DOI: 10.1111/j.1471-0528.2007.01481.x www.blackwellpublishing.

com/bjog

Epidemiology

Intimate partner violence and unwanted pregnancy, miscarriage, induced abortion, and stillbirth among a national sample of Bangladeshi women
JG Silverman,a J Gupta,a MR Decker,a N Kapur,b A Rajc
a Department of Society, Human Development and Health, Harvard School of Public Health, Boston, MA, USA b Department of Internal Medicine, Yale University School of Medicine, New Haven, CT, USA c Department of Social and Behavioral Sciences, Boston University of Public Health, Boston, MA, USA Correspondence: Dr JG Silverman, Department of Society, Human Development and Health, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115, USA. Email jsilverm@hsph.harvard.edu

Accepted 27 June 2007.

Objective To estimate (1) lifetime prevalence of physical and

sexual victimisation from husbands among a national sample of Bangladeshi women, (2) associations of unwanted pregnancy and experiences of husband violence, and (3) associations of miscarriage, induced abortion, and fetal death/stillbirth and such victimisation.
Design Cross-sectional, nationally representative study utilizing matched husband-wife data from the 2004 MEASURE Bangladesh Demographic Health Survey. Setting Bangladesh. Population Married Bangladeshi women ages 1340 years old

Muslim women were at greatest risk. Women experiencing violence from husbands were more likely to report both unwanted pregnancy (ORsadj 1.461.54) and a pregnancy loss in the form of miscarriage, induced abortion, or stillbirth (ORsadj 1.431.69). Assessed individually, miscarriage was more likely among victimised women (ORadj 1.81). A nonsignicant trend was detected for increased risk of induced abortion (ORadj 1.64); stillbirth was unrelated to violence from husbands.
Conclusion Intimate partner violence is extremely prevalent and

(n = 2677).
Methods Bivariate and multivariate logistic regression analysis. Main outcome measures Relations of intimate partner violence

to unwanted pregnancy, miscarriage, induced abortion and stillbirth.

Results Three out of four (75.6%) Bangladeshi women

relates to unwanted pregnancy and higher rates of pregnancy loss or termination, particularly miscarriages, among Bangladeshi women. Investigation of mechanisms responsible for these associations will be critical to developing interventions to improve maternal, fetal, and neonatal health. Such programmes may be vital to reducing the signicant health and social costs associated with both husband violence and unwanted and adverse pregnancy outcomes.
Keywords Induced abortion, intimate partner violence,

miscarriage, stillbirth, unwanted pregnancy.

experienced violence from husbands. Less educated, poorer, and

Please cite this paper as: Silverman J, Gupta J, Decker M, Kapur N, Raj A. Intimate partner violence and unwanted pregnancy, miscarriage, induced abortion, and stillbirth among a national sample of Bangladeshi women. BJOG 2007;114:12461252.

Introduction
Intimate partner violence (IPV) against women is a pervasive global public health concern, with 1571% of women experiencing such abuse worldwide.1 Among the numerous negative sexual and reproductive health outcomes associated with IPV (see Campbell2 and Heise et al.3 for review), unwanted/unintended pregnancy,410 induced abortion,1113

miscarriage,14,15 and fetal death11,16,17 are of particular concern among Bangladeshi women. Bangladeshi women experience 24 stillbirths and 36 neonatal deaths per 1000 deliveries,18 and an estimated 4070% of married Bangladeshi women experience IPV.1921 These rates are notably higher than those for IPV and stillbirth in the UK and in the USA, where approximately 57 deaths occur per 1000 deliveries,22,23 and an estimated 20% of women experience IPV victimisation.24,25

