Journal of Immigrant and Minority Health

, Volume 12, Issue 2, pp 242–248

Pacific Islands Families Study: Intimate Partner Violence and Postnatal Depression


    • National Institute for Public Health and Mental Health Research, Faculty of Health and Environmental SciencesAUT University
  • Janis Paterson
    • National Institute for Public Health and Mental Health Research, Faculty of Health and Environmental SciencesAUT University
  • Max Abbott
    • Faculty of Health and Environmental SciencesAUT University
  • Sarnia Carter
    • National Institute for Public Health and Mental Health Research, Faculty of Health and Environmental SciencesAUT University
  • Leon Iusitini
    • National Institute for Public Health and Mental Health Research, Faculty of Health and Environmental SciencesAUT University
Original Paper

DOI: 10.1007/s10903-008-9190-y

Cite this article as:
Gao, W., Paterson, J., Abbott, M. et al. J Immigrant Minority Health (2010) 12: 242. doi:10.1007/s10903-008-9190-y


Aim The present study examined the association between maternal intimate partner violence (IPV) and postnatal depression (PND) 6 weeks after giving birth. Study Design Data were gathered from the Pacific Islands Families Study. Mothers of a cohort of Pacific infants born in Auckland, New Zealand during 2000 were interviewed 6 weeks after giving birth. There were 1,085 mothers cohabiting in married or de-facto partnerships who completed measures of IPV and PND at the 6-week assessment point. Results Women who were victims of physical violence were more likely to report postnatal depressive symptoms than those who were not (29.6% vs. 10.9%, OR: 3.44, 95% CI: 2.42, 4.97). The adjusted odds remained statistically significant (OR: 2.34, 95% CI: 1.52, 3.60). Conclusion Findings suggest that being the victim of physical violence more than doubles the risk of PND. The results of the study may help to develop culturally appropriate social services and policies for Pacific women.


Intimate partner violenceFamily violenceMaternal healthPostnatal depressionPostpartum depression


The infliction of violence by intimate partners is common in many societies, affecting millions of women throughout the world each year [1]. In the last decade, research has identified some of the direct and indirect health consequences of abuse, including physical, mental, and reproductive health problems [26]. Depression is a common psychological reaction to intimate partner violence (IPV) [7, 8]. Among women who have been diagnosed with major depression, approximately 60% report histories of intimate partner abuse, a rate that is two times greater than the general population [9].

Studies conducted internationally indicate that postnatal depression (PND), is the most common mental disorder associated with childbirth and is a significant health issue [10, 11]. Previous New Zealand studies [12, 13], found that past psychiatric disorder, antenatal depression, poor partner relationships, and single parenthood are significant risk factors for PND.

While there is a substantial body of research examining the relationships between IPV and mental health [6, 9, 1418], only a few studies have specifically assessed the impact of partner relationships or conflict on PND. In an Australian study, Johnstone et al. (2001) [19] reported that ‘arguments with a partner’ was a significant risk factor for PND at 8 weeks postpartum, while a prospective study of Israeli women found that ‘marital disharmony’ was significantly related to PND [20]. A sample of US women reported that those who experienced ‘partner-associated stress’ were two times more likely to be very depressed in the months after delivery than those with no stress from their partner [21]. The Pacific Islands Families (PIF) study concluded that the strongest protective factor associated with PND was happiness in partner relationships [22]. A Canadian study reported that women with depressive symptoms at 8-weeks postpartum had significantly lower perceptions of relationship-specific and postpartum-specific partner support and significantly higher levels of relationship conflict than women with no depressive symptoms [23].

The Pacific population is a demographically young and rapidly growing ethnic group in New Zealand, particularly in Auckland [24]. At the time of the 2006 census, 265,974 Pacific people were resident in New Zealand, 6.6% of the total population [25]. Relative to the general population, Pacific people experience lower socioeconomic status, poorer health, and lower life expectancy [26].

Researchers have contended that in Pacific societies the likelihood that family violence is considered ‘normal’ or at the least acceptable in principle may be higher than in non-Pacific cultures [27]. Within a Pacific context it is postulated that violence within the home may be attributable to the stress of living in New Zealand and related to factors such as unemployment and associated financial concerns [28]. It has been suggested that such violence is exacerbated by the stresses associated with Western influence on traditional lifestyles [29].

