|Home | About | Journals | Submit | Contact Us | Français|
Violence, mostly partner violence, is now considered to be one of the most important health risk factors for women of reproductive age (Garcia-Moreno, Jansen, Ellsberg, Heise & Watts, 2006). Intimate Partner Violence (IPV) is a frequent occurrence: in a population-based study in the USA, 17.6% of women had experienced physical or sexual IPV in their lifetime (Coker, Davis, Arias et al., 2002). In Italy, population data are not available; studies among patients show that prevalence is lower but still substantial. In a sample of family doctors’ female patients (aged 18−85 years), lifetime prevalence of IPV was 18%; in the last year, 4% of women had suffered physical or sexual IPV, and 8.3% psychological violence (Romito, Molzan Turan & De Marchi, 2005); among hospital emergency department patients, 8.4% were currently experiencing physical or sexual violence by a partner or ex partner (Romito & Gerin, 2002).
Studies considering the role of violence in women's sexual and reproductive health have shown that IPV adversely affects sexual risk taking, contraceptive use, unplanned pregnancy, risk of sexually transmitted disease and sexual function (Gazmararian, Adams, Saltzman et al., 1995; Winghood & Di Clemente, 1997; Coker, 2007). Other studies have shown that violence during pregnancy may compromise maternal and infant health (Murphy, Schei, Myhr & Du Mont, 2001; Coker, Sanderson & Dong, 2004).
Yet, to date, fewer studies have analyzed the relationships between past and current violence and an Elective Abortion (EA). There are various possible links between these two events. Current violence limits the woman's agency concerning sexual intercourse and contraception (Fanslow, Whitehead, Silva & Robinson, 2008), making an unwanted pregnancy more likely; in some cases, an unwanted pregnancy may be the result of rape (Wu, Guo & Qu, 2005). In addition, a victim of violence may be more likely to terminate an unwanted or mistimed pregnancy than another woman, either by choice or forced by the aggressor (Bajos & Ferrand, 2003). Past violence as experienced in childhood or adolescence within the family context is associated with risky sexual behavior and with having a violent partner (Fergusson, Horwood & Lynskey, 1997), both circumstances increasing the probability of an EA.
The prevalence of IPV among women seeking an abortion is high. In the USA, Glander, Moore, Michielutte & Parsons (1998) found that 39.5% of these women had a history of IPV; in another study (Woo, Fine & Goetzl, 2005), 14% of EA patients reported partner abuse in the last year. Rates are higher among women with repeat terminations (Wu, Guo & Qu, 2005; Fisher, Singh & Shuper, et al., 2005). The few studies looking at the prevalence of violence in EA as compared to other patients show a higher prevalence among EA women. In a Canadian study (Bourassa & Bérubé, 2007), the probability of current IPV was much higher in the EA group (25.7%) than among women continuing the pregnancy (9.3%); the lifetime prevalence of violence was also significantly higher.
Although the role of violence in the decision to terminate a pregnancy is likely to be important, other social factors are involved. Compared with women continuing the pregnancy, women seeking an EA tend to be younger, have more children, and are more likely to be single, poorly educated and unemployed (Finer & Henshaw, 2006; Bankole, Singh & Haas, 1999). In various European countries, immigrants have higher abortion rates than non-immigrant women (Rash, Gammeltoft, Kudsen, Tobiassen, Ginzel & Kempf, 2007).
Women in Italy have had the right to elective abortion on demand, at no cost, since 1978, providing that the interruption is performed within the first 12 weeks of pregnancy. As in other countries, the abortion rate in Italy was highest in the years just after the act came into effect, and has steadily declined since: in the period from 2003 to 2006, it was 11.1 per 1000 women aged 15−44 years, one of the lowest in the industrialized world (Ministero della Salute, 2008). This low rate is most likely explained by fertility control through contraception by Italian women, rather than by any stigma associated with abortion or to religious prescriptions. Today, 31.6% of EAs performed in Italy are among immigrant women (Spinelli, Figà Talamanca & Lauria, 2000). In Italy, the total fertility rate is also very low, at 1.3 (the rate for Europe is 1.4, and for the developed world is 1.6) (Ashford & Clifton, 2005).
