OUP user menu

Intimate partner violence and sexual health outcomes: a population-based study among 16–44-year-old women in Estonia

Made Laanpere, Inge Ringmets, Kai Part, Helle Karro
DOI: http://dx.doi.org/10.1093/eurpub/cks144 688-693 First published online: 23 October 2012


Background: Violence against women as a public issue and health burden has only recently been acknowledged in Estonia. The aim of this study was to outline the prevalence and to explore the associations of intimate partner violence (IPV) and selected sexual health outcomes. Methods: Data from the population based cross-sectional survey among representative sample of 16–44-year-old non-pregnant women (n = 1966) were analysed using multivariate logistic regression. Results: Of 1966 respondents, 362 (18.4%) reported IPV during 12 months preceding the survey. Physical IPV was reported by 339 (17.2%), sexual by 80 (4.1%) and both physical and sexual by 35 (1.8%) of respondents. After controlling for significant confounding socio-demographic factors, exposure to IPV was found to be associated with contraception non-use [adjusted odds ratio (AOR) = 2.02, 95% confidence interval (CI) 1.44–2.82] or the use of unreliable contraceptive methods (AOR = 1.54, CI 1.16–2.04) during the most recent sexual intercourse, having never used a condom (AOR = 1.53, CI 1.12–2.10), repeat induced abortion (AOR = 1.72, CI 1.24–2.37), lifetime sexually transmitted infections (AOR = 2.05, CI 1.56–2.68) and dyspareunia (AOR = 2.14, CI 1.65–2.77). Conclusion: The exposure of IPV was an important contributor to sexual risk behaviour and adverse sexual health outcomes among women of reproductive age in Estonia. Any strategy to promote sexual health should include prevention of IPV and other forms of violence against women with the strengthening of women’s sexual and reproductive rights.


The most common form of violence against women is intimate partner violence (IPV), defined by the World Health Organisation as use of power or force, threatened or actual, by a husband or intimate male partner that results in injury, death or psychological harm.1 In Estonia, similarly to other East European countries, despite the fundamental political and economic changes that have taken place during the past two decades, IPV has only recently been acknowledged and discussed publicly.2,3 At the same time, IPV is still considered a private issue not only among the general population but also by governmental structures.2,3 Most of the programmatic efforts so far adopted are primarily the funding of victim support initiatives but not the development of legislative instruments. It was not until 2001 that the first violence-related survey was conducted in Estonia among 535 women. This showed that one-fifth of women aged 15–74 years were exposed to violence by someone during the previous year.3

Although physical injuries are the most visible signs of violence, data suggest that the relationship between IPV and women’s health may go far beyond direct physical consequences of violence.4–10 Sexual health can be even more seriously impaired by IPV.5 There is consistent evidence that IPV is significantly associated with women’s sexual risk behaviour leading to less contraceptive use, unintended pregnancies, human immunodeficiency virus/sexually transmitted infections (HIV/STIs), sexual dysfunction and gynaecological complaints.4–10

Estonia has a relatively high abortion rate (21.6 abortions per 1000 women of fertile age in 2010)11 and the highest reported prevalence of HIV (1.3%) in the European Union.12 In the absence of reliable crime statistics on IPV victims, violence screening in health care system and lack of research data, it is even more important to understand the risk factors for and consequences of IPV, including those related to sexual and reproductive health in Estonia.

The aim of this study was to estimate the prevalence of IPV during 12 months preceding the survey and explore the associations between IPV exposure and selected sexual health outcomes from a population-based sample of Estonian women.



