.001, respectively) and sexual IPV severity (rho = 0.22, p < 0.001; rho = 0.19, p = 0.003). Greater number of head injuries and LOC episodes correlated significantly with greater cognitive (rho = 0.33, p < 0.001; rho = 0.23, p < 0.001, respectively), physical (rho = 0.36, p < 0.001; rho = 0.31, p < 0.001), emotional (rho = 0.36, p < 0.001; rho = 0.18, p = 0.004), and total symptom severity (rho = 0.39, p < 0.001; rho = 0.26, p < 0.001). In group comparisons, participant groups stratified by number of head injuries (i.e., 0, 1-3, 4+) differed in total symptom severity (p < 0.001, η² = 0.15), with greater symptom burden associated with more head injuries. Participants with and without LOC differed in symptom severity: cognitive (p < 0.001, d = 0.45), physical (p < 0.001, d = 0.60), emotional (p = 0.004, d = 0.37), and total symptom severity (p < 0.001, d = 0.53). Group differences between participants with and without LOC remained significant after controlling for sociodemographic variables and IPV severity. There was no cumulative effect of LOC, in that participants with 1 LOC episode did not differ from participants with 2 + LOC episodes (p > 0.05). Based on hierarchical regression analyses, only physical symptoms were independently related to number of head injuries (p = 0.008, ΔR² = 0.011) and number of LOC episodes (p = 0.014, ΔR² = 0.021) after controlling for sociodemographic characteristics and IPV severity. Among women survivors of IPV, cumulative head injuries appear related to greater symptom severity. Greater head injury history was independently related to worse physical symptoms (e.g., headaches, dizziness, sleep problems), whereas cognitive and emotional symptoms were, in part, attributable to cumulative physical and emotional trauma due to IPV. Women survivors of IPV with repetitive head injuries have unmet neurobehavioral health needs that may benefit from targeted interventions.

Introduction

Approximately one in two women in the United States (i.e., 47.3%) experiences sexual or physical violence or stalking from an intimate partner in their lifetime.¹ Many injuries from intimate partner violence (IPV) occur to the head or neck,^2,3 which places survivors of IPV at risk for head injuries and traumatic brain injuries (TBIs).^4,5 IPV-related head injuries and TBIs and their sequelae remain underexamined in published research compared to sport- and combat-related neurotrauma in predominantly male samples.^6,7 Early research has associated IPV-related head injuries and TBIs with neurological changes,^8-10 worse cognitive functioning,^11,12 and greater post-concussion symptoms.^9,13–18 The estimated prevalence rates of TBIs in survivors of IPV range from 28-100%, with 7-58% of injuries involving loss of consciousness (LOC).¹⁹ Estimated rates of repetitive TBI among IPV survivors range from 31-75% for injuries involving AOC and 13-28% for injuries involving LOC.¹⁹ IPV-related head injuries encompass a broader spectrum of injury than TBI, involving insult to the head, but not always clinical signs of TBI, such as alteration of consciousness (AOC) or LOC. Researchers have focused on repetitive head impacts, which are operationalized as subconcussive injuries in addition to mild TBIs, as an exposure criterion of traumatic encephalopathy syndrome,²⁰ the clinical disorder associated with chronic traumatic encephalopathy,²¹ but thresholds of exposure for women survivors of IPV have not yet been proposed due to lack of quality research.

Researchers and the public have substantial interest in the effects of repetitive head injuries on health and functioning, with substantial concerns for the potential long-term neurodegenerative effects of cumulative injury.^22,23 Apart from long-term neurodegenerative effects, researchers also have interest in persistent post-concussion symptoms following repetitive head injuries. Post-concussion symptoms reflect common health problems experienced by individuals following a concussion or mild TBI, typically inclusive of cognitive, physical, and emotional symptoms.^24–29 That said, post-concussion symptoms are non-specific neurobehavioral problems, often experienced by populations without known head injury or TBI history, including healthy adults,^30-32 adults with chronic pain,^33,34 and adults with mental health problems, such as depression.³⁵ The relationship between repetitive head injuries and post-concussion symptoms have been examined in other populations at-risk for concussions and mild TBIs, such as athletes^36-38 and Veterans.^39,40 These studies have collectively indicated that post-concussion symptoms are often experienced at greater severity in individuals with multiple lifetime concussions, although not all studies have reached this conclusion.⁴¹ A meta-analysis found only a small, non-significant difference (d = 0.12) in symptom severity between participants with a single mild TBI versus two or more mild TBIs,⁴² and existing literature remains indeterminate on the effects of prior concussions on recovery following a subsequent injury.⁴³

In general, IPV-related head injuries represent an overlooked public health issue,⁴⁴ with prior researchers calling for further evaluation of the sequelae of head injury among survivors.^45-47 Women with IPV-related head injuries have been underrepresented in literature exploring the effects of repetitive head injuries on persistent post-concussion symptoms, despite a high prevalence of repetitive injuries in this population.¹⁹ An understanding of how repetitive head injuries are related to cognitive, physical, and emotional symptom severity in this population may help identify the specific healthcare needs of this population, and whether those needs differ based on lifetime number of head injuries. This study involved an examination of whether cognitive, physical, and emotional symptom severity differed by number of lifetime head injuries in women survivors of IPV, examining differences by lifetime count of head injuries and lifetime count of head injuries involving LOC. The hypotheses were that: 1) greater lifetime head injuries and LOC episodes would be associated with greater symptom severity across domains; and 2) these relationships would persist, even after controlling for relevant sociodemographic characteristics and lifetime IPV severity.