A number of factors linked to an EA are also associated with being a victim of IPV. Studies in Europe have shown that having many children, being young, unmarried, poor and poorly educated, unemployed, and immigrant are associated with experiencing violence by a partner or ex-partner (Walby & Allen, 2004; Condon & Schrottle, 2007). These factors are also linked to IPV during pregnancy (Saltzman, Johnson, Colley & Goodwin, 2003; Saurel-Cubizolles & Lelong, 2005).
A limitation of the studies showing an excess of violence among women seeking an EA lies in the fact that these social factors are not controlled, making it difficult to disentangle the respective role of these factors and of violence in the path leading to an EA. Another shortcoming is that most studies include only physical and sexual IPV, while psychological violence against women is frequent and may be strongly associated with adverse health outcomes (Coker, Davis & Arias, 2002). Additionally, usually only violence by a partner (IPV) is considered, while the perpetrators of violence against women during pregnancy and postpartum may be also other relatives (Gielen, O'Campo, Faden, Kass & Xue, 1994; Romito, Molzan Turan, Neilands, Lucchetta, Pomicino & Scrimin, 2009).
Age may be a key factor in the complex relationships between violence and abortion, as it is linked not only to EA and IPV, but also to violence by other relatives. In an Italian study (Romito & Gerin, 2002), younger women reported significantly more violence than older women, both by a partner or former partner and by other family members; violence by relatives was reported almost exclusively by women in the 18−24 years group.
The present study aims to analyze if the role of partner and family violence in the lives of women seeking an EA is different for women aged under 30 years and women aged 30 years or over, controlling for relevant socio-demographic variables, and hopes to overcome some of the aforementioned limits.
An unmatched case-control study was carried out in the only maternity hospital in Trieste (Italy), where all the births and abortions in the city take place. The cases comprised all consecutive elective abortions (445) occurring from March 2006 to July 2007, performed at less than 12 weeks of pregnancy. The unmatched control group included all consecutive live births (438), occurring from March 2006 to August 2006 in the same hospital.
In Trieste Maternity Hospital, EAs are performed on an out-patient basis. The information was collected from women by means of an anonymous, self-administered questionnaire during their stay in the maternity unit, two days after giving birth (controls), and in the afternoon after the EA (cases). All eligible women were approached in their hospital room at a time when they did not have visitors, and asked to participate. The study was presented as research on the health of women during pregnancy and confidentiality of the responses was assured. Verbal informed consent was obtained from each participant, and each was given a letter explaining the purposes of the study and information about support resources (telephone hotlines, shelters, other health and social services). The response rate was 93% among cases and 93% among controls. The study was approved by the Ethics Committee of the hospital.
Besides socio-demographic information (age, number of living children, education, employment, living or not with a partner, financial problems, and nationality), we asked questions about current and past violence. We included three questions to evaluate psychological, physical and sexual violence during the last 12 months, adapted from the Abuse Assessment Screen Questionnaire (McFarlane, Parker, Soeken & Bullock, 1992), asking the woman in each case to identify the aggressors. Similar questions were asked concerning violence in childhood or adolescence. For current partner violence, we constructed a variable which categorized women according to whether they had experienced no violence, only psychological violence, or physical/sexual violence. For current family violence, as most of it was psychological and no sexual violence was reported, we constructed a variable combining psychological and physical violence. For past violence, we constructed a synthetic variable including any kind of violence experienced in childhood or adolescence.
For women who had an EA, we asked three questions concerning their agency, both in the decision to become pregnant and in opting for an abortion. First, among the questions inquiring about the context in which the pregnancy occurred, women were asked if “it was the partner who wanted her to become pregnant”. Second, in a list of items concerning the reason for having an EA, women were asked if “the partner wanted a child, but they did not”. Third, another question inquired if the woman had been pressured not to have an abortion.
