A Four-Country Study of Strangulation-Related Alterations in Consciousness in Women Who Have Experienced Intimate Partner Violence: Co-Occurrence with Traumatic Brain Injuries and Measures of Psychological Distress

Abstract

At least one in three women experience intimate partner violence (IPV) in their lifetime. The most commonly sustained IPV-related brain injuries include strangulation-related alterations in consciousness (S-AICs) and traumatic brain injuries (TBIs). Moreover, survivors of IPV-related S-AICs and/or TBIs often demonstrate psychological distress such as depression, anxiety, and post-traumatic stress. However, the co-occurrence of S-AICs and TBIs, and whether such TBIs may be moderate to severe, has not been systematically examined, and most data have been collected from women in North America. The purpose of this study was to examine the co-occurrence of IPV-related S-AICs and TBIs across a range of geographical locations and to determine the extent to which these S-AICs are related to psychological distress. Women who had experienced physical IPV (n = 213) were included in this secondary analysis of retrospectively collected data across four countries (Canada, the United States, Spain, and Colombia). The Brain Injury Severity Assessment (BISA) was used to assess IPV-related BI across all sites. Because various questionnaires were employed to assess levels of depression, anxiety, and post-traumatic stress disorder at each site, we created a standardized composite score by converting raw scores into Z-scores for analysis. Mann–Whitney U tests and chi-square tests were conducted to examine differences between women with and without experience of S-AICs and to discover if there was a relationship between the occurrence of S-AICs and TBIs. Analysis of variance and analysis of covariance (to control for the potential confounding effects of age, education, and non IPV-related TBI) were used to compare levels of psychological distress in women who had or had not experienced S-AICs. Approximately, 67% of women sustained at least one IPV-related BI (i.e., TBI and/or S-AIC). In a subsample of women who sustained at least one IPV-related BI, approximately 37% sustained both S-AICs and TBIs, 2% sustained only S-AICs (with no TBIs), and 61% sustained TBIs exclusively (with no S-AICs). Furthermore, women who had sustained S-AICs (with or without a TBI) were more likely to have experienced a moderate-to-severe BI than those who had not sustained an S-AIC (BISA severity subscale: U = 3939, p = 0.006). In addition, women who experienced S-AICs (with or without a TBI) reported higher levels of psychological distress compared with women who never experienced S-AICs, irrespective of whether they occurred once or multiple times. These data underscore the importance of assessing for S-AIC in women who have experienced IPV and when present, to also assess for TBIs and the presence of psychological distress. Unfortunately, there were methodological differences across sites precluding cross-site comparisons. Nonetheless, data were collected across four culturally and geographically diverse countries and, therefore, highlight IPV-related BIs as a global issue that needs to be aggressively studied with policies established and then implemented to address findings.

Introduction

It is estimated at least one in three women globally experience intimate partner violence (IPV) in their lifetime.¹ Widely recognized as a devastating public health problem, IPV occurs across a range of ages, ethnicities, socioeconomic statuses, and education levels,² yet its effects are disproportionately felt by women of color—particularly Black and Indigenous women^3,4—and women of lower socioeconomic status.⁵

An unfortunate consequence of physical IPV is brain injury (BI), hereafter termed as “IPV-related BI”, either via blunt force trauma, namely traumatic brain injuries (TBIs), or inferred from strangulation-related alterations in consciousness (S-AICs).^2–4,6–10 A scoping review¹¹ found that between 28%¹² and 100%^13,14 of women who had experienced IPV had also endured one or more BIs. The rates of strangulation (a distinct form of violence that is characterized by the external compression of the airway and blood vessels in the neck¹⁵), irrespective of AIC, have been reported to be as high as 68% in women who have experienced partner violence.⁸ Valera and Berenbaum⁶ reported that 27% of women sustained at least one S-AIC and 12% sustained repetitive S-AICs in a sample of 99 women who had experienced physical IPV. Furthermore, many women reported both S-AICs and TBIs. However, the co-occurrence of S-AICs and TBIs has not been systematically examined. It has also never been examined whether women who experience S-AICs also experience higher rates of moderate-to-severe BIs relative to women who do not experience S-AICs.

