Abstract
Background:
Stalking is one of the most common forms of interpersonal violence, with nearly one in three women experiencing it in their lifetime. Restraining orders are a legal remedy aimed at protecting women who experience the most severe forms of stalking and interpersonal violence. Prior work has shown associations between experiences of violence and cardiovascular health among women, but little is known about the link between stalking or obtaining a restraining order history and risk of cardiovascular disease (CVD).
Methods:
We examined whether experiences of stalking or obtaining a restraining order (considered an indicator of severe violence) were each associated with a greater risk of incident CVD over 20 years in the Nurses’ Health Study II cohort. We conducted Cox proportional hazards regression models among women free of CVD in 2001 when their exposure status was ascertained (n=66,270). CVD events included self-reported myocardial infarction (MI) or stroke through 2021. Analyses were adjusted for a range of covariates. Secondary analyses included examining stroke and MI separately and confirmed (via medical records) or corroborated cases of CVD as the outcomes.
Results:
Mean age in our sample was 46.3 (SD=4.7) years in 2001. Overall, 7,721 women (11.7%) reported experiencing stalking, and 3,686 women (5.6%) reported obtaining a restraining order. Over follow-up (median 19.9 years (interquartile range:19.5, 20.0)), a total of 1,879 women (2.8%) reported new onset CVD. Compared to women who had not been stalked, women who reported experiencing stalking had an increased risk of self-reported incident CVD (hazard ratio (HR):1.41, 95% CI:1.24–1.60), adjusting for sociodemographic and childhood/family factors. Women who had, versus had not, obtained a restraining order also had an increased risk of self-reported incident CVD (adjusted HR:1.70, 95% CI:1.44–1.98). Results were similar when considering self-reported MI and stroke as separate outcomes and confirmed/corroborated CVD. Results were slightly attenuated but remained statistically significant when adjusted for health behaviors, medications, health conditions, childhood abuse, and depressive symptoms.
Conclusions:
Experiences of stalking and obtaining a restraining order are each associated with an increased risk of developing CVD in women. Common experiences of violence among women may impact cardiovascular health and warrant consideration alongside more traditional risk factors.
Keywords: violence, sexual violence, gender-based violence, women’s health, cardiovascular disease, stalking, restraining order
Introduction
Cardiovascular disease (CVD) is the leading cause of death among women in the United States (US) and worldwide (1,2). Past research has shown that women’s experiences of violence are associated with an increased risk of CVD and other cardiovascular conditions such as hypertension (3). However, prior work has predominantly examined experiences of physical or sexual intimate partner violence (IPV), sexual assault, or childhood abuse (3–5). In contrast, some of the most common forms of violence women experience, including stalking, have received little attention in the cardiovascular literature (3). Stalking can be defined as someone’s repeated use of unwanted harassment or threatening tactics that also cause fear or safety concerns and is reported by nearly 1 in 3 women in the US (6). The most common perpetrators of stalking reported in the US are current or former intimate partners, and the second most common are acquaintances (6). Given the high prevalence of stalking in women’s experiences and prior work linking violence exposure to adverse health outcomes, examining its relation to CVD risk in women is paramount.
To the best of our knowledge, no studies have investigated the relationship between stalking and the risk of incident CVD, although a limited number of studies have examined other physical health outcomes. For example, using cross-sectional data from a national US survey (the National Violence Against Women Survey (NVAWS)), it was found that women with a history of stalking were more likely to report poor physical health as well as report developing a chronic disease (which included hypertension and heart disease) since being stalked (4). Another study using the NVAWS data did not find an association between women’s history of stalking by an intimate partner and poor perceived health (7). The most recent National Intimate Partner and Sexual Violence Survey (NISVS) report on stalking (a large epidemiological survey of women in the US) found that women who reported experiencing stalking were more likely to report conditions such as chronic pain and sleep difficulties but were less likely to report diabetes and hypertension (6). However, inferences from the NISVS analysis are limited by the study’s cross-sectional nature, the lack of adjustment for relevant demographic and health risk factors, and the reliance on self-reported outcomes only (6).
Restraining orders are a legal means of protection from violence that aim to limit the contact between a victim and perpetrator (8) and can be issued to prevent women from severe stalking, harassment, and IPV (9). Restraining orders may, therefore, provide an alternative assessment of women’s experience of severe violence history. Some evidence suggests restraining orders protect women from experiencing repeated violence (8–10). However, there have been reported associations between obtaining restraining orders and poorer health (11), likely suggesting that the violence that led to taking out a restraining order, or additionally, any potential violation or worsening of violence following a restraining order (10,12,13), had negative health consequences.
Taken together, women’s experiences of stalking and severe violence (as indicated by obtaining a restraining order) may be related to adverse physical health outcomes, including behavioral CVD risk factors such as sleep problems (14). In addition to behavioral risk factors, a relationship between stalking and CVD may act via psychological distress, including depression (4,15). In the present study, we investigated the prospective relationships between a history of experiencing stalking, a history of obtaining a restraining order, and subsequent risk of incident CVD over 20 years in the Nurses’ Health Study II (NHSII), a longitudinal sample of civilian women aged 36–56 years at the start of the present follow-up in 2001. We hypothesized that women with a history of stalking (versus no history) and women who had (versus had not) obtained a restraining order would have a higher risk of developing CVD over time. We accounted for a wide range of covariates including childhood abuse and depressive symptoms, given prior work has shown links between these and both violence and CVD (4,15–17).
Methods
Data availability
NHSII data is not publicly available but may be made available upon request to the Channing Division of Network Medicine at the Brigham and Women’s Hospital. Further information on data access is described at https://www.nurseshealthstudy.org/researchers (email: nhsaccess@channing.harvard.edu).
Sample
NHSII is an ongoing cohort study of US women, which invited participants through mailings and enrolled 116,429 nurses aged 25 to 42 years in 1989. An extensive range of sociodemographic, medical, and behavioral variables have been measured in this cohort, primarily via biennial questionnaires. The study protocol was approved by the institutional review boards of the Brigham and Women’s Hospital and the Harvard T.H. Chan School of Public Health, and the institutional review boards allowed participants’ completion of questionnaires to be considered as implied consent. In 2001, a sub-study questionnaire including questions about participant’s history of stalking and having obtained a restraining order was mailed to 91,297 participants who had responded within three of the previous biennial questionnaires and who had not requested short questionnaires; the violence sub-study questionnaire was returned by 68,376 women.
Figure 1 details our inclusions and exclusions. We excluded women missing information on race/ethnicity, a key covariate, given known differences in women’s experience of violence by race (6). Our main analytic sample consisted of 66,270 women, followed from return of the 2001 biennial questionnaire to return of the 2019 questionnaire (between 2019–2021).
Figure 1.

Inclusions and exclusions for analytic sample.