Methods

Participants

Participants included 641 cisgender women from Kentucky who were recruited from courts in urban and rural jurisdictions after receiving a protective order against an intimate partner. Women were eligible if they were: 1) age 18 or older or 17 and emancipated; 2) experienced IPV from a male perpetrator with whom they were married, cohabitated, or shared a child; 3) had a protective order issued; and 4) did not plan on moving out of Kentucky in the next 12 months. Prior research has involved secondary analysis of this sample.^18,48

These women were asked about their history of head injuries, with 268 reporting no head injury history and 373 (58.2%) reporting a lifetime history of one or more head injuries. Among those reporting head injuries, 255 (39.8%) reported experiencing one or more lifetime IPV-related head injuries. Women were excluded from analyses if they: 1) reported prior head injuries, but none of these injuries were due to IPV (n = 118); 2) reported having experienced a coma due to a head injury (n = 4), because a coma indicated a more severe brain injury than most of the sample; or 3) were missing on number of estimated head injuries (n = 1). This resulted in a sample of 268 women survivors of IPV with no reported lifetime head injuries and 250 women survivors with one or more IPV-related head injuries. These women could also report head injuries unrelated to IPV, but at least one lifetime head injury needed to be due to IPV. Participants reported, on average, 8.3 head injuries (SD = 36.1, Mdn = 0, range: 0 to 515). Among participants with one or more head injuries, participants reported, on average, 17.2 head injuries (SD = 50.5; Mdn = 4, interquartile range: 2 to 12, full range: 1 to 515), and 1.8 LOC episodes due to head injuries (SD = 4.3; Mdn = 1, interquartile range: 0 to 2, full range: 0 to 35). A pair of participants (n = 2) were missing on their count of LOC episodes. Full sociodemographic characteristics, stratified by number of head injuries and LOC episodes, are provided in Table 1.

Table 1.

Demographic Characteristics

	No HIs (n = 268)	1-3 HIs (n = 110)	4+ HIs (n = 140)	F/χ2	p	η2/ϕ	No LOC (n = 109)	1 LOC (n = 72)	2 + LOC (n = 67)	F/χ2	p	η2/ϕ
Age, M (SD)	31.8 (9.8)	33.5 (9.1)	34.1 (8.9)	3.11	.046	.01	33.4 (9.0)	33.8 (9.3)	34.9 (8.5)	.58	.559	.01
Race/ethnicity, n (%)
White	207 (77.2%)	87 (79.1%)	133 (95.0%)	21.11	<.001	.20	94 (85.5%)	64 (88.9%)	62 (92.5%)	2.06	.382	.09
Black	53 (19.8%)	18 (16.4%)	3 (2.1%)				13 (11.8%)	6 (8.3%)	1 (1.5%)
Hispanic/Latina	3 (1.1%)	2 (1.8%)	1 (0.7%)				3 (2.7%)	0 (0.0%)	0 (0.0%)
Asian	2 (0.7%)	0 (0.0%)	0 (0.0%)				0 (0.0%)	0 (0.0%)	0 (0.0%)
Native American	2 (0.7%)	0 (0.0%)	1 (0.7%)				0 (0.0%)	0 (0.0%)	1 (1.5%)
Puerto Rican	0 (0.0%)	0 (0.0%)	1 (0.7%)				0 (0.0%)	0 (0.0%)	1 (1.5%)
Biracial	1 (0.4%)	3 (2.7%)	1 (0.7%)				0 (0.0%)	2 (2.8%)	2 (3.0%)
Education, n (%)
Less than high school	66 (24.6%)	31 (28.2%)	55 (39.3%)	9.62	.008	.14	26 (32.7%)	24 (33.3%)	25 (37.3%)	.42	.833	.04
High school or greater	202 (75.4%)	79 (71.8%)	85 (60.7%)				74 (67.3%)	48 (66.7%)	42 (62.7%)
Employment, n (%)
Unemployed	131 (49.2%)	60 (54.5%)	95 (67.9%)	13.42	.001	.16	68 (61.8%)	39 (54.2%)	47 (70.1%)	3.76	.153	.12
Part-time employment	37 (13.9%)	12 (10.9%)	11 (7.9%)				11 (10.0%)	6 (8.3%)	6 (9.0%)
Full-time employment	98 (36.8%)	38 (34.5%)	34 (24.3%)				31 (28.2%)	27 (37.5%)	14 (20.9%)
Area, n (%)
Urban	153 (57.1%)	54 (49.1%)	29 (20.7%)	49.76	<.001	.31	38 (34.5%)	27 (37.5%)	17 (25.4%)	2.54	.280	.10
Rural	115 (42.9%)	56 (50.9%)	111 (79.3%)				72 65.5%)	45 (62.5%)	50 (74.6%)
Income (US$), Mdna	$12.5k	$12.5k	$12.5k	.613	.542	.00	$12.5k	$12.5k	$12.5k	.285	.752	.00

The sum of column sample sizes differ slightly between HI and LOC stratifications because five participants were missing on LOC data. ^aTwo participants were missing on Income, which was the total self-reported annual income from all sources in the past year. The Income range was the same for all groups (< $1,000 to >50,000). χ² test for race was based on percentage White versus percentage other racial/ethnic identities; for education was based on high school degree versus less than a high school degree; and for employment was based on unemployed versus part-time or full-time employed.

HIs, head injuries, LOC, loss of consciousness; M, mean; SD, standard deviation; Mdn, median.

Materials

As part of a larger study related to health outcomes among women with recent protective orders against an intimate partner, participants completed an in-person interview through which they were asked about sociodemographic characteristics, lifetime head injury history, and self-reported problems related to physical and psychological health across multiple domains. Sociodemographic information included age, race/ethnicity, highest level of education, employment status, area of residence (i.e., urban or rural), and annual household income.

Head injury history

The interview questions pertaining to head injury contextualized head injuries as injuries that caused severe pain, being knocked out, comas, hospitalizations, or problems with memory, vision, difficult recovery, or rehabilitation. These head injury responses did not include minor bumps to the head or injuries that did not result in severe pain, actual injury, or LOC. Participants answered whether their head injuries were from the following causes: automobile crash, motorcycle crash, all-terrain vehicle crash, bicycle crash, other vehicle crash, assault/abuse, gunshot wound, sports injury, fall/being pushed or shoved, near drowning, or another mechanism of injury. They indicated the number of times they experienced a head injury due to each mechanism, whether an intimate partner ever caused an injury due to that mechanism, and how many head injuries within each mechanism category were caused by an intimate partner. Participants were also asked whether they ever experienced LOC or a coma due to a head injury, and, if so, how many times. The interview questions on head injury history are provided as an Appendix in prior published research on this cohort.¹⁸

