Increased Risk of Mortality Due to Interpersonal Violence in Foreign-Born Women of Reproductive Age

Abstract

Violence against women is an increasing public health concern, with assault leading to death as the most extreme outcome. Previous findings indicate that foreign-born women living in Sweden are more exposed to interpersonal violence than Swedish-born women. The current study investigates mortality due to interpersonal violence in comparison with other external causes of death among women of reproductive age in Sweden, with focus on country of birth. Foreign-born women and especially those from countries with low and very low gender equity levels had increased risk of mortality due to interpersonal violence, thus implicating lack of empowerment as a contributing factor.

Keywords

foreign-born women violence against women interpersonal violence mortality gender equity

Introduction

An increasing body of evidence shows that violence against women is a major global public health concern (Campbell, 2002; Ellsberg, Jansen, Heise, Watts, & Garcia-Moreno, 2008; Jewkes, 2002; Krantz, 2002), with mortality due to violence as the most severe outcome (Campbell, Glass, Sharps, Laughon, & Bloom, 2007). Gender inequalities such as those fostered by patriarchal structures sanctioning women’s subordination to men are strong contributors to interpersonal violence (Flood & Pease, 2009; Krantz & Garcia-Moreno, 2005) and to intimate partner violence (Abu-Ras, 2007; Chang, Shen, & Takeuchi, 2009). Our previous research has shown that foreign-born women are more exposed to violence than Swedish-born women (Fernbrant, Essén, Östergren, & Cantor-Graae, 2011), thus raising the question of whether foreign-born women have a greater risk of premature death due to violence. Such violence can take place either within an interpersonal or an intimate partner context.

Premature death is defined as mortality occurring before life expectancy at birth (Sullivan & Sheffrin, 2003). Previous research shows that persons with low education and low socioeconomic status are at greater risk of premature death due to greater exposure to risk factors for disease and injuries, and low levels of health care seeking (Marmot, 2005; Wilkinson & Pickett, 2008). External causes of death–for example, injuries, suicide, and interpersonal violence–are the second most common causes of premature death among women in Sweden (National Board of Health and Welfare, Sweden, 2012).

To our knowledge, no previous research exists on the possible relationship between mortality due to interpersonal violence and foreign country of birth in Sweden. Ethnic minorities in the United States (Sanford et al., 2006) and immigrants in the United Kingdom (Marmot, Adelstein, & Bulusu, 1984) have been shown to have a greater risk of mortality due to interpersonal violence. Such findings, however, are not entirely applicable to a Swedish setting, where large-scale migration has been a more recent phenomenon and, during the past 30 years, has been dominated by asylum-seeking refugees (Bevelander, 2004). Many such refugees originate from countries where the use of violence in the family is sanctioned, due to factors such as the influence of patriarchal structures and honor culture (Abu-Ras, 2007).

Access to Swedish register data on cause of death provided the current opportunity to investigate risk of mortality due to interpersonal violence in relation to other external causes of death among women of reproductive age (per definition 15-49 years old; World Health Organization [WHO], 2010), with a focus on country of birth. Due to the fact that information concerning the identity of the perpetrator is lacking in the register, uncertainty exists regarding the gender of the perpetrator and the relationship between the perpetrator and the victim. Thus, the term interpersonal violence is used here to denote the type of violence under study, and acknowledges this potential limitation. The study intentionally targeted women of reproductive age based on the assumption that women within this age range would be at greater risk of exposure to violence due to several reasons. Such women are at risk of pregnancy, and pregnancy is a risk factor for abuse (McFarlane, Campbell, Sharps, & Watson, 2002). Furthermore, within certain cultures, women of reproductive age may potentially be at increased risk of interpersonal violence due to the value placed on female sexuality and purity as an embodiment of honor (Abu-Ras, 2007). Our previous research conducted in Sweden showed that foreign-born women from low- and middle-income countries had a higher prevalence of interpersonal violence than women from high-income countries (Fernbrant et al., 2011), thus implicating a possible role for low levels of gender equity. Women in such countries more often have lower education levels and labor market participation compared with women in high-income countries. A secondary aim, therefore, was to explore the hypothesis that the putative increased risk of mortality due to interpersonal violence among foreign-born women is related to low levels of gender equity, that is, lack of empowerment, in country of origin. Information on socioeconomic and demographic background (i.e., age, marital status, educational level, employment, and disposable income) was also examined to explore potential contributors to mortality due to interpersonal violence.

Materials and Method

Study Population

This is a Swedish register-based study of 6,124 women of reproductive age who were identified as having an external cause of death during the period 1991-2007. Data were obtained from the Swedish National Cause of Death Register, which includes all residents, whether or not the person in question was a citizen or was present in Sweden at the time of death. Asylum seekers and temporary residents in Sweden are not included in the register, in that such persons lack a personal identification number. The information recorded in the register is based on death certificates, and such certificates were either completed by the attending physician, or in cases of violent or unclear death, the coroner conducting the forensic examination. Measures included were age at death, date of death, causes of death, and country of birth. Information on country of birth, and for foreign-born women, on duration of residence in Sweden was obtained through record linkage with the Swedish Population Register by means of each person’s unique personal identification number. Additional demographic and socioeconomic information for each deceased woman regarding marital status, educational level, employment, and disposable monthly income was obtained by linkage with the Longitudinal Integration Database for Health Insurance and Labor Market Studies (LISA), at Statistics Sweden. Ethical approval for this study was not needed according to Swedish laws concerning ethical review, as all participants were deceased.