In addition, IPV and IPV-related BIs have been associated with a multitude of physical,¹⁴ psychological,¹⁶ neurological,^14,17 cognitive,^14,17 and emotional outcomes, most notably, anxiety, depression, and post-traumatic stress disorder (PTSD).^13,16 A recent systematic review of papers published between 1980 and 2020 across 10 countries¹⁵ synthesized 30 empirical and peer-reviewed studies on non-fatal strangulation in IPV and sexual assault. The authors found a link between strangulation and psychological distress including depression and PTSD. Two recent studies,^16,18 based on samples in the United States and Spain included in this report, found associations between S-AICs and psychopathology. In the data collected from the United States, Valera and colleagues¹⁶ demonstrated that women who experienced S-AICs had higher levels of anhedonic depression and PTSD symptomatology compared with those who had not experienced S-AICs. From Spain, Daugherty and colleagues¹⁸ demonstrated an association between strangulation and PTSD, depression, and anxiety. These findings are not surprising, as women not only report S-AICs but also report strangulation to be one of the most terrifying forms of abuse they experience.^19,20

In this report, our first objective is to examine the co-occurrence of S-AICs and TBIs across samples of women from four culturally and geographically diverse countries, namely Canada, the United States, Colombia, and Spain. Second, we aim to examine whether S-AICs are associated with other outcomes (e.g., mild BIs vs. moderate-to-severe BIs). Our final objective is to determine the degree to which S-AICs are associated with psychological distress (PTSD, depression, and anxiety) across cultures. Specifically, we predict a high co-occurrence of S-AICs with TBIs, as well as an increased rate for moderate-to-severe BIs for women who have experienced S-AICs relative to women who have not experienced S-AICs. Consistent with, and extending findings of the studies by Valera and colleagues¹⁶ and Daugherty and colleagues,¹⁸ we anticipate greater levels of psychological distress (i.e., depression, anxiety, and PTSD symptoms) in women who have experienced S-AICs.

Methods

Design, participants, and study settings

This is a secondary analysis of prospectively collected data across four countries—Canada, the United States, Spain, and Colombia. The data were originally collected as part of studies conducted by several of the co-authors (M.P.-G., N.Q.M., E.M.V., and P.V.D.), who were the lead investigators on the original studies. The common inclusion criterion across all studies is the enrollment of women who had experienced at least one instance of IPV in the past. Only women who experienced physical IPV (with or without sexual or psychological IPV) were included in the current analyses.

At the Canadian site,¹⁴ women between the ages 18 and 50 years were recruited from a local women’s shelter and other women-serving organizations. With an aim to enroll with a spectrum of previous TBIs, women were recruited irrespective of previously reported head trauma. However, they were excluded if they had a history of neurological disease/disorder (e.g., stroke and seizures). Demographic, psychopathological, and BI-related assessments (all in English) were conducted on two separate consecutive days by a trained research assistant.

At the site in the United States,⁶ women between the ages 18 and 55 years were recruited from two women’s shelters, as well as through different educational and women supporting programs. Five additional women were also recruited who heard about the study from a friend. Two sessions of assessment (all in English) were conducted in succession or on different days (within 1 week of each other) by the principal investigator of the study and two trained female research assistants.

At the Spanish site,¹⁸ women between the ages 18 and 67 years were recruited through various information centers for women (governmental or nongovernmental) and through flyers and word of mouth. The cross-culturally adapted Spanish version of the Composite Abuse Scale-Short Form (CAS) was used to measure physical IPV (including strangulation) throughout the entire lifespan as well as in the past 12 months. As assessments were conducted in Spanish, women were required to speak fluent and proficient Spanish. The mental health questionnaires, the BISA and CAS, were administered in one session by a trained doctorate student in psychology.

At the Colombian site,²¹ women between the ages 18 and 57 years were recruited from different non-governmental organizations (e.g., shelters) and through flyers. Women who reported neurological or psychiatric problems were excluded. All assessments were conducted in Spanish (the local language) by a trained assessor.

At all sites, women were informed about the study either directly from the organization, the PI of the study, or through flyers and word of mouth. Women were able to stop their participation at any moment. They were also informed about their anonymity and that responses would not be analyzed individually but rather at a group level. At the U.S. site, as the PI was a volunteer at the shelter, this helped with establishing rapport with the women participating in the study. In addition, at the Spanish site, the majority of women received information about the study from a support center where the women were receiving psychological treatment.

Demographic data, including age and education, were also collected. Approval to conduct each study was received from the local research ethics boards, and all participants provided written informed consent prior to participation in the study. Further details are available in the previously published articles.^6,14,21,22

Assessment of strangulation-related AICs and IPV-related TBIs

The Brain Injury Severity Assessment (BISA) tool⁶ was used (consistent with numerous previous studies)^{6,13,14,17,18,21,23,24} to assess IPV-related BI. The BISA tool was designed specifically to identify the types (i.e., TBIs and S-AICs) and severities of brain injuries sustained from IPV. It is a semi-structured interview containing a series of questions about AICs, including loss of consciousness, dizziness, memory loss, feeling stunned or disoriented, or seeing stars or spots following potential traumas to the brain. The first half of the BISA focuses on BIs sustained from a partner, and the second half focuses on all other types (e.g., accident related) of acquired BIs. The BISA tool does not include any questions to specifically distinguish TBI from strangulation-related AIC. Rather the questions about AICs (i.e., loss of consciousness, dizziness, memory loss, feeling stunned or disoriented, or seeing stars or spots) were asked, and once an AIC was endorsed, the woman was asked what event caused the AIC. A blunt force trauma would indicate a TBI, whereas a strangulation (“choking” or something pressing against the neck to block air or blood flow) was considered a strangulation-related AIC.