Note: no further exclusions were made for missing covariate information, see Supplemental Methods for further information on how missing information was handled for other covariates.
Measures
Stalking and restraining order
Lifetime history of stalking and having obtained a restraining order were assessed separately in 2001. Stalking exposure was assessed by an item asking women to indicate whether they had ever been stalked (‘Have you ever been stalked by anyone?’), which we then categorized as a binary variable for our main stalking exposure (yes versus no [reference group]). For our main restraining order exposure, women were asked whether they had ever taken out a restraining order (‘Have you ever taken out a restraining order?’) and we categorized responses into yes versus no [reference group]. These exposures were not mutually exclusive.
We included a secondary measure of stalking using an alternate, narrow definition where we classified women following the NVAWS (18), which was the basis of the stalking items in NHSII. This measure was more narrow as it included multiple requirements to be considered exposed; only women who (a) reported experiencing one or more stalking harassment behaviors (e.g., followed or spied on), as well as (b) reported having had these behaviors done on more than one occasion, and (c) additionally reported feeling very frightened or feared bodily harm to them or a loved one as a result of these behaviors were considered exposed for this narrow measure (yes versus no [reference group]).
Further details on exposures are provided in Table S1.
Cardiovascular Disease (CVD)
CVD was defined as self-reported fatal or non-fatal myocardial infarction (MI) or stroke. At NHSII baseline in 1989, women were asked with separate items about lifetime physician-diagnosed “MI or angina” and “stroke (CVA [cerebrovascular Accident] or TIA [transient ischemic attack])”; women were then asked on each subsequent biennial questionnaire (1991–2019) about diagnoses in the previous two years. After 1989, angina and MI were asked separately; we considered responses for MI only. From the 2005 biennial questionnaire onwards, stroke (CVA) and TIA were asked independently, and we only considered stroke (CVA), following previous work (19), with only one woman in our analytic sample having confirmed TIA alone. For all analyses, we excluded women who reported having CVD on questionnaires from 1989 to 2001 (inclusive) and included only CVD events after the return of the 2001 questionnaire.
We included a secondary measure for confirmed or corroborated CVD (henceforth termed adjudicated CVD). For women who self-reported MI or stroke, permission was requested from them (or next of kin if the participant had died) to allow physicians, who were unaware of the participants’ exposure status, to review their medical records. Approximately 53% of women in our analytic sample who self-reported a CVD event had adjudicated CVD. For participants who had died, disease status was reviewed from death certificates or autopsy reports. Physicians used the World Health Organization criteria (20) to confirm MI: typical symptoms plus either elevated cardiac enzyme levels or diagnostic electrocardiographic findings. To confirm MI deaths, either enzyme electrocardiographic changes before death or autopsy reports showing evidence of acute infarction or thrombus were required. For stroke, physicians used the National Survey of Stroke criteria (21), requiring evidence of a neurological deficit with a sudden or rapid onset that persisted for more than 24 hours or until death. Cerebrovascular pathology resulting from infection, trauma, or malignancy and “silent” strokes discovered only by radiology were excluded. In circumstances where medical records were inadequate or inaccessible, participants were asked to corroborate their self-reported diagnoses on questionnaires via letter or interview. Adjudication is ongoing and may not include all cases (incomplete record review for questionnaires 2015–2019 for stroke and 2017–2019 for MI).
Covariates
We considered a range of potential confounders and/or mediators of the relationship between stalking, restraining orders, and CVD using measures of sociodemographic, childhood and family factors, and adult health behaviors and conditions, all self-reported in questionnaires. Time invariant covariates included: race/ethnicity (non-Hispanic White, African-American/Hispanic/Asian/Other; measured in 1989; due to relatively small sample sizes, Asian, Hispanic, African-American, and ‘other’ responses were combined); maximum parental education in participants’ infancy (high school or less, some college, 4+ years of college; measured in 2005); maternal and paternal history of MI (assessed in 1989, 1997 and 2001); maternal and paternal occupation in participants’ infancy (see Supplemental Methods; assessed in 2005); parental home ownership in participants’ infancy (ever/never; assessed in 2005); maternal and paternal history of stroke (assessed in 1997 and 2001); and somatotype at age 5 (see Supplemental Methods; queried in 1989) (22,23).
Time-varying factors were updated every 2 or 4 years over the follow-up period and included: physical activity (<3, 3–<9, 9–<18, 18–<27, 27+ metabolic equivalent hours per week); alcohol intake (none, 1–<5, 5– <10, 10+ grams per day); smoking status (non-smokers, former smokers, or current smokers of 1–14, 15–24, or 25+ cigarettes/day); diet quality (continuous); adult body mass index (BMI) (in kg/m2; continuous); hypertension (ever/never); type 2 diabetes (ever/never); and menopausal hormone therapy use (never or premenopausal, current user, past user). BMI was derived using self-reported height in inches from 1989 and current weight in pounds on each biennial questionnaire, transformed to kg/m2. Physical activity, diet quality, and alcohol intake were assessed via validated measures (24–27), including the Alternative Healthy Eating Index (28,29). Missing data information is provided in Supplemental Methods and Table S2.
In sensitivity analysis, we included childhood abuse and depressive symptoms as covariates, given prior work has shown independent links between these experiences and violence and CVD (4,15–17). Childhood abuse was assessed in 2001 with the Physical/Emotional Abuse Subscale of the Childhood Trauma Questionnaire (30,31) and the Sexual Maltreatment Scale of the Conflict Tactics Scales (32,33), and coded following previous work in this cohort (34). We included childhood abuse as a continuous measure ranging from 0 (no abuse) to 5 (severe combined abuse), marking the prevalence and severity of childhood physical, emotional, and sexual abuse. Depressive symptoms were measured in 2001 using the validated Mental Health Inventory-5 (past-month symptoms; possible scores range from 0 (high depression) to 100) (35).
Statistical analysis
We examined the prevalence and overlap of exposures and then assessed sample characteristics as well as the distribution of covariates by history of stalking and obtaining a restraining order (separately). We explored the prevalence of women who also reported IPV (physical, emotional, or sexual) in 2001 among our sample and among the subsample of women with stalking history. We further examined the prevalence of women who indicated a partner was the reason for restraining orders, or the source of harassment behavior included in our narrow measure of stalking. We additionally examined the distribution of key covariates for our analytic sample compared to the rest of the NHSII cohort.
To determine whether these experiences of violence were prospectively associated with incident CVD, we used Cox proportional regression models (computing hazard ratios (HR) and 95% confidence intervals (CI)), after first verifying the proportional hazards assumption was met. The follow-up period started with the return of the 2001 violence questionnaire and extended through to the 2019 questionnaire return (between 2019–2021; mean:18.4 years (standard deviation [SD]=3.6); median:19.9 years (interquartile range=19.5, 20.0)). Separate models were conducted for the two main exposures of reported experience(s) of stalking and having obtained a restraining order. Women contributed person-time from the start of follow-up until the onset of CVD or censoring. Women were censored at the end of follow-up, their last questionnaire return date if they did not return the 2019 questionnaire (n=8,125), or death (non-cases, all causes n=2,322).