Outcome Measures

The outcome variable was defined as the injury that initiated the medical condition that resulted in death, or the circumstances involving the accident or the act of violence that caused the fatal injury. The causes of death are coded according to the ICD-9 (International Statistical Classification of Disease and Related Health Problems – ninth revision, until 1996) and the ICD-10 (International Statistical Classification of Disease and Related Health Problems – tenth revision, from 1997; WHO, 2010) in the Cause of Death Register, due to the fact that the study spanned a period (1991-2007) during which both coding systems were used. Causes of death used in this study were external causes of morbidity and mortality: transport accident injury (ICD-9: E800-E849; ICD-10: V01-V99); other external causes of accidental injury, for example, falls, accidental drowning, accidental poisoning (ICD-9: E850-E928; ICD-10: W00-X59); intentional self-harm, including suicide (ICD-9: E950-E959; ICD-10: X60-X84); assault, including homicide (ICD-9: E960-E969; ICD-10: X85-Y09); and event of undetermined intent (ICD-9: E980-E989; ICD-10: Y10-Y34). For the purpose of data analysis, the code “death due to assault” (ICD-9: E960-E969; ICD-10: X85-Y09) was operationally chosen to represent death due to interpersonal violence. Assault was defined in the register as homicide, including manslaughter, or other death by assault, by a spouse or a partner, a parent, an acquaintance or a friend, official authorities, or other specified and unspecified persons (WHO, 2010). Two categories of external cause of death were excluded–complications of medical and surgical care (n = 17) and sequelae of external causes of morbidity and mortality (n = 24)–due to few individuals in these categories. Also, causes of death not included in the study were internal causes of death, such as disease and other health-related problems.

Classification of Country of Birth by Gender Equity Index (GEI)

Among the 6,124 deceased women included in this sample, 5,129 (83.8%) were born in Sweden and 995 (16.2%) were born in a foreign country. The deceased foreign-born women in this sample originated from 85 different countries. Women from the Nordic countries constituted 38.9% of the foreign-born women, followed by Poland (9%), former Yugoslavia (8%), and Iran, Iraq, and Turkey (3% each). For the purpose of analysis, all foreign-born women were aggregated into one group, due to small numbers in some groups when divided by country. Foreign-born status was thus categorized as foreign-born or Swedish-born.

Country of birth was classified according to gender equity level based on information obtained from the GEI (Social Watch, 2012). The GEI is based on information from the United Nations Educational, Scientific, and Cultural Organization (UNESCO) and includes in total 155 countries. Of the 85 countries in the current study, 79 were included in the GEI; information for women from Barbados (n = 1), Iraq (n = 26), Libya (n = 1), Somalia (n = 7), Taiwan (n = 2), and Tunisia (n = 1) was lacking. For further information on classification of country of birth, see Table 1.

Table 1.

Countries of Birth^a of Women of Reproductive Age Who Died in Sweden 1991-2007, Grouped According to the Gender Equity Index 2012.^b

Medium gender equity (n = 5,520)	Low gender equity (n = 382)	Very low/critical gender equity (n = 184)
Denmark, Finland, Iceland, Norway, Spain, Sweden	Australia, Austria, Bolivia, Botswana, Brazil, Bulgaria, Burundi, Canada, Cape Verde, Chile, China, Colombia, Costa Rica, Croatia, Czech Republic, Ecuador, El Salvador, Estonia, France, Georgia, Germany, Ghana, Greece, Hungary, Indonesia, Ireland, Italy, Mexico, Moldova, the Netherlands, Peru, the Philippines, Poland, Portugal, Romania, Russian Federation, Singapore, Serbia, South Africa, Sri Lanka, Thailand, Uganda, Ukraine, the United Kingdom, the United States, Uruguay, Venezuela, Vietnam	Afghanistan, Arab Rep. Egypt, Bangladesh, Bosnia-Herzegovina, Dem. Rep. Congo, Eritrea, Ethiopia, Gambia, India, Islamic Rep. Iran, Japan, Kenya, Rep. Korea, Lao PDR, Lebanon, Liberia, Morocco, Pakistan, Papua New Guinea, Syrian Arab Rep., Tanzania, Turkey, Uzbekistan, Zambia, Zimbabwe

Data are missing for women from Barbados, Iraq, Libya, Taiwan, Tunisia, and Somalia (n = 38).

Data were obtained from the Gender Equity Index, Social Watch.

The assessment of gender equity level provided by the GEI is based on three components representing social, economic, and political dimensions of gender equity. The measures that comprise these three components are education, economic activity, and empowerment, and the gender gaps in these measures are presented in terms of female/male ratios. The gap in education is based on literacy rate, and enrollment rate in primary, secondary, and tertiary education. Economic activity includes estimated perceived income. Women’s political empowerment is based on the following indicators: % of women in technical positions, % of women in management and government positions, % of women in parliament, and % of women in ministerial level positions (Social Watch, 2009). The final score is the average of the three dimensions, and the categories yielded are “acceptable,” “medium,” “low,” “very low,” and “critical” levels of gender equity. There are no countries in the category “acceptable” GEI level; thus, the countries with the highest levels are in the “medium” category. For this study, the categories “very low” and “critical” gender equity were aggregated due to small numbers of women in the critical category, resulting in three gender equity categories of deceased women: “medium” gender equity, 5,520 (90.1%); “low” gender equity, 382 (6.2%); and “very low/critical” gender equity, 184 (3.0%; Table 1).

Demographic and Socioeconomic Variables

Age at death was stratified into two groups, that is, 15-29 years and 30-49 years, based on findings from previous studies showing that women under the age of 30 are more at risk of interpersonal violence than older women (van Wijk & de Brunijn, 2012; WHO, 2006). Marital status was dichotomized as unmarried/divorced/widowed or married/registered partnership/cohabitation. Duration of residence was defined as the year the deceased woman received residence in Sweden and stratified into two groups: years 1954-1979 and 1980-2007. The stratification was based on changes in immigration patterns in Sweden. Prior to 1980, immigrants in Sweden consisted primarily of labor market immigrants, and from 1980 onward, the majority of immigrants have been refugees (Bevelander, 2004). The socioeconomic measures were educational level, employment status, and disposable income at the time of death. Educational level was dichotomized as low (≤9 years or less) or middle/high education (>9 years), and employment status was defined as being employed or not. Disposable income was based on yearly household income and divided by 12 to obtain monthly disposable income. Monthly income level was dichotomized as high (≥9,000 SEK ≈ ≥1,300 US$) and low (<9,000 SEK ≈ <1,300 US$).