1246

2007 The Authors Journal compilation RCOG 2007 BJOG An International Journal of Obstetrics and Gynaecology

IPV, unwanted and terminated pregnancy in Bangladesh

Studies of Indian samples indicate associations of IPV and unwanted pregnancy9,10 as well as fetal death.17,26 However, as well as lack of distinction between stillbirth and neonatal death,26 as well as among other forms of adverse pregnancy outcomes (i.e. grouped assessment of abortion and miscarriage17), limits clarication of the impact of IPV within this context. The only Bangladeshi investigation of the relations of IPV to induced abortion and miscarriage identied an association of IPV and induced abortion among an urban sample but not among rural women; no associations were observed between IPV and miscarriage.27 Neither this nor other studies have assessed the relation of IPV to stillbirth/fetal death among a Bangladeshi sample, and no previous research on violence from male partners, regardless of national context, has included assessments of induced abortion, miscarriage, and stillbirth. The present study builds upon previous research by using a nationally representative sample of women at relatively high risk for both IPV and reproductive health concerns (married Bangladeshi women) to assess relations of IPV to unwanted pregnancy and multiple forms of pregnancy termination. Investigation of such patterns is critical to advancing our understanding of mechanisms by which violence from partners may impact specic forms of reproductive outcomes (e.g. induced abortion due to male partner coercion versus miscarriage or stillbirth due to injury, malnutrition, or other morbidity potentially associated with violence from a male partner). Specically, the present analyses provide (1) estimates of lifetime physical and sexual IPV victimisation among married Bangladeshi women based on husband reports of such violence, (2) the relative likelihood of unwanted pregnancy based on experiences of IPV, and (3) the relative likelihood of induced abortion, miscarriage, and fetal death/ stillbirth based on this victimisation.

participated (95.7% response). Details of data collection and management procedures were described elsewhere.28 Observations from the mens data were matched to those of their married female partners, and a merged dataset of husbandwife dyads was created. For purposes of the present analyses, only those women of childbearing age (ages 1340 years) who were currently married and living with their husband (n = 2677) were included.

Measures
The BDHS questionnaire was drafted in English and then translated into Bangla, the national language of Bangladesh. All variables were assessed through self-report, with the exception of household wealth. Demographics (age, education, polygamy, and religion) were assessed through single items. A relative index of household wealth was calculated based on interviewer-observed assets, including ownership of consumer items and dwelling characteristics; individual household wealth scores were grouped by quintile, with 1 = poorest and 5 = wealthiest 20% of households. Husband perpetration of IPV was assessed through six items included in the mens survey. A positive response to any one of the following behaviours indicated physical IPV perpetration: pushing or shaking your wife or throwing something at her, slapping her or twisting her arm, punching her with your st or something that could hurt her, kicking her or dragging her, or trying to strangle her or kill her by burning her. A positive response to physically forcing her to have sexual intercourse even when she didnt want to indicated sexual IPV perpetration. These assessments were recoded to create a four-level categorical variable reecting experiences of three categories of IPV (physical IPV only, sexual IPV only, and both physical and sexual IPV) and a referent group of no IPV perpetration of either form. Mens IPV perpetration was equated to IPV victimisation among married female partners of these men for purposes of the present study; no assessment of womens reports of IPV victimisation was conducted within the BDHS. Reproductive health outcomes were assessed through the womens questionnaire. For each birth within the past 5 years, women reported whether they wanted to become pregnant at that time, become pregnant but at a later date, or did not want to have any (more) children, with the latter two responses indicating an unwanted pregnancy. These assessments were recoded to create dichotomous variables for unwanted pregnancy at the most recent pregnancy and unwanted pregnancy during the past 5 years. Lifetime history of miscarriage, induced abortion, and stillbirth was assessed for all women through a single item asking if they had ever had a pregnancy that ended in miscarriage, ended due to an induced abortion or menstrual regulation (regional term referring to clinical procedures to induce abortion), or ended in the stillbirth of a child. Respondents who reported ever having experienced

Materials and methods

Sample
The present study used the 2004 MEASURE Bangladesh Demographic Health Survey (BDHS) conducted by the National Institute for Population Research and Training of the Ministry of Health and Family Welfare of Bangladesh from January to May 2004. The BDHS sample was drawn from all Bangladeshi adults, residing in private dwellings. A stratied, multi-stage cluster sample of 361 primary sampling units, 122 in urban areas and 239 in rural areas, was constructed. Of the 11 601 women deemed eligible to participate in the survey, 11 440 participated (98.6% response). After developing the sampling frame for the womens questionnaire, 50% of selected households were randomly targeted for implementation of the mens questionnaire. One man between 15 and 54 years was chosen at random from each of these households. Of 4490 eligible men identied, 4297

2007 The Authors Journal compilation RCOG 2007 BJOG An International Journal of Obstetrics and Gynaecology

1247

Silverman et al.

a miscarriage, induced abortion, or stillbirth were asked whether such an event had occurred within the past 5 years and, if such an event had occurred in this time frame, whether the most recent pregnancy termination or loss was a miscarriage, induced abortion, or stillbirth.

5 years (ORadj 1.54), and their most recent pregnancy resulting in a live birth being unwanted (ORadj 1.46) (Table 2). No associations of sexual IPV and unwanted pregnancy outcomes were detected.