The BISA provides a summary score ranging from 0 to 8, as well as subscores related to the frequency (0–4 scale), recency (0–3 scale), and severity (0–1 scale) of the IPV-related BIs (see Table 1 for further details). The frequency score provides an estimate of the number of previous BIs, the recency score probes the time since the most recent event resulting in a BI, and the severity score indicates whether a moderate-to-severe BI was ever sustained.

Table 1.

Scoring Brain Injury Severity Assessment Tool

Subscales	Score	Description
Frequency	0	There have been no previously reported brain injuries
	1	There have been 1–5 previously reported brain injuries
	2	There have been 6–10 previously reported brain injuries
	3	There have been 11–15 previously reported brain injuries
	4	There have been 16 or more previously reported brain injuries
Recency	0	The event occurred more than 52 weeks ago
	1	The event occurred between 27 and 52 weeks ago
	2	The event occurred between 14 and 26 weeks ago
	3	The event occurred within the past 13 weeks
Severity	0	The presence of exclusively mild brain injuries
	1	At least one episode involving loss of consciousness for over 30 min or a period of post-traumatic amnesia for over 24 hours had occurred (indicating presence of moderate-to-severe injuries)

Based on the fact that there is typically very little further recovery a year or more after a BI,^25,26 participants for whom the most recent BI occurred a year or more ago were given a recency score of 0. Therefore, any incident occurring past 1 year was collapsed into a single score. Moreover, given that recall accuracy diminishes with the passage of time from the injury, this approach prevents making assumptions about precision. Finally, the prevailing method for identifying BI in research and clinical contexts involves relying on an individual’s recall of exposure to potentially brain damaging events, despite its imperfections.^27,28

The severity score is based on the definition for mild traumatic BI provided by the American Congress of Rehabilitation Medicine (1993)—in particular, injuries resulting in LOC less than or equal to 30 min or post-traumatic amnesia (PTA) less than or equal to 24 h were defined as mild BI (severity score: 0), whereas injuries resulting in Loss Of Consciousness (LOC) greater than 30 min and PTA greater than 24 h were defined as moderate-severe BI (severity score: 1). In the presence of other alterations in consciousness (e.g., confusion and dizziness), the injury was always considered to be mild. Mild BI was only determined if the aforementioned symptoms resulted from a plausible biomechanical force or act that could result in brain damage (e.g., hit to the head or strangulation). For further information on the BISA, please refer to Valera and Berenbaum.⁶

Although the BISA tool has not yet been formally tested for reliability and validity, it has been shown to capture the effects of IPV-related BI on resting state brain activation,¹³ white matter diffusions,²³ brain morphology,¹⁸ cognitive-motor function,¹⁷ and BI symptoms.¹⁴ Thus, the literature suggests that it is sensitive enough to previous BI to be a robust and reliable subjective measure and allow us to make strong inferences about the effects on brain function and structure. The BISA score has been shown to be predictive of a range of negative outcomes among women who have experienced IPV.^{6,13,14,17,18,21,24}

For use in local language (at Spanish and Colombian sites), the BISA was cross-culturally adapted by conducting a forward translation and backtranslation following International Test Commission guidelines. Translators and assessors were fluent and/or native speakers in both English and Spanish.

Assessment of depression, anxiety, and PTSD

Various questionnaires were used to assess depression, anxiety, and PTSD at each site. At the Canadian site, depression, anxiety, and PTSD were measured using Beck’s Depression Inventory (BDI),²⁹ Beck’s Anxiety Inventory (BAI),³⁰ and the Clinician-Administered PTSD scale (CAPS),³¹ respectively. At the site in the United States, depression and anxiety were measured using the Mood and Anxiety Symptom Questionnaire-Short Form (MASQ),³² and PTSD was measured using the CAPS.³¹ At the sites in Spain and Colombia, the GAD-7 questionnaire,³³ the Patient Health Questionnaire depression subscale (PHQ-9),³⁴ and the PTSD Checklist for DSM-5 (PCL-5)³⁵ were used to measure depression, anxiety, and PTSD, respectively.