All analyses were adjusted for the same sets of covariates. In model 1, we included adjustment for age (models were stratified jointly by age in months and follow-up cycle to finely control for potential confounding by age and calendar time). In model 2, we additionally included sociodemographic, childhood, and family factors: race/ethnicity, parental education, somatotype at age 5, maternal and paternal occupation, parental homeownership, maternal and paternal history of MI, and maternal and paternal history of stroke. In model 3, we additionally adjusted for adult BMI and health behaviors (alcohol intake, smoking status, physical activity, and diet quality). In model 4, we further adjusted for adult health conditions and medications: type 2 diabetes, hypertension, and menopausal hormone therapy use. Covariates added in model 2 measure factors that likely occur before violence and may be confounders, with parental CVD history included for model efficiency. We considered covariates in models 3 and 4 separately as they could be potential mediators of associations and were measured after violence.
In secondary analyses, we used our narrow stalking measure as the exposure. We also conducted analyses examining outcomes of self-reported MI and stroke (separately). Additionally, we examined a combined exposure where we compared women who reported experiencing both stalking and obtaining a restraining order, or women who reported one of these types of exposures, to women who reported neither. To explore potential associations of violence committed by non-partners, we examined the risk of CVD for reporting an ‘other’ person for whom the restraining order was against, and separately, reporting an ‘other’ person as committing at least one of the harassment behaviors included in our narrow measure of stalking. Next, we used adjudicated CVD as our outcome and, when doing so, censored for self-reported CVD that was not confirmed/corroborated. We additionally examined whether adjudication for CVD was associated with exposure status for stalking and obtaining a restraining order.
In sensitivity analysis, we included models for our main exposures of stalking and obtaining a restraining order with self-reported CVD that also adjusted for childhood abuse and depressive symptoms in addition to the covariates in model 2. All analyses were performed using SAS software Version 9.4 (SAS Institute Inc) and an alpha value of 0.05 was used to determine statistical significance. We adhered to the STROBE reporting guidelines (see Supplemental Material).
Results
Overall, 7,721 women (11.7%) reported experiences of stalking and 3,686 women (5.6%) reported obtaining a restraining order. There were 1,895 women who reported both stalking and obtaining a restraining order (24.5% of women stalked; 51.4% of women who obtained a restraining order). When defining stalking according to our narrow definition following the NVAWS (18), 4,710 women (10.3%) were classified as having been stalked. The mean age at the start of follow-up was 46.3 (SD=4.7) years and most women in our sample were non-Hispanic White (n=61,822; 93.3%). Table 1 presents the distribution of covariates at the start of follow-up by having been stalked or obtained a restraining order. Among our sample, 28,443 women (42.9%) reported IPV. Of women who reported stalking, 5,500 (71.2%) also reported IPV. Of women who met the narrow definition of stalking, 2,448 (52.0%) reported that at least one of the contributing harassment behaviors was committed by a current or former spouse/significant other. Furthermore, we found that, of women who reported having obtained a restraining order, 3,127 (84.8%) reported that it was against a spouse/significant other, 595 (16.1%) reported an ‘other’ person, and 36 (1.0%) reported restraining orders against both. See Table S3 for key covariate distributions in our analytic sample compared to the rest of the NHSII cohort.
Table 1.
Age-standardized characteristics by exposures among 66,270 women in the Nurses’ Health Study II at start of follow-up in 2001.
| Stalked | Restraining Order | |||
|---|---|---|---|---|
| No (n=58549) |
Yes (n=7721) |
No (n=62584) |
Yes (n=3686) |
|
|
| ||||
| Sociodemographic, childhood and family factors | ||||
|
| ||||
| Age (years)a | 46.3 (4.7) | 46.2 (4.5) | 46.3 (4.7) | 46.6 (4.5) |
| Race/ethnicity | ||||
| Non-Hispanic White, %(n) | 93.4 (54699) | 92.3 (7123) | 93.4 (58448) | 91.6 (3376) |
| African-American/Asian/Hispanic/Otherb, %(n) | 6.6 (3850) | 7.7 (598) | 6.6 (4136) | 8.4 (310) |
| Parental education | ||||
| High school or less, %(n) | 51.8 (27344) | 52.5 (3553) | 51.7 (29100) | 56.0 (1797) |
| Some college, %(n) | 24.4 (12889) | 25.1 (1699) | 24.5 (13792) | 24.9 (798) |
| 4+ years of college, %(n) | 23.7 (12516) | 22.4 (1514) | 23.8 (13417) | 19.1 (612) |
| Maternal Occupationc | ||||
| Homemaker, %(n) | 69.5 (36249) | 67.1 (4493) | 69.3 (38613) | 67.3 (2131) |
| Unskilled labor/service, %(n) | 5.0 (2606) | 5.7 (383) | 5.0 (2802) | 5.9 (188) |
| Farming, %(n) | 1.5 (779) | 1.4 (96) | 1.5 (838) | 1.2 (37) |
| Mechanic/machine/military/sales, %(n) | 11.8 (6175) | 13.1 (876) | 11.9 (6607) | 13.9 (439) |
| Professional or manager, %(n) | 12.2 (6339) | 12.7 (851) | 12.3 (6822) | 11.7 (370) |
| Paternal Occupationc | ||||
| Homemaker, %(n) | 0.2 (112) | 0.3 (23) | 0.2 (130) | 0.1 (5) |
| Unskilled labor/service, %(n) | 14.9 (7716) | 14.4 (958) | 14.7 (8174) | 16.0 (502) |
| Farming, %(n) | 7.7 (3979) | 6.4 (423) | 7.6 (4198) | 6.5 (205) |
| Mechanic/machine/military/sales, %(n) | 47.7 (24752) | 49.1 (3261) | 47.7 (26412) | 51.0 (1602) |
| Professional or manager, %(n) | 29.6 (15362) | 29.8 (1980) | 29.8 (16514) | 26.3 (824) |
| Parental homeownership, %(n) | 54.3 (28550) | 53.2 (3585) | 54.3 (30491) | 51.6 (1645) |
| High somatotype at age 5d, %(n) | 6.8 (3968) | 7.1 (548) | 6.8 (4242) | 7.5 (277) |
| Maternal history of MI, %(n) | 10.8 (6309) | 12.7 (980) | 10.9 (6838) | 12.2 (448) |