Analyses
Prevalence estimates for experience of lifetime physical IPV (without sexual IPV), sexual IPV (without physical IPV), both physical and sexual IPV, and any form of IPV were calculated for the total sample of married Bangladeshi women and by demographic characteristics. Demographic differences in IPV victimisation were assessed through chi-square analyses; signicance for all analyses was set at P < 0.05. Adjusted logistic regression models were constructed to estimate odds ratios and 95% CIs for the association of forms of IPV victimisation and unwanted pregnancy, miscarriage, induced abortion, and stillbirth. Variables that either altered crude point estimates more than 10% or were signicant predictors at alpha = 0.20 were included as potential confounders.29 Analyses were conducted using STATA version 9 (College Station, TX, USA)30 so as to appropriately consider sample weighting.

Relations of IPV victimisation and pregnancy termination or loss

Bangladeshi women experiencing physical and not sexual IPV were at increased risk of experiencing a miscarriage, induced abortion, or stillbirth, both ever (ORadj 1.58) in the past 5 years (ORadj 1.69) (Table 2). Women experiencing both physical and sexual IPV were at increased risk of ever experiencing a miscarriage, induced abortion, or stillbirth (ORadj 1.43), with a nonsignicant trend towards elevated risk for any of these outcomes in the past 5 years (ORadj 1.47). No associations of IPV and stillbirth as the most recent form of pregnancy termination was detected. Miscarriage as the most recent form of pregnancy termination was more likely among women victimised by physical and not sexual IPV (ORadj 1.81). Among women experiencing this same form of IPV, a nonsignicant trend was detected for increased risk of induced abortion as the most recent form of termination (ORadj 1.64).

Results
Descriptive statistics on IPV victimisation and associations with demographics of married Bangladeshi women
Three out of four (75.6%) married Bangladeshi women experienced violence from their husbands based on husbands reports of perpetration of such abuse (Table 1). The most common form of IPV reported was physical violence in the absence of sexual violence (47.4%), followed by both physical and sexual violence (24.0%). Sexual violence without physical violence was experienced by 4.2% of women based on husband reports. Women older than 20 years were more likely to experience physical violence in the absence of sexual violence, while women younger than the age of 30 years were more likely to experience both physical and sexual violence. Women with lower educational attainment and occupying the lowest wealth quintile were more likely to experience victimisation of all forms of IPV with the exception of sexual IPV in the absence of physical violence. Both any form of IPV and physical and not sexual IPV were found to be more prevalent among Muslim women compared with Hindu Bangladeshi women.

Discussion
Over 75% of married Bangladeshi women in this nationally representative sample experienced physical and/or sexual violence from their husbands based on these mens self-reported perpetration of such abuse. This extremely high lifetime prevalence rate is consistent with other recent studies conducted among married Bangladeshi women21,24 and conrms that IPV is alarmingly commonplace in this impoverished South Asian nation, potentially affecting the health of a large majority of Bangladeshi women. Furthermore, consistent with prior IPV research,31 including that conducted in Bangladesh,21 IPV was most prevalent among the most disadvantaged groups; however, IPV also occurred at high rates among even the most advantaged strata of Bangladeshi society. Notably, Muslim women experienced higher rates of IPV than Hindu women (77 versus 62%) in this majority Muslim country. Prior work conducted among South Asian samples has yielded mixed results regarding religion-based differences in experiences of partner violence.10,17,32 While exploring the inuence of religion in cases of IPV is beyond the scope of the current study, further research into the potential role of religion and other cultural practices and institutions in violence against wives in South Asia is warranted, both based on the current ndings and the need to consider participation of religious institutions in efforts to reduce IPV. In this rst study of the relation of IPV to unwanted pregnancy among Bangladeshi women, unwanted pregnancy within the past 5 years and regarding the most recent live

Relations of IPV victimisation and unwanted pregnancy

In adjusted logistic regression analyses, women experiencing physical and not sexual IPV were more likely to report an unwanted pregnancy resulting in a live birth during the past

1248

2007 The Authors Journal compilation RCOG 2007 BJOG An International Journal of Obstetrics and Gynaecology

IPV, unwanted and terminated pregnancy in Bangladesh

Table 1. Demographic characteristics and IPV victimisation of married Bangladeshi women ages 1340 years (n = 2677) % (95% CI) Any form of IPV, % (95% CI) Physical IPV only, % (95% CI) Sexual IPV only, % (95% CI) Both physical and sexual IPV, % (95% CI)