Post-concussion symptoms

Questions related to specific constructs of cognitive, physical, and emotional symptoms were drawn from a longer interview. The selected questions on cognitive, physical, and emotional symptoms did not follow any existing post-concussion symptom questionnaire, but many of the questions overlapped with constructs measured through established instruments.^24–29 Per prior research, the three-factor model had acceptable model fit and each subscale had acceptable reliability: cognitive (ω = 0.88), physical (ω = 0.74), and emotional (ω = 0.88).¹⁸ The Cognitive items included Trouble remembering things, Trouble concentrating, Difficulty making decisions, Your mind going blank, and The idea that something is wrong with your mind. The Physical items included Headaches and head pain, Faintness or dizziness, Nausea or upset stomach, Numbness or tingling in parts of your body, and Trouble falling asleep. The Emotional items included Nervousness or shakiness inside, Feeling easily annoyed or irritated, Temper outbursts that you could not control, Feeling blue, and Feeling no interest in things. One Physical item (i.e., Headaches and head pain) was rated on frequency at which participants experienced these symptoms within the past 30 days, based on a 5-point scale: Never (0), Once or twice (1), A few times (2), Fairly often (3), and Very often (4). All other items were rated on how much a specific symptom bothered participants in the past 7 days, based on a different 5-point scale: Not at all (0), A little bit (1), Moderately (2), Quite a bit (3), and Extremely (4). The exact interview questions on post-concussion symptoms are provided as an Appendix in prior published research on this cohort.¹⁸

Sexual and physical IPV history

Participants responded to questions focused on physical and sexual IPV that were adapted from the Conflict Tactics Scales^49,50 per pilot research with the target population.⁵¹ Participants were asked whether they had specific experiences of physical or sexual violence, with each specific experience listed in Table 2. IPV items were assigned weights based on tactic severity, aligning with prior research.^{18,48,52–56} For physical violence, the items were as follows, with weights listed in parentheses: twist arm/hair (1), push/shove (1), grab (1), slap (1), kick (2), bite (2), punch/hit with something that could hurt (2), slam against the wall (2), beat up (5), burn/scald (5), choke/strangle (5), threaten with knife/gun (6), try to run down with car (6), and use knife/gun (8). For sexual violence, the items were as follows, with weights listed in parentheses: sexual insistence without force (1); used threats to do sexual things, including intercourse (2); used physical force to do sexual things, including intercourse (3).

Table 2.

Intimate Partner Violence History by Head Injury Group

Variable (weight in parentheses)	No HIs (n = 268)		1-3 HIs (n = 110)		4+ HIs (n = 140)		F/χ2	p	η2/ϕ
Physical Violence, MW (SD)	17.1a,b	9.8	20.6a,c	10.2	26.1b,c	10.9	35.59	<.001	.12
Twist arm or hair (1), n (%)	138	51.7%	72	65.5%	96	68.6%	13.11	<.001	.16
Push or shove (1), n (%)	215	80.5%	94	85.5%	123	87.9%	3.96	.137	.09
Grab (1), n (%)	214	80.1%	95	86.4%	127	90.7%	8.19	.017	.13
Slap (1), n (%)	121	45.3%	58	53.2%	98	70.0%	22.51	<.001	.21
Kick (2), n (%)	78	29.2%	50	45.9%	83	59.3%	35.78	<.001	.26
Bit (2), n (%)	33	12.4%	10	9.2%	33	23.6%	12.59	.002	.16
Punch or hit with something (2), n (%)	131	49.1%	74	67.3%	106	75.7%	30.17	<.001	.24
Slam against the wall (2), n (%)	135	50.6%	66	60.0%	94	67.1%	10.80	.004	.15
Beat up (5), n (%)	153	57.1%	81	73.6%	119	85.0%	34.94	<.001	.26
Burn or scald on purpose (5), n (%)	10	3.7%	9	8.3%	20	14.3%	14.70	<.001	.17
Choke/strangle (5), n (%)	152	56.9%	70	63.6%	107	76.4%	15.09	<.001	.17
Threaten with a knife or gun (6), n (%)	132	49.3%	60	55.6%	96	68.6%	13.92	<.001	.16
Try to run down with a car (6), n (%)	34	12.7%	23	21.1%	47	33.6%	25.04	<.001	.22
Used a knife or fired a gun on (8), n (%)	95	35.4%	43	40.2%	82	58.6%	20.45	<.001	.20
Sexual Violence, MW (SD)	1.2b	1.8	1.6c	2.1	2.5b,c	2.5	14.76	<.001	.07
Sexual insistence (1), n (%)	127	47.4%	58	53.2%	87	62.1%	8.05	.018	.13
Threats to do sexual things including intercourse (2), n (%)	36	13.4%	21	19.4%	46	32.9%	21.74	<.001	.21
Physical force to do sexual things including intercourse (3), n (%)	45	16.8%	23	21.3%	59	42.1%	32.66	<.001	.25

For the Sexual Violence severity comparison, Levene's test for equality of variance was significant (p < .05), with equal variance not assumed and the F-value for this test being asymptotically distributed and drawn from Welch robust test of equality of means. ^aSignificant difference between No HIs and 1-3 HIs, ^bSignificant difference between No HIs and 4+ HIs, ^cSignificant difference between 1-3 HIs and 4+ HIs.

HIs, head injuries; M_W, weighted mean; SD, standard deviation.

Procedure

Women survivors of IPV were recruited from one urban and three rural court jurisdictions where they received a civil protective order for partner abuse.⁵⁶ The research team did not have a partnership with any entities related to the court jurisdiction, but were guests within each court. The team met with the judges of each court and were granted permission to conduct the study. The interviewers attended domestic violence courts in four Kentucky counties. At the conclusion of each case that met eligibility criteria, a project interviewer approached the petitioner and told her about the study using a standardized protocol. All women were provided with study information that contained the study phone numbers and were asked to call if they wished to participate. They could also provide their contact information if preferred.

With each woman's consent, the interviewer contacted the woman within 4 weeks of the court date. When potential participants were contacted, the interviewer explained the study in detail, with particular emphasis on the risks and protections, particularly confidentiality, associated with participation in the study. If the participant wished to participate in the study, an interview time and place was scheduled. Of those that had valid contact information and that were actively pursued to participate, 70% completed the interview, 7.7% decided not to participate, and 22.3% were never successfully contacted and/or scheduled. Each participant completed a 2-h interview between October 2001 and December 2004. Participants chose public settings (e.g., libraries and hospitals) for the interviews and provided written informed consent prior to study participation. Although the interviews occurred in public settings, they were conducted in private spaces, such as a reserved private room. The interviews occurred away from others who could hear responses, ensuring the public setting did not influence responses to sensitive questions or compromise confidentiality. The study was approved by a local institutional review board.