Data and Statistical Analysis

Statistical calculations were carried out using SPSS computer software, Version 20.0. Background characteristics, such as age, marital status, educational level, unemployment, disposable monthly income, GEI level, and duration of residence in Sweden were analyzed in relation to country of birth, using independent t tests for the analysis of numeric measures and chi-square tests for the analysis of category variables. Age-standardized mortality rates (MR) for external cause of death were calculated separately for Swedish-born and foreign-born women, using the female population of Sweden for the year 2000 as the reference group, and the following age group stratifications: 15-19 years, 20-24 years, 25-29 years, 30-34 years, 35-39 years, 40-44 years, and 45-49 years. The mortality rate ratio was expressed in terms of relative risk (RR) with 95% confidence intervals (CIs). Multivariate logistic regression analyses were used to examine mortality due to interpersonal violence in relation to foreign-born status and in relation to the GEI level of country of birth, adjusted stepwise for the potential confounders represented by age, marital status, educational level, employment, and disposable monthly income. Differences in marital status regarding foreign-born status were examined by two-way interaction terms with death due to interpersonal violence as dependent variable. Finally, bivariate logistic regression analysis was used to calculate odds ratios (OR) with 95% CIs for external cause of death in relation to GEI level of country of birth. Statistical significance was accepted at p < .05.

Results

Mortality due to external causes was examined first in relation to the underlying population at risk, here defined as female residents of Sweden between the ages of 15 and 49 during the period 1991-2007. Using population statistics for the year 2000 (Statistics Sweden) as the standard, calculation of age-standardized mortality rates for external causes showed that foreign-born women (vs. Swedish-born) had a slightly higher risk of mortality due to external causes of death in general (RR = 1.15, 95% CI = [1.08, 1.23]), and more specifically, a higher risk of mortality due to intentional self-harm, events of undetermined intent, and interpersonal violence (Table 2). Foreign-born women had a significantly lower risk of death due to transport accidents. The greatest risks of mortality due to external causes in foreign-born women were those for death due to interpersonal violence (RR = 2.97, 95% CI = [2.36, 3.74]; Table 2).

Table 2.

MR and RR With 95% CI for External Causes of Death in Swedish-Born^a and Foreign-Born^b Women Aged 15-49 Years.^c

Cause of death	Country of birth	No. of deceased women (%)	MR^d	RR^e	(95% CI)
All external causes of death	Swedish-born	5,129 (84.8)	17.4	1.00	[1.08, 1.23]
All external causes of death	Foreign-born	995 (16.2)	19.6	1.15	[1.08, 1.23]
Intentional self-harm	Swedish-born	2,382 (46.4)	8.13	1.00	[1.09, 1.32]
Intentional self-harm	Foreign-born	481(48.3)	9.49	1.20	[1.09, 1.32]
Transport accidents	Swedish-born	1,107 (21.6)	3.72	1.00	[0.69, 0.97]
Transport accidents	Foreign-born	153 (15.4)	3.08	0.82	[0.69, 0.97]
Accidents (excl transport)	Swedish-born	763 (14.9)	2.60	1.00	[0.72, 1.07]
Accidents (excl transport)	Foreign-born	113 (11.4)	2.25	0.88	[0.72, 1.07]
Event of undetermined intent	Swedish-born	661 (12.9)	2.26	1.00	[1.05, 1.51]
Event of undetermined intent	Foreign-born	140 (14.1)	2.67	1.26	[1.05, 1.51]
Interpersonal violence	Swedish-born	216 (4.2)	0.73	1.00	[2.36, 3.74
Interpersonal violence	Foreign-born	108 (10.9)	2.11	2.97	[2.36, 3.74

Note. MR = mortality rate; RR = relative risk; CI = confidence interval.

Swedish-born women are persons born in Sweden to Swedish-born parents.

Foreign-born women are persons born abroad to foreign-born parents.

Data were obtained from the Swedish National Cause of Death Registry, 1991-2007 (n = 6,124).

Age-standardized mortality rates per 100,000 person years, based on the female population (15-49 years) year 2000 (n = 1,981,853).

Relative risk: Swedish-born population (n = 1,730,792); foreign-born population (n = 291,518).

Table 3 shows the distribution of the background characteristics for foreign-born and Swedish-born women separately. Foreign-born women did not differ from Swedish-born women with regard to mean age at time of death. However, stratification of the sample by age group yielded a significantly different age distribution with fewer younger (vs. older) women among the foreign-born (χ² = 32.9, p < .001). Foreign-born women were more frequently married or cohabiting at time of death (χ² = 56.5, p < .001), had lower educational level (χ² = 10.9, p = .001), higher prevalence of unemployment (χ² = 4.2, p = .04), lower disposable income (χ² = 17.0, p < .001), and lower levels of gender equity (χ² = 3,344.5, p < .001) than the Swedish-born women (Table 3).

Table 3.

Background Characteristics for Swedish-Born^a and Foreign-Born^b Women Aged 15-49 Years^c Deceased Due to External Causes of Death.

	Swedish-born	Foreign-born	χ² test	p value^d
M age (SD)	34.29^e (10.25)	35.96 (9.23)
Age, n (%)
15-29 years	1,790 (34.9)	254 (25.5)	32.9	<.001
30-49 years	3,339 (65.1)	741 (74.5)
Total	5,129	995
Married/cohabitation, n (%)
No	3,846 (77.0)	619 (65.4)	56.5	<.001
Yes	1,152 (23.0)	327 (34.6)
Total	4,998	946
Educational level, n (%)
≤9 years	1,525 (30.9)	317 (36.6)	10.9	.001
>9 years	3,408 (69.1)	546 (63.4)
Total	4,933	861
Employment, n (%)
No	692 (13.8)	155 (16.4)	4.2	.04
Yes	4,306 (86.2)	791 (83.6)
Total	4,998	946
Disposable income^f, n (%)
<9,000	3,409 (68.2)	709 (74.9)	17.0	<.001
≥9,000	1,589 (31.8)	237 (25.1)
Total	4,998	946
GEI level^g, n (%)
Medium	5,129 (100.0)	391 (40.9)	3,344.5	<.001
Low	0	382 (39.9)
Very low/critical	0	184 (19.2)
Total	5,129	957

Note. GEI = Gender Equity Index.