Age (years) 1320 2130 3140 P value Residence Urban Rural P value Education No education Primary Secondary Higher P value Wealth index Poorest Poorer Middle Richer Richest P value Religion Muslim Hindu P value Number of wives One More than one P value Total

20.81 (19.1122.62) 49.37 (47.0851.67) 29.82 (27.9231.78)

73.20 (68.3677.54) 77.33 (74.6179.84) 74.45 (71.1177.53) 0.164 74.53 (70.0278.56) 75.93 (73.4578.25) 0.571 81.88 (79.2284.27) 78.82 (75.1882.06) 66.86 (62.8970.61) 43.89 (34.7453.47) ,0.001 81.99 (77.8285.52) 77.56 (72.9881.56) 78.25 (73.5282.33) 72.64 (67.9676.87) 63.99 (58.8368.85) ,0.001 76.92 (74.7478.96) 61.72 (55.0867.94) ,0.001 75.88 (73.7877.85) 69.94 (57.4180.07) 0.322 75.61 (73.4677.64)

41.19 (37.2445.27) 47.12 (43.7950.48) 52.2 (48.1056.26) ,0.001 48.73 (44.2653.21) 47.01 (43.8950.15) 0.538 55.03 (51.3158.70) 47.31 (42.8951.77) 38.40 (34.4142.55) 28.83 (21.4437.54) ,0.001 50.97 (46.2555.67) 48.5 (43.7153.31) 47.30 (41.9152.75) 47.12 (41.4252.90) 41.34 (37.0645.76) 0.053 48.61 (45.9551.28) 35.26 (27.8743.43) 0.006 47.86 (45.2350.51) 37.66 (27.4249.13) 0.103 47.4 (44.7850.03)

6.31 (4.428.93) 3.71 (2.715.05) 3.56 (2.385.29) 0.085 4.05 (2.835.76) 4.25 (3.365.37) 0.820 3.43 (2.434.82) 3.95 (2.765.64) 6.05 (4.188.68) 2.43 (0.708.05) 0.190 4.39 (2.786.87) 3.78 (2.316.11) 4.19 (2.716.43) 5.01 (3.187.80) 3.76 (2.425.79) 0.905 4.16 (3.375.12) 5.02 (2.729.08) 0.590 4.24 (3.465.19) 3.42 (1.258.99) 0.643 4.20 (3.445.13)

25.70 (22.1329.63) 26.50 (23.5329.70) 18.69 (15.9621.77) ,0.001 21.75 (18.4425.48) 24.67 (22.1127.43) 0.191 23.42 (20.2626.90) 27.56 (23.9931.45) 22.42 (18.9826.27) 12.64 (7.5620.37) 0.002 26.63 (22.5331.18) 25.29 (21.5729.40) 26.76 (22.3131.73) 20.51 (16.0625.82) 18.89 (15.2823.14) 0.033 24.15 (22.0126.43) 21.43 (14.929.84) 0.470 23.78 (21.6126.09) 28.87 (18.7541.65) 0.406 24.01 (21.8726.28)

22.83 (21.0124.75) 77.17 (75.2578.99)

39.93 (37.5742.31) 30.99 (29.0732.98) 24.94 (22.9827.0) 4.15 (3.435.01)

23.5 (21.0226.17) 23.24 (21.2825.32) 19.53 (17.7521.45) 16.80 (15.0318.74) 16.93 (15.0119.04)

90.62 (87.6292.95) 9.38 (7.0512.38)

95.51 (94.0296.65) 4.49 (3.355.98)

birth was found to be approximately 50% more likely among those physically abused by their husbands than their nonabused counterparts. These ndings are consistent with research in other national contexts,33 including India,9,10 underlining the reduced control abused women appear to possess over their reproductive choices and, potentially, reduced access to family planning or other fertility control resources. Although preventing such violence is the obligation of all governments, present ndings suggest that reducing IPV may signicantly reduce numbers of unwanted pregnancies, likely enhancing development; according to the 2005 UN World Population Report,34 womens lack of reproductive control is a key limiting factor in major goals of economic development. Thus, prevention of violence from husbands, and thereby promoting womens control over their own reproduction, may be considered both an economic as well as social imperative.