Data for the present study were drawn from a population-based cross-sectional Estonian Women’s Health Survey conducted in 2004–05. In March 2004, a computer-generalized age-stratified random sample was obtained from the Estonian Population Registry, which covers all residents of Estonia (territory: 45 200 sq km, 1.3 million population). Mailing address and birth data were abstracted from the registry. Total sample size was 5190 women (by age groups: 2515 respondents aged 16–25 years, 1240 aged 26–35 years and 1435 aged 36–44 years).Questionnaires, 3472 in the Estonian and 1718 in the Russian language, were posted with an information letter and prepaid reply envelopes. For non-responders (n = 3113), a second mailing of the study materials was sent after 10–12 weeks. Of the 5087 eligible respondents, 2335 did not return the questionnaire, 13 declined to participate and 4 questionnaires were completed unsatisfactorily. Total response rate was 53.8% (n = 2735). Study participation was anonymous with a study code being assigned to each study participant to prevent personal identification. The respondent sent the study code back to the research team separately from the questionnaire. We used information from the self-administered questionnaire, in which women were asked about their socio-demographic background, reproductive history, use of and satisfaction with sexual health care services, contraception, future plans for having children and self-rated health. The study was not specially designed for assessing violence, but a violence module with clearly worded questions about the respondent’s exposure of specific violent acts was included. The survey has been described in more detail in the study report.13 Ethical approval was granted by the Ethics Review Committee on Human Research of the University of Tartu, Estonia (107/65 26.08.2002).

In this study, among a sample of 2533 non-pregnant women aged 16–44 years, 2487 (98.2%) completed the violence module, of them, 2207 (87.1%) had sexual intercourse experience. Among the respondents with sexual intercourse experience, 603 (27.3%) answered affirmatively to at least one question referring to violence exposure during the preceding 12 months. We excluded from further analysis 241 (10.9%) respondents who had reported violence perpetrated by someone other than their intimate partner. Final total sample size was 1966.

Intimate partner violence

In this study, IPV was defined as physical or sexual violence by an intimate partner during the last 12 months before the survey. We categorized responses indicating either physical or sexual violence or both in one group because the frequent overlap of these types of violence makes it difficult to attribute specific sexual risk behaviour or adverse sexual health outcome to one specific type of violence. Intimate partner was defined as current/former dating partner or married/cohabiting partner. The violence module contained seven questions. A respondent was categorized as having exposure to physical violence if the answer to one of the following questions was affirmative: that she had been exposed to threats of physical violence; pushed, shaken or had something thrown at her; hit with something that caused/could have caused physical injury; threatened with a knife/gun/or other object by her partner. A respondent was categorized as having exposure to sexual violence if she stated that, against her will, she had been physically forced to have sexual intercourse, threatened/frightened into having sexual intercourse or forced to participate in other sexual acts. We created an aggregate measure of violence, which was coded positive for respondents who reported either physical or sexual violence or both (IPV+) and negative (IPV−) if no exposure during the last year was reported. The survey had additional questions concerning physical injuries after IPV (pain/bruises/wounds/broken bones/injuries that needed medical attention).

Socio-demographic variables and reproductive history

We describe the selected background characteristics, all self-reported, of the three groups in terms of age (16–24, 25–34 and 35–44 years), marital status (married/cohabiting, divorced/separated and single), completed educational level (basic/less, secondary/secondary special and university) and employment (employed, student and unemployed/housewife/retired). The ethnicity of respondent was based on self-reported native language, and participants were categorized as Estonian and non-Estonian. For 97.3% of non-Estonian respondents, Russian was their native language. Economic subsistence, which was assessed with the question ‘Do you have difficulties with paying bills for housing, electricity, heating etc?’, was divided into two: often/always and never/sometimes. Respondents were asked to report a lifetime number of births and lifetime number of induced abortions.

Sexual health outcomes

The following sexual health outcomes were used: contraceptive method used during the most recent sexual intercourse, condom use ever, repeat induced abortions, lifetime HIV/STIs and dyspareunia. From all eligible respondents, 95% had their most recent sexual intercourse during the preceding year. When answering about the contraceptive method they had used, respondents could choose a contraceptive method from a list of different methods. Based on available data about the effectiveness of contraceptive methods,14 all respondents who reported the use of rhythm method, withdrawal, spermicide or vaginal douching were considered to have used an unreliable contraceptive method. Those who had used hormonal contraceptive pill, patch, injectable hormonal contraceptive, emergency contraceptive pill, condom, intrauterine device (IUD) or had undergone sterilization were considered to have used a reliable contraceptive method. Although women were allowed to indicate more than one contraceptive method, for the purposes of this analysis, the categorization of contraceptive method was based on the most reliable method reported. Category ‘contraceptive non-use’ included women who did not use contraception during the most recent sexual intercourse. Respondents who reported more than one lifetime induced abortion were categorized as having a repeat induced abortion. Those reporting having had HIV/STIs during their lifetime were asked to choose from the following list of infections: genital herpes, condylomas, trichomoniasis, chlamydiosis, gonorrhoea, syphilis and HIV/acquired immune deficiency syndrome. Dyspareunia was assessed by the question: ‘How often have you experienced pain/discomfort during/after sexual intercourse in your genitals (vaginal introitus, vagina) or abdomen?’. The variable was dichotomized into categories: never/rarely and always/often. The respondents were also asked about the most probable decision in case they get pregnant, and the answer options were as follows: would have a baby, would have an abortion and do not know (including missing answers).