Statistical analysis

Participants were categorized based on number of lifetime head injuries (i.e., 0, 1-3, and 4+) and number of lifetime head injuries involving LOC (i.e., 0, 1, and 2+) and were compared on sociodemographic variables using one-way analysis of variance (ANOVA) tests for age and income and χ² tests for race/ethnicity, education, employment status, and urban/rural area or residence. Lifetime history of physical and sexual IPV severity and each symptom domain were correlated with lifetime number of head injuries and lifetime number of LOC episodes due to head injury. Spearman rho correlations were used due to the non-normality of the head injury frequency (i.e., skewness = 11.1, kurtosis = 145.2) and LOC frequency variables (i.e., skewness = 4.8, kurtosis = 27.3). Participants with and without LOC were compared on cognitive, physical, and emotional symptoms using t-tests and analysis of covariance, controlling for age, race/ethnicity, education, employment status, urban/rural area, and physical and sexual IPV severity, which were all associated with symptom severity in prior research.¹⁸

A series of analyses of variance (ANOVAs) compared cognitive, physical, emotional, and total symptoms across participant groupings based on number of lifetime head injuries (i.e., 0, 1-3, and 4+) and number of lifetime head injuries involving LOC (i.e., 0, 1, and 2+). Endorsement frequencies of individual symptoms were also compared across groups using χ² tests. An individual symptom was considered endorsed if it was rated as 1 or greater on a 5-point scale (range: 0 to 4). Hierarchical regression analyses examined whether number of lifetime head injuries or number of lifetime LOC episodes were associated with each symptom domain and total symptoms after controlling for age, race/ethnicity, education, employment status, urban/rural residence, and physical and sexual IPV severity. Model 1 included all covariates and Model 2 introduced either the number of lifetime head injuries or the number of lifetime LOC episodes, with change in R² evaluated for significance. For all analyses, p < 0.05 indicated significance, with a Bonferroni adjustment for post hoc analyses following the one-way ANOVA analyses. Per sensitivity analysis, the sample had sufficient power (1-β = 0.80) to detect a small-to-medium effect sizes for t-tests, ANOVAs, and change in R².⁵⁷

Results

Sociodemographic characteristics

Demographic comparisons across head injury and LOC groups are provided in Table 1. There were significant differences across head injury groups in age, race/ethnicity, education, employment status, and urban/rural residence, but no differences in income. A higher number of head injuries was associated with greater age, White race, less than high school education, unemployment, and rural residence. None of these sociodemographic variables differed based on number of LOC episodes. Lifetime number of head injuries and LOC episodes were moderately correlated (rho = 0.43 [95% CI: 0.32, 0.53], p < 0.001).

Physical and sexual IPV severity

As shown in Tables 2 and 3, lifetime physical and sexual IPV severity differed by number of head injuries and number of LOC episodes, in that greater lifetime head injuries and LOC were associated with greater IPV severity. Participants with no prior head injuries had the lowest severity of physical IPV. The 4+ head injury group had significantly greater physical IPV severity than the 1-3 head injury group. Regarding sexual IPV severity, the 4+ head injury group had significantly greater severity than both the no head injury and 1-3 head injury group, but these two latter groups did not differ significantly from each other. Participants with no LOC and 1 LOC did not differ from each other in terms of physical IPV history, but participants with 2 + LOC had significantly greater physical IPV severity than both other groups. For sexual IPV severity, the 2 + LOC group had significantly greater severity than the no LOC group, but the 1 LOC group did not significantly differ from either group. The comparison of 2 + LOC to 1 LOC groups neared significance but was above the threshold for significance after Bonferroni adjustment controlling for multiple comparisons (p = 0.052).

Table 3.

Intimate Partner Violence History by LOC Group

Variable (weight in parentheses)	No LOC (n = 109)		1 LOC (n = 72)		2 + LOC (n = 67)		F/χ2	p	η2/ϕ
Physical Violence, MW (SD)	21.0b	10.1	23.2c	10.6	28.9b,c	10.9	12.19	<.001	.09
Twist arm or hair (1), n (%)	67	60.9%	50	69.4%	50	74.6%	3.81	.152	.12
Push or shove (1), n (%)	92	83.6%	63	87.5%	61	91.0%	2.04	.377	.09
Grab (1), n (%)	98	89.1%	59	81.9%	64	95.5%	6.43	.040	.16
Slap (1), n (%)	62	56.4%	41	57.7%	51	76.1%	7.70	.021	.18
Kick (2), n (%)	52	47.3%	36	50.7%	44	65.7%	5.92	.056	.15
Bit (2), n (%)	15	13.6%	8	11.3%	21	31.3%	11.80	.003	.22
Punch or hit with something (2), n (%)	75	68.2%	46	63.9%	57	85.1%	8.70	.013	.19
Slam against the wall (2), n (%)	67	60.9%	45	62.5%	49	73.1%	2.93	.223	.11
Beat up (5), n (%)	84	76.4%	53	73.6%	61	91.0%	7.68	.020	.18
Burn or scald on purpose (5), n (%)	11	10.0%	6	8.5%	13	19.4%	4.71	.099	.14
Choke (5), n (%)	72	65.5%	49	68.1%	55	82.1%	5.90	.053	.15
Threaten with a knife or gun (6), n (%)	61	55.5%	49	70.0%	46	68.7%	5.08	.085	.14
Try to run down with a car (6), n (%)	24	21.8%	17	23.9%	29	43.3%	10.37	.005	.20
Used a knife or fired a gun on (8), n (%)	39	35.8%	40	57.1%	46	68.7%	19.51	<.001	.28
Sexual Violence, MW (SD)	1.6b	2.1	2.0	2.3	3.0b	2.6	6.33	<.001	.06
Sexual insistence (1), n (%)	62	56.4%	38	53.5%	45	67.2%	3.00	.225	.11
Threats to do sexual things including intercourse (2), n (%)	23	20.9%	18	25.7%	26	38.8%	6.85	.032	.17
Physical force to do sexual things including intercourse (3), n (%)	24	21.8%	23	32.9%	34	50.7%	15.81	<.001	.25

For the Sexual Violence severity comparison, Levene's test for equality of variance was significant (p < 0.05), with equal variance not assumed and the F-value for this test being asymptotically distributed and drawn from Welch robust test of equality of means. ^aSignificant difference between No LOC and 1 LOC, ^bSignificant difference between No LOC and 2+ LOC, ^cSignificant difference between 1 LOC and 2+ LOC.