The current investigation yielded notable differences in patterns of the pregnancy-related outcomes based on IPV experiences. Married Bangladeshi women victimised by their husbands were slightly less than twice as likely to report miscarriage as the most recent form of pregnancy termination or loss than nonabused women. Mechanisms underpinning these associations may include poorer general health found among abused women,3537 poor nutrition as a potential aspect of abuse,20,38 poor or delayed antenatal care received by abused women,10,3941 injuries during pregnancy that directly or indirectly compromise fetal health, or excessive labour demands that may be placed upon abused wives. Because current ndings are cross-sectional in nature, it is also possible that the chronology involved is reversed; for example, women may experience IPV subsequent to miscarriages of pregnancy based on husbands blaming of wives for their failure to produce

2007 The Authors Journal compilation RCOG 2007 BJOG An International Journal of Obstetrics and Gynaecology

1249

Silverman et al.

Table 2. Associations of lifetime IPV victimisation with unwanted pregnancy, miscarriage, induced abortion, and stillbirth among married Bangladeshi women ages 1340 (n = 2677) % Any violence (95% CI) % No violence (95% CI) Physical IPV only OR (95% CI) Sexual IPV only OR (95% CI) Both physical and sexual IPV OR (95% CI) 1.14 (0.731.79) 1.13 (0.791.63) 1.43 (1.091.90) 1.47 (0.992.18) 1.53 (0.773.04) 1.47 (0.832.60) 1.16 (0.552.45)

Unwanted pregnancypast 5 years*,**,1 23.46 (20.6826.49) 17.54 (13.5922.35) Most recent pregnancy unwanted*,***,1 31.57 (28.4234.9) 25.06 (20.4730.3) Miscarriage, induced abortion, or 26.22 (24.0628.5) 19.82 (16.7423.3) stillbirthever2 Miscarriage, induced abortion, or 12.9 (11.3214.65) 9.20 (7.1211.82) stillbirthpast 5 years3 Stillbirth (most recent termination)4 2.76 (2.143.54) 0.65 (0.401.06) Miscarriage (most recent termination)5 4.23 (3.455.17) 0.88 (0.581.34) Abortion (most recent termination)6 2.77 (2.163.54) 0.71 (0.431.16) *Analysis limited to women who had given birth in the past 5 years, n 5 1651. **Ever experienced any unwanted pregnancy that resulted in live birth. ***Most recent pregnancy that resulted in a live birth was unwanted. 1Analyses adjusted for age, education, and religion. 2Analyses adjusted for age, wealth, and urban/rural. 3Analyses adjusted for age, education, wealth and urban/rural. 4Analyses adjusted for age. 5Analyses adjusted for age, wealth, urban, religion, and number of wives. 6Analyses adjusted for age, education, wealth, and urban/rural.

1.54 (1.102.15) 1.40 (0.623.15) 1.46 (1.072.00) 1.24 (0.592.94) 1.58 (1.212.06) 1.22 (0.732.05) 1.69 (1.202.38) 1.17 (0.552.45) 1.30 (0.752.33) 1.18 (0.373.79) 1.81 (1.053.12) 0.84 (0.292.48) 1.64 (0.972.78) 1.50 (0.504.53)

a child.17 Further research is needed regarding both miscarriage as a potential consequence of IPV and clarication of mechanisms responsible for the relations of IPV to this outcome. Stillbirth was not associated with IPV experiences in the current data, contrasting with prior research conducted in India, which demonstrated associations of IPV with fetal and neonatal death.17,26 However, these prior studies included deaths occurring shortly after birth (within 1 week in Ahmed et al.; period not specied in Jejeebhoy). Thus, it may be that there exists a critical period earlier in pregnancy within which a developing fetus is placed at risk based on the mothers experience of abuse, and then a period of relative invulnerability to the effects of abuse that continues until birth. However, immediately following birth, survival may, again, be impaired based on the presence of the abusive father. The overlap between mens abuse of wives and children has been demonstrated in North American populations;42 similar investigation is needed in a South Asian context. How IPV relates to death of neonates and infants is a critically important topic deserving of increased attention, particularly in South Asia where child mortality is relatively high and, notably, where girls suffer higher child mortality than boys, a disparity rarely seen globally.43 Future research regarding the potential role of abusive male partners in the death of infants and children should also consider additional factors inuencing such mortality previously found to relate to IPV (e.g. low birthweight).44