| Paternal history of MI, %(n) | 29.1 (17018) | 31.5 (2435) | 29.3 (18344) | 30.1 (1108) |
| Maternal history of stroke, %(n) | 7.9 (4648) | 9.0 (694) | 8.0 (4991) | 9.5 (349) |
| Paternal history of stroke, %(n) | 10.7 (6288) | 11.9 (916) | 10.9 (6814) | 10.7 (393) |
|
| ||||
| Adult health behaviors and body mass index | ||||
|
| ||||
| Body mass index (kg/m2) | 26.7 (6.2) | 26.6 (6.3) | 26.7 (6.2) | 26.8 (6.2) |
| Current smokerd, %(n) | 7.6 (4427) | 13.0 (1000) | 7.8 (4880) | 15.0 (553) |
| Physical activity (MET-hrs/wk) | ||||
| <3, %(n) | 19.4 (11381) | 17.0 (1316) | 19.2 (11987) | 19.1 (705) |
| 3-<9 %(n) | 20.9 (12249) | 19.2 (1482) | 20.8 (13021) | 19.2 (709) |
| 9-<18, %(n) | 20.1 (11785) | 18.8 (1450) | 20.0 (12518) | 19.6 (721) |
| 18-<27, %(n) | 13.3 (7803) | 13.2 (1017) | 13.4 (8388) | 11.7 (432) |
| 27+, %(n) | 26.2 (15330) | 31.8 (2456) | 26.6 (16671) | 30.3 (1118) |
| Alcohol intake (g/d)e | ||||
| None, %(n) | 38.3 (22412) | 37.1 (2867) | 38.0 (23811) | 39.7 (1463) |
| 1-<5, %(n) | 32.4 (18991) | 31.8 (2455) | 32.5 (20314) | 30.8 (1137) |
| 5-<10, %(n) | 12.3 (7213) | 12.2 (944) | 12.3 (7696) | 12.5 (460) |
| 10+, %(n) | 17.0 (9932) | 18.8 (1455) | 17.2 (10763) | 17.0 (627) |
| Diet qualityf | 51.0 (12.0) | 52.2 (12.1) | 51.2 (12) | 51.3 (12.1) |
|
| ||||
| Adult health conditions and medications | ||||
|
| ||||
| Hypertension, %(n) | 17.3 (10110) | 19.0 (1468) | 17.4 (10876) | 19.1 (704) |
| Type 2 diabetes, %(n) | 2.0 (1160) | 2.1 (166) | 2.0 (1234) | 2.5 (92) |
| Menopausal hormone therapy | ||||
| Never or premenopausal, %(n) | 82.1 (48063) | 79.9 (6170) | 82.0 (51301) | 79.4 (2928) |
| Current user, %(n) | 14.5 (8489) | 15.9 (1229) | 14.6 (9128) | 16.2 (597) |
| Past user, %(n) | 3.4 (1997) | 4.2 (322) | 3.4 (2155) | 4.4 (160) |
Note: Values are means (standard deviations) for continuous variables; percentages (n’s) for categorical variables, and are standardized to the age distribution of the study population. Values of polytomous variables may not sum to 100% due to rounding. Variables are presented after missing has been accounted for (e.g., with mean imputation). Abbreviations: MI: myocardial infarction; MET-hrs/wk: metabolic equivalent task hours per week; kg/m2: kilogram per (meter squared); g/d: grams per day.
Value is not age-adjusted.
Other was a response option for the questions regarding race/ethnicity.
See Supplemental Methods for further details on response categories for maternal and paternal occupations.
For brevity, only certain categories for somatotype at age 5 and smoking status are presented. Highest somatotype refers to the highest category of somatotype pictogram responses.
Alcohol intake was assessed in 2003, the cycle closest to 2001 where alcohol was assessed. Diet quality was assessed in 2003 (the closest cycle to 2001) and based on the Alternative Healthy Eating Index. A higher diet quality score indicates a higher quality diet.
Over follow-up, 1,879 women (2.8%) self-reported developing CVD. We verified that the proportional hazards assumption was satisfied for mean age at the start of follow-up (younger than 46 years versus 46 years or older) and follow-up period (pre versus post 2013) using likelihood ratio tests for analyses reported in Table 2 (p-values>0.10 for all tests). We found that compared with women who reported no stalking history, women who experienced stalking had a higher risk of developing CVD (Table 2; Figure S1). For example, in model 2, women who reported experiencing stalking had a 41% increased risk of self-reported incident CVD (HR:1.41, 95% CI:1.24–1.60; Table 2). Women who reported having obtained a restraining order also had a higher risk of developing CVD relative to those who had not (Table 2; Figure S2), such that women who reported having obtained a restraining order showed 70% increased risk of self-reported incident CVD (model 2 HR:1.70, 95% CI:1.44–1.98). Results were attenuated but remained statistically significant when including adjustment for adult BMI, health behaviors, medications, and health conditions.
Table 2.
Cox proportional hazards estimates for the association between stalking or obtaining a restraining order and self-reported CVD events among 66,270 women in the Nurses’ Health Study II for follow-up between 2001–2019 cycles.
| N=1,879 events among 1,220,565 person-years | Model 1 | Model 2 | Model 3 | Model 4 | |
|---|---|---|---|---|---|
|
| |||||
| Cases/person-years | HR (95% CI) | HR (95% CI) | HR (95% CI) | HR (95% CI) | |
| Stalked | |||||
| No | 1584/1080993 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Yes | 295/139572 | 1.46 (1.29, 1.65) | 1.41 (1.24, 1.60) | 1.37 (1.21, 1.55) | 1.35 (1.19, 1.53) |
| Restraining order | |||||
| No | 1704/1,154,698 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Yes | 175/65,866 | 1.75 (1.49, 2.04) | 1.70 (1.44, 1.98) | 1.61 (1.38, 1.89) | 1.58 (1.35, 1.85) |
Note: Age (in months) and follow-up cycle were stratified. Covariates were time-updated apart from race/ethnicity, somatotype at age 5 years, parental education, maternal and paternal occupation separately, parental homeownership, maternal and paternal history of myocardial infarction, and maternal and paternal history of stroke. Abbreviations: HR: hazard ratio; CI: confidence interval; CVD: cardiovascular disease.
Model 1 adjusted for age.
Model 2 also adjusted for socioeconomic/childhood/family factors: race/ethnicity, somatotype at age 5 years, parental education, maternal and paternal occupation separately, parental homeownership, maternal and paternal history of myocardial infarction, and maternal and paternal history of stroke.
Model 3 also adjusted for adulthood health behaviors and body mass index: smoking status, alcohol intake, physical activity, diet quality, and body mass index.
Model 4 also adjusted for adulthood health conditions and medications: hypertension, type 2 diabetes, and menopausal hormone therapy use.