Although different measures were used to examine symptom severity for different types of psychopathology, the scales used for each psychopathology are highly correlated, demonstrating convergent validity. The PCL and CAPS scales are considered gold-standard measures for the assessment of PTSD.³⁶ Both scales were based on the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders,^35,37 and they are both intended to measure the severity of PTSD symptoms. There is also evidence that PTSD symptoms measured using the PCL highly correlate with the CAPS scores, across all CAPS subscales.³⁸ Similarly, the BAI and GAD-7 both measure anxiety and show adequate convergent validity (r = 0.72).³³ The BAI is also significantly correlated with the MASQ with a medium effect size (η², Eta squared: 0.062).³⁹ The MASQ is also significantly correlated with the BDI, showing a large effect size (η² = 0.192).³⁹ Finally, the two measures for depression (BDI and PHQ-9) are also highly correlated (r = 0.77) and are essentially interchangeable.⁴⁰ Brief descriptions of each of these questionnaires are also available in the previously published articles.^6,14,21,22

Because these different questionnaires were used to assess depression, anxiety, and PTSD at each site, we created a standardized composite score for each site by converting raw scores into Z-scores. Converting raw to Z-scores is a valid method of normalization and is widely employed, including in the World Health Organization’s child growth standard index.^41–43 Z-scores were calculated by subtracting the site mean from the subject’s value and then dividing by the standard deviation (SD) of that variable. After normalizing the values at each site, all data shared a common scale with a 0 mean and SD of 1,⁴² allowing us to combine the Z-scores of the four different sites/datasets for subsequent analysis.^42,43

Datasets for analysis

Although we had initially hoped to compare data across sites, because our recruitment methods varied at each site, this was not possible (i.e., scientifically valid). That said, the fact that we had such different recruitment methods is a strength of the study in other ways as it increases the generalizability of these data to a wide range of women experiencing IPV. These data clearly indicate that these issues are important to address on a global level. We analyzed the combined data from the four sites for a total N of 213: Canada, n = 42; United States, n = 87; Spain, n = 42; Colombia, n = 42.

Owing to changes in the personnel collecting the data at the Canadian site, the strangulation versus TBI-specific information stopped being collected in detail after the first 18 participants. Therefore, for 18 of the Canadian women, we had detailed information on TBIs and S-AICs. However, for the other 22 women, we only acquired information regarding the experience of S-AIC (yes or no response) with or without TBIs (i.e., we did not know how many of them experienced only S-AIC, only TBIs, or both AIC and TBIs). This latter group of women, therefore, was excluded from analyses that required that information. As such, the co-occurrence of TBIs and S-AICs was determined from the subsample of 120 women for whom the BIs were broken down into TBIs and S-AICs (Canada, n = 18; United States, n = 64; Spain, n = 19; Colombia, n = 19). However, we had information regarding the experience of S-AIC (yes or no response) with or without TBIs on the full sample (213 women). Therefore, we considered the full sample to determine whether there is an effect of S-AIC (independent of the presence or absence of TBIs) on psychological distress. We did not compare psychological distress between TBI and S-AIC groups.

Statistical analysis

Descriptive statistics were conducted to analyze the demographic and clinical characteristics of the cohort. The Kolmogorov–Smirnov (K-S) test was used to examine the distribution of the data. For our first objective, descriptive statistics were used to assess the co-occurrence of sustaining S-AICs and TBIs. Chi-square and Mann–Whitney U tests were conducted for the second objective, which was to examine differences in TBI severities among women with versus without S-AICs. For our final objective, the outcome variables were normally distributed (K-S test, p > 0.05) and also met the assumptions of homogeneity (Levene’s test, p > 0.05). As such, an analysis of variance (ANOVA) was used to compare levels of psychological distress in women who had or had not experienced S-AICs.¹⁶ Analysis of covariance (ANCOVA) was used to control for the potential confounding effects of age, education, non IPV-related TBIs, and moderate-severe TBI. Effect size (η², Eta squared) was calculated to determine whether the effects were small (0.01), medium (0.06), or large (0.14).^44,45 Significance was set at p < 0.05. All statistical analyses were performed using Statistical Package for the Social Sciences Statistics version 29.0 (IBM Corp, Armonk, NY).

Results

A sample of 213 women ages 18–67 years was included in this study. Table 2 outlines the demographic and clinical characteristics of the cohort. There were no significant differences in age (p = 0.59) or education (p = 0.83) for the two subgroups (with or without S-AICs) or across the three groups (TBIs only or S-AICs only or TBIs + AICs). In the full sample (N = 213), 32.86% women (n = 70) experienced one to five previous BIs, 86.85% (n = 185) women had experienced a mild BI, and 66.67% (n = 142) women experienced the most recent IPV incident more than 52 weeks prior to the evaluation. Descriptive findings regarding the frequency, recency, and severity in various groups are shown in Table 2.