Descriptive statistics, such as frequencies and relative frequencies, were used to describe socio-demographic and reproductive history characteristics of study respondents and prevalence of violence. Differences between respondents having and having not been exposed to IPV were tested by chi-square test. Significance level of P = 0.05 was used. Socio-demographic factors that were statistically significantly related to IPV exposure (age, education, economic subsistence and ethnicity) were entered into five multivariate logistic regression models exploring associations between IPV exposure and selected sexual health outcomes. Associations are presented as crude odds ratios (ORs) and adjusted ORs (AORs) with 95% confidence intervals (95% CIs). Only women who had answered all the questions used in the logistic regression model were taken into account when exploring these associations. The data were analysed using the statistical package Stata 10.


Self-reported socio-demographic and reproductive history characteristics by IPV exposure are presented in table 1. Factors associated with IPV exposure were younger age, lower educational level, non-Estonian ethnicity and economical subsistence. In addition, in the IPV exposure group, more women had had induced abortions, but the number of births was similar in the two study groups.

View this table:
Table 1

Selected self-reported socio-demographic and reproductive history characteristics by IPV exposure during the last 12 monthsa

Socio-demographic and reproductive history characteristicsTotal (n = 1966)IPV+ (n = 362)IPV− (n = 1604)
n (%)n (%)
Age (years)*
    16–24755155 (42.8)600 (37.4)
    25–34585112 (30.9)473 (29.5)
    35–4462695 (26.2)531 (33.1)
Marital status
    Married/cohabiting1280248 (68.5)1032 (64.3)
    Divorced/separated18939 (10.8)150 (9.4)
    Single49375 (20.7)418 (26.1)
    Missing404 (0.3)
    University44458 (16.0)386 (24.1)
    Secondary/secondary special1317245 (67.8)1072 (66.8)
    Basic/less19558 (16.0)137 (8.5)
    Missing101 (0.3)9 (0.6)
    Estonian1391216 (59.7)1175 (73.3)
    Non-Estonian570145 (40.1)425 (26.5)
    Missing51 (0.3)4 (0.3)
Employment status
    Employed1026184 (50.8)842 (52.5)
    Student53494 (26.0)440 (27.4)
    Other40284 (23.2)318 (19.8)
    Missing404 (0.3)
Difficulties with paying bills**
    Often/always329101 (27.9)228 (14.2)
    Never/sometimes1588252 (69.6)1336 (83.3)
    Missing499 (2.5)40 (2.5)
Number of births
    0794133 (36.7)661 (41.2)
    1505106 (29.3)399 (24.9)
    247288 (24.3)384 (23.9)
    ≥319535 (9.7)160 (10.0)
Number of abortions*
    01212212 (58.6)1000 (62.3)
    136058 (16.0)302 (18.8)
    221341 (11.3)172 (10.7)
    ≥318151 (14.1)130 (8.1)
  • a: Chi-square test was applied only for non-missing values.

  • *P < 0.05, **P < 0.001.

Of 1966 respondents, 362 (18.4%) women reported physical or sexual IPV or both during the last 12 months. Table 2 shows different types of IPV. The most common type of IPV was physical violence (17.2% of all respondents and 93.6% of IPV survivors), followed by sexual violence (4.1 and 22.1%, respectively) and 1.8% of all respondents and 9.7% of IPV survivors had been exposed to both types of violence (n = 35).