Swedish-born women are persons born in Sweden by Swedish-born parents (n = 5,129).

Foreign-born women are persons born abroad to foreign-born parents (n = 995).

Data were obtained from the Swedish National Cause of Death Registry, 1991-2007 (n = 6,124).

Analysis was conducted by chi-square test.

Comparisons for age were conducted by independent t test, n.s.

Disposable monthly household income: ≥9,000 SEK ≈ ≥1,300 US$, <9,000 SEK ≈ <1,300 US$.

Gender Equity Index (GEI): groups; medium, low, very low/critical.

Potential contributing factors to the increased risk of mortality due to interpersonal violence found among foreign-born women were then further examined using multivariate logistic regression analyses (Tables 4 and 5).

Table 4.

Interpersonal Violence as a Cause of Death in Relation to Swedish Versus Foreign Country of birth.

	Crude OR	Model 1	Model 2
	OR (95% CI)	OR (95% CI)	OR (95% CI)
Foreign-born^a	2.77 [2.18, 3.53]	2.83 [2.22, 3.61]	2.45 [1.88, 3.20]
Age^b	—	1.23 [0.97, 1.56]	1.25 [0.95, 1.62]
Marital status^c	—	—	1.40 [1.07, 1.83]
Educational level^d	—	—	1.29 [1.00, 1.65]
Employment^e	—	—	1.39 [1.03, 1.88]
Disposable income^f	—	—	1.12 [0.86, 1.47]

Note. OR = odds ratio; CI = confidence interval.

Country of birth: foreign-born, Swedish-born (ref).

Age groups: 15-29 years, 30-49 years (ref).

Married/registered partnership/cohabitation, unmarried/divorced/widowed (ref).

Educational level: ≤9 years, >9 years (ref).

Employment: yes, no (ref).

Disposable monthly income: ≤9,000 SEK, >9,000 SEK (ref).

Table 5.

Interpersonal Violence as a Cause of Death in Relation to GEI Level of Country of Birth.

	Crude OR	Model 1	Model 2
	OR (95% CI)	OR (95% CI)	OR (95% CI)
Low GEI level^a	3.26 [2.35, 4.52]	3.26 [2.35, 4.53]	3.14 [2.20, 4.49]
Very low/critical GEI level	3.58 [2.32, 5.52]	3.59 [2.33, 5.55]	3.72 [2.31, 6.00]
Age^b	—	1.13 [0.89, 1.44]	1.16 [0.89, 1.52]
Marital status^c	—	—	1.29 [0.98, 1.70]
Educational level^d	—	—	1.34 [1.04, 1.72]
Employment^e	—	—	1.34 [0.98, 1.82]
Disposable income^f	—	—	1.10 [0.84, 1.45]

Note. GEI = Gender Equity Index; OR = odds ratio; CI = confidence interval.

Gender Equity Index level: low/very low/critical, medium (ref).

Age groups: 15-29 years, 30-49 years (ref).

Married/registered partnership/cohabitation, unmarried/divorced/widowed (ref).

Educational level: ≤9 years, >9 years (ref).

Employment: yes, no (ref).

Disposable monthly income: ≤9,000 SEK, >9,000 SEK (ref).

Table 4 shows the results of the multivariate logistic regression analysis of mortality due to interpersonal violence in relation to foreign-born status, with Swedish-born as the reference category. In the crude model, significantly greater odds ratios for mortality due to interpersonal violence were obtained in foreign-born women compared with Swedish-born women. These results remained significant (OR = 2.45; 95% CI = [1.88, 3.20]) after stepwise adjustment for the potential confounders represented by age, marital status, educational level, employment, and disposable household income. Also, marriage/cohabitation, low educational level, and unemployment were significantly related to mortality due to interpersonal violence (Table 4). Next, a possible interaction between foreign-born status and marital status in relation to mortality due to interpersonal violence was examined, based on the assumption that the majority of interpersonal violence occurs within the context of a relationship, and that abusive relationships may be more common among foreign-born than among Swedish-born women. A statistically significant interaction was found between foreign country of birth and marriage/cohabitation, after adjustment for all factors included in Model 2, Table 4 (p = .007, data not shown).

Among foreign-born women, analysis of duration of residence in relation to external cause of death showed that women who had migrated to Sweden in 1980 or later had significantly greater odds ratios for mortality due to interpersonal violence compared with foreign-born women who had migrated earlier (OR = 3.2; 95% CI = [2.0, 5.2]; data not shown).

The role of level of gender equity of country of birth, regardless of foreign-born status, was then examined in relation to death due to interpersonal violence (Table 5). In the crude models, significantly greater odds ratios for mortality due to interpersonal violence were obtained in women with low and very low/critical levels of gender equity compared with women with medium level of gender equity. These results remained significant for both groups (OR = 3.14; 95% CI = [2.20, 4.49]; OR = 3.72; 95% CI = [2.31, 6.00], respectively) after stepwise adjustment for the potential confounders. Low educational level was also significantly related to mortality due to interpersonal violence (Table 5). When Swedish-born women were omitted from the reference group (medium GEI level) in the above analyses, similar results, albeit slightly weaker in magnitude, were obtained (data not shown). Furthermore, the results of an interaction analysis between GEI level and marital status in relation to mortality due to interpersonal violence, with simultaneous adjustment for all factors shown in Model 2, Table 5, yielded a statistically significant interaction between low GEI level of country of birth and marriage/cohabitation (p = .03, data not shown). The strength of the interaction was even greater between very low/critical GEI level country of birth and marriage/cohabitation (p = .006, data not shown).

To examine the specificity of the increased risk of interpersonal violence associated with low and very low/critical levels of gender equity shown in Table 5, the three levels of gender equity were investigated further in relation to the remaining categories of external cause of death. Results of bivariate logistic regression analyses when performed separately for intentional self-harm, transport accidents, other accidents, and events of undetermined intent showed that women from countries with low or very low/critical levels of gender equity did not have significantly increased odds ratios for death due to any other external cause compared with women with medium GEI levels (data not shown). However, odds ratios for transport accidents (OR = 0.75; 95% CI = [0.56, 0.98]) and other accidental injuries (OR = 0.82; 95% CI = [0.48, 0.95]) were significantly decreased for women with low GEI levels compared with women with medium GEI levels (data not shown).