A previous qualitative investigation identified PND as a concern within Pacific communities [30]. The authors suggested that Pacific women are likely to be at high risk of PND, but that risk factors identified in Western populations may not apply to them, and concluded that traditional family and cultural supports, affordable childcare and antenatal education are probably the most important protective factors.

The PIF study is a prospective longitudinal study that follows a cohort of Pacific children and their families living in New Zealand. At the 6-week data collection point, the Conflict Tactics Scale [31] and the Edinburgh Postnatal Depression Scale (EPDS) [32] were included. Findings from the PIF study demonstrated that IPV prevalence rates were high for a general married or cohabiting female sample with overall violence victimization rates of 22.9% [33]. Concurrently, 16.4% of mothers were assessed as experiencing symptoms of PND, a figure at the upper end of the range typically reported [22]. The purpose of the present study was to examine the association between maternal IPV victimization and postnatal depressive symptoms in this cohort of Pacific mothers resident in Auckland, New Zealand.



Data were collected as part of the Pacific Islands Families (PIF) study. This study follows a cohort of Pacific Islands infants born at Middlemore Hospital between 15 March and 17 December 2000. All potential participants were selected from birth where at least one parent was identified as being of a Pacific Islands ethnicity and a New Zealand permanent resident. Participants were identified through the Birthing Unit, in conjunction with the Pacific Islands Cultural Resource Unit. Information about the study was provided and consent was sought to make a home visit.

Approximately 6-weeks after the birth of their child, Pacific interviewers, fluent in both English and a Pacific language, visited the mothers in their homes. Once eligibility criteria were confirmed and informed consent was obtained, mothers participated in 1-h interviews concerning the health and development of the child and family functioning. Of the eligible mothers, 1,376 (93%) participated at 6 weeks. This interview was carried out in the preferred language of the mother. Detailed information about the cohort and procedures is described elsewhere [34]. The analyses in the present study were based on the 1,085 women who were biological mothers, cohabiting in married or de-facto partnerships and who completed the measure of IPV and PND. The demographic profile of these mothers is similar to that of the full cohort of 1,376 mothers.


Intimate Partner Relationships

Intimate Partner Violence (IPV) was measured using Form R of the Conflict Tactics Scale (CTS) developed by Straus [35]. Participants were asked to enter their responses on an answer sheet while the interviewer read the questions out loud. Mothers reported on their behaviour towards their partner (“perpetration”) and their partner’s behaviour towards them (“victimization”). The CTS measure of verbal aggression includes six items; the minor physical violence scale includes three items; and the scale of severe physical violence includes six items. An individual was considered to be a perpetrator of verbal aggression, minor physical violence, or severe physical violence if she reported any of the behaviors during the past 12 months. Likewise, an individual was considered to be a victim of verbal aggression, minor physical violence, or severe physical violence if she reported that a partner had perpetrated any of the behaviors towards her during the past 12 months. Psychometric properties of the CTS scales are robust and have been described by Straus [36]. Responses were included in these analyses if participants completed 17 of the 19 CTS items. Cronbach’s coefficient alpha reliability values for the victimization and perpetration scales, respectively, were: verbal aggression, 0.77, 0.80; minor violence, 0.84, 0.81; and severe violence, 0.79, 0.81 [33]. Only IPV victimization was used for analysis as the purpose of this study is to explore the impact of experiencing IPV victimization on PND.

Physical violence was defined as a dummy variable with the values ‘no violence’ or ‘any violence (minor and/or severe)’. In addition, to examine the gradient relationship, it was also classified as a categorical variable with three levels: (1) no violence, (2) minor violence only, and (3) severe violence.

Postnatal Depressive Symptoms

The depressive symptoms of the mothers were screened with the Edinburgh Postnatal Depression Scale (EPDS). In the EPDS, the mothers are asked to choose from the given options those that best describe their feelings during the previous week. Each of the items is scored on a four-point scale from 0 to 3, with a total score ranging from 0 to 30. A cut-point of 12/13 is recommended for screening major depression [32]. In this study, mothers who scored above the cut-point are referred to as experiencing ‘probable depression’ and mothers who scored below the cut-point as ‘non-depression’. Although the reliability and validity of the EPDS has been established in a variety of populations and settings [37], its use with Pacific Island mothers is limited [38]. The reliability coefficient for mothers in the PIF Study at 6 weeks was 0.86.