After descriptive analyses, we fitted logistic regression models to estimate the associations between violence and elective abortion vs. childbirth. Because in a previous study (Romito & Gerin, 2002), we found that the patterns of family violence were different according to women's age, we ran separate multivariate logistic regressions for women under 30 and women 30 years and older, adjusting for the other demographic and social factors known to be associated with both violence and EA: education, employment, economic problems, living alone, number of children, and nationality (Saltzman, Johnson, Colley & Goodwin, 2003; Walby & Allen, 2004; Saurel-Cubizolles & Lelong, 2005; Condon & Schrottle, 2007). We chose these age categories on theoretical and statistical grounds: in Italy, the mean age for having the first child is over 29 years (Istat, 2008); with a lower age cut-point, there were not enough post-partum (PP) women in this age group to run meaningful statistical analyses.
Adjusted Odds ratios (OR) and 95% confidence intervals (CI) were the measures of association obtained after fitting these models. Analyses were conducted using SPSS software version 15.
The social and demographic characteristics of the women are presented in Table 1. As compared to PP women, EA patients were significantly more likely to be young, have no previous births, have a low educational level, to lack regular employment, to report financial problems, to be living alone, and to be non-Italians.
The social characteristics of the EA and the PP samples correspond to national data concerning, respectively, women having an abortion and those giving birth in Northern Italy (Ministero della Salute, 2008; Istat, 2008).
Non-Italian women in this sample belonged to 38 different nationalities; most from Eastern European countries.
The distributions of women according to current and past violence are shown in Table 2. Women seeking an EA were more likely to report any kind of violence. The percentages of women exposed to psychological violence by a partner in the last 12 months were 11.0% in the EA group and 2,5% among post-partum women; 4.6% vs 0.9% respectively experienced physical violence, and 1.8 vs 0.5 experienced sexual violence. Using the current definition of IPV, including physical or sexual violence, 5.5% of women with an EA were exposed to IPV compared to 1.1% of PP women (data not shown).
The percentages of women reporting current psychological family violence were 6.2% among EA women and 1.1% among PP women; 1.6 of EA and no PP women reported family physical violence. The percentages of women who experienced violence in childhood or adolescence were also higher among women seeking an EA.
Table 3 shows the relationships of the synthetic violence variables with EA vs. PP status, stratified by women's age. Among women under 30 years of age, current partner violence, current family violence, and violence in childhood were all significantly more frequent among women with an EA than among PP women. Among older women, only partner violence was significantly more frequent among women with an EA; family violence was rare in this age group.
Concerning EA women's agency in getting pregnant and having an abortion: 2% of those without partner violence, 7% of those with psychological violence and 13% of those with physical or sexual violence reported that the conception had occurred because “the partner wanted her to become pregnant” (p=.002); 4.5%, 3.6% and 21.7% respectively said that they decided to have an abortion because “the partner wanted a child, but they did not” (p=.002).
The multivariate analyses revealed different patterns concerning violence for younger and older women (Table 4).
In the under 30 years age group, the adjusted odds of having an EA were significantly higher, 15.3, for women experiencing current psychological partner violence (p<0.05), while partner physical/sexual violence was no longer associated with EA after adjustment. Women experiencing current family violence, mostly psychological, had also significantly higher odds, 15.1, of having an EA (p<0.05).
Education, employment, nationality and number of previous live births were also associated with the probability of an EA.
In the 30 years and over age group, neither current partner violence nor family violence was associated with EA after adjustment. All the social factors considered - education, employment, financial problems, living alone, nationality and number of previous live births – increased the probability of an EA.
Results from this Italian study are unique in showing the role of partner and family violence in the lives of women – and especially of younger women- seeking an abortion.
Considering the whole sample, women having an EA are significantly more likely than women giving birth to report any current or past violence. They were also more likely to be poorly educated, without stable employment, with financial difficulties, living alone, and immigrants.
Analysis of the relationships between violence and EA vs. birth, stratified for women's age, reveals a more complex picture. Only among women under thirty years of age, are the differences between EA and PP women significant, with EA women reporting more partner, family, and childhood violence than new mothers. Among the women 30 years of age or older, only partner violence is significantly more frequent among women with an abortion. Noticeably, family violence is a frequent occurrence only among younger women with an abortion (13.7%), while it is almost non-existent among older women and women giving birth.