HIs, head injuries; LOC, loss of consciousness; M_W, weighted mean; SD, standard deviation.

As a continuous variable, lifetime physical IPV severity had significant moderate correlations with number of head injuries (rho = 0.35 [0.27, 0.43], p < 0.001) and LOC episodes (rho = 0.33 [0.21, 0.44], p < 0.001) and sexual IPV severity had significant small correlations with number of head injuries (rho = 0.22 [0.13, 0.30], p < 0.001) and LOC episodes (rho = 0.19 [0.06, 0.31], p = 0.003). As shown in Table 2, the frequencies of all individual physical and sexual IPV experiences showed significant differences across head injury groups, as all increased with greater number of head injuries. As shown in Table 3, for LOC groupings, some individual IPV experiences differed significantly across groups, including being grabbed, slapped, bit, beat up, run down with a car, having a knife or gun used on them, and threats or force for sexual acts. In nearly all instances, women with 2 + LOC episodes reported the highest frequency of each individual experience, but for some experiences, women with no LOC history reported higher frequencies than women with 1 LOC episode.

Physical, cognitive, and emotional symptom severity

All symptom domains correlated significantly with lifetime number of head injuries and lifetime number of LOCs. The correlation between head injury count and symptom severities were moderate in magnitude: cognitive (rho = 0.33 [0.25, 0.41], p < 0.001), physical (rho = 0.36 [0.28, 0.43], p < 0.001), emotional (rho = 0.36 [0.29, 0.44], p < 0.001), and total (rho = 0.39 [0.32, 0.47], p < 0.001). The correlations between LOC count and symptom severities were small in magnitude, apart from physical symptoms: cognitive (rho = 0.23 [0.10, 0.35], p < 0.001), physical (rho = 0.31 [0.19, 0.42], p < 0.001), emotional (rho = 0.18 [0.06, 0.30], p = 0.004), and total (rho = 0.26 [0.14, 0.38], p < 0.001). Participants with head injuries involving LOC had significantly higher symptoms across domains than participants with head injuries not involving LOC: cognitive (No LOC: mean [M](SD) = 5.7(5.2) vs. LOC M(SD) = 8.2(5.7), t = 3.53, p < 0.001, d = 0.45 [0.20, 0.70]), physical (No LOC: M(SD) = 6.5(3.9) vs. LOC M(SD) = 9.1(4.7), t = 4.78, p < 0.001, d = 0.60 [0.34, 0.85]), emotional (No LOC: M(SD) = 7.3(4.9) vs. LOC M(SD) = 9.3(5.6), t = 2.92, p = 0.004, d = 0.37 [0.12, 0.62]), and total (No LOC: M(SD) = 19.4(11.8) vs. LOC M(SD) = 26.5(14.2), t = 4.27, p < 0.001, d = 0.53 [0.28, 0.79]). These group differences were still present in similar magnitude after adjusting for sociodemographic variables and physical and sexual IPV severity: cognitive (F = 9.72, p = 0.002, η_p² = 0.04), physical (F = 17.91, p < 0.001, η_p² = 0.07), emotional (F = 7.63, p = 0.006, η_p² = 0.03), and total (F = 14.74, p < 0.001, η_p² = 0.06).

Participants were categorized based on number of head injuries and compared using ANOVA with results reported in Table 4. The ANOVAs indicated significant group differences associated with medium-to-large effect sizes (η² range: 0.11 to 0.15) for all symptom domains and total symptom severity. For cognitive, physical, and total symptoms, symptom severity increased with number of head injuries, with significant differences between all groups (i.e., 1-3 head injuries had greater symptom severity than no head injuries, and 4+ head injuries had greater symptom severity than 1-3 head injuries). For emotional symptom severity, the 1-3 and 4+ head injury groups did not significantly differ from one another. As shown in Table 4, all rates of individual symptom endorsement significantly differed across groups, apart from Trouble remembering. The rate of endorsement was lowest for the no head injury group for all individual symptoms. The 4+ head injury group had the highest rates of endorsement for all individual symptoms apart from two emotional symptoms associated with anger and irritability (i.e., Easily annoyed or irritated and Temper outbursts), for which women with 1-3 symptoms had slightly higher but comparable endorsement rates to the 4+ head injury group.

Table 4.

Comparison of Symptom Severity and Individual Symptom Endorsement by Head Injury History

	No HIs (n = 268)		1-3 HIs (n = 110)		4+ HIs (n = 140)		F/χ2	p	η2/ϕ
Cognitive symptoms, M (SD)	3.8a,b	4.0	6.2a,c	5.3	7.8b,c	5.7	30.63	<.001	.11
Trouble remembering things, n (%)	171	63.8%	77	70.0%	105	75.0%	5.53	.063	.10
Trouble concentrating, n (%)	139	51.9%	74	67.3%	108	77.1%	26.60	<.001	.23
Difficulty making decisions, n (%)	117	43.7%	63	57.3%	93	66.4%	20.30	<.001	.20
Mind going blank, n (%)	119	44.4%	67	60.9%	101	72.1%	30.35	<.001	.24
Something wrong with mind, n (%)	66	24.6%	47	42.7%	68	48.6%	26.92	<.001	.23
Physical symptoms, M (SD)	5.0a,b	3.9	6.9a,c	4.1	8.7b,c	4.7	32.73	<.001	.12
Headaches and head pain, n (%)	218	81.3%	96	87.3%	128	91.4%	7.89	.019	.12
Faintness or dizziness, n (%)	70	26.1%	45	40.9%	67	47.9%	21.11	<.001	.20
Nausea or upset stomach, n (%)	126	47.0%	70	63.6%	96	68.6%	20.37	<.001	.20
Numbness or tingling, n (%)	88	32.8%	49	44.5%	81	57.9%	23.97	<.001	.22
Trouble falling asleep, n (%)	153	57.1%	84	76.4%	116	82.9%	32.47	<.001	.25
Emotional symptoms, M (SD)	4.6a,b	4.6	7.7a	5.4	9.0b	5.4	38.34	<.001	.13
Nervousness or shakiness, n (%)	120	44.8%	73	66.4%	118	84.3%	62.16	<.001	.35
Easily annoyed or irritated, n (%)	203	75.7%	97	88.2%	120	85.7%	10.55	.005	.14
Temper outbursts, n (%)	68	25.4%	59	53.6%	72	51.4%	40.06	<.001	.28
Feeling blue, n (%)	147	54.9%	83	75.5%	110	78.6%	28.91	<.001	.24
Feeling no interest in things, n (%)	119	44.4%	66	60.0%	103	73.6%	32.67	<.001	.25
Total symptoms, M (SD)	13.5a,b	11.2	20.8a,c	12.8	25.4b,c	13.9	43.28	<.001	.15