Other Socio-demographic and Maternal Factors

Maternal age, ethnicity, education level, marital status, country of birth, years resident in New Zealand, household income, parity, cultural orientation [39], maternal health problems, being stressed due to insufficient money for food, happiness in relationship, satisfaction with home, home overcrowding, difficulty with transport, whether pregnancy was planned, reaction to pregnancy, satisfaction with birth experience, satisfaction with infant’s sleep patterns, infant health problems, manageability of infant, and whether mother is enjoying infant, were collected at the baseline 6-week interview.


The study was approved by Auckland Branch of the National Ethics Committee, the Royal New Zealand Plunket Society and the South Auckland Health Clinical Board.

Statistical Analysis

Chi-squared tests were employed to investigate associations between IPV variables and possible postnatal depression. The associations between happiness in partner relationships and IPV variables were tested using the Cochran-Armitage Trend Test. Multiple logistic regression analyses were used to examine the variation in risk of experiencing possible PND for IPV variables controlling for confounding factors. In these analyses, verbal aggression, physical violence, and other socio-demographic and maternal factors (listed in the above section of measures) that were identified as potential risk factors for PND in another PIF study [22], were entered to the original models and then the backward elimination method was used to select significant variables that were retained in the final models.

All analyses were performed using the SAS version 9.1 statistical software package and a significance level of α = 0.05 was used to determine statistical significance for all calculations. The burden of PND attributable to IPV was calculated using population attributable risk (PAR) involving the prevalence of IPV (P) and the relative risk (RR) of the outcome: PAR = P(RR − 1)/(1 + P(RR − 1). The relative risk here was estimated by adjusted odds ratio.


Of the 1,376 mothers who were interviewed at the 6-week assessment point, there were 1,100 mothers who were cohabiting in married or de-facto partnerships; of these, 1,085 mothers completed measures of IPV and the questions regarding postnatal depressive symptoms and were used as the cohort for the analyses reported in this paper. Details of the socio-demographic characteristics of this sample are presented in Table 1.
Table 1

Frequencies (percentages) of socio-demographic characteristics for participating mothers (N = 1,085)

Socio-demographic characteristics



Age (years)













Highest educational qualification

    No formal qualifications
















    Cook Island Maori






    Other Pacific






Years lived in New Zealandb










Marital status




















Household income (NZD) at 6 weeks interview













NZ born







Cultural orientation

    High NZ, Low Pacific



    Low NZ, High Pacific



    High NZ, High Pacific



    Low NZ, Low Pacific



aIncludes mothers identifying equally with two or more Pacific Island groups, equally with Pacific Island and non-Pacific Island groups, or with Pacific Island groups other than Tongan, Samoan, Cook Island Maori or Niuean

bThree observations missing

cEleven observations missing

Prevalence of IPV and Probable Depression

Within this cohort of 1,085 mothers, 836 (77.1%) reported being the victim of verbal aggression and 250 (23.0%) reported physical violence victimization. Of the 250, 120 (48%) reported that the violence was severe and 130 (52%) reported the violence to be minor only. About 165/1,085 (15.2%) women reported postnatal depression, a prevalence rate significantly lower than that of 276 non-partnered women (21.2%, P = 0.02). However, no significant difference in the prevalence of depression was found (P = 0.08) between this cohort (N = 1,085) and those cohabiting but having missing IPV variables (N = 15).

Associations Between IPV and Partner Relationships

There were statistically significant associations between partner relationships and verbal aggression and any physical violence (P value for both trend tests <0.0001). The happier the mothers were in their relationships, the less likely they were to be victims of verbal aggression (the prevalence of victimization of verbal aggression was 97.8% for those being a little/fairly/extremely unhappy with their relationships, 82.8% for being happy with their relationships, and 74.8% for being very happy/extremely happy/perfect with their relationships, respectively). The corresponding prevalence of victimization of physical violence was 60.0%, 41.9%, and 17.3%, respectively.