A similar pattern of results was found for our narrow measure of stalking exposure (Table S4). Results for MI and stroke as separate outcomes also showed similar results (Table S5). In analyses examining the combined exposure for stalking and obtaining a restraining order, we found a dose-response relationship whereby women who reported both history of stalking and obtaining a restraining order showed the highest risk of CVD compared to women who reported neither exposure (e.g., model 2 HR: 2.16, 95% CI:1.77–2.62) (Table S6). Results for analyses examining violence by non-partners (i.e., ‘other’ person) are presented in Table S7, suggesting that women who experience violence by non-partners appear at increased risk of developing CVD, although case numbers were small and limited statistical power for restraining orders against non-partners.
In our other secondary analyses examining adjudicated CVD, 990 CVD events (1.5%) occurred over follow-up, of which 49 (4.9%) were fatal. To note, adjudication of self-reported CVD varied as a function of stalking and restraining order status, whereby women with these exposures were less likely to have adjudicated CVD (chi-square p-values each <0.01). Regardless, we found a similar pattern of results when considering adjudicated CVD, whereby a history of stalking or having obtained a restraining order were associated with a higher risk of developing the disease (e.g., restraining order HR: 1.60, 95% CI:1.28–2.00 in model 2) (Table 3).
Table 3.
Cox proportional hazards estimates for the association between stalking or obtaining a restraining order and adjudicated CVD events among 66,270 women in the Nurses’ Health Study II for follow-up between 2001–2019 cycles.
| N=990 events among 1,220,604 person-years | Model 1 | Model 2 | Model 3 | Model 4 | |
|---|---|---|---|---|---|
|
| |||||
| Cases/person-years | HR (95% CI) | HR (95% CI) | HR (95% CI) | HR (95% CI) | |
| Stalked | |||||
| No | 840/1,081,025 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Yes | 150/139,579 | 1.41 (1.18, 1.68) | 1.37 (1.15, 1.63) | 1.33 (1.11, 1.58) | 1.31 (1.10, 1.56) |
| Restraining order | |||||
| No | 903/1,154,738 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Yes | 87/65,866 | 1.64 (1.31, 2.05) | 1.60 (1.28, 2.00) | 1.53 (1.22, 1.91) | 1.50 (1.20, 1.88) |
Note: Age (in months) and follow-up cycle were stratified. Covariates were time-updated apart from race/ethnicity, somatotype at age 5 years, parental education, maternal and paternal occupation separately, parental homeownership, maternal and paternal history of myocardial infarction, and maternal and paternal history of stroke. Adjudicated CVD was defined as confirmed (using medical records) or corroborated. Abbreviations: HR: hazard ratio; CI: confidence interval; CVD: cardiovascular disease.
Model 1 adjusted for age.
Model 2 also adjusted for socioeconomic/childhood/family factors: race/ethnicity, somatotype at age 5 years, parental education, maternal and paternal occupation separately, parental homeownership, maternal and paternal history of myocardial infarction, and maternal and paternal history of stroke.
Model 3 also adjusted for adulthood health behaviors and body mass index: smoking status, alcohol intake, physical activity, diet quality, and body mass index.
Model 4 also adjusted for adulthood health conditions and medications: hypertension, type 2 diabetes, and menopausal hormone therapy use.
In our sensitivity analysis that additionally accounted for childhood abuse and depressive symptoms, results were attenuated but in line with our main findings (e.g. restraining order HR:1.58, 95% CI:1.35–1.85 when adjusting for model 2 covariates and childhood abuse and depressive symptoms) (Table S8).
Discussion
We examined whether experiences of stalking, and separately whether obtaining a restraining order, were prospectively associated with the risk of developing CVD over 20 years in a large, population-based cohort of women. Stalking history was associated with a 41% increased risk of developing CVD even after adjusting for risk factors such as parental CVD history. For women who took out a restraining order, their risk of developing CVD was 70% higher than women who had not taken out a restraining order (whether they had experienced violence or not). These relationships remained similar when using adjudicated CVD, where case status was corroborated or determined via medical records. Results were also similar overall when examining self-reported MI and stroke separately. Our secondary analysis highlighted how women who experienced both stalking and who took out a restraining order had the highest risk of developing CVD, which may reflect either the severity of the stalking experience or serve as a marker of polyvictimization. Therefore, through leveraging a richly characterized cohort of women, we rigorously showed that stalking, one of the most common experiences of violence for women in the US (6), is associated with a subsequent risk of developing CVD. We believe this study is unique in examining the long-term risk of incident CVD for women whose experience of violence was severe enough that they took out a restraining order, with associations supporting this likely severity. We extend previous findings of CVD risk by related exposures (e.g. IPV, sexual assault) to stalking, a direct (often non-contact) form of violence, and having obtained a restraining order, an indicator of severe violence experiences (3). Consequently, our results further document the devastating consequences of violence on women’s cardiovascular health.
Non-contact forms of violence have previously been shown to have cardiovascular health consequences, such as workplace sexual harassment (36), yet are less studied compared to contact forms (3) such as sexual assault. In the present study, we did not have detailed information on the type of violence or context that may have co-occurred with being stalked or led to obtaining a restraining order. However, our narrow measure of stalking, based on multiple behavioral requirements and defined similarly to the NVAWS (18), assessed predominantly non-contact behaviors (e.g., being sent unsolicited letters or written correspondence) (6) and results showed that this definition of stalking was associated with an increased risk of developing CVD. Non-contact forms of violence may be viewed as less severe yet are prevalent and can be chronic, with prior work showing that stalking lasts an average of 22 months (37). It is therefore important to study these forms of violence, both in the context of cardiovascular healthcare and to ensure that women receive appropriate help, especially when stalking can be a misunderstood crime and, at times, mistaken by others as innocent romantic gestures (6).
Perpetrators of stalking or violence resulting in a restraining order may be persons known or unknown to the victim. National US survey data reveal that women who are stalked often knew their perpetrators beforehand as most are intimate partners; the second most common perpetrators are acquaintances (6). Stalking is considered a form of IPV when committed by a former, current, or prospective partner (38), and stalking has previously been linked with other forms of IPV and, in combination, an increased risk of intimate partner homicide (39,40). Restraining orders are commonly obtained against abusive intimate partners or ex-partners (8). Here, we found substantial overlap between stalking and IPV, although we cannot ascertain whether reports refer to the same relationship or perpetrator. Therefore, our findings likely represent a form of IPV for many women in our sample, and our results are in line with prior reports on the cardiovascular health implications of IPV (3,41). However, we did find that many women reported stalking harassment behaviors and restraining orders for non-partners. Overall, stalking is still a relatively understudied form of IPV compared to physical or sexual forms and, to our knowledge, has not been specifically examined with women’s later risk of CVD. Moreover, aside from intimate partner stalking, stalking by an acquaintance has been linked with psychological factors and chronic diseases; therefore, even though prior work focuses on intimate partner stalking, it is also important to study stalking broadly when considering it as a potential CVD risk factor (4,42). Indeed, we showed that stalking by non-partners was associated with higher risk of CVD among women who met our narrow definition of stalking; statistical power for examining obtaining a restraining order against non-partners was limited.