Table 2.

Summary of Demographic and Clinical Information of the Women by Full and Subsample

Characteristics	Experience of S-AICsa		Co-occurrence of TBIs and S-AICsb
	(Full sample, n = 213)		(Subsample, n = 120)
	Who experienced S-AICs (with or without TBIs; n = 60)	Who had not experienced S-AICs (with or without TBIs; n = 153)	Only TBIs (n = 74)	Only S-AICs (n = 2)	TBIs + S-AICs (n = 44)
Age (Y): Mean ± SD	35.10 ± 9.91	34.56 ± 11.06	32.61 ± 9.36	31.0 ± 2.83	34.48 ± 9.70
Education (# of Y): Mean ± SD	13.09 ± 3.21	13.10 ± 4.60	12.55 ± 3.12	17.0 ± 0.0	12.93 ± 3.03
Non IPV-related TBIs; n (%)
Yes	28 (46.7)	40 (26.1)	17 (22.97)	0 (0.0)	19 (43.18)
No	32 (53.3)	113 (73.9)	57 (77.03)	2 (100)	25 (56.82)
Number (percentage) of women according to the scores on the BISA subscalesc
Frequency score; n (%)
0	0	71 (46.41)	0	0	0
1	20 (33.33)	50 (32.67)	48 (64.86)	2 (100)	16 (36.36)
2	12 (20.0)	10 (6.54)	9 (12.16)	0	10 (22.73)
3	2 (3.33)	2 (1.31)	1 (1.35)	0	0
4	26 (43.34)	20 (13.07)	18 (24.33)	0	18 (40.91)
Recency score; n (%)
0	26 (43.33)	116 (75.81)	37 (50)	1 (50)	18 (40.91)
1	7 (11.67)	11 (7.19)	9 (12.16)	1 (50)	4 (9.09)
2	7 (11.67)	7 (4.58)	10 (13.51)	0	6 (13.64)
3	20 (33.33)	19 (12.42)	18 (24.33)	0	16 (36.36)
Severity score; n (%)
0	46 (76.67)	139 (90.85)	68 (91.89)	2 (100)	34 (77.27)
1	14 (23.33)	14 (9.15)	6 (8.11)	0	10 (22.73)

^a The Mann–Whitney U test was used to analyze differences between the women who had and had not experienced strangulation-related alterations in consciousness.

^b One-way analysis of variance was used to analyze differences across three groups; significance level was set at p < 0.05; n, number of women in a subgroup; N, total number of women in the full sample.

^c Scoring of the Brain Injury Severity Assessment tool in terms of frequency, recency, and severity is presented in Table 1.

IPV, intimate partner violence; S-AICs, strangulation-related alterations in consciousness; SD, standard deviation; TBIs, traumatic brain injuries; Y, year.

Out of 213 women, 66.67% (n = 142) sustained partner-related BIs (i.e., TBIs and/or S-AICs). Of the 142 women who experienced BIs (Canada, n = 40; United States, n = 64; Spain, n =19; Colombia, n = 19), the data of 120 women (Canada, n = 18; United States, n = 64; Spain, n = 19; Colombia, n = 19) were broken down into whether the BIs were S-AICs only or TBIs only or both. In this subsample, the results for our first objective demonstrate that approximately 37% (n = 44) sustained both S-AICs as well as TBIs, approximately 61% (n = 74) sustained TBIs only, and approximately 2% (n = 2) sustained only S-AICs.

With regard to our second objective, we found that women who had sustained S-AICs (n = 60) were more likely to have experienced TBI-related LOC exceeding 30 min and/or TBI-related PTA surpassing 24 h (i.e., a moderate-to-severe BI) compared with women who had never experienced S-AICs (n = 82; BISA severity subscale: n = 142, U = 2060, p = 0.02). Incidentally, approximately one-third of women (n = 68) also experienced non IPV-related TBIs in the past, with a larger percentage being in the group of women who experienced S-AICs (χ² = 8.35, df = 1, p = 0.004).

With respect to our third objective regarding psychological distress, ANOVAs run on the full sample (n = 213) revealed that women who had experienced S-AICs had higher scores on PTSD and depression symptomatology compared with women who had never experienced S-AICs (see Table 3), with a small to medium effect size (η² = 0.02 to 0.04). The ANCOVAs further demonstrated that the relationship between PTSD and experience of S-AIC remained significant even after controlling for the potential confounding effects of age, education, non IPV-related TBIs, and moderate-to-severe TBI (see Table 4). The association between depression and experience of S-AIC was not statistically significant after controlling for potential confounding factors; however, the effect size (η² = 0.02, small to medium) did not change. We did not find any significant relationships between anxiety and the experience of S-AIC (see Tables 3 and 4). The Z-scores for PTSD, depression, and anxiety are plotted as a function of group (i.e., women who had versus had not experienced S-AIC) and presented in Figure 1. Z-scores were significantly different between the groups for PTSD (p = 0.0069) and depression (p = 0.0485) but not for anxiety (p = 0.6855). PTSD, depression, and anxiety scores have also been plotted against the frequency and recency scores following one-way ANOVA and are presented in Figure 2. The Z-scores for PTSD, depression, and anxiety showed significant differences between the highest versus lowest scores in frequency as well as recency. In other words, PTSD, depression, and anxiety scores were higher when frequency for BI was higher or more recent.