Multivariate results confirm that violence may play a different role according to women's age. Among the younger women, taking into account demographic and social factors, psychological partner violence and family violence, mostly psychological, remained associated with EA status: the adjusted odds of having an EA vs. giving birth were 15.3 for women exposed to psychological partner violence, and 15.1 for those exposed to family violence, while partner physical/sexual violence was no longer associated with EA after adjustment. Education, employment, nationality and number of previous live births also increased the probability of an EA; the odds of having an EA were higher also for women with financial problems and living alone, although, probably due to the small numbers of women involved, the associations became non-significant after adjustment.
Among older women, neither current partner violence nor family violence were associated with EA after adjustment, while all the demographic and social factors included in the model increased the probability of an EA.
These results are striking for several reasons. While it is known that victimization is more frequent among young women (Walby and Allen, 2004), and that young women are more likely to seek an EA (Finer & Henshaw, 2006), our results indicate that for young women, violence may have an important role in the process of opting for an abortion.
Equally noticeably, these results reveal the importance of psychological violence and violence by relatives other than the partner, phenomena that have been neglected in previous studies, which have mostly focused on physical and sexual partner violence. Yet, there is evidence that psychological violence – being called names, humiliated, denigrated, controlled - may have an effect on women's health as great or even greater then physical or sexual violence (Coker, Davis & Arias, 2002; Romito, Molzan Turan & De Marchi, 2005), and it is not surprising that it can also affect the abortion decision.
As far as abuse by other relatives is concerned, when it is asked about, results show that it is frequent and can be harmful, and not only in a family-oriented country such as Italy (Gielen, O'Campo, Faden, Kass & Xue, 1994; Pinnelli, Racioppi & Terzera, 2007). Romito et al. (Romito, Molzan Turan, Neilands, Lucchetta, Pomicino & Scrimin, 2009) found that family violence was even more strongly associated with postpartum depression than partner violence. John et al. (John, Johnson, Kukreja, Found & Lindow, 2004), in Great Britain, inquiring about violence among gynecology patients, found that the father ranked high in the list of perpetrators, higher than the husband (the most frequent perpetrators being ex husbands or ex boyfriends). In this study as well, perpetrators of family violence were mostly the woman's own parents.
There are several possible explanations for the links between partner and family violence and abortion among younger women. Women's agency in sexual and contraceptive choices is limited by partner violence, and it is not surprising that women resort to abortion when intercourse or the pregnancy have been imposed, or when they know that the partner is not caring nor accountable. Moreover, while from our data it seems that violent men tend to force the women to continue the pregnancy, other studies suggest that some violent men may, on the contrary, force the women to get a termination (Bajos, Ferrand et al., 2002). Concerning family violence, a woman who is maltreated as a daughter may prefer to avoid the risk of saddling a child with the unhappiness she is facing; parental abuse may also shatter a woman's self-esteem, leading her to doubt her capacities as a mother (Phillips, 2005; Romito, Crisma & Saurel-Cubizolles, 2003); it is also possible that, for some very young women, violence by relatives is the consequence, and not an antecedent, of an untimely pregnancy.
The prevalence of partner physical or sexual violence (IPV) against women in the EA group in the last year found in this study, 5.5%, was near to what found in Canada, 7% (Bourassa & Bérubé, 2007), and lower than what found in Texas, 14% (Woo, Fine & Goetzl, 2005). Among post-partum women, only 1.1% reported IPV, a figure similar to the 1% found in Central and South European countries and in Japan (Saurel-Cubizolles & Lelong, 2005; Garcia-Moreno, Jansen, Ellsberg, Heise & Watts, 2006). Since violence is more common among younger women, the comparison of rates of violence in pregnancy should, to be meaningful, take into account the age structure of the samples.
Another, more alarming, hypothesis is that some victims of violence were “switched” from the EA to the PP group - therefore diminishing the strength of the associations between violence and EA- because they were forced, against their wishes, to continue the pregnancy. Forcing the pregnancy and then impeding the abortion may represent an abusive man's strategy to control the woman (Bajos, Ferrand et al., 2002). Among the EA-seeking women in this study, those with partner violence, and especially physical or sexual violence, were more likely to say that the pregnancy was imposed on her by the partner. While the small numbers impose great caution in interpreting these data, they suggest that the more seriously victimized women may experience problems in obtaining an abortion.