Our results appeared to be independent of sociodemographic, childhood, and family factors (e.g., parental education, somatotype at age 5, parental CVD history). Although results were attenuated, substantive associations remained evident even when accounting for health behaviors (e.g., smoking status), health conditions (e.g., type 2 diabetes), menopausal hormone therapy use, or when adjusting for childhood abuse and depressive symptoms. The average length of stalking exposure, as well as the often-arduous nature of obtaining a restraining order on top of the violence itself, may equate to a substantial period where women may feel a lack of control over their lives and fear for their safety. Psychological distress likely lies on the pathway between experiences of stalking or having obtained a restraining order and CVD, with prior work documenting an increased likelihood of distress with violence, including stalking (4), as well as associations between distress and CVD (15,19). Psychological distress may affect women’s risk of CVD directly via biological mechanisms such as dysregulation of the hypothalamic-pituitary-adrenal axis, chronic sympathetic nervous system activation, and impaired endothelial function (43). Additionally, deleterious health behaviors may lie on the violence-CVD pathway and can follow psychological distress and violence (43–45). Although we considered measures of both psychological distress (depressive symptoms) and health behaviors, further research should examine relationships in data with greater detail on the timing of violence and potential mediators, including posttraumatic stress disorder, to better understand the pathways.
There are at least five primary limitations of our study. First, we included only self-reported measures of violence. Lifetime exposure to violence was queried retrospectively and may, therefore, suffer from misclassification due to recall bias, although underestimates of violence prevalence would likely attenuate our results towards the null. Stalking prevalence in our study assessed in 2001 (10–12%) was comparable to national estimates from 1995 (8%) (18); national estimates reported for restraining order prevalence among stalking cases (28%) was likewise comparable to the prevalence of ever obtaining a restraining order among women reporting stalking exposure in the present study (25%), though our data does not allow us to directly link stalking experiences to obtaining a restraining order (13). Future research should explore multi-record and real-time reports of violence, perhaps through the linkage of legal and health records. Second, our primary outcome was self-reported CVD. However, we further included a measure of adjudicated CVD that was verified using medical records or follow-up. Self-reported versus adjudicated CVD each suffer from different biases (see Supplemental Material for details). Overall, our results did not appear to be heavily biased by the type of CVD report given our findings were similar when using either self-reported or adjudicated CVD as the outcome. Third, our assessments of violence were limited in detail. For example, we lacked information on timing, although recent national estimates suggest the majority of women are first stalked in adulthood before age 25 (6,13,18), prior to our CVD outcomes. We do not know the perpetrator relationship (for our main stalking exposure), or number of times experienced (for having obtained a restraining order, including whether restraining orders were violated or not which may result in further violence exposure (8,46)). Research examining the impact of stalking or obtaining a restraining order with CVD using more detailed and modern exposure measures (e.g., tracking technology (38)) may improve our understanding of the relationship and magnitude of associations with CVD, increasing understanding of these potential CVD risk factors. Fourth, our sample mainly consisted of non-Hispanic White women and the NHSII is an occupational cohort that enrolled registered nurses and required at least a two-year Associates degree at study entry. Evidence shows that violence history, including stalking, is more common among women from minority racial/ethnic backgrounds and those with low income (38,47). Prior work has shown that obtaining a restraining order is also socially stratified (8,48), and structural inequalities may impact women’s abilities to obtain a restraining order. For example, women with higher income and educational attainment are reportedly more likely to use restraining orders (8,48). Challenges in obtaining restraining orders may further contribute to our results not being generalizable to other populations, and further studies should seek to examine the risk of CVD risk following experiences of stalking and obtaining restraining orders in more diverse and understudied groups, considering structural barriers and exploring potential effect modification of such factors. Finally, the observational nature of our data means that the possibility of bias due to unmeasured confounding remains.
Our study included several strengths. We included a large sample of midlife and older women, allowing us to look beyond subclinical CVD factors and examine risk of manifest disease (3). Furthermore, we included a prospective design with around 20 years of follow-up. The extensive data available on covariates allowed us to consider measures of childhood and family factors as potential confounders, as well as potential mediators including health factors and depressive symptoms (3). There is a scarcity of large prospective aging cohorts that assess health outcomes and violence (49) however, we managed to include two assessments of stalking history from the NHSII. One perception-based measure queried whether women had been stalked, whereas a second, narrow, measure was derived to identy whether women had been stalked based on objective violent behaviors queried in national surveys and may allow results to be compared across studies (6,18,38). Lastly, we included a measure of adjudicated CVD based on corroboration or medical record review. Regardless of these exposure or outcome definitions used, we consistently found associations between stalking or obtaining a restraining order and risk of developing CVD, further highlighting the need to not overlook such types of violence as potential CVD risk factors.
Conclusions
Our results demonstrated relationships between experiences of stalking and having obtained a restraining order in women and increased risk of developing CVD. Whilst obtaining a restraining order may help reduce violence in some circumstances (8), the history and severity of violence marked by a restraining order, similar to a history of stalking, appears to have long-term cardiovascular consequences. Through obtaining a restraining order, women who have experienced violence are identified by agencies; this interaction provides an opportunity to support these women and possibly intervene to limit potential long-term cardiovascular health consequences. Furthermore, stalking experiences are common among women but are unlikely to be disclosed to healthcare providers given the lack of screening (50), yet may impact their cardiovascular health. Therefore, attempts to understand women’s lives beyond traditional risk factors when evaluating CVD risk likely warrants further attention for healthcare. Overall, our findings highlight the urgent need to consider violence against women, a prevalent yet underexplored risk factor for CVD, in understanding women’s health and orienting priorities for public health interventions.
Supplementary Material
Supplemental Materials
Supplemental Methods
Tables S1-S8
Figures S1-S2
Strobe Checklist
References 51-54
Clinical perspective.
What is new?
Women’s experiences of being stalked and women’s history of obtaining a restraining order were each prospectively associated with a greater risk of developing cardiovascular disease over a 20-year follow-up period in a sample of civilian women in midlife.
Women who reported obtaining a restraining order showed a particularly increased risk of developing cardiovascular disease.
What are the clinical implications?
Common, yet poorly understood, forms of violence appear to represent a risk factor for cardiovascular disease in women and should be considered in future research and healthcare settings.
Women’s lifetime cardiovascular health may be improved by preventing experiences of stalking and violence leading to obtaining restraining orders.
Acknowledgements
We thank the participants and staff of the NHSII for their valuable contributions and acknowledge the Channing Division of Network Medicine, Department of Medicine, Brigham and Women’s Hospital for its management of the NHSII.