Table 3.

Analysis of Variance to Compare Mental Health Comorbidities for Women Who Had Versus Had Not Experienced Strangulation-Related Alterations in Consciousness (N = 213)

Outcome variables	Mean (SD)		F	p	η2
	Who experienced S-AICs (with or without TBIs)	Who had not experienced S-AICs (with or without TBIs)
	(n = 60)	(n = 153)
PTSD	0.31 (0.86)	−0.10 (1.02)	7.66	0.006*	0.04
Depression	0.22 (1.02)	−0.09 (0.97)	4.01	0.04*	0.02
Anxiety	−0.03 (0.88)	0.01 (1.05)	0.06	0.81	0.001

^* Significant at p < 0.05, η²: Eta squared for the effect size.

AICs, alterations in consciousness; PTSD, post-traumatic stress disorder; SD, standard deviation; TBIs, traumatic brain injuries.

Table 4.

Analysis of Covariance to Compare Mental Health Comorbidities for Women Who Had (n = 60) Versus Had Not Experienced Strangulation-Related Alterations in Consciousness (n = 153) While Controlling for the Potential Covariates (Age, Education, Non IPV-Related TBI, and Moderate-to-Severe TBI), (n = 213)

Exposure variables	Outcome variables
	PTSD			Depression			Anxiety
	F	p	η2	F	p	η2	F	p	η2
History of strangulation	5.26	0.02*	0.03	2.58	0.11	0.02	0.23	0.63	0.001
Covariates
Age	1.09	0.30	0.005	0.002	0.97	0.001	4.56	0.03*	0.022
Education	3.68	0.06	0.017	4.98	0.03*	0.023	3.37	0.07	0.016
Non IPV-related TBI	0.41	0.52	0.002	0.81	0.37	0.004	0.65	0.42	0.003
Moderate-to-severe TBI	1.70	0.19	0.008	0.74	0.39	0.004	0.41	0.53	0.002

^* Significant at p < 0.05, η²: Eta squared for the effect size.

IPV, intimate partner violence; PTSD, post-traumatic stress disorder; TBI, traumatic brain injury.

FIG. 1.

Z-scores plotted as a function of group (had vs. had not experienced S-AIC). A: Z-scores for PTSD; B: Z-scores for depression; C: Z-scores for anxiety. PTSD, post-traumatic stress disorder; S-AIC, strangulation-related alteration in consciousness.

FIG. 2.

Z-scores plotted against frequency and recency scores of the BISA. A: Z-scores for PTSD against frequency score; B: Z-scores for depression against frequency score; C: Z-scores for anxiety against frequency score; D: Z-scores for PTSD against recency score; E: Z-scores for depression against recency score; F: Z-scores for anxiety against recency score. BISA, Brain Injury Severity Assessment; PTSD, post-traumatic stress disorder.

Discussion

This is the first study to systematically examine the co-occurrence of S-AICs and IPV-related TBI. Data were combined from four culturally and geographically diverse countries to increase generalizability and to expand our knowledge in this area beyond what is known almost exclusively in North America. Across our sample of 213 women, approximately 67% (142 women) sustained either a TBI and/or an S-AIC. In a subset of those women (120 women) for whom we had details regarding the type of BI (TBI or S-AIC), we found that approximately 37% of women experienced both S-AICs and IPV-related TBIs. When distinguishing between the different types of potential BI, results indicate that 61% sustained a partner-related TBI while only two women sustained an S-AIC without TBI. This last finding is critical, as it indicates that if a woman has been strangled by her partner, there is a reasonable likelihood that she has also acquired a partner-related TBI. Notably, as S-AIC and TBI frequently co-occur, it is difficult to disentangle their relative contributions to symptom presentation. Results from our second objective demonstrated that women with S-AICs were more likely to sustain a moderate-to-severe BI than those who had not experienced S-AIC. Thus, strangulation with AIC may serve as a red flag for co-occurring TBI, and practitioners should consider screening for BI caused by traumatic hits to the head if strangulation has been detected.