Discussion

This study shows that foreign-born women had significantly higher mortality rates for interpersonal violence, intentional self-harm, and events of undetermined intent, compared with Swedish-born women. By far, the largest differences in mortality rates between Swedish-born and foreign-born were those concerning interpersonal violence. In contrast, mortality due to transport accidents was significantly lower among foreign-born women compared with Swedish-born women. The association between foreign country of birth and mortality due to interpersonal violence remained significant after adjusting for potential confounders. Moreover, marriage/cohabitation, low levels of education, and unemployment were significantly associated with mortality due to interpersonal violence in the total sample. Regardless of foreign-born status, low and very low/critical GEI levels of country of birth were also significantly related to mortality due to interpersonal violence.

The most striking finding is the relationship between foreign country of birth and risk of mortality due to interpersonal violence. It may be noted that the current findings pertain to foreign-born women and not ethnic minorities, albeit previous research from the United States also shows that mortality due to interpersonal violence is twice as high among ethnic minorities compared with the native-born population (Sanford et al., 2006). Although no information was currently available concerning the perpetrator, the association between mortality due to interpersonal violence and marriage/cohabitation suggests that such violence may have occurred within the context of a relationship and thus may represent mortality due to intimate partner violence, as do the significant interactions found between marriage/cohabitation and both foreign country of birth, and low/very low/critical GEI levels, respectively. According to a Swedish report by the Swedish National Council for Crime Prevention (2007), approximately 35 women are killed due to interpersonal violence in Sweden every year. The majority of the women (85%) during the reported period (1990-2004) knew their perpetrators, and in half of the cases, the perpetrator was an intimate partner. Mortality due to intimate partner violence most often (78%) takes place in the woman’s home (Campbell et al., 2003; Sanford et al., 2006). Also, our recent study showed that foreign-born women had a higher exposure to violence in the home (Fernbrant et al., 2011). Living in an abusive relationship has been shown to be the primary risk factor for mortality due to intimate partner violence (Aldridge & Browne, 2003; Campbell et al., 2003). Previous studies show that migration to a more gender-equal society may affect the internal balance of the relationship, and may initiate or exacerbate intimate partner violence due to the man’s loss of status and social networks, together with the woman’s increased status and increased access to education and the labor market (Jewkes, 2002; Jin & Keat, 2009).

Several findings indicate that lack of empowerment, as hypothesized, may be an important risk factor for mortality due to interpersonal violence. Lower educational level and unemployment (Table 4) as well as low to critically low GEI levels of country of birth (Table 5) are factors that may compromise a woman’s ability to cope with exposure to interpersonal violence, especially within the context of a relationship. It may be noted that educational level and unemployment pertain to the woman’s own circumstances rather than to her country of birth, and thus these measures may provide a more precise indication of the woman’s possible lack of empowerment than the GEI level of country of birth. Also, women having the opportunity to immigrate to Sweden may have had a more empowered position in their country of birth than the index suggests. However, the current findings are in line with previous research showing that uneducated and less empowered women are at greater risk of interpersonal violence due to patriarchal structures sanctioning women’s subordination to men (Abu-Ras, 2007). Furthermore, although Sweden is one of the countries with the highest GEI level in the world, adjustment to these norms by those who are foreign-born might take time, and the norms of the country of origin may continue to be more of a predominant influence (Marmot et al., 1984). Thus, the shorter length of residence associated with mortality due to interpersonal violence among foreign-born women suggests that these women may not yet have acquired sufficient knowledge of Swedish legislation or Swedish gender equity norms. It is also a challenge for Swedish society to ensure that these women receive adequate information and support during their integration process. An alternative explanation may be that the deceased women’s country of birth differs between the two periods (before and after 1980). During the 1970s and the 1980s, the immigration pattern in Sweden changed from primarily labor immigration to primarily immigration on humanitarian grounds (Bevelander, 2004). Thus, the women who arrived after 1980 consist primarily of refugees from countries with a lower GEI level, such as Ethiopia, former Yugoslavia, Iran, and Lebanon (Social Watch, 2012).

It may be noted that the current results represent the most extreme outcome of interpersonal violence, that is, death. However, the extent to which this most extreme outcome was preceded by a prolonged period of threat and violence is unknown. Nevertheless, the results are in line with our previous findings showing that foreign-born women from middle- and low-income countries have a higher prevalence of interpersonal violence per se compared with women from high-income countries (Fernbrant et al., 2011), results that may be attributable to a more tolerant attitude toward violence against women in such countries. A recent study on women in sub-Saharan Africa showed that women in a more subordinated position relative to their husbands, due to polygamy and/or the husbands’ superior level of education, had a higher tolerance for intimate partner violence compared with other women (Uthman, Lawoko, & Moradi, 2010). Previous research also shows that countries with low gender equity have higher prevalence of intimate partner violence, unwanted pregnancies due to intimate partner violence (Pallitto, Campbell, & O’Campo, 2005), and mortality due to violence against women (Palma-Solis, Vives-Cases, & Alvares-Dardet, 2008). The current results suggest that gender equity is a protective factor for violence against women. In countries with low to critically low gender equity, legislation promoting equal rights for men and women, as well as the prohibition of violence against women, including intimate partner violence and marital rape, is an important development area for increased gender equity. The implementation of women’s rights varies globally and is closely linked to women’s empowerment as well as the prevalence of violence against women (Pallitto et al., 2005). Violence against women may be further sanctioned by certain immigrant groups, due to fewer rights for women in the country of birth and traditional values such as honor culture (Abu-Ras, 2007). In an honor culture, the woman’s sexuality is owned and governed by her family, such that if she, for example, has pre-marital sex and thereby brings shame on her family, she needs to be punished to restore the family’s honor, with the most extreme outcome being honor killing (Kulwicki & Miller, 1999). Iraq, Iran, and Turkey are among those countries having especially strong honor cultures (Kandiyoti, 1988), and women from these countries were among the largest groups of deceased women with mortality due to interpersonal violence in the current study.