Funding Sources
This research is funded in part by: NIH grants U01CA176726 (for Nurses’ Health Study II infrastructure), K24HL123565 (to Thurston), K23HL159293 (to Jakubowski); National Institute of Mental Health grant R01 MH101269 (to Koenen and Kubzansky); National Institute on Aging grant F31AG080869 (to Murchland); the Broad Trauma Initiative at the Broad Institute of MIT and Harvard. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Abbreviations and Acronyms:
- BMI
Body mass index
- CI
Confidence interval
- CVD
Cardiovascular disease
- HR
Hazard ratio
- NHSII
Nurses’ Health Study II
- SD
Standard deviation
- SX
Symptoms
- US
United States
- NVAWS
National Violence Against Women Survey
- NISVS
National Intimate Partner and Sexual Violence Survey
- MI
Myocardial infarction
- IPV
Intimate partner violence
- MET
Metabolic equivalent task
- TIA
Transient ischemic attack
- CVA
Cerebrovascular Accident
Footnotes
Conflict of Interest Disclosures
In the past three years, Dr. Koenen has served as a paid consultant for the US Department of Justice and Covington and Burling, LLP. Dr. Koenen receives royalties from Guilford Press and Oxford University Press. Dr. Thurston is on the medical advisory board of Astellas Pharma, Bayer, and Hello Therapeutics. Dr. Thurston is a past consultant for Happify Health. No other authors have disclosures.
References
- 1.Patel SA, Winkel M, Ali MK, Narayan KMV, Mehta NK. Cardiovascular mortality associated with 5 leading risk factors: National and state preventable fractions estimated from survey data. Ann Intern Med. 2015;163:245–253. [DOI] [PubMed] [Google Scholar]
- 2.Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, et al. Heart disease and stroke statistics—2020 update: A report from the American Heart Association. Circulation. 2020;141:E139–596. [DOI] [PubMed] [Google Scholar]
- 3.Jakubowski KP, Murray V, Stokes N, Thurston RC. Sexual violence and cardiovascular disease risk: A systematic review and meta-analysis. Maturitas. 2021;153:48–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Davis KE, Coker AL, Sanderson M. Physical and Mental Health Effects of Being Stalked for Men and Women. Violence Vict. 2002;17:429–443. [DOI] [PubMed] [Google Scholar]
- 5.Spencer CN, Khalil M, Herbert M, Aravkin AY, Arrieta A, Baeza MJ, et al. Health effects associated with exposure to intimate partner violence against women and childhood sexual abuse: a Burden of Proof study. Nat Med. 2023;29:3243–3258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Smith SG, Basile KC, Kresnow M. The National Intimate Partner and Sexual Violence Survey: 2016/2017 Report on Stalking — Updated Release. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention; 2022. [Google Scholar]
- 7.Lacey KK, McPherson MD, Samuel PS, Powell Sears K, Head D. The Impact of Different Types of Intimate Partner Violence on the Mental and Physical Health of Women in Different Ethnic Groups. J Interpers Violence. 2013;28:359–385. [DOI] [PubMed] [Google Scholar]
- 8.Cordier R, Chung D, Wilkes-Gillan S, Speyer R. The Effectiveness of Protection Orders in Reducing Recidivism in Domestic Violence: A Systematic Review and Meta-Analysis. Trauma Violence Abuse. 2021;22:804–828. [DOI] [PubMed] [Google Scholar]
- 9.Benitez CT, McNiel DE, Binder RL. Do protection orders protect? J Am Acad Psychiatry Law. 2010;38:376–385. [PubMed] [Google Scholar]
- 10.Holt VL, Kernic MA, Lumley T, Wolf ME, Rivara FP. Civil protection orders and risk of subsequent police-reported violence. JAMA. 2002;288:589–594. [DOI] [PubMed] [Google Scholar]
- 11.Linares LO, Groves BM, Greenberg J, Bronfman E, Augustyn M, Zuckerman B. Restraining orders: a frequent marker of adverse maternal health. Pediatrics. 1999;104:249–257. [DOI] [PubMed] [Google Scholar]
- 12.Acquadro Maran D, Varetto A, Zedda M, Franscini M. Health Care Professionals as Victims of Stalking: Characteristics of the Stalking Campaign, Consequences, and Motivation in Italy. J Interpers Violence. 2017;32:2605–2625. [DOI] [PubMed] [Google Scholar]
- 13.Tjaden P, Thoennes N. Stalking in America: Findings From the National Violence Against Women Survey. U.S. Department of Justice, Office of Justice Programs, National Institute of Justice; 1998. Report No.: 169592. [Google Scholar]
- 14.Makarem N, Castro‐Diehl C, St‐Onge M, Redline S, Shea S, Lloyd‐Jones D, et al. Redefining Cardiovascular Health to Include Sleep: Prospective Associations With Cardiovascular Disease in the MESA Sleep Study. J Am Heart Assoc. 2022;11:e025252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Krittanawong C, Maitra NS, Qadeer YK, Wang Z, Fogg S, Storch EA, et al. Association of Depression and Cardiovascular Disease. Am J Med. 2023;136:881–895. [DOI] [PubMed] [Google Scholar]
- 16.Rich-Edwards JW, Mason S, Rexrode K, Spiegelman D, Hibert E, Kawachi I, et al. Physical and sexual abuse in childhood as predictors of early-onset cardiovascular events in women. Circulation. 2012;126:920–927. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bonagura AG, Widom CS. Child Maltreatment and Psychiatric Disorders Increase Risk for Stalking Victimization. J Interpers Violence. 2023;38:NP60–83. [DOI] [PubMed] [Google Scholar]
- 18.Tjaden P, Thoennes N. Full report of the prevalence, incidence, and consequences of violence against women: findings from the National Violence Against Women Survey. 2000. [Google Scholar]
- 19.Sumner JA, Kubzansky LD, Elkind MSV, Roberts AL, Agnew-Blais J, Chen Q, et al. Trauma Exposure and Posttraumatic Stress Disorder Symptoms Predict Onset of Cardiovascular Events in Women. Circulation. 2015;132:251–259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Rose G, Blackburn H. Cardiovascular Survey Methods Geneva, Switzerland: World Health Organization; 1968. p. 1–188. [PubMed] [Google Scholar]
- 21.Walker A, Robins M, Weinfeld F. The National Survey of Stroke. Clinical findings. Stroke. 1981;12:13–44. [PubMed] [Google Scholar]
- 22.Must A, Willett WC, Dietz WH. Remote recall of childhood height, weight, and body build by elderly subjects. Am J Epidemiol. 1993;138:56–64. [DOI] [PubMed] [Google Scholar]
- 23.Magnusson C, Baron J, Persson I, Wolk A, Bergström R, Trichopoulos D, et al. Body size in different periods of life and breast cancer risk in post-menopausal women. Int J Cancer. 1998;76:29–34. [DOI] [PubMed] [Google Scholar]