Associations Between IPV and PND

Univariate Analyses

The prevalence of possible PND and the unadjusted odds ratios (OR) and 95% confidence intervals (CI) by IPV victimization in terms of verbal aggression and physical violence are presented in Table 2.
Table 2

Numbers (row percentages) and univariate odds ratios of possible PND by IPV



Possible PND

Univariate odds ratio








95% CI

Victim of verbal aggression














1.10, 2.67

Victim of physical violence (two levels)














2.42, 4.97

Victim of physical violence (three levels)














1.62, 4.05







2.96, 7.03

Statistically significant differences were found in the prevalence of PND for both verbal aggression and physical violence victimization. Women who were victims of IPV were more likely to report probable depression than those who were not (16.6% vs. 10.4, OR: 1.71, 95% CI: 1.10, 2.67 for verbal aggression; 29.6% vs. 10.9, OR: 3.44, 95% CI: 2.42, 4.97 for physical violence). Further classification of the violence by severity suggested a moderate dose-response relationship with the severe physical violence group having a higher proportion of reports of PND than the minor physical violence group (35.8% vs. 23.9%).

Multiple Logistic Regression Analyses

The results from the two multiple logistic regression models including all possible socio-demographic and maternal risk factors and IPV for predicting depression are presented in Table 3. Verbal aggression was removed from both models. The odds of reporting depression for physical violence victimization remained statistically significant (OR: 2.34, 95% CI: 1.52, 3.60). Model 2 suggested that the moderate dose-response relationship identified from the univariate analyses (severe, minor, and no violence) became minor and effects tended to merge (2.00 for minor and 2.80 for severe violence).
Table 3

The adjusted odds ratios and 95% confidence intervals of reporting possible PND by physical violence victimizationa

Victim of physical violence



95% CI

Model 1: two levels






1.52, 3.60

Model 2: three levels






1.17, 3.42



1.61, 4.86

aThe values were adjusted by maternal ethnicity, household income, parity, satisfaction with home, difficulty with transport, reaction to pregnancy, satisfaction with infant’s sleep patterns, happiness in relationship, and whether stressed due to insufficient money for food that were collected at the baseline 6-week interview. Other variables included for consideration and were then removed from the final models are victim of verbal aggression, maternal age, marital status (Legally married vs. Partnered/defacto), educational qualifications, whether born in New Zealand, length of time lived in New Zealand, home overcrowded, cultural orientation, maternal health problems, pregnancy planned, satisfaction with birth experience, infant health problems, manageability of infant, and whether enjoying infant

In addition to physical violence victimization, the following variables were selected into the two final models with similar effects across the models: maternal ethnicity, household income, parity, satisfaction with home, difficulty with transport, reaction to pregnancy, satisfaction with infant’s sleep patterns, happiness in relationship, and whether stressed due to insufficient money for food.


Overall, the PAR of PND by partner physical violence victimization was 23.6%, suggesting that PND would possibly decrease by 24% if partner physical violence victimization was eliminated from the Pacific Islands populations.


The aim of the present study was to examine the relationships between maternal IPV victimization and symptoms of PND among women of a Pacific birth cohort. The findings suggested that mothers of Pacific children who were victims of any intimate partner physical violence in the previous 12 months were more than twice as likely as those who were not victims to report postnatal depressive symptoms at 6 weeks postpartum, after controlling for a whole range of socio-demographic and maternal risk factors. There was a weak trend for a gradient relationship between severity of physical violence and depressive symptoms. The magnitude of the IPV effect found in our study is comparable to a number of national and international studies [6, 15, 17], in which an odds ratio of 2–3 was reported to predict women’s depression. Using burden of disease methodology to estimate the health risks of IPV among women in Victoria, Australia, Vos et al. reported that the largest contribution to the burden of disease associated with IPV was poor mental health. Depression, anxiety, and suicide together contributed to 73% of the total disease burden associated with IPV [15]. A study of women attending general practice suggested a strong association between depression and physical, emotional, and sexual abuse by partners or ex-partners and it was found that depressed women were about six times more likely to have experienced severe combined abuse than women who were not depressed after adjusting for other significant socio-demographic variables [40]. International population-based studies of other ethnic groups have found similar or stronger associations between IPV and depression, reporting odds ratios of 3.4 and 4.1, respectively [14, 16].

Although depression at any time during a woman’s lifetime is of concern, depression around the time of pregnancy and postpartum is of special significance as it may adversely affect the health and wellbeing of the mother and infant [10, 41]. Very little research has examined the association between IPV and PND, making comparison of our findings difficult. An investigation of women interviewed 3–5 days postpartum found that 83% of 41 women who were battered during pregnancy were depressed, compared to 57% of women who had never been victimized [42]; however, the assessment of depression was somewhat limited because it was based on only one question. PIF study findings [22], consistent with international studies [1921, 23], demonstrated that mothers who reported feeling unhappy in their partner relationships were more likely to report depressive symptoms; however, it is clear that violence within a partner relationship has a detrimental effect on PND and should be considered as an independent risk factor. Our study showed that after controlling for confounding factors, both IPV and happiness in partner relationships were independently associated with PND.