View this table:
Table 2

IPV during the last 12 months grouped according to type of violence

Type of IPV violencenPercentage of all respondents (n = 1966)Percentage of IPV survivors (n = 362)
Physical violence33917.293.6
    Threats of physical violence1598.143.9
    Pushed, shaken, had something thrown28014.277.3
    Hit with something that caused/could have caused physical injury1035.228.5
    Threatened with a knife/gun/ other object201.05.5
Sexual violence804.122.1
    Physically forced to have sexual intercourse341.79.4
    Threatened/frightened into having sexual intercourse211.15.8
    Forced to participate in other sexual acts422.111.6

Of all IPV survivors, 209 (64.1%) reported pain or physical injuries, 30 (8.3%) of them wounds or broken bones or injuries that needed medical attention.

A significant difference was found between the two study groups regarding intentions in the hypothetical case of pregnancy. More IPV-exposed women compared with non-exposed ones reported that they would opt for an abortion (26.2 vs. 22.1%, P < 0.001) or expressed ambivalence about the decision/not respond (34.8 vs. 26.3%, P < 0.001). The proportion of respondents who did not use contraception because they were planning a pregnancy was higher among IPV-exposed group compared with the respondents with no IPV exposure (27.9 vs. 20.3%, P < 0.001). The vast majority of the respondents reported using some contraceptive method during their most recent sexual intercourse (table 3). The pattern of contraceptive use differed by IPV exposure—reported use of hormonal contraceptive methods and condom was lower, and contraceptive non-use and the use of unreliable contraceptive methods were higher among respondents exposed to IPV compared with those who not been exposed to IPV.

View this table:
Table 3

Contraceptive method during the most recent sexual intercourse by IPV exposure during the last 12 monthsa

Contraceptive methodsTotal (n = 1966)IPV+ (n = 362)IPV− (n = 1604)
n (%)n (%)n (%)
Hormonal methods486 (24.7)71 (19.6)415 (25.9)
IUD300 (15.3)58 (16)242 (15.1)
Condom463 (23.6)64 (17.7)399 (24.9)
Sterilization15 (0.8)1 (0.3)14 (0.9)
Emergency contraception21 (1.2)2 (0.6)19 (1.2)
Unreliable methods427 (21.7)101 (27.9)326 (20.3)
None113 (5.8)31 (8.6)82 (5.1)
Missing8 (0.4)08 (0.5)
  • a: Respondents could report several contraceptive methods; classification of method effectiveness was based on the most reliable method reported.

Table 4 shows selected sexual health outcomes in the two groups of respondents by IPV exposure. Multivariate logistic analysis showed that IPV during the last 12 months, after adjustment for potential confounders, was significantly associated with contraception non-use (AOR = 2.02, 95% CI 1.44–2.82), use of unreliable contraceptive method (AOR = 1.54, CI 1.16–2.04), having never used a condom (AOR = 1.53, CI 1.12–2.10), having a repeat induced abortion (AOR = 1.72, CI 1.24–2.37), HIV/STIs (AOR = 2.05, CI 1.56–2.68) and dyspareunia (AOR = 2.14, CI 1.65–2.77).

View this table:
Table 4

Crude and adjusted OR for selected sexual health outcomes associated with IPV exposure during the last 12 monthsa

Sexual health outcomesTotal (n)IPV+ (n)Crude OR (95% CI)Adjusted ORb (95% CI)
Contraceptive method during the most recent sexual intercourse
    Unreliabled419981.70 (1.29–2.24)1.54 (1.16–2.04)
    Contraceptive non-use243652.03 (1.47–2.81)2.02 (1.44–2.82)
Condom use ever
    No289721.60 (1.18–2.15)1.53 (1.12–2.10)
Repeat induced abortion
    Yes391921.49 (1.14–1.95)1.72 (1.24–2.37)
    Yes6381461.64 (1.28–2.09)2.05 (1.56–2.68)
    Always/often4681322.20 (1.72–2.82)2.14 (1.65–2.77)
  • a: Only respondents who had non-missing answers for the question of interest and for all adjusting variables are presented in the table.

  • b: Adjusted for age, education, economic subsistance and ethnicity.