Levene's test for equality of variance was significant (p < 0.05) for all comparisons, with equal variance not assumed for all subscale and total symptom comparisons. F-values were asymptotically distributed and drawn from Welch robust test of equality of means. Post hoc comparisons after Bonferroni adjustment: ^aSignificant difference between No HIs and 1-3 HIs, ^bSignificant difference between No HIs and 4+ HIs, ^cSignificant difference between 1-3 HIs and 4+ HIs. An individual symptom was considered endorsed if it was rated as 1 or greater on a 5-point scale (range: 0 to 4).

HIs, head injuries; M, mean; SD, standard deviation.

Participants reporting one or more head injuries were also categorized based on lifetime LOC due to head injury. As presented in Table 5, women with no LOC, 1 LOC, and 2 + LOC differed significantly in cognitive, physical, and total symptom severity, with small-to-medium effect sizes (η² range: 0.04 to 0.10). For cognitive, physical, and total symptom severity, the 1 LOC and 2 + LOC groups had significantly greater symptom severity than the no LOC group, but did not differ from each other. For emotional symptom severity, only the 2 + LOC group had significantly greater severity than the no LOC group. With regard to individual symptom endorsement, six symptoms differed significantly in endorsement rates, including three cognitive symptoms (i.e., Trouble concentrating, Mind going blank, Something wrong with mind), two physical symptoms (i.e., Numbness or tingling, Trouble falling asleep), and one emotional symptom (i.e., Nervousness or shakiness). For these individual symptoms, the no LOC group had the lowest endorsement rate. The 2 + LOC group had the highest endorsement rate for all symptoms apart from Trouble falling asleep, for which the 1 LOC (i.e., 86.1%) and 2 + LOC (i.e., 85.1%) had similar endorsement rates.

Table 5.

Comparison of Symptom Severity and Individual Symptom Endorsement by LOC History

	No LOC (n = 109)		1 LOC (n = 72)		2 + LOC (n = 67)		F/χ2	p	η2/ϕ
Cognitive symptoms, M (SD)	5.7a,b	5.2	7.8a	5.6	8.5b	5.7	6.34	.002	.05
Trouble remembering things, n (%)	79	71.8%	52	72.2%	51	76.1%	0.43	.806	.04
Trouble concentrating, n (%)	68	61.8%	55	76.4%	57	85.1%	12.09	.002	.22
Difficulty making decisions, n (%)	62	56.4%	50	69.4%	44	65.7%	3.54	.170	.12
Mind going blank, n (%)	64	58.2%	49	68.1%	54	80.6%	9.52	.009	.20
Something wrong with mind, n (%)	41	37.3%	34	47.2%	39	58.2%	7.44	.024	.17
Physical symptoms, M (SD)	6.5a,b	3.9	8.3a	4.4	9.8b	4.8	12.92	<.001	.10
Headaches and head pain, n (%)	96	87.3%	65	90.3%	63	94.0%	2.12	.347	.09
Faintness or dizziness, n (%)	41	37.3%	35	48.6%	36	53.7%	5.10	.078	.14
Nausea or upset stomach, n (%)	67	60.9%	50	69.4%	48	71.6%	2.60	.272	.10
Numbness or tingling, n (%)	46	41.8%	43	59.7%	41	61.2%	8.56	.014	.19
Trouble falling asleep, n (%)	80	72.7%	62	86.1%	57	85.1%	6.38	.041	.16
Emotional symptoms, M (SD)	7.3b	4.9	8.8	5.3	9.7b	5.9	4.83	.009	.04
Nervousness or shakiness, n (%)	75	68.2%	58	80.6%	58	86.6%	8.72	.013	.19
Easily annoyed or irritated, n (%)	94	85.5%	63	87.5%	59	88.1%	0.30	.863	.03
Temper outbursts, n (%)	51	46.4%	40	55.6%	38	56.7%	2.36	.308	.10
Feeling blue, n (%)	82	74.5%	59	81.9%	51	76.1%	1.40	.496	.08
Feeling no interest in things, n (%)	70	63.6%	49	68.1%	49	73.1%	1.73	.422	.08
Total symptoms, M (SD)	19.4a,b	11.8	24.9a	14.0	28.1b	14.4	9.61	<.001	.07

Levene's test for equality of variance was non-significant (p > .05) for all comparisons, with equal variance assumed for all subscale and total symptom comparisons. Post hoc comparisons after Bonferroni adjustment: ^aSignificant difference between No LOC and 1 LOC, ^bSignificant difference between No LOC and 2 + LOC. There were no significant differences between the 1 LOC and 2+ LOC groups. An individual symptom was considered endorsed if it was rated as 1 or greater on a 5-point scale (range: 0 to 4).

LOC, loss of consciousness; M, mean; SD, standard deviation.

The results of the hierarchical regression analyses are presented in Table 6 for the models including lifetime number of head injuries and lifetime number of LOC episodes. The results indicated that lifetime number of head injuries was independently predictive of physical symptoms (β = 0.111, p = 0.008), accounting for significant additional variance (ΔR² = 0.011, p = 0.008), but was predictive of no other symptom domain or total symptom severity. Similarly, lifetime number of head injuries involving LOC was only independently predictive of physical symptoms (β = 0.154, p = 0.014) and accounted for significant additional variance in this symptom domain (ΔR² = 0.021, p = 0.014).

Table 6.