- 24.Chasan-Taber S, Rimm EB, Stampfer MJ, Spiegelman D, Colditz GA, Giavanucci E, et al. Reproducibility and validity of a self-administered physical activity questionnaire for male health professionals. Epidemiology. 1996;7:81–86. [DOI] [PubMed] [Google Scholar]
- 25.Giovannucci E, Colditz G, Stampfer MJ, Rimm EB, Litin L, Sampson L, et al. The assessment of alcohol consumption by a simple self-administered questionnaire. Am J Epidemiol. 1991;133:810–817. [DOI] [PubMed] [Google Scholar]
- 26.Schrieks IC, Wei MY, Rimm EB, Okereke OI, Kawachi I, Hendriks HFJ, et al. Bidirectional associations between alcohol consumption and health-related quality of life among young and middle-aged women. J Intern Med. 2015;279:376–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yue Y, Yuan C, Wang DD, Wang M, Song M, Shan Z, et al. Reproducibility and validity of diet quality scores derived from food-frequency questionnaires. Am J Clin Nutr. 2021;115:843–853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Chiuve SE, Fung TT, Rimm EB, Hu FB, McCullough ML, Wang M, et al. Alternative dietary indices both strongly predict risk of chronic disease. J Nutr. 2012;142:1009–1018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Willet WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, et al. Reproduciblity and Validity of a Semiquantitative Food Frequency Questionnaire. Am J Epidemiol. 1985;122:51–65. [DOI] [PubMed] [Google Scholar]
- 30.Bernstein DP, Stein JA, Newcomb MD, Walker E, Pogge D, Ahluvalia T, et al. Development and validation of a brief screening version of the Childhood Trauma Questionnaire. Child Abuse Negl. 2003;27:169–190. [DOI] [PubMed] [Google Scholar]
- 31.Bernstein P, Ph D, Fink L, Ph D, Foote J, Ph D, et al. Initial Reliability and Validity of a New Retrospective Measure of Child Abuse and Neglect. Am J Psychiatry. 1994;151:1132–1136. [DOI] [PubMed] [Google Scholar]
- 32.Straus MA. Measuring intrafamily conflict and violence: The conflict tactics (CT) scales. J Marriage Fam. 1979;41:75–88. [Google Scholar]
- 33.Straus MA, Hamby SL, Finkelhor D, Moore DW, Runyan D. Identification of child maltreatment with the Parent-Child Conflict Tactics Scales: development and psychometric data for a national sample of American parents. Child Abuse Negl. 1998;22:249–270. [DOI] [PubMed] [Google Scholar]
- 34.Roberts AL, Sumner JA, Koenen K, Kubzansky L, Grodstein F, Rich-Edwards J, et al. Childhood Abuse and Cognitive Function in a Large Cohort of Middle-Aged Women. Child Maltreat. 2020;27:100–113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Berwick D, Murphy J, Goldman P, Ware JJ, Barsky A, Weinstein M. Performance of a five-item mental health screening test. Med Care. 1991;29:169–176. [DOI] [PubMed] [Google Scholar]
- 36.Lawn RB, Nishimi KM, Sumner JA, Chibnik LB, Roberts AL, Kubzansky LD, et al. Sexual violence and risk of hypertension in women in the Nurses’ Health Study II: a 7-year prospective analysis. J Am Heart Assoc. 2022;11:e023015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Spitzberg BH, Cupach WR. The state of the art of stalking: Taking stock of the emerging literature. Aggress Violent Behav. 2007;12:64–86. [Google Scholar]
- 38.Black MC, Basile KC, Breiding MJ, Smith SG, Walters ML, Merrick MT, et al. The National Intimate Partner and Sexual Violence Survey (NISVS): 2010 Summary Report. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention; 2011. [Google Scholar]
- 39.Rai A, Villarreal-Otálora T, Blackburn J, Choi YJ. Correlates of Intimate Partner Stalking Precipitated Homicides in the United States. J Fam Violence. 2020;35:705–716. [Google Scholar]
- 40.Campbell JC, Webster D, Koziol-Mclain J, Block C, Campbell D, Curry MA, et al. Risk Factors for Femicide in Abusive Relationships: Results From a Multisite Case Control Study Femicide Cases. Vol. 93, American Journal of Public Health Public Health. 2003. p. 1089–1097. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.El-Serag R, Thurston RC. Matters of the Heart and Mind: Interpersonal Violence and Cardiovascular Disease in Women. J Am Heart Assoc. 2020;9:1–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Logan TK. Examining Factors Associated with Stalking-Related Fears Among Men and Women Stalked by Male and Female Acquaintances. J Interpers Violence. 2022;37:NP6958–87. [DOI] [PubMed] [Google Scholar]
- 43.Ebong IA, Quesada O, Fonkoue IT, Mattina D, Sullivan S, Oliveira GMM de, et al. The Role of Psychosocial Stress on Cardiovascular Disease in Women: JACC State-of-the-Art Review. J Am Coll Cardiol. 2024;84:298–314. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Lloyd-Jones DM, Ning H, Labarthe D, Brewer L, Sharma G, Rosamond W, et al. Status of Cardiovascular Health in US Adults and Children Using the American Heart Association’s New “Life’s Essential 8” Metrics: Prevalence Estimates From the National Health and Nutrition Examination Survey (NHANES), 2013 Through 2018. Circulation. 2022;146:822–835. [DOI] [PubMed] [Google Scholar]
- 45.Strauss CV, Haynes EE, Cornelius TL, Shorey RC. Stalking Victimization and Substance Use in College Dating Relationships: An Exploratory Analysis. J Interpers Violence. 2019;34:2878–2896. [DOI] [PubMed] [Google Scholar]
- 46.Logan TK, Walker R. Civil protective order effectiveness: justice or just a piece of paper? Violence Vict. 2010;25:332–48. [DOI] [PubMed] [Google Scholar]
- 47.Güler A, Lambert J, Rojas-Guyler L, Lee RC, Smith CR. Shared Risk Factors Among Women for Intimate Partner Violence in the United States: A Secondary Analysis. Violence Women. 2024;30:3–30. [DOI] [PubMed] [Google Scholar]
- 48.Durfee A, Messing JT. Characteristics related to protection order use among victims of intimate partner violence. Violence Women. 2012;18:701–710. [DOI] [PubMed] [Google Scholar]
- 49.Murchland AR, Koenen KarestanC, Hayes-Larson E. The link between traumatic experiences and health in late life: challenges and opportunities for US-based longitudinal aging cohorts. Am J Epidemiol. 2025;194:565–572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Lawn RB, Jakubowski K. Violence against women and girls. BMJ. 2024;386:q1791. [DOI] [PubMed] [Google Scholar]
- 51.von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
NHSII data is not publicly available but may be made available upon request to the Channing Division of Network Medicine at the Brigham and Women’s Hospital. Further information on data access is described at https://www.nurseshealthstudy.org/researchers (email: nhsaccess@channing.harvard.edu).