Experiences of IPV did not signicantly increase risk for induced abortion as the most recent form of pregnancy termination or loss, although a trend towards increased risk for this outcome was detected among women experiencing only physical IPV. These ndings are partially consistent with previous research that demonstrated an association of IPV with induced abortion among an urban but not among a rural sample of Bangladeshi women.27 Current analyses did not stratify the population based on urban or rural residence but adjusted for differences in the predicted outcomes in single regression equations to present a nationally representative estimate. Bangladeshi laws banning abortion, likely resulting in underreporting of induced abortion, must be considered in interpreting the present ndings. While the BDHS attempts to facilitate the reporting of abortions by including events described as menstrual regulation, a term used in Bangladesh to describe administration of legal abortive procedures by a clinician, stigma may have resulted in abortions being misreported as other forms of termination, for example miscarriage, to avoid stigmatisation. Furthermore, illegal abortion procedures may result in future fertility concerns, thus introducing the possibility that miscarriage as the most recent termination event may relate to a womans history of abortion, data not available in the current study. Future research conducted in such national settings should make efforts to ensure reliable reporting of induced abortion and adjust for histories of abortion in estimating risks for miscarriage. Such improved data collection may provide

1250

2007 The Authors Journal compilation RCOG 2007 BJOG An International Journal of Obstetrics and Gynaecology

IPV, unwanted and terminated pregnancy in Bangladesh

additional and more precise insights into relationships between IPV, miscarriage, and induced abortion. The present ndings should be considered in light of several limitations beyond those previously mentioned. Data regarding unwanted pregnancy were only available for those women reporting a history of live births. Thus, we were unable to assess the impact of IPV on unwanted pregnancy among those who had been pregnant but had experienced miscarriage, induced abortion, or stillbirth and never delivered a live infant. Most notably, the present study was only able to assess these outcomes individually within the context of the most recent form of pregnancy loss or termination among those experiencing such events in the past 5 years, eliminating the ability to detect relations of IPV to either experiences of these outcomes occurring previous to this recent event within the past 5 years, or to such outcomes that may have occurred earlier than 5 years ago. Thus, the nature of the assessment limits the scope of our estimates and likely biases results to the null compared with prior studies that more broadly assessed the prevalence of these pregnancyrelated outcomes. The current analyses were also not able to account for contraceptive use due to inability to match such behaviour to assessed periods; although prior research has found that contraceptive use did not affect the documented association between IPV and unwanted pregnancy,7 future research should consider such behavioural factors when possible, including womens ability to negotiate condom use. Regarding miscarriage, prior work has shown that women who experience adverse pregnancy outcomes are more likely to recall and report harmful exposures.45 However, reports of IPV for the current study were collected in the form of husband rather than wife reports. It may also be possible that, for similar reasons, women who have been abused are more likely to recall adverse events, such as miscarriage, potentially biasing current ndings away from the null; biased recall in this context has yet to be assessed. Reliance on mens reports of IPV perpetration is noteworthy and may have underestimated sexual violence, in particular, based on both the narrow measure used (i.e. forced sex) as well as norms of limited communication regarding sex in the South Asian marital context.46 Such measurement issues may account for the lack of ndings regarding sexual IPV and any outcomes assessed in the current study. It may also be possible that men over-reported IPV perpetration, possibly as a boast regarding their power over their wives. However, such over-reporting has not been documented and is unlikely to be differentially related to womens reporting of adverse pregnancy outcomes; further, such reduced precision in measurement is likely to bias results towards the null. Although interviewer effects may have also affected study ndings, it is difcult to estimate the magnitude or direction of such bias, if existent, given the multiple sources of possible variance (e.g. variation in training, presence of others during data collection). Underreporting of

violence victimisation by women based on presence of husbands during the interview has been documented,47 but no such interviewer-related biases have been reported regarding husband reports of IPV perpetration. Finally, as discussed above, the current analyses are cross-sectional and, thus, do not allow for assessment of the chronology of the associated events or inferences regarding causality. Longitudinal research regarding the relations of IPV to unwanted pregnancy and pregnancy outcomes is needed to provide clarity regarding these concerns. The current ndings conrm that violence against wives by their husbands is a shockingly prevalent problem in Bangladesh, likely affecting the health of tens of millions of adolescent and adult women. The current study also indicates that abused married Bangladeshi women are both less able to control their reproduction, as represented by higher rates of unwanted pregnancy, and more likely to experience termination or loss of pregnancy, particularly in the form of miscarriages, compared with those not experiencing violence from their husbands. Investigation of mechanisms responsible for these associations will be critical to developing interventions to improve maternal, fetal, and neonatal health, and should be considered a public health research priority. Violence against wives represents a health and human rights concern of the highest level in Bangladesh and elsewhere, and greater resources must be dedicated to prevent such abuse based both on its pervasiveness and on the potential of such prevention efforts to reduce the signicant health and social costs associated with large numbers of additional unwanted pregnancies and losses of pregnancy.