Hierarchical Regression Analyses: Model Set 1 – Lifetime Number of Head Injuries and Model Set 2 – Lifetime Number of Head Injuries Involving Loss of Consciousness

	Cognitive symptoms		Physical symptoms		Emotional symptoms		Total symptoms
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Model Set 1: Head Injury Count
Age	.113**	.110**	.103*	.089*	.077	.074	.108**	.101*
Race/ethnicity	.015	.015	.021	.020	-.006	-.006	.010	.010
Education	-.012	-.010	-.042	-.034	-.067	-.065	-.045	-.041
Unemployed	.210***	.210***	.181***	.183***	.160***	.160***	.205***	.205***
Rural/urban	.005	.003	.160***	.149**	.072	.069	.084	.078
Physical IPV severity	.151**	.148**	.191***	.177***	.150**	.147**	.181***	.175***
Sexual IPV severity	.134**	.131**	.107*	.094*	.092	.089	.123**	.117*
Number of Head Injuries	–	.026	–	.111**	–	.025	–	.057
F	10.27***	9.02***	16.37***	15.38***	8.67***	7.62***	13.94***	12.44***
R2	.124	.125	.185	.196	.107	.108	.162	.165
ΔR2	–	.001	–	.011**	–	.001	–	.003
Model Set 2: LOC Count
Age	.135*	.128*	.174**	.162**	.143*	.138*	.170**	.161**
Race/ethnicity	.037	.030	.017	.006	-.072	-.077	-.008	-.016
Education	.035	.026	-.078	-.090	-.042	-.047	-.028	-.038
Unemployed	.282***	.272***	.235***	.220***	.211**	.205**	.278***	.266***
Rural/urban	-.005	-.014	.139*	.127	.029	.024	.056	.046
Physical IPV severity	.009	-.017	.107	.068	-.014	-.032	.034	.003
Sexual IPV severity	.168*	.158*	.090	.075	.116	.110	.145*	.133*
Number of LOC	–	.104	–	.154*	–	.069	–	.121
F	4.99***	4.71***	6.99***	6.09*	3.27**	3.00**	5.87***	5.65***
R2	.128	.137	.171	.191	.088	.092	.147	.160
ΔR2	–	.009	–	.021*	–	.004	–	.013

N = 514 for Model Set 1: Head Injury Count; N = 246 for Model Set 2: LOC Count; ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. Race/ethnicity was coded White = 1 and Other racial/ethnic identities = 0. Education was coded High school diploma/GED or greater = 1 and Less than high school = 0. Unemployed was coded 1 = Unemployed and 0 = Employed either part-time or full-time. Rural/urban was coded 1 = Rural residence and 0 = Urban residence. Physical and Sexual IPV severity were calculated as weighted sums based on prior experiences. Dependent variables were symptom severity scores, based on summed severity ratings for items aligning with each symptom domain.

IPV, intimate partner violence; LOC, loss of consciousness.

Discussion

This study involved the examination of repetitive head injuries and LOC in relation to sociodemographic characteristics, lifetime IPV severity, and self-reported cognitive, physical, and emotional symptom severity among women who recently received a protective order against a domestic partner. A higher number of head injuries was associated with several sociodemographic characteristics, including greater age, White race, less educational attainment, unemployment, and rural residence. These results likely indicate that women who were older had more years of exposure to head injuries, and greater head injury exposure was associated with socioeconomic disadvantage, including residing in a rural setting with less educational and employment opportunity. White race substantially overlapped with rural residence (i.e., 98.2% of participants with rural residences identified as White). Women without employment and residing in rural settings likely experience greater degrees of isolation, have less access to services for IPV survivors, and may experience more chronic and severe IPV,⁵⁸ resulting in repeated victimization involving injuries to the head. Group comparisons based on LOC exposure mostly followed these same patterns, but these group differences were associated with minimal-to-small effect sizes and did not reach statistical significance.

Women recalling a higher count of lifetime head injuries and episodes of LOC tended to report greater lifetime physical and sexual IPV severity as well, indicating that greater lifetime violence exposure corresponds with increased risk for a higher number of head and brain injuries. Higher lifetime counts of head injuries were associated with greater symptom severity across domains, but the relationship between LOC and symptom severity followed a slightly different pattern. Although having experienced LOC was associated with greater symptom severity, there was not an observed cumulative effect of LOC on symptom severity, in that the 1 and 2 + LOC groups did not differ in symptom severity for any domain per pairwise group comparisons, and rates of individual symptom endorsement were very similar between these two groups.

A series of regression analyses examined whether lifetime head injuries and LOC episodes were related to symptom severity after controlling for sociodemographic characteristics and physical and sexual IPV severity. These analyses indicated that number of head injuries and LOC episodes were independently associated with only physical symptom severity, whereas individual differences in cognitive and emotional symptoms were explained by the other covariates in the model. This finding indicates that stressors associated with socioeconomic disadvantage and cumulative IPV-related trauma exposure may explain ongoing problems with thinking (e.g., inattention, forgetfulness) and affect (e.g., anxiety, depression, anger), but do not fully account for physical symptoms that commonly occur following head injuries (e.g., headaches, dizziness, nausea, numbness, tingling, sleep problems).

There are meaningful differences between the findings of head injury versus LOC counts in relation to symptom severity: 1) the correlations between LOC count and symptom severity were smaller in magnitude than the correlations between head injury count and symptom severity; and 2) in group comparisons, there was not an apparent cumulative effect of LOC, which was observed for head injury. Although LOC as a binary variable had a medium association with total symptom severity (d = 0.53), there were no differences between 2 + LOC and single LOC groups for any symptom domain. These findings indicate that lifetime count of head injuries had a stronger association with symptom severity than LOC, but the relationship between LOC and symptoms was likely attenuated by the fact that all participants with no LOC still reported a head injury, meaning the no LOC group still had a possible mild TBI that potentially contributed to current symptoms.