In the current study, the household’s disposable income was not related to mortality due to interpersonal violence. However, in our recent study, we found that women’s exposure to interpersonal violence was associated with being socioeconomically disadvantaged (Fernbrant et al., 2011). Also, previous studies report that social isolation (Raj & Silverman, 2003) and poverty (Jewkes, 2002; Krishnan et al., 2010) are risk factors for interpersonal violence among foreign-born women. The reasons for the current discrepant findings are unclear; however, disposable household income may not necessarily represent the woman’s actual access to resources, because she can be disadvantaged and/or impoverished within the household.

It may be noted that foreign-born women also had increased risk of mortality due to intentional self-harm compared with Swedish-born women (Table 2). Previous findings indicate that foreign-born women in Sweden are at risk of suicide (Westman, Sundquist, Johansson, Johansson, & Sundquist, 2006), as well as serious mental illness (Hjern, Wicks, & Dalman, 2004). Although exposure to interpersonal violence might lead to intentional self-harm, the possible contribution of such exposure to the current findings is entirely unknown. Foreign-born women also had increased risk of mortality due to events of undetermined intent compared with Swedish-born women (Table 2). Such findings are difficult to interpret because they de facto represent cases where sufficient information was lacking for a more specific categorization.

Methodological Considerations

This is the first study to examine the possible relationship between mortality due to interpersonal violence and foreign country of birth in Sweden. One of the strengths of our study is the use of data obtained from the Swedish Cause of Death Register, which is a comprehensive register with national coverage. This register holds an internationally high standard, including all deceased persons registered in Sweden, and uses the same classification system as the WHO (Johansson, Björkenstam, & Westerling, 2009; Nyström et al., 1995). Previous research on the accuracy of the register has demonstrated that this national database shows good validity (Nyström et al., 1995). A recent study on mortality rates among female immigrants also utilizes information from the same register (Esscher, Haglund, Högberg, & Essén, 2013).

An additional strength in the current study is the use of the GEI to further examine factors related to mortality due to interpersonal violence. The GEI is a prominent and highly valid system for evaluating both socioeconomic and political aspects of gender equality in any given country (Mills, 2010), and is therefore more relevant for examining the potential role of empowerment in relation to violence against women rather than any single indicator such as a country’s “income ranking.” The GEI has a broad coverage of countries and includes all but six countries in the current study. However, it is a comparative index on a national level; thus, countries with a high score may still have inequity issues (Mills, 2010; Permanyer, 2010). It may be noted that odds ratios for mortality due to interpersonal violence were somewhat greater when GEI level was used as the independent variable rather than foreign-born status. Nevertheless, caution may be warranted in the interpretation of the relative importance of gender equity versus foreign-born status, in that these measures do overlap as both are based on country of birth. Also, it is entirely unknown whether the deceased woman herself would be characterized by a low GEI level, regardless of the level ascribed to her country of origin. However, the highly significant relationship found here between low GEI level and mortality due to interpersonal violence indirectly supports the notion of a correspondence between the woman’s own level and the GEI level of her country of birth. Although information on GEI was lacking for Iraq, according to other GEIs Iraq has a low level of gender equity (Hausmann, Tyson, & Zahidi, 2011). Thus, the current results may potentially represent an underestimate of the relationship between mortality due to interpersonal violence and low gender equity.

An important limitation of the study with regard to mortality due to interpersonal violence is that the identity of the perpetrator is unknown; consequently, the relationship between the perpetrator and the victim is entirely unknown. Therefore, it is not possible to establish whether the perpetrator of mortality due to interpersonal violence was an intimate partner, a relative, casual acquaintance, or person unknown to the victim. The extent to which the violence that took place was in fact carried out by an intimate partner cannot be determined. Also, because the perpetrator is unknown, no information is available concerning the possible GEI level of the perpetrator.

A possible limitation is the risk of misclassification errors with regard to the accuracy with which causes of death are reported, due to insufficient information of the fatal event. There may be an underreporting of interpersonal violence due to surrounding circumstances, as opposed to cases concerning fall accidents and other accidents in the home. However, when a criminal offense is suspected, as for mortality due to interpersonal violence, it is reported to the police, and a forensic examination is conducted by the coroner who decides the cause of death. Mortality due to interpersonal violence is followed by a police investigation that may have different outcomes; for example, if a criminal act cannot be proven, the case is dismissed. Previous research shows that most cases of interpersonal violence are never reported and approximately 80% of all reported cases are dismissed, although violence against women has increased during the last 10 years in Sweden, due to, among other factors, better reporting (Ingemann-Hansen, Brink, Sabroe, Sørensen, & Charles, 2008; Stene, Ormstad, & Schei, 2010; Swedish National Council for Crime Prevention, 2012). It should be noted, however, that the outcome of the police investigations of the deceased women in the current study does not affect the results of this study, because cause of death in the register is a medical assessment.

Although the Swedish Cause of Death Register has a high standard and accuracy, and no systematic bias has been found with regard to the coding of causes of death (Nyström et al., 1995), it is unknown whether mortality among foreign-born women due to external causes is more likely to be reported as due to interpersonal violence than to other causes. A potential limitation is the extent to which the foreign-born population in Sweden may be overestimated, due to the fact that some foreign-born do not report their emigration to the registration authorities when they leave Sweden (Weitoft, Gullberg, Hjern, & Rosen, 1999). An overestimation of the foreign-born population in Sweden would result in an underestimation of the population-based risk of mortality due to interpersonal violence associated with being foreign-born. The net effect of these two sources of bias–that is, differential reporting of mortality due to interpersonal violence and a possible overestimation of the foreign-born population of Sweden–cannot be determined. Also, asylum seekers are not included in the register; these were 36,207 persons in 2007. Furthermore, some foreign-born women included in the study population may have been international adoptees. Some 17,000 foreign-born girls were adopted by Swedish before 1993; they were primarily from South Korea (high-income country) Colombia, India, and Sri Lanka (all middle-income countries) (MIA [Myndigheten för internationella adoptionsfrågor], Swedish Intercountry Adoptions Authority, 2009). Also, the study was limited to women 15-49 years of age to be able to examine women of reproductive age due to previous findings indicating that being pregnant (McFarlane et al., 2002), having a child living in the home who is not the biological child of the perpetrator, and separating or talking about leaving the perpetrator (Tjaden & Thoennes, 1998) are all risk factors for mortality due to interpersonal violence. The chosen age group may be a potential weakness in that women above age 50 were excluded.