Acknowledgement
No external funding was provided to the authors to support completion of this study. j

References
1 Garcia-Moreno C, Jansen HA, Ellsberg M, Heise L, Watts CH; WHO Multi-country Study on Womens Health and Domestic Violence against Women Study Team. Prevalence of intimate partner violence: ndings from the WHO multi-country study on womens health and domestic violence. Lancet 2006;368:12609. 2 Campbell JC. Health consequences of intimate partner violence. Lancet 2002;359:13316. 3 Heise L, Ellsberg M, Gottmoeller M. Ending violence against women. Popul Rep L 1999;27:1318. 4 Amaro H, Fried LE, Cabral H, Zuckerman B. Violence during pregnancy and substance use. Am J Public Health 1990;80:5759. 5 Hillard PJ. Physical abuse in pregnancy. Obstet Gynecol 1985;66: 18590. 6 Stewart DE, Cecutti A. Physical abuse in pregnancy. CMAJ 1993; 149:125763. 7 Pallitto CC, OCampo P. The relationship between intimate partner violence and unintended pregnancy: analysis of a national sample from Colombia. Int Fam Plan Perspect 2004;30:16573.

2007 The Authors Journal compilation RCOG 2007 BJOG An International Journal of Obstetrics and Gynaecology

1251

Silverman et al.

8 Kaye DK, Mirembe FM, Bantebya G, Johansson A, Ekstrom AM. Domestic violence as risk factor for unwanted pregnancy and induced abortion in Mulago Hospital, Kampala, Uganda. Trop Med Int Health 2006;11:90101. 9 Martin SL, Kilgallen B, Tsui AO, Maitra K, Singh KK, Kupper LL. Sexual behaviors and reproductive health outcomes: associations with wife abuse in India. JAMA 1999;282:196772. 10 Muthal-Rathore A, Tripathi R, Arora R. Domestic violence against pregnant women interviewed at a hospital in New Delhi. Int J Gynaecol Obstet 2002;76:835. 11 Kaye DK, Mirembe F, Bantebya G, Johansson A, Ekstrom AM. Reasons, methods used and decision-making for pregnancy termination among adolescents and older women in Mulago Hospital, Uganda. East Afr Med J 2005;82:57985. 12 Wu J, Guo S, Qu C. Domestic violence against women seeking induced abortion in China. Contraception 2005;72:11721. 13 Fisher WA, Singh SS, Shuper PA, Carey M, Otchet F, MacLean-Brine D, et al. Characteristics of women undergoing repeat induced abortion. CMAJ 2005;172:63741. 14 Taft AJ, Watson LF, Lee C. Violence against young Australian women and association with reproductive events: a cross-sectional analysis of a national population sample. Aust N Z J Public Health 2004; 28:3249. 15 Fonck K, Leye E, Kidula N, Ndinya-Achola J, Temmerman M. Increased risk of HIV in women experiencing physical partner violence in Nairobi, Kenya. AIDS Behav 2005;9:3359. 16 Lipsky S, Holt VL, Easterling TR, Critchlow CW. Impact of policereported intimate partner violence during pregnancy on birth outcomes. Obstet Gynecol 2003;102:55764. 17 Jejeebhoy SJ. Associations between wife-beating and fetal and infant death: impressions from a survey in rural India. Stud Fam Plann 1998; 29:3008. 18 WHO. World Health Report: Make Every Mother and Child Count. Geneva, Switzerland: World Health Organization, 2005. 19 Bhuiya A, Sharmin T, Hani SM. Nature of domestic violence against women in a rural area of Bangladesh: implication for preventive interventions. J Health Popul Nutr 2003;21:4854. 20 Naved RT, Azim S, Bhuiya A, Persson LA. Physical violence by husbands: magnitude, disclosure and help-seeking behavior of women in Bangladesh. Soc Sci Med 2006;62:291729. 21 Bates LM, Schuler SR, Islam F, Islam MK. Socioeconomic factors and processes associated wtih domestic violence in rural Bangladesh. Int Fam Plan Perspect 2004;30:1909. 22 Ofce for National Statistics. Investigation in to the increase in stillbirth rate in 2002 in England and Wales. Health Statistics Quarterly 2003; 20:667. 23 National Center for Health Studies. Health, United States, 2006 with chartbook on trends in the health of Americans. US Department of Health and Human Services: Centres for Disease Control and Prevention; November 2006. DHHS Publication No. 2006-1232. 24 Tjaden P, Thoennes N. Extent, Nature, and Consequences of Intimate Partner Violence: Findings from the National Violence Against Women Survey. Washington, DC: Department of Justice, National Institute of Justice, 2000. 25 Nicholas S, Povey D, Walker A. Crime in England and Wales 2004/ 2005. Home Ofce Statistics Bulletin. Crown copyright, 2006, London. 26 Ahmed S, Koenig MA, Stephenson R. Effects of domestic violence on perinatal and early-childhood mortality: evidence from north India. Am J Public Health 2006;96:14238.