The results of our third objective demonstrated higher levels of psychological distress (PTSD and depression) for women sustaining S-AIC relative to women who did not sustain S-AIC. Specifically, women who sustained S-AICs reported higher levels of PTSD and depression compared with women who had never sustained an S-AIC. The association with PTSD could not be accounted for by age, education, non IPV-related TBIs, or moderate-to-severe TBI. However, the association with depression was strongly influenced by education to the degree that the effect was no longer statistically significant. Nonetheless, the effect size suggested that there could be a potentially meaningful association that we were underpowered to detect. Overall these are not only consistent with results from the individual studies using the United States¹⁶ and Spain¹⁸ cohorts but also further extend their findings by providing a larger sample and including data from two additional countries, namely Canada and Colombia.

In addition, the findings of our third objective fall in line with the results of a recent review¹⁵ and other existing literatures,^46–49 which found higher levels of depression and PTSD symptomology, as well as suicide ideation and nightmares, among strangulation survivors. Beyond the physical trauma, strangulation is often used as a coercive control tactic.^50–52 Indeed, a singular incident of strangulation can be enough to instill paralyzing fear in women, further strengthening the perpetrator’s power and reinforcing the victim’s compliance and isolation.⁵³ Thus, it is unsurprising that several studies, including this one, have shown women who experience strangulation also exhibit many symptoms consistent with depression and PTSD.^{8,46–49,51,52,54}

Using an international sample, the current study has demonstrated that IPV-related BIs and their negative associations are broad, across several culturally diverse locations. In particular, our data were heterogeneous in that it was collected (1) across four culturally diverse countries and (2) from women who had a wide range of previous BIs (0 to >16) and post-injury durations (incidents had occurred within the past 13 weeks to more than a year ago). In addition, our study demonstrated a relationship between the PTSD and S-AIC, even after controlling for the potential confounding effects of age, education, non IPV-related TBIs, and moderate-to-severe TBIs. However, the present study did not find any association between the experience of strangulation and anxiety, which is in contrast with the findings of previous work.^16,18 This variation may stem from methodological differences among studies (e.g., how BI is defined), variations between research sites (e.g., recruitment strategies), as well as the diverse cultural interpretations and definitions of anxiety and its dimensions.⁵⁵ This highlights the necessity for further research to better understand the complex relationship between psychopathology and nonfatal strangulation.

Limitations of the study

We did not consistently collect data on various dimensions of identity (e.g., ethnicity, gender, sexual orientation, socioeconomic status, and citizenship status⁵⁶). Canada, the United States, Spain, and Colombia are diverse, culturally heterogenous countries with populations that are made up of many ethnicities, languages, and religions. That we observed S-AICs and IPV-related TBIs across these complex sociodemographic and sociopolitical contexts underscore the need for intersectional approaches that consider the social location of an individual in the design and implementation of resources for women who have experienced IPV. An intersectional approach exploring the impact of various dimensions of identity (e.g., ethnicity, socioeconomic status, and citizenship status) could allow for a more nuanced analysis of S-AICs and IPV-related TBIs and inform the design of effective and culturally sensitive resources and supports. Unfortunately, because of methodological limitations (e.g., differences in how and where women were recruited across the sites), we were not able to compare across sites, but these data clearly indicate that these issues are important to address on a global level.

Methodological heterogeneity between sites very likely affected our results. For instance, each site used a different clinical questionnaire to assess anxiety and depression. In addition, the BISA tool is limited by the self-reported and retrospective nature of its assessment. The BISA tool has not yet been validated, and this may lead to variations in administration. Nonetheless, at the time these data were collected, there were no other measures developed for the assessment of BIs in samples of women across a range of countries. Furthermore, the BISA has been used successfully in a range of studies showing associations with cognitive and psychological functioning,⁶ white matter fractional anisotropy,²³ brain morphology,¹⁸ cognitive motor dysfunction,¹⁶ and TBI symptoms.¹⁴ Despite these potential limitations, the current study revealed that the BISA is able to elicit important information on both TBIs and S-AICs across cultures, which are consistently linked to psychological distress in survivors of IPV.

Furthermore, in theory, using medical records to validate responses regarding incidents of TBI or S-AIC could be useful. However, studies, for example, Valera and Berenbaum,⁶ have shown that women do not typically seek medical attention for these injuries (for a host of reasons including fear of or being prevented by the abuser). As such, there would be no such medical documentation for many or most women, resulting in high rates of false negatives.