Finally, two classification systems–ICD 9 (until 1996) and ICD 10 (from 1997)– have been used during the time period of the current study. The ICD 10 is an updated version of ICD 9 and contains a greater detail. In the current study, 60% of the women died after 1996 and are classified according to ICD 10. With regard to the more specific death causes, some disparities may exist between the two coding periods; however, due to the broader causes of death utilized in this study, such as traffic accidents and interpersonal violence, the use of two classification systems is likely to have had a very minimal effect on the data.

Implications and Future Research

The increased risk among foreign-born women for the most extreme outcome of interpersonal violence, namely, physical assault leading to death, as reported in the current study indicates a public health problem of great importance. The tragedy is not only in the loss of a life, but also in the considerable cost to society due to the premature death of women in the work force and/or other productive activities. The current study highlights the need for the development of prevention strategies to enable the detection of interpersonal violence at an early stage such that women’s lives can be saved.

Because mortality due to interpersonal violence is, to our knowledge, relatively unexplored in Sweden, it is important to gain further knowledge about the potential determinants and factors that may be targeted for prevention. Gender equity may also be a challenge for policy makers in Sweden in light of the increasing influx of foreign-born women who, despite residence in Sweden, may be especially vulnerable to exposure to interpersonal violence, and thus in need of support and information regarding their rights. Further studies of mortality due to interpersonal violence may be potentially relevant for a deeper understanding of the broader underlying problem of violence against women as well as the nature of the perpetrators.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a grant from the Swedish Council for Working Life and Social Research (FAS 2007-2026) and by the European Refugee Fund.

Author Biographies

Cecilia Fernbrant, MSc, PhD, is a researcher at Social Medicine and Global Health in the Department of Clinical Sciences Malmö at Lund University, Sweden.

Birgitta Essén, MD, PhD, is a professor in the Department of Women’s and Children’s Health, International and Maternal Health at Uppsala University, Sweden.

Annika Esscher, MD, PhD, is a senior consultant at Uppsala University Hospital, Sweden.

Per-Olof Östergren, MD, PhD, is a professor in the Division of Social Medicine and Global Health in the Department of Clinical Sciences Malmö at Lund University, Sweden.

Elizabeth Cantor-Graae, MA, PhD, is an associate professor in the Division of Social Medicine and Global Health in the Department of Clinical Sciences Malmö at Lund University, Sweden.

References

Abu-Ras

(2007). Cultural beliefs and service utilization by battered Arab immigrant women. Violence Against Women, 13, 1002-1028.

Aldridge

M. L.

Browne

K. D.

(2003). Perpetrators of spousal homicide: A review. Trauma Violence Abuse, 4, 265-276.

Bevelander

(2004). Immigration patterns, economic integration and residential segregation: Sweden in the late 20th century. Malmö, Sweden: International Migration and Ethnic Relations (IMER), Malmö University.

Campbell

J. C.

(2002). Health consequences of intimate partner violence. Lancet, 359, 1331-1336.

Campbell

J. C.

Glass

Sharps

P. W.

Laughon

Bloom

(2007). Intimate partner homicide: Review and implications of research and policy. Trauma, Violence, & Abuse, 8, 246-269.

Campbell

J. C.

Webster

Koziol-McLain

Block

Campbell

Curry

M. A.

. (2003). Risk factors for femicide in abusive relationships: Results from a multisite case control study. Research and Practice, 93, 1089-1097.

Chang

D. F.

Shen

B. J.

Takeuchi

D. T.

(2009). Prevalence and demographic correlates of intimate partner violence in Asian Americans. International Journal of Law and Psychiatry, 32, 167-175.

Ellsberg

Jansen

H. A. F. M.

Heise

Watts

C. H.

Garcia-Moreno

(2008). Intimate partner violence and women’s physical and mental health in the WHO multi-country study on women’s health and domestic violence: An observational study. Lancet, 371, 1165-1172.

Esscher

Haglund

Högberg

Essén

(2013). Excess mortality in women of reproductive age from low-income countries: A Swedish National Register study. European Journal of Public Health, 23, 274-279.

10.

Fernbrant

Essén

Östergren

P. O.

Cantor-Graae

(2011). Perceived threat of violence and exposure to physical violence against foreign-born women: A Swedish population-based study. Women’s Health Issues, 21, 206-213.

11.

Flood

Pease

(2009). Factors influencing attitudes to violence against women. Trauma, Violence, & Abuse, 10, 125-142.

12.

Hausmann

Tyson

Zahidi

(2011). The global gender gap report 2011. Geneva, Switzerland: World Economic Forum. Retrieved from http://www3.weforum.org/docs/WEF_GenderGap_Report_2011.pdf

13.

Hjern

Wicks

Dalman

(2004). Social adversity contributes to high morbidity in psychoses in immigrants—A national cohort study in two generations of Swedish residents. Psychological Medicine, 34, 1025-1033.

14.

Ingemann-Hansen

Brink

Sabroe

Sørensen

Charles

A. V.

(2008). Legal aspects of sexual violence—Does forensic evidence make a difference? Forensic Science International, 180, 98-104.

15.

Jewkes

(2002). Intimate partner violence: Causes and prevention. Lancet, 359, 1423-1429.

16.

Jin

Keat

J. E.

(2009). The effects of change in spousal power on intimate partner violence among Chinese immigrants. Journal of Interpersonal Violence, 10(2), 1-16.

17.

Johansson

L. A.

Björkenstam

Westerling

(2009). Unexplained differences between hospital and mortality data indicated mistakes in death certification: An investigation of 1,094 deaths in Sweden during 1995. Journal of Clinical Epidemiology, 62, 1202-1209.