27 WHO. WHO Multi-country Study on Womens Health and Domestic Violence Against Women: Summary Report of Initial Results on Prevalence, Health Outcomes and Womens Responses. Geneva, Switzerland: World Health Organization; 2005. 28 NIPORT MaA. Bangladesh Demographic and Health Survey 2004. Dhaka, Bangladesh and Calverton, MD: National Institute of Population Research and Training, Mitra and Associates, and ORC Macro, 2005. 29 Rothman KJ, Greenland S. Modern Epidemiology. Philadelphia, PA: Lippincott-Raven, 1998. 30 StataCorp. Stata Statistical Software: Release 9. College Station, TX: StataCorp LP, 2005. 31 WHO. World Report on Health and Violence. Geneva, Switzerland: World Health Organization; 2002. 32 Koenig MA, Ahmed S, Hossain MB, Khorshed Alam Mozumder AB. Womens status and domestic violence in rural Bangladesh: individual and community-level effects. Demography 2003;40:26988. 33 Pallitto CC, Campbell JC, OCampo P. Is intimate partner violence associated with unintended pregnancy? A review of the literature. Trauma Violence Abuse 2005;6:21735. 34 UN Department of Economic and Social Affairs. Population Challenges and Development Goals. New York, NY: United Nations, 2005. 35 Leung TW, Leung WC, Ng EH, Ho PC. Quality of life of victims of intimate partner violence. Int J Gynaecol Obstet 2005;90:25862. 36 Parish WL, Wang T, Laumann EO, Pan S, Luo Y. Intimate partner violence in China: national prevalence, risk factors and associated health problems. Int Fam Plan Perspect 2004;30:17481. 37 Hurwitz EJ, Gupta J, Liu R, Silverman JG, Raj A. Intimate partner violence associated with poor health outcomes in U.S. South Asian women. J Immigr Minor Health 2006;8:25161. 38 Panchanadeswaran S, Koverola C. The voices of battered women in India. Violence Against Women 2005;11:73658. 39 Bloom SS, Lippeveld T, Wypij D. Does antenatal care make a difference to safe delivery? A study in urban Uttar Pradesh, India. Health Policy Plan 1999;14:3848. 40 Kapoor RK, Srivastava AK, Misra PK, Sharma B, Thakur S, Srivastava KI, et al. Perinatal mortality in urban slums in Lucknow. Indian Pediatr 1996;33:1923. 41 Purwar MB, Jeyaseelan L, Varhadpande U, Motghare V, Pimplakute S. Survey of physical abuse during pregnancy GMCH, Nagpur, India. J Obstet Gynaecol Res 1999;25:16571. 42 Edleson JE. The overlap between child maltreatment and woman battering. Violence Against Women 1999;5:13454. 43 Filmer D, King EM, LP. Gender disparity in South Asia: comparisons between and within countries. 1996 [www.worldbank.org/research/ projects/gender/sasgdis.pdf]. Accessed 19 October 2006. 44 Silverman JG, Decker MR, Reed E, Raj A. Intimate partner violence victimization prior to and during pregnancy among women residing in 26 U.S. states: associations with maternal and neonatal health. Am J Obstet Gynecol 2006;195:1408. 45 Neutra RR, Swan SH, Hertz-Picciotto I, Windham GC, Wrensch M, Shaw GM, et al. Potential sources of bias and confounding in environmental epidemiologic studies of pregnancy outcomes. Epidemiology 1992;3:13442. 46 Lambert H, Wood K. A comparative analysis of communication about sex, health and sexual health in India and South Africa: implications for HIV prevention. Cult Health Sex 2005;7:52741. 47 Ellsberg M, Heise L, Pena R, Agurto S, Winkvist A. Researching domestic violence against women: methodological and ethical considerations. Stud Fam Plann 2001;32:116.

1252

2007 The Authors Journal compilation RCOG 2007 BJOG An International Journal of Obstetrics and Gynaecology