This study has limitations and strengths. The risk of under reporting of violence cannot be excluded. This risk may be higher in the post-partum group, due to the stigma and shame associated with being a victim of violence during pregnancy and a desire to deny negative events leading to the birth of a child. Hopefully, the procedure for data collection – anonymous and self-administered questionnaires, completed when the woman had no visitors or staff in the room - should have minimized this possibility. In addition, the smaller sample sizes for within age-group analyses, leading to some unreliability of estimates and wide confidence intervals, represents another limitation of the study.
One challenge of case-control studies lies in the difficulty of finding appropriate, unbiased controls. Women who have just given birth may not be an unbiased control group. In the Trieste region, where the research was conducted, the total fertility rate is even lower than in the rest of the country (Istat, 2008), and abortion services are freely accessible. It is possible that, in a social context where not having children is both culturally acceptable and practically possible, only women who are in non violent couple relationships decide to have a baby; this would explain the low rate of couple violence among post-partum women found in this as well in a previous Italian study (Romito, Molzan Turan, Lucchetta, Scrimin, 2007). However, there is no other potential control group in which it would be possible to examine the factors associated with violence during pregnancy.
Strengths of our study include the fact that cases and controls were, respectively, consecutive EA and consecutive births; both coming from the same hospital – the only maternity hospital in the city - and were interviewed in the same time period; in addition, the response rate was the same in the two groups, and very high (93%).
Unique strengths of the study consist in having examined the role of psychological violence and of violence by other relatives, in having explored the differences between younger and older women, and in having controlled for women's social characteristics. Moreover, this is one of the very few studies on violence and abortion carried out in Southern Europe. The generalization of the results, and more particularly the interactions between violence, age and abortion, to women in other countries can be only tentative. However, the results of all studies on phenomena shaped by the social, cultural and psychological environment, such as violence against women and elective abortion, are at least partly context-dependent: while most studies find some common elements – such as young age as a risk factor – it is possible that other elements and relationships vary according the specific context of the study.
Many women who seek an EA live in difficult personal and social circumstances, in which violence often has a central role. They may seek abortion as a last resort, a rational and at times empowering choice when they evaluate that safe parenting is precluded (Jacoby, Gorenflo, Black, Wunderlich & Eyler, 1999; Russo & Denious, 2001; Fergusson, Boden & Horwood, 2007). In other cases they may be forced under threat of violence by their partners or other relatives either to terminate the pregnancy, or to continue it against their wishes. Limited ability of health professionals to recognize such situations can have life threatening implications for the women involved (Campbell, 1998).
The public health implications of this study are several. There is a need for sensitive screening for partner and family violence especially among women seeking an EA and among young women (Parsons, Goodwin & Petersen, 2000). These women often interact with a variety of health providers: social workers, psychologists, nurses, family doctors, anesthesiologists, and gynecologists. All these professionals should be adequately informed and trained to be able to support them, helping them to escape violence, prevent further harm, and maintain or regain control over their lives.
We wish to record out thanks to the women who participated in the study; to the Hospital personnel; to Lisa Bacchia, Irene Badina, Erinda Bartoli and Elisa Lombardi, who participated in data collection; and to the two reviewers, for their insightful and useful comments.
Funding: The study was supported by a grant from the Istituto di Ricovero e Cura Burlo Garofolo, Trieste. Janet Molzan Turan's work on Ethis article was also supported, in part, by grant # K01MH081777 from the U.S. National Institute of Mental Health (NIMH).
Competing interests: None.
Vicenta Escribà-Agüir, Centre for Public Health Research, Valencia, Spain. CIBER Epidemiología y Salud Pública (CIBERESP), Spain.
Laura Pomicino, University of Trieste, Trieste, Italy.
Chiara Lucchetta, IRCCS Maternity Hospital Burlo Garofolo, Trieste, Italy.
Federica Scrimin, IRCCS Maternity Hospital Burlo Garofolo, Trieste, Italy.
Janet Molzan Turan, Department of Obstetrics, Gynecology, and Reproductive Sciences, University of California, San Francisco, U.S.A.