18.

Kandiyoti

(1988). Bargaining with patriarchy. Gender & Society, 2, 274-290.

19.

Krantz

(2002). Violence against women: A global public health issue! Journal of Epidemiology & Community Health, 56, 242-243.

20.

Krantz

Garcia-Moreno

(2005). Violence against women. Journal of Epidemiology & Community Health, 59, 818-821.

21.

Krishnan

Rocca

C. H.

Hubbard

A. E.

Subbiah

Edmeades

Padian

N. S.

(2010). Do changes in spousal employment status lead to domestic violence? Insights from a prospective study in Bangalore, India. Social Science & Medicine, 70, 136-143.

22.

Kulwicki

A. D.

Miller

(1999). Domestic violence in the Arab American population: Transforming environmental conditions through community education. Issues in Mental Health Nursing, 20, 199-215.

23.

Marmot

(2005). Social determinants of health inequalities. Lancet, 365, 1099-1104.

24.

Marmot

Adelstein

A. M.

Bulusu

(1984). Lessons from the study of immigrant mortality. Lancet, 30, 1455-1457.

25.

McFarlane

Campbell

J. C.

Sharps

Watson

(2002). Abuse during pregnancy and femicide: Urgent implications for women’s health. Obstetrics & Gynecology, 100(1), 27-36.

26.

MIA [Myndigheten för internationella adoptionsfrågor], Swedish Intercountry Adoptions Authority. (2009). Arrived non-Nordic adoptive children (aged 0-10 years) [Anlända utomnordiska adoptivbarn (ålder 0-10 år)]. Retrieved from http://www.mia.eu/Documents/Statistik/scb/grund2009.pdf

27.

Mills

(2010). Gender roles, gender (in) equality and fertility: An empirical test of five gender equity indices. Canadian Studies in Population, 37, 445-474.

28.

National Board of Health and Welfare, Sweden. (2012). Statistik databas för dödsorsaker [Statistical database of causes of death]. Retrieved from http://192.137.163.49/sdb/dor/val.aspx

29.

Nyström

Larsson

L. G.

Rutqvist

L. E.

Lindgren

Lindqvist

Ryden

. (1995). Determination of cause of death among breast cancer cases in the Swedish randomized mammography screening trials: A comparison between official statistics and validation by an endpoint committee. Acta Oncologica, 34, 145-152.

30.

Pallitto

C. C.

Campbell

J. C.

O’Campo

(2005). Is intimate partner violence associated with unintended pregnancy? A review of the literature. Trauma, Violence, & Abuse, 6, 217-235.

31.

Palma-Solis

Vives-Cases

Alvares-Dardet

(2008). Gender progress and government expenditure as determinants of femicide. Annals of Epidemiology, 18, 322-329.

32.

Permanyer

(2010). The measurement of multidimensional gender inequality: Continuing the debate. Social Indictors Research, 95, 181-198.

33.

Raj

Silverman

J. G.

(2003). Immigrant south Asian women at greater risk for injury from intimate partner violence. American Journal of Public Health, 93, 435-437.

34.

Sanford

Marshall

S. W.

Martin

S. L.

Coyne-Beasley

Waller

A. E.

Cook

P. J.

. (2006). Deaths from violence in North Carolina, 2004: How deaths differ in females and males. Injury Prevention, 12, ii10-ii16.

35.

Social Watch. (2009). Gender Equity Index 2009. The three gaps. Retrieved from http://www.socialwatch.org/sites/default/files/GEI2009_eng_0.pdf

36.

Social Watch. (2012). Social Watch report 2012. Sustainable development: The right to a future. Gender Equity Index 2012. Retrieved from http://www.socialwatch.org/sites/default/files/SW_Overview2012_eng.pdf

37.

Stene

L. E.

Ormstad

Schei

(2010). Implementation of medical examination and forensic analyses in the investigation of sexual assaults against adult women: A retrospective study of police files and medical journals. Forensic Science International, 199, 79-84.

38.

Sullivan

Sheffrin

S. M.

(2003). Economics: Principles in action (ISBN 0-13-063085-3). Upper Saddle River, NJ: Prentice Hall.

39.

Swedish National Council for Crime Prevention. (2007). Utvecklingen av dödligt våld mot kvinnor i nära relationer. Rapport 2007:6 [The development of fatal intimate partner violence against women. Report 2007:6]. Stockholm, Sweden: Author.

40.

Swedish National Council for Crime Prevention. (2012). Confirmed cases of lethal violence. Statistics for 2011 [Konstaterade fall av dödligt våld. Statistik för 2011]. Stockholm, Sweden: Author.

41.

Tjaden

Thoennes

(1998). Prevalence, incidence and consequences of violence against women: Findings from the National Violence Against Women Survey (NCJ 1728370). Washington, DC: National Institute of Justice.

42.

Uthman

O. A.

Lawoko

Moradi

(2010). The role of individual, community and societal gender inequality in forming women’s attitudes toward intimate-partner violence against women: A multilevel analysis. World Health & Population, 12(2), 5-17.

43.

van Wijk

N. P.

de Brunijn

J. G.

(2012). Risk factors for domestic violence in Curacao. Journal of Interpersonal Violence, 27, 3032-3053.

44.

Weitoft

G. R.

Gullberg

Hjern

Rosen

(1999). Mortality statistics in immigrant research: Method for adjusting underestimation of mortality. International Journal of Epidemiology, 28, 756-763.

45.

Westman

Sundquist

Johansson

L. M.

Johansson

S. E.

Sundquist

(2006). Country of birth and suicide: A follow-up study of a national cohort in Sweden. Archives of Suicide Research, 10, 239-248.

46.

Wilkinson

R. G.

Pickett

K. E.

(2008). Income inequality and socioeconomic gradients in mortality. American Journal of Public Health, 98, 699-704.

47.

World Health Organization. (2006). WHO multi-country study on women’s health and domestic violence against women. Geneva, Switzerland: Author.

48.

World Health Organization. (2010). International statistical classification of diseases and related health problems (10th Revision). Geneva, Switzerland: Author.