Conclusion

Harnessing international data collected across four countries, this study provides important information regarding the co-occurrence of TBIs and S-AICs in women who have experienced physical IPV. More specifically, women experiencing S-AICs have a higher likelihood of sustaining an IPV-related TBI. Furthermore, the present study found that women who experienced S-AICs had higher likelihoods of sustaining moderate-to-severe IPV-BIs as well as higher levels of PTSD and depression compared with women who had never experienced strangulation-related AICs. These findings underscore the urgency of conceptualizing IPV-related BIs as a global issue and highlights the need to consider strangulation as an important risk factor for both TBIs and potentially moderate-to-severe TBIs. Additional studies are needed to better understand the interaction between IPV, BIs, and mental health as both IPV and BI contribute to mental health issues observed in women who experience IPV.

Transparency, Rigor, and Reproducibility Statement

This is a secondary analysis of prospectively collected data across four countries—Canada, the United States, Spain, and Colombia. Neither the study nor the analysis plan was preregistered. The materials and methods are described in detail in the articles that are already published using these four datasets (one each from four countries).^6,14,21,22 Participants’ selection, settings, measurements, and procedures in those articles are described in a simpler and clearer form to enhance the reproducibility of the studies. Trained assessors collected the data in the local language.

Those women survivors of IPV who met eligibility criteria were recruited at each site. Women between the ages of 18 and 67 years who had experienced at least one instance of IPV in the past were recruited from various women-serving organizations in the community. Only women who experienced physical IPV (with or without sexual/psychological IPV) and had complete data (BISA score, depression, anxiety, and PTSD scores) were included in the current analysis. Women were excluded if they had a history of neurological disease/disorder (e.g., stroke and seizures) or reported current illicit substance abuse. This resulted in a total sample size of 213 (Canada, n = 42; United States, n = 87; Spain, n = 42; Colombia, n = 42). The data of 120 women (Canada, n = 18; United States, n = 64; Spain, n = 19; Colombia, n = 19) for whom there were complete datasets were broken down into whether they experienced traumatic BI or strangulation-related alteration in consciousness. Therefore, this subsample of 120 women was used to calculate the co-occurrence of traumatic BI and strangulation-related alteration in consciousness.

The statistical analysis conducted in this secondary analysis is consistent with the statistical method applied in the existing literature.^15,18 Co-occurrence of traumatic BI and strangulation-related alteration in consciousness was calculated using descriptive statistics (number and percentage). Chi-square test and Mann–Whitney U tests were conducted to examine differences in TBI severities among women with versus without strangulation-related alteration in consciousness. ANOVA was used to compare levels of psychological distress in women who had or had not experienced strangulation-related alteration in consciousness. ANCOVA was used to control for the potential confounding effects of age, education, and non IPV-related TBIs. Effect size (r) was calculated to determine whether the effects were small (0.1 to <0.3), medium (0.3 to <0.5), or large (0.5–1.0).^44,45

Approval to conduct each study was received from the local research ethics boards, and all participants provided written informed consent prior to participation in the study. Funding sources and acknowledgments are included. Data from each site are securely stored by the principal investigator of each study.^6,14,21,22 Anonymized data are available upon request from the Canadian site. However, as such, sharing was not required by the funding agencies, and given the sensitivity of the data, this type of data sharing was not requested in the informed consent for the protocols used in the United States, Spain, and Colombia. As such, these data are not allowed to be shared freely.

Authors’ Contributions

S.P.A.: Conceptualization, analysis, methodology, writing (original draft), and writing (reviewing and editing); N.M.-R., C.W., and J.S.: Methodology, data collection, and writing (reviewing and editing); M.P.-G., C.J.D.L.R., and N.H.-R.: Conceptualization and writing (reviewing and editing); P.V.D., E.M.V., J.C.D., and N.Q.M.: Conceptualization, methodology, supervision, data collection, and writing (reviewing and editing).

Author Disclosure Statement

No competing financial interests exist.

Funding Information

EV.M. would like to acknowledge current funding (R01 NS112694-01). This research was funded by National Research Service Award 1 F31 MH11763-01A2, from the National Institutes of Health, and by three grants from the University of Illinois, the Graduate College On-Campus Dissertation Research Grant, Women’s Studies Funding for Feminist Scholarship, and the Graduate College Thesis Project Grant. This work was supported at the Canadian site by grants from the Department of Women and Gender Equality (GV18315-01), Canadian Institutes of Health Research (013902), and Canadian Foundation for Innovation, and an anonymous donor (012153). N.M.-R. was supported by a Canada Graduate Scholarship—Master’s award. The study at the Spanish site was funded by the I+D+I Operative Program FEDER project of the Andalusian Government in 2018 (B-CTS493-UGR18). J.C.D. was funded by a predoctoral fellowship of the Ministry of Education and Professional Training in Spain as well as the postdoctoral Puente contract at the University of Granada (Spain). The study at Colombian site was supported by the Minciencias (grant number 727 from 2015).