The present study is confined to cross-sectional data and thus a causal relationship cannot be ascertained. It is possible that IPV was a consequence of depressive symptoms that occurred before or during pregnancy or that depressed women tended to be more likely to perceive themselves as victims of IPV than non-depressed women. However, the fact that mothers were asked about their IPV experience in the last 12 months and depressive symptoms in the last 7 days suggested a temporal order, that is, most IPV occurred before postnatal depressive symptoms. Previously, some socio-demographic factors such as ethnicity and household income were found to be associated with both IPV and PND [22, 33]. The significant association between IPV and PND even after adjusting for these common underlying causes places more weight on a causal relationship. Researchers have demonstrated that in a perinatal sample, changes in marital adjustment (as measured using the Dyadic Adjustment Scale) were evident in a depressed group even when the effects of depression were controlled for [43]. Inclusion of assessment of past history of partner violence and enquiry into the timing, duration, persistence, or recurrence of depressive episodes would have helped to identify the roles played by IPV and a history of depression in developing PND. No such information was collected in the PIF study; however, subsequent measurement cycles in this longitudinal study will include measures of IPV and maternal health. This will enable us to revisit the two domains to establish the temporal sequence of events and identify the associations between IPV and depressive symptoms beyond the postpartum period.

Although our findings suggested that women who experienced verbal IPV victimization had an elevated risk of PND compared to those who did not experience such victimization, the association was no longer significant after adjusting for confounding factors. This is in line with other findings that have shown that although women are highly exposed to verbal aggression, it is the acts of physical violence that are the most influential [40].

Previous PIF study findings report a high degree of overlap between IPV victimization and perpetration [33]. CTS did not examine the context of the violence, thus no deep understanding could be gained of the roles that each partner played in the eventual outcome of IPV, e.g., who initiated the violence, for what reason, how the partner reacted, etc. Based on the assumption that IPV victimization is more likely to cause PND than perpetration and the collinearity between IPV victimization and perpetration, we have only included IPV victimization in the present paper. More qualitative research to explore the cultural contexts of male–female relationships and perpetration and victimization may help us understand the mechanisms through which IPV is associated with postnatal depression.

The following limitations must be noted when interpreting the findings from our study. Given that the data are based on maternal report the possibility of bias from self-reported IPV victimization and PND remains. Depressed women may be oversensitive to IPV, leading to an exaggeration of their experience of IPV. Paternal reports that will be collected in later assessment points may clarify this possible bias. It is also acknowledged that Pacific women may have a different interpretation of violence, in particular severe violence with their intimate partner, and that the cultural context must be taken into consideration. Such issues are important to follow up with more in-depth qualitative interviews with samples both within and outside the longitudinal cohort, further exploring the reasons for IPV and how women are affected within this context.

Despite these limitations, this study contributes to the limited data available worldwide on IPV and PND from different ethnic groups. There is a growing body of research showing that many women, including those who are pregnant, have been victims of IPV, including physical and sexual assault [44]. Reducing partner violence by Pacific peoples has been identified by Pacific communities and government agencies as a priority issue [45]. The findings reinforce the importance of providing screening for both IPV and depression within the context of women’s health care. The results of the study may help to develop culturally appropriate social services and policies for Pacific women. From the practice perspective, women’s care providers should be cross-trained, that is, reproductive health providers should be aware of signs of women’s depressive symptoms and refer clients to appropriate services; professionals who provide mental health services should screen their female patients for violence; likewise, specialists who provide services to abused women should assess their clients for depression and other health problems. By working together, the women, their families, and indeed the whole community will benefit greatly.


The Pacific Islands Families (PIF) Study is funded by grants awarded from the Foundation for Research, Science & Technology, the Health Research Council of New Zealand and the Maurice & Phyllis Paykel Trust. The authors gratefully acknowledge the families who participate in the study as well as other members of the research team. In addition we wish to express our thanks to the PIF Advisory Board for their guidance and support.

Copyright information

© Springer Science+Business Media, LLC 2008