  • c: Hormonal contraceptive pill, patch, injectable hormonal contraceptive, emergency contraceptive pill, condom, IUD or sterilization.

  • d: Rhythm method, withdrawal, spermicide or vaginal douching.


This is the first study in Estonia to provide population-based data on women’s exposure of physical and sexual violence by their intimate partner and to explore associations with sexual health outcomes. Among the 16–44-year-old women in Estonia surveyed, 18.4% reported physical or sexual IPV. IPV exposure during the previous year was associated with sexual risk behaviour and adverse sexual health outcomes.

Contextual and methodological differences across violence studies do not allow direct comparisons of prevalences of violence.15,16 However, population-based studies have shown lower 1-year IPV prevalence than the prevalence found in our study: in Eastern European countries like Poland and the Czech Republic 3 and 9%,17 respectively, in Norway 5.5%9 and in Switzerland 6%,1 in New Zealand’s urban areas 5.2%10 and in Canada 3%.1 Our data are in line with previous data from Estonia,3 as well as the results from a Norwegian study, which found that women living in Norway, but born in Eastern European countries, were significantly more likely to be exposed to IPV compared with women born in Norway.

Several factors might influence the high prevalence of IPV in our study. First, some violence studies do not include threats of violence or other levels of violence (pushing, shaking and throwing things), which were the most common types of violence in our study. Second, the relatively low mean age of our respondents (29.4 years) might have influenced the overall IPV prevalence. An inverse relationship between violence and age has been described in previous research,1,17,18 as well as the phenomenon that younger women are more prone to report violence compared with older women.19 Without research data in Estonia, one can only hypothesize that the higher prevalence of IPV among younger women may be because of the real higher prevalence of IPV or because of changing social attitudes that allow more frequent reporting of violence. On the other hand, high prevalence of IPV may be linked to the relatively high prevalence of general violence in society. For example, during the period 2005–08 Estonia had by far the highest incidence of homicides in the European Union, with >6 cases annually per 100,000 population on average.20

Research findings concerning the associations between socio-economic status and risk of IPV are mixed and have cross-national variations.1,17,21 In addition, there is often uncertainty about the direction of these relationships.17,21,22 We found that women exposed to violence did not differ in their marital status from the reference group, but had lower educational attainment, lower economic subsistence and non-Estonian ethnicity. IPV survivors did not differ in regard to their employment status, but were two times more likely to have difficulties with paying bills. This might be the result of overall unequal income distribution, as well as the limited control over financial resources by the survivor in the violent relationship. Although women’s participation in working life is high in Estonia (74% in 2010), the gap between the salaries of women and men is the largest in the European Union.20 In our study population, 70% of respondents stated their native language to be Estonian, which correlates with the official country statistics.23 Despite the same socio-political settings, non-Estonian respondents (mostly Russians) reported IPV significantly more frequently compared with Estonians. This finding supports the hypothesis that cultural differences may be one of the key players regarding violence against women.6,22,24,25 Studies carried out in Russia have noted that Russian family structure has been a highly patriarchal institution for centuries, giving rise to a legacy of wife abuse.26–28 Furthermore, alcohol abuse has been found to be an important correlate of male violence against women.1,17,21 Research findings have confirmed that Russian culture is rather tolerant of certain behaviours when individuals are under the influence of alcohol.29 In Estonia, more frequent higher heavy alcohol consumption among Russian compared with Estonian men has been reported.30

IPV influences women’s ability to take control of their own fertility. Coerced sex can lead directly to an unintended pregnancy, but any type of violence can interfere with contraceptive choice.7,8,31 The paradoxical gap between contraception use and fertility desire was detected among our study respondents. More IPV survivors reported that they did not use any contraceptive method because they were planning a pregnancy; nevertheless, in the hypothetical case of becoming pregnant, they were significantly more likely to opt for an abortion rather than have the baby, or expressed their ambivalence about the decision. One reason for this discrepancy might also be reproductive coercion—male behaviours to promote pregnancy that is unwanted by the women—in an abusive relationship.7,32 In many studies, IPV was associated with higher parity.1,7,9 In Estonia, where abortion services are safe and easily accessible, it might be that in the case of unintended pregnancy, women choose an abortion, which could explain why we did not find differences in the number of births between IPV-exposed women compared with the reference group. IPV survivors had elevated risk for a repeat abortion, and this association was even stronger after adjustment.