This study involved limitations. This sample involved predominantly White women from a single geographical region in Kentucky, although they spanned multiple urban and rural jurisdictions. These women received protective orders, which many women survivors of IPV do not obtain, indicating that the findings may not fully generalize to the population of women with IPV-related head injuries. Participants with head injuries unrelated to IPV were excluded because these participants were younger at their time of injury and experienced fewer injuries with lower degrees of lifetime violence,⁴⁸ but, as a consequence, the findings may not apply to survivors with prior head injuries unrelated to IPV. The analyses accounted for sociodemographic variables that may explain some individual differences in symptom severity, but did not account for psychological (e.g., depression) or physical conditions (e.g., chronic pain) that have been related to neurobehavioral symptoms among participants without head injuries or TBIs in prior research.^33-35

The cognitive, physical, and emotional symptoms measured herein were non-specific, in that many participants without head injuries endorsed these symptoms. These symptoms may also be related to nonfatal strangulation, which was experienced by 56.9% of women with no prior head injuries, 63.6% of women with 1-3 prior head injuries, and 76.4% of women with 4 or more prior head injuries. Nonfatal strangulation may result in anoxic-hypoxic brain injuries and possible cervical injuries, with a growing body of research examining injury sequelae.⁵⁹ Prior research with this sample indicated a small but significant effect of nonfatal strangulation on emotional symptom severity among women without head injuries (d = 0.26), but no group differences in symptom severity attributable to nonfatal strangulation among women with IPV-related head injuries.¹⁸

This study relied solely on self-report data collected during an in-person interview. There was no standardized method for determining symptom validity, either under or overreporting, and there was no collateral information to verify the presence of a head or brain injury (e.g., informant-report, hospital records). This study did not apply a standard definition for head injury or TBI,^60-62 although LOC would indicate a brain injury occurred. The operational definition of a head injury provided to participants reflected a broad definition that included unconventional criteria (e.g., severe pain) and excluded more minor injuries (e.g., minor bumps). Because these analyses were based on archival data, a more recently published or established definition^63,64 could not be applied to the current study. Severity could not be determined because information such as Glasgow Coma Scale (GCS), duration of LOC, and duration of post-traumatic amnesia were not available. Future research would benefit from applying standardized operational criteria for TBI diagnosis and classification of severity among women survivors of IPV.

Women survivors of IPV are a unique population in terms of examining head injury and TBI history, and only recently have tailored instruments been designed to capture unique mechanisms of injury in this population (e.g., being violently shaken, strangulation).^5,65 There are several reasons that recollection of head injury information may be difficult for women survivors of IPV, in comparison to different populations. In sport or military settings, the injury may be observed by teammates or fellow servicemembers, but in the context of IPV, the perpetrator of the injury may be the only person observing the woman after the assault. She may have difficulty differentiating LOC from post-traumatic amnesia and may have no reference for how long either clinical sign of injury may have lasted. She also may have limited access to medical care, meaning information that would conventionally be gathered in an emergency department, such as GCS, may not be available. Further, considering the mere number of head injuries reported, an exact count over the course of a lifespan may be uniquely difficulty to recall.

In military contexts, neurotrauma may be restricted to unique events, such as blast exposure, and may primarily occur during deployment; and in sports, concussions may be restricted to athletic seasons and occur during practices or games. In contrast, women survivors of IPV may experience head injuries in the context of continuous abuse that may last years, making a specific number of injuries more difficult to recall. Although a limitation, the reliance on self-report data of an estimated number of head injuries with limited information pertaining to injury characteristics and severity may be a common issue when conducting neurotrauma research in this population.

As with research on other populations,⁴² the effects of multiple head injury exposure on post-concussion symptoms remains equivocal based on the current findings, in that cumulative head injury appears related to worse symptom severity, but cumulative LOC, which more definitively indicates brain injury, does not necessarily relate to greater symptom severity. Further, sociodemographic characteristics and IPV severity accounted for the relationship between number of head injuries and LOC episodes and cognitive and emotional symptoms. Among women survivors of IPV, physical symptoms may be uniquely associated with head injury or TBI history, whereas cognitive and emotional symptoms may be more closely related to cumulative emotional trauma. Interestingly, headache was fairly common in women without head injuries, despite post-traumatic headache being an area of consistent investigation in TBI literature.⁶⁶ The symptoms that appeared to increase the most in endorsement frequency with more head injuries were faintness and dizziness, numbness or tingling, nausea, and trouble falling asleep, which may benefit from clinical evaluation and potential interventions (e.g., sleep medication or psychotherapy, vestibular-ocular rehabilitation for dizziness).^67–70 Women survivors of IPV with repetitive head injuries have many current symptoms that reflect unmet healthcare needs. These symptoms deserve attention in future research investigations on targeted interventions addressing neurobehavioral symptoms in this population.

Transparency, Rigor, and Reproducibility Summary

This study, based on archival data, and the analysis plan were not pre-registered. The sample included 518 adult cisgender women from Kentucky, USA, who were asked about head injury history during an in-person interview. Participants were excluded if they reported prior head injuries, but none of these injuries were due to IPV (n = 118); reported having experienced a coma due to a head injury (n = 4); or were missing on number of estimated head injuries (n = 1). The final sample included 250 women with one or more IPV-related head injuries and 268 women with no prior head injuries. Before analyses, a sensitivity analysis was conducted to estimate power using the existing sample,⁵⁷ which indicated sufficient power (1-β = 0.80) to detect a small-to-medium effect size for t-tests, ANOVAs, and change in R². Participants completed in-person interviews from 2001 to 2004, with all head injury history, IPV history, and symptom reporting based on self-report. The statistical analyses were conducted in SPSS and the investigator conducting the statistical analyses was aware of all characteristics of the participants. The reliability and validity of symptom reporting following head injury has been well-established using multiple similar instruments in prior research.⁷¹ Reliability estimates for symptom severity scores were reported based on prior analyses with the current sample, finding acceptable reliability for the scores. Missing data was minimal, with listwise deletion used in some analyses. Effect sizes and confidence intervals have been reported in the abstract and main text for all outcomes reported. Bonferroni corrections were made for post hoc multiple comparisons for ANOVAs. At this time, no replication or external validation studies have been performed or planned. Participants did not consent to data sharing and the data from this study will not be shared. The authors agree to provide the full content of the manuscript on request by contacting the corresponding author.

Authors' Contributions

Justin E. Karr conceptualized the study, conducted the statistical analyses, and wrote the manuscript. Sharon E. Leong assisted with statistical analyses, wrote portions of the manuscript, and reviewed and edited the manuscript. Eric O. Ingram wrote portions of the manuscript and reviewed and edited the manuscript. TK Logan conceptualized the study, led data collection, and reviewed and edited the manuscript.

Funding Information

This work was supported, in part, by a Building Interdisciplinary Research Careers in Women's Health (BIRCWH) grant (#K12-DA035150) from the National Institute on Drug Abuse (NIDA) of the National Institutes of Health (NIH).

The data collection was supported by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) grant (#AA12735-01) and the University of Kentucky General Clinical Research Organization funded by the NIH (#M01RR02602).

Author Disclosure Statement

No competing financial interests exist.