Among our study respondents, violence by an intimate partner was associated with higher rate of lifetime STI. IPV victims were significantly less likely to have used a condom during their lifetime compared with the reference group. At the same time, data suggest that abusive men are more likely to have an STI history.33 In the context of Estonia, which has witnessed one of the most rapidly expanding HIV epidemics in Europe,12 certain subgroups—substance users and their sexual partners—are at considerable risk of contracting HIV. The link between IPV and HIV/STIs supports the increasing call for emphasis on IPV in local HIV prevention strategies. Additionally, IPV exposure was significantly associated with experiencing pain during or after intercourse, a complaint that could help to identify and address violent relationships.

Several practical implications arise from the findings presented here, including the potential for health care providers to recognize and screen for violence. General screening for IPV by health care providers is often recommended in the literature 5,15,21 and has been seen as a core part of women‘s preventive health visits.34 A study in five Nordic countries showed that patients themselves seldom report their abusive experiences to gynaecologists.35 The high prevalence and health burden of IPV among our respondents clearly supports the need for IPV screening in the health care system in Estonia. This may also lead to increased public awareness of this public health issue and accordingly the adoption of supportive legislative instruments. Currently, it is important that health care providers consciously avoid blaming the survivors of IPV and reduce the danger of re-victimization in the health care system. A sympathetic attitude, help in planning pregnancies and protecting against HIV/STIs in a violent relationship and offering partner-independent and low-cost contraceptive methods should be the essential minimum. To ask about violence in the case of a repeat induced abortion, HIV/STIs and dyspareunia, as well as in cases where there are no other causes for symptoms, may be helpful to prevent gynaecological problems and long-lasting burden of ill-health among IPV survivors.

Our findings should be interpreted in the context of certain limitations. Data were obtained from the self-reported measures in a women’s sexual health study with a response rate of 54%, which leaves the prevalence of violence unknown in almost half of the potential respondents. It is likely that risk groups of violence may be excluded in the survey, as women in violent relationships may fear for their safety when answering questions about violence, and the prevalence of violence is not overestimated.21 At the same time, a high proportion of women (98.2%) responded to the violence module. The findings are somewhat limited by the fact that data were used from a cross-sectional study, which only provides evidence of correlations, but often does not allow the direction of causation to be determined. However, previous evidence suggests that health problems are mainly outcomes of violence, rather than triggers.1,4,5 It should be noted that respondents were asked about their exposure of IPV in the last 12 months preceding the study and not during their lifetime. Therefore, we cannot determine the temporal sequence between the timing of sexual health outcomes and violence. Also, we are aware that other types of IPV besides physical and sexual violence may also influence sexual health outcomes; however, our questionnaire did not allow detailed analysis of other types of violence. Additionally, we did not obtain any information about childhood abuse, which increases the risk of poor sexual health in adult life8 and means that our results about IPV might be biased by the fact that some women had been survivors of violence in childhood. However, we believe that our results are reliable providing robust associations consistent with data from a large variety of samples between IPV and adverse sexual health outcomes.4–10


This work was supported by the Estonian Science Foundation [5456], Finnish National Research and Development Centre for Welfare and Health, the Estonian Ministry of Education and Science [SF0182641s04 and SF0940026s07] and the Baltic Sea Task Force [PHC030].

Conflicts of interest: None declared.

Key points

  • The study adds evidence for and underlines the importance of physical and sexual IPV as a contributing factor in women’s sexual behaviour and sexual health outcomes: contraception patterns, repeat induced abortion, HIV/STIs and sexual dysfunction.

  • Any strategy and interventions to promote sexual health should include prevention of violence against women and the strengthening of women’s sexual rights.

  • The development and implementation of a strategy for dealing with IPV in the health care system, involving different stakeholders, should be urgently addressed in Estonia.


The authors acknowledge Kaja Rahu for her valuable work in data management.


View Abstract